of the _ Missou П Volume 78 A Volume 78, Number 1 Winter 1994 Annals of the Missouri Botanical Garden The Annals, published quarterly, contains papers, primarily in systematic botany, con- tributed from the Missouri Botanical Garden, St. Louis. Papers originating outside the Garden will also be accepted. Authors should write the Editor for information concerning arrangements for publishing in the ANNALS. Instructions to Authors are printed in the bei of the last issue of each volume. Editorial Committee Marshall R. Crosby Editor, Missouri Botanical Garden Amy Scheuler | Managing Editor, Missouri Botanical Cu Geri Dolan Magdalen Lampe онен ES Missouri Botanical Garden & | Henk = der Werff Gerrit Davidse Missouri Botanical Garden John D. Dwyer Saint Louis University — Peter Goldblatt Missouri Botanical Garden Dale E. Јакино Missouri Botanical Garden: У Missouri Botanical Garden Ф 7 | | For subscription information contact Department . Eleven, P.O. Box 299, St. Louis, MO 63166-0299. _ Subscription price is $75 рег volume U.S., $80 .. Canada and Mexico, $100 all other countries. Four _ issues per volume. Beginning in 1991, recipients of _ the ANNALS will automatically receive Novon, the Garden’ s newly established scientific quarter! у © Missouri Botarical Garden 1991 - (ISSN 0026-6493) is published quarterly by the - Missouri Botanical Garden, 2345 Tower Grove А\ THE Missourt. BOTANICAL GARDEN, Department Eleven, P.O. Box 299, St. Louis, мо 63166-0299 The ANNALS OF THE MISSOURI BOTANICAL GARDE enue, St. Louis, МО 5» 10. etd у paid at St. Louis, MO : Send address changes to ANNALS oF Volume 78 Annals Number 1 of the 1991 Missouri Botanical Garden ONE HUNDRED YEARS Emanuel D. Rudolph? OF THE MISSOURI BOTANICAL GARDEN'? ABSTRACT In the hundred years since the Board of Trustees of the Missouri Botanical Garden implemented the stipulations of Henry Shaw's will, which prescribed functions for his public garden and provided an appropriate endowment, six directors have guided the Garden's development. Henry Shaw's (1800-1889) vision of the Garden was based upon his British background and the European botanical garden tradition. William Trelease, the first director (from 1889 to 1912), set his highest n on building the herbarium and library and publishing А dem. annual volumes, which he believed would m e Garden a major center for taxonomic research. George T. Moore, director for 40 years (from 1912 to 1953), 4 the research into other areas of botany, applied botany, and е Кееп оп оаа display yet facing ан finances toward the latter part of his tenure, he began to seek private support rograms, including a new arboretum ا‎ =: outside the city. The short duration of Edgar Anderson's directorship (1954-1957) was focused on Garden improvements with further attempts to raise funds a find support mong garden clubs. Frits Went, the fourth deles (1958-1963), saw the Garden as a potential multifaceted cultural center for St. Louis. The geodesic domed greenhouse, named the Climatron, opened in 1960 as a ma pi implement in concentrating local and national on on the Garden. The fifth director, David Gates (1965-1971), moved the Garden into ecological research and found support for a new research building that had been sought by Went. protect it. Each director has guided the Missouri i Garden in certain directions during the last hundred years to make it the major botanical Marien that it no Thinking sentimentally back to his youth at the "Peter Collinson ... had a house and p оке Mill Hill School in England, Henry Shaw, then in ill ... ten miles from London, and was one о chief encouragers of gardening ^ е in his time his seventies, sat in Tower Grove House looking а Ded ke 1258 at the age of 74 years, out over his St. Louis Public Botanical Garden and аһ some time after his death ы house and grounds wrote: were purchased by the Protestant Dissenters Associa- ' This and the five articles that follow it are the роса of the 36th Annual Systematics Symposium of the Missouri Botanical Garden, А Century of Botany: 1889-1989. The symposium was held 6-7 October 1989 at the Missouri Botanical Garden in St. Louis, Missouri, U.S.A. ? The staff of the library and archives of the Missouri Botanical Garden are gratefully acknowledged for unfailing assistance and hospitality. The use of the archives and special collections of the Royal Botanic Gardens, Kew, the Gray-Arnold Arboretum and Farlow libraries of Harvard University, and Cornell University Library are gratefully acknowledge 3 Пена of Botany, The Ohio State University, Columbus, Ohio 43210, U.S.A. ANN. Missouri Вот. Garp. 78: 1-18. 1991. Annals of the Missouri Botanical Garden tion for a school for classics and mathematics grew in the garden, with the Cypress of Linnaeus; but the most admired trees were the noble Cedars of Leb- 181 2 to 1815 thus releived fugging [sic] at Latin Greek and Euclid, by growing pinks and geraniums, and re- tiring to a seat in a corner of his little domain, shaded by a broad trained Irish yew, purchased from a оин . The field in the Great Northern Mail coach, spanking grays, bedecked with greens and ribbons that carried the glorious news of victory to the joyful people of Scotland and North Britain." The wealthy bachelor Shaw (Fig. 1) had worked for more than 40 years before his death in 1889 to perfect his gift of a garden for the people of St. Louis, Missouri, and the nation.? Perhaps he thought about Peter Collinson's garden, which he knew was mostly lost to posterity. After retirement from his active hardware and dry goods business in 1839, Shaw invested his funds in sound St. Louis real estate and other business ventures, which increased his wealth.* Between 1840 and 1852 he made three extensive European trips, which included visiting gardens in different countries. In time, he settle upon making a botanical garden, rather than a pleasure garden, on part of his country land in St. Botanical gardens have a long and noble tra- dition dating from Renaissance times.’ Their early particular function was to cultivate plants with me- dicinal uses and those that served as living texts for learning about materia medica. Thus, they were mostly connected with university medical schools or apothecary guilds. The earliest botanical gardens at Padua and Leyden were laid out in the form of four large beds making a cross pattern with their major paths. The cross is significant not only for its religious connotations, but also because it di- vided the garden into the four great world divisions then known: Europe, Asia, Africa, and the Amer- icas. Here then was the world encompassed; God’s munificence to humans from all corners of the globe displayed (Fig. 2). In time, it became evident that not all plants did well under the same conditions. Later botanical gardens, such as the Jardin Du Roi in Paris, had varying habitats, some wet, dry, sunny, and shad- ed. In fact, it was what we call today an ecological approach to plant cultivation which by this time included all sorts of plants, not only medicinal ones (Fig. 3). By the eighteenth century, learning about God’s creation was a worthy endeavor in itself, one that could only magnify the deity.* When Henry Shaw traveled in Europe, there were numerous botanical gardens that served as models. = surprisingly, his British background led him in 1854 to contact Sir William Jackson Hooker, the en of the Royal Botanic Gardens, Kew.? In time, Shaw emulated some aspects of Kew, for example, by building a museum oe and plant houses, and later an orangery—the Lin nean House—based upon Kew models. However, as he wrote to Hooker, he did not have in mind the extensive size and broad activities of Kew, but rather a garden of more modest dimensions such as those of Glasgow or Liverpool." Hooker put Shaw in contact with Asa Gray at Cambridge and George Engelmann at St. Louis. They gave him advice and encouragement.’ On a European trip in 1857 Engelmann pur- chased the herbarium of some 40,000 specimens from the estate of the recently deceased Johann Jacob Bernhardi, and found botanical books for Shaw." These botanical resources, later including the herbarium and library of Engelmann after his death, were dutifully kept, but little used, in the Museum Building (Fig. 4), which then displayed stuffed animals and oil portraits of important bot- anists. These portraits can now be seen on the Garden's library walls. haw concentrated his efforts on the horticul- tural aspects of the garden. With the aid of his Kew-trained gardener m Gurney, he ike outdoor beds of ornamentals shrubs and fruit trees, and an Kuna of trees. The plan was to divide the garden into the three great groups of plants; herbs, shrubs, and trees. In ad- dition, plants needing protection were housed under glass (Fig. 5). By the time of this death in 1889, Henry Shaw had established a considerable public rurticet 11m Е 1. id Shaw at about age 87, from an oil painting at the Missouri Botanical Garden, painted by Fic Fairchild nd Fox Volume 78, Number 1 Rudolph 1991 100 Years of the Missouri Botanical Garden 4 Annals of the Missouri Botanical Garden cm | пр ~ fy ТАКИ BU'lAITJICO DI рапта RE 2. Plan of the botanical garden at Padua, from R. de Visiani, L'Orto botanico dei Padova nell'anno MDCCCXLII, Padua, Angelo Sicca, 1842. (Courtesy of Hunt Institute for Botanical Documentation, Carnegie Mellon University, Pittsburgh, Pennsylvania.) Rue de Seno Chanters + 3 М ; | I < > М N S 5 | ЕЕ | \ | "nell | ФА | T m I. (ШШЩ | AE WR MAU | 555 м жү ж, > Q 61000008 оваа LI. En 4D 909400004 023209907 20. к s ола e | 233: | Rue de ЖЕ Н } - ан — —— — ————————— A Shin ha garidi des plantar CME ла GURE 3. Plan of Jardin des Plantes when the Jardin du Roi, from J. B. Ријошх, Promenades au Jardin des рш, а la Ménagerie et dans les Galeries du Muséum d'Histoire cla, . Troisieme édition, Paris, Librairie Economique, 1803. (Tome premier.) (Courtesy of Hunt Institute for Botanical Documentation, Carnegie Mellon University, Pittsburgh, Pennsylvania.) Volume 78, Number 1 1991 Rudolph 5 100 Years of the Missouri Botanical Garden 5 |“ | ГУТ: пая а п съ EUN д "i a || || er Зли THE MUSEUM—SHAW'S GARDEN. FIGURE 4. The Museum Building at the Missouri Botanical Garden during the lifetime of Henry Shaw, from J. A. Dacus & James W. Buell, 4 Tour of St. Louis; or, the Inside Life of a Great City, Western Publishing Co., St. Louis, 1878. botanical NEN adjacent to his country villa, Tow- er Grove.'* Property nearby had been given to the city to develop, with his advice and assistance, an elaborate pleasure park with walks and roads, Tow- er Grove Park.'* With significant input from Asa Gray, who vis- ited Henry Shaw in 1884 and again in 1885, aiding him in his plans, Shaw drafted a will in 1885.'° It provided that upon his death the Missouri Botanical Garden, founded by him and approved by the Gen- eral Assembly of The State of Missouri in 1859,” would have income from an endowment to be ad- ministered by a self-perpetuating Board of Trust- ees, which appoints the director to oversee the running of the Garden. It states that the Garden, "should be forever kept up and maintained for the cultivation and propagation of plants, flowers, fruit and forest trees, and other productions of the veg- etable kingdom; and a museum and library con- nected therewith, and devoted to the same and to the science of botany, horticulture, and allied ob- jects ....”'* At Gray’s suggestion, Shaw decided a set yearly stipend for support of a botany school at Washington University, to be known as The Henry Shaw School of Botany, and a professorship, to be known as the Engelmann Professor.' William Trelease, then a professor of botany at the University of Wisconsin, was appointed the first Engelmann Professor and Director of the School in 1885. When the Trustees met upon Shaw’s death in 1889 to carry out the terms of his will, Trelease was also immediately selected to be the first director of the Missouri Botanical Garden (Fig. 6)." A northeasterner with an undergraduate де- gree in entomology from Cornell University under John Henry Comstock (1880) and a doctorate from Harvard University under William Gilson Farlow 1884), Trelease was recommended by Gray and others as the most suitable person to lead the new school of botany. Trelease had already made a name for himself in published studies of insects, pollination, fungi, and plant morphology. His doc- toral thesis on bacteria was the first on this group in America, and his bacteriology course at Wis- consin was perhaps the first one offered there. He had strong connections with the Eastern Coast bo- ~ 6 Annals of the Missouri Botanical Garden N s е AE Leo y, : tans D | D A LE P ay Pigilan сав уз gt LSU ду У 2 Ф INTERIOR OF PLANT HOUSE. FIGURE 5. Inside view of the Plant House during Henry Shaw's lifetime, from J. A. Dacus & James W. Buell, A Tour of St. Louis; or, the Inside Life of a Great City, Western Publishing Co., St. Louis, 1878. tanical establishment, and through his Wisconsin experience, some midwestern ties. Trelease proved to the ideal person to implement the spirit and substance of the Shaw-Engelmann- Gray vision of the Missouri Botanical Garden. Once moved into the Tower Grove House, Trelease quickly realized the serious problems with the Gar- den's physical plant. He at once began to improve the grounds and buildings with the help of James Gurney. Trelease also initiated a strong research and teaching program. He built a substantial her- barium of plants from the southwestern states by collection and purchase, and made monographic studies, as did others at the Garden, on plants of the region. Imagine the Garden in the year of the Louisiana Purchase International Exposition in 1904. This largest and most elaborate World's Fair up till that time? drew many to St. Louis and to the Garden. Trelease, after his 14 years as the director, was proud of the Missouri Botanical Garden. Once in- side the gate in the high stone wall around the Garden visitors would find that “11,357 species and varieties of plants were in cultivation.” After Volume 78, Number 1 1991 Rudolph 7 100 Years of the Missouri Botanical Garden visiting the outside garden, lily pools, some with the large Victorias, a large area with native Amer- ican flowers, an orchard, a fructicetum, and a large arboretum, one could enter a number of recent interconnected greenhouses. Some outside plant- ings and greenhouses were devoted to yuccas and agaves, which both Engelmann and Trelease stud- ied. Tower Grove House, with a wing added in 1890, was the director's residence. Shaw's city house, rebuilt at the Garden in 1891, contained the offices, and much of the herbarium and library; an added wing was to be built in 1909. Some of the herbarium and library still were housed in the Museum Building, which also included a few lab- oratories. Trelease had put most of his energy into building a sound foundation for the future research and teaching programs. The herbarium in 1889 con- sisted of about 160,000 specimens, many unpro- cessed, from the Bernhardi, Engelmann, and Nich- olas Riehl collections; in 1904 it had five times the number, or about 490,000 specimens.” Included were purchases or gifts of some major collections made by Ferdinand Blanchard, Farland Carr Broadhead, Samuel Botsford Buckley, Benjamin Franklin Bush, Alvan Wentworth Chapman, J. Q. Adams Fritchey, Gustav Jermy, Joseph Finley Joor, Sadie F. Price, John Howard Redfield, the Rev- erend Ernest C. Smith, E. Lewis Sturtevant, George Thurber, as well as Trelease's own collections. The library in 1889 had about 428 volumes from Henry Shaw's purchases and other books, some from George Engelmann's library, totaling about 3,000 volumes; in 1904 it had 45,815 volumes or an increase of over 15 times.” But numbers alone do not tell the whole story. In 1892 E. Lewis Stur- tevant donated an important collection of 463 pre- Linnaean books and his own manuscripts and notes. Then in 1902, a large collection of pre-Linnaean books was purchased from a German bookseller. Together, these collections made the Garden a significant repository of early botanical literature. Jens Christian Bay, newly arrived from Denmark, was hired to prepare a catalog of the Sturtevant Collection.” Bay was later to become the librarian at the Crerar Library in Chicago. The catalog was published in the Missouri Botanical Garden An- nual Report, a publication that served from 1890 until 1912 as an outlet for publishing Garden re- search results. Scientific research was the primary goal of Trelease's plans, and he himself had begun major monographs of Yucca and Agave done when he had time to travel and do research. Specialization in herbarium collections had be- gun. Because the Engelmann herbarium was strong William Trelease in 1903 at age 46. (From IGUR the archives of the Missouri Botanical Garden.) in plants of Texas and the Southwest, these col- lections were selectively increased by collecting expeditions. It is interesting to note that a start was made on an African collection by paying S. P. Verner, who had brought African Pygmies to the World's Fair, to collect plants in the Upper Congo. Bright young students were employed as assis- tants by Trelease. Some of them were to become important botanists: Albert Spear Hitchcock, Lewis Hermann Pammel, Frederick Hinsdale Horsford, and J. Arthur Harris. He encouraged Hermann von Schrenk, a plant pathologist associated with the United States Department of Agriculture, to locate his laboratory at the Garden, where he main- tained it for 50 years.” Though it was difficult to get a new major grad- uate teaching program instituted, a school for gar- deners was quickly established. This was one of Shaw's designated programs, probably because he and Asa Gray had so much trouble finding well- trained American gardeners.” This significant school was to last into the 1930s. The successful 1904 World's Fair led to op- Annals of the Missouri Botanical Garden = to a ori a nt d 5 а - am ча ча a = а " y | П СОВЕ 7. The Missouri Botanical Garden/Mill Hill School exhibit at the Louisiana Purchase International Exposition, St. Louis, 1904. (From the archives of the Missouri Botanical Garden.) portunities for the Garden.” The Garden had ап exhibition in association with the Mill Hill School, which showed living plants, preserved local mush- rooms, pictures of the Garden and Shaw, and Gar- den publications (Fig. 7). This exhibit, together with various living flowers that won awards in compe- titions, gave the Garden favorable publicity. At the close of the Fair, some of the exhibitors provided exotic living plants, thus increasing the Garden's holdings from such countries as Brazil, China, Cuba, and the Philippines.” Trelease was successful in establishing a superb reputation for the Garden in the botanical scientific community. He was less successful in convincing the Board of Trustees that his was the right direc- tion. In 1912, as a result of increasing disagree- ment with the Board, Trelease resigned. The dis- agreements had become acute a year before when a new president of the Board was elected and, as Trelease wrote to his mentor Farlow, he was ad- vised that “we're going to have more garden and less science, now."*' Trelease was not willing to compromise his principles, and his resignation by agreement was announced as а “retirement” to do research, few knowing the real reasons. George T. Moore, who had been brought in by Trelease in 1909 as professor of plant physiology in the plans to expand the graduate program, became the new director. Trelease also had hired George Pring, a Kew-trained gardener who was later to make a reputation as a water lily breeder, to be in charge of orchids and exotics.” Some able people were still in charge, but programs were in need of further implementation. George Thomas Moore (1871-1956; see Fig. 8) was a midwesterner by birth and early education, with an undergraduate degree from Wabash Col- lege in Indiana in 1894.* His graduate training was from Harvard University, working with Wil- liam G. Farlow (A.M. 1896, Ph.D. 1900), and then teaching at Dartmouth College in New Hamp- shire (1899-1901). He was algologist and plant physiologist with the Bureau of Plant Industry, United States Department of Agriculture, from 1901 to 1909, and was in charge of its plant physiology laboratory since 1902. Moore's back- ground made him an ideal selection for Trelease to put in charge of the graduate program at the Garden and Washington University in 1909. Moore's research was with algae and bacteria, both Volume 78, Number 1 1991 Rudolph 9 100 Years of the Missouri Botanical Garden FIGURE 8. George Thomas Moore after becoming director of the Missouri Botanical Garden. Photographed by S. D. Whiting. (From the archives of the Missouri Botanical Garden.) pure and applied. At the Bureau of Plant Industry he perfected methods to rid water supplies of pesty algae, and to inoculate plants with nitrogen-fixing bacteria. He was a skillful speaker, writer, teacher, and promoter of botany. It was no wonder that Moore impressed the director's search committee of the Board with his clear descriptions of what needed to be done at the Garden.* He had the ability of being able to speak to businessmen on their own terms. For example, after being the director for three years, he told the Board that **. . . the increase in the service of the activities, such as the herbarium, laboratories, etc., has been quite as great as the increased number of visitors. Аз a matter of fact, the development of what may be termed the strictly scientific side of the Garden shows more tangible results for the amount of money which is annually devoted to this purpose than the Garden proper; and so far as a return upon the investment is concerned, may al- ways be expected to be the most fruitful of the Garden activities.” He at once set out to improve the show aspects with frequent floral displays, well- kept grounds, and innumerable press releases pro- ducing publicity that attracted large crowds. By using comparisons with other institutions in St. Louis and elsewhere, by stressing practical uses of research, for example cancer research and water and air pollution research, and by persuasion, Moore also was able to expand the scientific and educa- tional programs, something that Trelease had been unable to accomplish. When Jesse M. Greenman was hired by Moore in 1913 to be curator of the herbarium, he was charged with preparing a south- western flora, a task never completed. Moore soon got on an equal footing with the Trustees by being first ап ex-officio member of the Board as the president of the St. Louis Academy of Science in 1918, and then in 1928 by being elected as a lifetime regular Trustee. Further, because the oard had so much confidence in his abilities, he was able to eliminate the required monthly budget requests, to curtail the written monthly reports of the director, and dispense with the oversight of Garden operations by the Garden committee of the Board. In short, Moore was able to run the Garden with less and less direct supervision by the Trustees. To examine Moore's achievements, imagine how it would have been to visit the Garden in the middle of his 41-year directorship, in the year 1926.” Upon entering through the recently enlarged and redesigned gate, one would find lily pools leading to new large greenhouses, one of which was a floral display house with continuous ornamental displays. In the rear of the greenhouses was an elaborate sunken Italian garden with pergola. А new rose Annals of the Missouri Botanical Garden garden, economic garden, and many simplified manicured ornamental flower beds were well cared for. A director's residence had been built south of the Administration Building The scientific staff in 1926 included: Benjamin M. Duggar, plant physiologist; Jesse M. Greenman, curator of the herbarium; Edward А. Burt, librarian (in name only) and mycologist; Edgar Anderson, geneticist, also in charge of the Gardeners' School; ermann von Schrenk, plant pathologist; and a second mycologist, David H. L nephew, soon to be replaced by Carroll W. Dodge from Harvard. Scientific programs were active, with results being published in the quarterly Annals of the Missouri Botanical Garden initiated in 1914 to replace that function in the old Annual Reports. Information and popular articles now appeared in the Bulletin, started in 1913. An active research project was collecting and studying the South- western flora, concentrated on the plants of Mis- inder, Farlow's souri, Arkansas, and Texas. Julian Steyemark, a graduate student, was to study the Missouri flora then, and for years to come. A newly opened tropical station of the Garden in Balboa, Panama Canal Zone, encouraged plant collecting and the flora of Panama project later to develop.” Research in plant physiology, particularly of fungi, plant pathology, mycology, and plant breeding was in progress. Moore gave up his active research to spend time in administration, teaching, and public speaking. The Garden was becoming noted in horticultural circles for its water lily and orchid hybridization. Collected and donated orchids had made the Gar- den preeminent in this country for its extensive living collection of orchid species." James Gurney and George Pring were active in the hybridizing program. Unfortunately, the excessive smog of soft coal particles was damaging outside plantings and indoor orchids. The orchids were not only impor- tant for displays, but also for income by supplying local florists. Moore campaigned with those in the city who sought smoke abatement laws, but prog- ress was slow.*' As a result of the smoke problem, the Trustees, with the court's approval, sold land adjacent to the Garden in order to purchase an out-of-town site for plants, particularly for trees, and greenhouses for orchids. In 1925, extensive property on the Meramec River at Gray Summit was purchased, to be called at a later date the Shaw Arboretum.*? In 1926 the future for the Garden seemed bright. However, in a few years that was to change dras- tically. The Great Depression affected value and income from the properties from which the Gar- den’s operating funds came. Even after the De- pression had abated, these properties did not re- cover enough to help the Garden recover. Moore was getting older and less energetic. He kept up a good front at the same time that he was econ- omizing by moving the plant physiology to the university, decreasing the number of gardeners, cutting the number of floral displays, and limiting the budget for library, herbarium, and grounds. Edgar Anderson was to write of this period that Moore ^... retrenched as skillfully as he ad- vanced,” which meant that many did not realize the seriousness of the Garden's problems. Moore attempted to raise funds through his extensive so- cial contacts, and by the formation in 1939 of a Friends group. However, because of the immense needs for repairs, replacement, and catching-up, these efforts turned out to be insufficient. Then, too, Moore's health deteriorated and Edgar An- derson was made assistant director in 1951. The Trustees seemed to be little involved with solving the problems. In short, the Missouri Botanical Garden was in the doldrums in 1953 when a new Board president, John S. Lehmann, was elected and when Moore became Director Emeritus. The scientific com- munity scarcely realized the situation because many of the research programs and their publications continued, including the Flora of Panama, even though the tropical station had closed some years earlier. But the physical plant and gardens were in disrepair. The directorship of Edgar Anderson (Fig. 9), which began in 1954, was to last only a few years; however, Anderson and his assistant Hugh Cutler made significant changes. Anderson's long con- nection with the Garden began in 1922 when he came as geneticist and assistant professor at Wash- ington University, fresh from the Bussey Institution of Harvard University with a doctor of science degree. He was also in charge of the gardeners’ school. Growing up in East Lansing, Michigan, where his father was a professor of animal hus- bandry at Michigan State College, Anderson re- ceived his undergraduate degree there in 1918. In 1931 he returned to Harvard's Arnold Arboretum as arborist for four years, and then explored the Balkans in 1934-1935 to collect horticulturally promising strains of ivy, boxwood, and holly. He returned to the Garden in 1935 as geneticist and in a few years was also the Engelmann Professor of Botany at Washington University. Moore had high regard for Anderson's abilities, and recogni- tion was to come to Anderson for his innovative research on plant hybridization and evolution. He Volume 78, Number 1 1991 Rudolph 11 100 Years of the Missouri Botanical Garden ~ Edgar Anderson in 1950, a few years before becoming director of the Missouri Botanical Garden. -) FIGURE 9. Edga Photograph by mandi Dorrill Phu radii: (From the archives of the Missouri Botanical Garden described himself as ““blunt and unconventional,” a description with which most agreed. He was an outstanding and original thinker. As Ledyard Steb- bins wrote, “No plant scientist known to me was more able then Edgar Anderson to compare in their natural habitat the leaves, flowers, and wonderfully symmetrical architecture of plants and to express their interrelationships in a meaningful way.’ Although Anderson had become the assistant director in 1951, even earlier he had taken par- ticular interest in the horticultural aspects of the Garden and Arboretum by working with garden clubs and other special interest groups. By tem- perament and desire he was ill-suited to be an administrator. However, he and some of the Trust- ees believed that the Garden needed new direction, so he reluctantly agreed to be director for a while. With Cutler's help he tried to revitalize the Gar- den's programs. Їп the process of trying to make rapid changes, he managed to alienate many of the older staff members. Robert Woodson, who had been the curator of the herbarium since 1948 and coordinator of the Flora of Panama and had pushed the building up of African collections, re- treated to the university and disassociated himself from the herbarium administration. Carroll Dodge was upset by the sale of the fungi collections and the transfer of portions of the library and came less and less to the Garden. А number of senior Garden and Arboretum employees either left or were unhappy with the changes. These changes included moving the orchids back to the Garden with elimination of many duplicates, and the dis- mantling of the greenhouses at the Arboretum. A part of the Garden's land on Magnolia Avenue was given to the National Council of State Garden Clubs to build its headquarters. А strong emphasis was given to involving garden clubs and interested am- ateurs in helping the Garden in practical and mon- etary ways. Physical plant improvements were pushed and a National Science Foundation grant was obtained to improve the herbarium and library facilities.** Fern specialist Rolla M. Tryon, who came in 1948 as the assistant curator of the her- barium and associate professor at Washington Uni- versity, organized a small systematics symposium for faculty and students of midwestern institutions іп 1953.” The following year, this was expanded with the help of a National Science Foundation grant, and the annual symposium became a tra- dition. Progress was slow and attendance at the Garden did not begin to increase until 1956. Because of ill health and frustration with administrative pro- cedures, Anderson soon resigned to become Cu- rator of Useful Plants and spend his time with his beloved research and teaching. He continued to be awarded research grants and recognitions, includ- ing the prestigious Darwin- Wallace Medal from the Linnean Society of London in 1958. The Board was now searching for a decisive leader to be di- rector. Frits Warmolt Went, California Institute of Technology Professor, Director of the Earhart Lab- oratory Phytotron, which he helped design, and eminent botanist, was selected to be director in Annals of the Missouri Botanical Garden KU NN Se ЛО OON d FIGURE 10. Frits Warmolt Went inside the Climatron while director of the Missouri Botanical Garden, about 1969. Photograph by Arthur Fillmore. (From the archives of the Missouri Botanical Garden.) 1958.** At the time he wanted to take a short leave from Cal Tech, and the Garden directorship offered him an opportunity to try out some of his ideas about botanical garden functions. He had impres- sive international experience with botanical gar- dens, having grown up in one in Utrecht, where his father was director and where he did his doc- torate on pioneering studies of plant hormones. He had been associated some years with the Royal Botanical Garden in Bogor, Java, and had con- nections with several West Coast United States gardens. An urgent need in St. Louis was the re- placement of large deteriorating greenhouses, a challenge that especially intrigued him. Went believed that the Garden's prime aim was to develop a public awareness of plants by inno- vative educational displays. By doing this, public support and interest would be assured and would help to develop an increased research program. At first Went had thought that the research should be done at the University, but soon decided that some research should be done at the Garden in a future new building. In the five years of his directorship, most of Went's efforts went into implementing this plan. At the suggestion of the firm of Murphy and Mack- ey, a greenhouse in the form of a Buckminster Fuller geodesic dome was decided upon rather than a traditional greenhouse. From his previous ex- periences, Went knew that many tropical plants needed summer cooling to do well in a greenhouse. Thus, air conditioning as well as heating was needed to provide year-round temperature control. Went named the new greenhouse the **Climatron," be- cause the dome could allow several types of tropical climates by gradient conditions of temperature and moisture. The Climatron, opened in October 1960, sparked much public interest and attracted large paying crowds. (Fig. 10). It served to educate the public on aspects of plant growth under varying conditions, even though the anticipated climatic conditions were not completely controllable. As a building, it received international recognition for its design and concept.“ With the aid of Board President Robert Brook- ings Smith, and the help of the Friends groups, substantial contributions were received to help build the Climatron and to make other improvements. Other innovations were made: several concerts were held; East Indian food was served at a food stand; and special programs in botany for children were instituted. Exotic plants, particularly orchids, were obtained by Went on his many travels. After a while, Went envisioned making the Gar- den the cultural center of St. Louis. It was to be a floral pavilion with places for musical events, theater, art, lectures, and restaurants. The city was considering a new cultural center development and thus the Garden became a contender for the site.” Some Trustees did not believe that this was an appropriate function for the Garden, nor did they think it affordable. Disagreement led to the resignation of the Board president, even as the City Planning Commission put the Garden fourth on its list of favored sites for a cultural center. Eventually frustration led Went to resign in 1963. Earlier he ad begun to obtain funding for a new research building, but much more money was still needed." Went had revitalized the Garden in the eyes of the public. He had drawn attention to its potential as a cultural force in St. Louis. However, he had not been as able an administrator nor as successful in promoting research as he wanted. The Board of Volume 78, Number 1 1991 Rudolph 13 100 Years of the Missouri Botanical Garden Trustees was now suspicious of grandiose plans for future deficit development and sought the assis- tance of outside scientific reviewers for guidance. In 1965 David M. Gates (Fig. 11) became its new director after two years of indecisiveness and deteriorating research at the Garden. Gates, trained as a physicist but interested in nature since child- hood, was in 1965 well known for his original program of applying physics to ecological problems. He had recently become Professor of Natural His- tory at the University of Colorado, a position es- pecially designed for his unique qualifications. Gates was an outstanding scientist who had administrative experience with the American Embassy and the fice of Naval Research in London. He felt that the Garden needed firm direction to overcome the deteriorating horticultural and research programs. He believed that his ecological knowledge would be useful as a research direction for the Garden.” Born a midwesterner, with all of his degrees from the University of Michigan, Gates seemed able to understand St. Louis better than Frits Went had. His first goals were to reestablish credibility with the Board, to promote fiscal responsibility and a balanced budget, to seek new sources of funding, and to rebuild the research and graduate programs at the Garden. With his assistant, Mark Paddock, he was able in his six years of directorship to achieve these interconnected goals. He instituted a general admission charge. He hired new re- searchers and staff. With their help, he was able to obtain grants for support of Garden activities. He initiated a $3 million Capital Fund Drive to help fund a new herbarium and library building, make other improvements, and obtain a develop- ment plan. In his first three years he had increased the total staff by 50%, and the scientific staff by 200%.* Budgets had been balanced each year. Research in systematics had been revitalized by the new Director of the Herbarium, Walter H. Lewis, and the new Curator of Cryptogams, Mar- shall R. Crosby. Grants were obtained for support of herbarium and library improvements and for significant research. A major grant was obtained to help with the construction of a new herbarium and library building, to be named the John 5. Lehmann Building, which was completed in 1972. The herbarium passed its two millionth specimen mark, and the Flora of Panama project was mov- ing ahead with new and old staff, including John D. Dwyer, a professor at St. Louis University, who had been involved since Woodson's time. A new herbarium and library were established in the Pan- ama Canal Zone as well.** Director David Murray Gates measuring the Missouri Botanical Gar- den. Photograph by Claude Johnston. (From the archives of the Missouri Botanical Garden.) FIGURE 11. the temperature of a rose at The Gates administration was to write that “The inertia gripping this institution for the last 30 years as been overcome. We have taken positive steps to fulfill our proper role in this community and in the significant world." ** Optimism was returning to the staff and all seemed on track when Gates left to become director of the University of Michi- gan's Biological Station in 1971. He had built up a large, funded, biophysical ecology program at the Garden and Arboretum. This sort of ecological research was to leave with Gates, at the same time that the revitalized systematics program was to persist and grow when systematist Peter H. Raven came immediately to become director. Raven (Fig. 12) who grew up and went to school in California, was an outstanding young botanist teaching at Stanford University. He had diverse interests in systematics, ethnobotany, ecology, evo- lutionary biology, and general natural history.^ In the 18 years of Raven's directorship, with the help Annals of the Missouri Botanical Garden FicunE 12. Peter Hamilton € shortly after be- coming ie director of the Missouri Botanical Garden (From the archives of the Missouri Botanical Garden ) of his wife Tamra, and for some years of William Klein, the Garden has moved forward on many fronts. Some new focus gardens were designed: an English woodland garden; a new rose garden, and a Japanese garden. These, together with expanded floral displays, added outdoor sculptures, and spe- cial events have quadrupled attendance since 1971, resulting in more than 800,000 visitors a year with a staff of RUE 250 people. The Ridgway Center, a new ilding containing various facilities and the public education programs, opened in 1982. The Lehmann Building was renovated in 1987 to make it more ‚ useful for its growing herbarium and It is 16 area of research, particularly in sys- tematics and tropical ecology, where the Missouri Botanical Garden has taken leadership. With a scientific staff of about 40 scattered throughout the world, and a herbarium of 3.6 million speci- mens, the Garden has accomplished much. It has been able to receive tax support for the first time, which together with an expanded volunteers pro- gram, allows it to do much. It continues to have a significant interest in the flora of Central America, particularly Panama and Nicaragua, and be a ma- jor focus for the Flora Mesoamericana project. It has been investigating the South American floras, particularly those of the northern part. Africa, particularly the rapidly changing flora of Mada- gascar and southern Africa, has been an interest of the Garden for many years and continues to engage researchers. The North American flora proj- ect is now housed at the Garden, and a new flora of China project is developing. Very active research continues on bryophytes and ferns. The Garden is a major center for study of the plant kingdom at the organismal and higher levels. Went, Gates, and Raven have been actively involved in national and international botanical activities. So we come to the end of this brief history of the Missouri Botanical Garden. The eccentric es- sayist-historian Lytton Strachey once wrote that "ignorance is the first requisite of the historian, "^? because the historian must be selective and inter- pretive and must ignore much. Perhaps it would be more accurate to say that the historian is more like a maker of myths, myths that have one or more morals. What then is the mythology of the Missouri Botanical Garden? In my view, one moral is that an institution with a divided administrative structure, Board and Director, can thrive, though at times there may be strong conflicts. Another moral is that a botanical garden, with strong traditions, can be an effective force for good in a community and in society. The Missouri Botanical Garden, for example, has helped to beautify St. Louis, and helped to clean up its air, has educated Missourians about their natural and horticultural heritage, and has helped expand people's view of the importance of vegetation for human survival. It has shown that botanical and horticultural re- search can be made meaningful to the general public. I have reviewed how the founder and six direc- tors, each with differing views of its mission but all concerned with education and research, have mold- ed the Garden. The ordinary citizen and the sci- entific community have been educated in the im- portance of the Garden's research, whether it be through a flora of Missouri or Panama, a new hybrid water lily, an understanding of the genetics of significant crop plants, an experiment with an ecologically designed greenhouse, or a race to know the plants of the disappearing tropical rainforests. The mythology of the Missouri Botanical Garden Volume 78, Number 1 1991 Rudolph 100 Years of the Missouri Botanical Garden is one that is real, perhaps more real than fact or history. NOTES 4. [21] i - со . For example, ee: N. S. B. Gra Henry Shaw manuscripts, M.B.C. archives, box 3 of botanical manuscripts, 1876 or later, beginning * American Trees." . The best source of information about Shaw is the recent biography by William Barnaby Faherty, S.J., Henry Shaw, His Life and Legacies, Univ. Missouri Botanical Garden [la] St. Louis, 1890, pp. 7-28 (Reprinted as pamphlet n.d. but probably late 1930s, 30 pp.); E[dgar] A[nderson], “Henry Shaw—a pic- torial biography," M.B.G. Bull. 42: 103-118, 1954; Edgar Anderson & Dorothy Brockhoff, “Непгу Shaw, a pictorial biography," M.B.G. Bull. 55(6): 1-16, 67; and Jean Marie Denken & James R. Reed, “Henry Shaw, his life and legacy," [unpaged pam- phlet, M.B.C., St. Louis] n.d. (probably 1977). s, “Ап early sedentary merchant in m Middle West, records available for a study of the ud of Henry Shaw," Bull. Business History Soc. 18: 1-9, 1944; Вала О. Jensen, “The enigma of Mr. Shaw, " Missouri Historical Soc. Bull. E 1959, pp. 311-318; James Neal Primm, **Hen- y Shaw, Merchant Capitalist,” EMT Heritage 5(1): 2-9, 1984; and Faherty, 1987. . The literature on the history ^t ы gardens is extensive. A few examples are: Wolfgang Born, “Во- tanical gardens," Ciba Symposia 11: 1093-1 1949; Frans Verdoorn, “Botanical gardens and аг- boretums of the past and their reconstruction, " An- nee Biol. 29: 275-282, 1953; W. B. T Thomas Stern, “Botanical gardens and botanical lit- erature in the eighteenth century,” in Allan Steven- son (compiler), Catalogue of Botanical Books in the Collection of Rachel oe rs Hunt, 21): xli-cxl, 1961; Edward Hyams, ал dens of the World, Macmillan Now York, 1969, 288 pp.; Н. R. Fletcher & W. Н. Brown, The Royal Botanic Garden, Edinburgh, 1670-1970, H.M. Stationery Ofhce, Edinburgh, 1970; C. D. Darling- on, “The Oxford Botanic ign. 1621-1971," Nature 233: 455-456, 1971; W. T. Stern, “Sources Es information about Ioni gardens and herbaria,” Biol. J. Linnean Soc. 3 233, 1971; Wilfrid Blunt, /n For a Penny; а Бо ect ој Kew Gardens: Their Flora and Falballas, Hamish one, The oe don, 1978, 218 pp.; and A. The Los Gardens of Glasgow ети, м Helm. London, 1988, 298 pp. Information aes the garden at Padua can be found in: Davi “Padua American Horticulturist 59(4): 18- 21 1980; and Lucia Rossetti, The University of Padua; an Outline History, Trans. > x M. Hargraves, Edizioni Lint, Trieste, 1983, Я See, for example Ер C. Ritterbush, Overtures to Speculations of Eighteenth bord niv. Press, New Haven, 1964, 7 pp. Lorin Anderson, “Charles Bonnel s tax- — — — N — © onomy and chain of being," J. Hist. Idea 37: 45- 5 . Shaw letter to Hooker, Aug. 10, 1886 (Book 64, # . Some examples are: "Sir William, kind, active and 4 , Kew, Aug. 11, 1857 (M.B.C. archives, Director's Office Scrapbook 1857-1890); “Аз your endowment is quite unprec- edented in this country, if not anywhere, at least i in its future, and that Ї am very confident of its complete success and enduring usefulness." A. Gray to H. Shaw, Cambridge, May 29, 1884 (M.B.C. archives, Director" 5 Scrapbook 1857- 1890); "Mr. Shaw took made will hinder the ds some matters, there will be a grand foundation laid." A. Gra D. Hooker, Cambridge, June 9, 1884, in Jane Loring Pe (ed- itor), Letters of Asa Gray, Boston, 1894, p. 753. . "Another thing that I have at heart is the offer that has been mad e of an excellent and very rich Her- such a treasure—and what of library?" С. Engelmann to H. Shaw 1857- 1890). See also: У | Louis 1: 316, 1856-1860; “Early collections | in the Garden herbarium," M.B.G. Bull. 7: 29-35, 1919; and William G. D'Arcy, Mysteries and treasures in Bernhardi’s herbarium,” M.B.G. Bull. 59(1): 20- 25, l. A photograph of Henry Shaw lying in state 1 M the M.B. useum is in in Gateway pipa a 5: 9, can be made collected by Mr. Shaw for the purpose have suffered to such an extent as to be for the most part worthless, owing to the ravages of insects and the accumulation n th E n andi is она. 16 Annals of the Missouri Botanical Garden t during many years." 2nd Ann. Report M.B.G., Louis, Missouri, Pruett Publishing, Boulder, Colo- rado, 1981, pp. 401-418. 14. The ка published guide to the Garden was that 22. Missouri Botanical Garden, Board of Trustees, St. of souri Botanical Garden, Board of Louis, 1904, p. Trustees, (S. pe is], 9 pp., 11 pls., 1 folding map. 23. 16th Ann. Report M.B.G., 1905, p. ther guides were published as they were revised а: 24. The figure for Shaw’s botanical library c can be ob- - Oo — ~] 18. 19. 20. N = . Much ha later dates. . H. MacAdam, dard Grove Park of the City of ouis, 1883, am], “Tower ve Park. A careful review of its foundation and istory," The Republic, June 26, Meis "Shaw's living ЕМ Victorian pou mix Bull. 648) u. torian Strolling Park, 2nd ойноп бы Норрег Press, St. Louis, 1983, 105 . *Henry Shaw's will. Мати p Richa at St. Louis, Sept. 2, 1889." [ 1st] Ann. Report M.B.G., pp. 29- 55, 1890; this has been reprinted i in pamphlet form Wil of Henry Shaw Establishing the Missouri Botan- ical Garden and Other Documents Pertaining Thereto, M.B.G., St. Louis, 1986, 45 pp. . "An act to enable Henry Shaw to convey or devise to trustees certain lands." (Reprinted from Missouri Sessions | of 1859, p. 434) [1st] Ann. Report M.B.G., pp. 26-28, 1890. Ibid. p. 26. Ibid. pp. 36-37 No definitive biography of Trelease is available. The best published sources are: L. H. Pammel, Prominent Men 1 Have Met, Ш, Dr. William Trelease, the Author, Ames, Iowa, 1927, 84 pp.; J. Christian Bay, William Trelease 1857—1945, des Reminis- cences, the Author, Chicago, 1945, JT. Buchholz, “William aes 1857- Ter i а 101: 192-193, 19 . М. Greenman, “William Trelease," M.B.G. rx 33: 71-72, 1945; Louis Otto Kunkel, “William Trelease, Scientific Biography 13: 4 s Scribner's Sons, New York, 1976. The papers of release are in the archives at Cornell University and t ni- versity of Illinois. An extensive bibliography is г ш Кипкеј, 1945. Тће опју перапуе assessment of Trelease known to me is by Marcus E. Jones: m cus Е. Jones, Western Geologist, Mining En- and and Вион Rancho Santa Ana Botanic Garden, Claremont, 1986, p. 205 (I thank Joseph Ewan for calling my attention to this). s been written about the Fair. Some ex- amples are: The Greatest of Expositions, Completely ыо Official Publication, Official Р and James Neal Primm, Lion of the Valley, St. ~ o о o о — . [J. С. Bay], “The Sturtevant prelinnean library,” . For further information about von Schrenk, see: s). . Board of Trustees Poser Feb. 10 tained from a manuscript “Catalogue of Books be- longing to the Library of the Mo. Bot. Gardens" written by J. Monell probably about 1876, and from the inventory of the Town House and Tower Grove House prepared at Shaw's death, both in the M.B.G. P Board of Trustees” Minutes, Apr. 10, 1912 (M.B.G. archives). 7th Ann. Report M.B.G., 1896, pp. 123-209; C. E. Hutchings, “А supplementary catalogue of the Stur- tevant prelinnean library," 14th Ann. Report M.B.G. 1904, pp. 233-316; Board of Trustees’ Minutes, May 14, 1902 (M.B.C. archive s). . Board of Trustees’ Minutes, Dec. 14, 1904 (M.B.C. archives). Board of Trustees' Minutes, Mar. 9, 1904; Apr. 12, 1905; June 12, 1907; (M.B.G. archives); el Forshey, р. 5, 9; August P. Beilmann, “Herm (1873- 1953)," M.B.G. Bull. 41: 50- 53 1953; and Man James E. Cronin, Herman von Sc : The Who was Timber, a Biography, Кош Chicago, 1960, 257 pp. . have you got a man to spare— or r do you know ci &c things. If you can light on a а ea “~ get a worker — very well. If not I suppose t have a botanical assistant a to do “м 1 have always to do— If we put all in с r, Cambridge, Dec. 27, 1871 (Asa Gra y Kew o 1839-1873, Royal а Саг- dens, Kew, archive 4; May 4 (M.B.C. d таи 16th "1906. p. 29-30. 11, 1904, Nov. Ann. Report MB. c. . Board of Trustees’ Minutes, Nov. 9, 1904; Jan. 11, 1905 (M.B.G. archives). . “The work has been hard and uphill, but I hope that the results are really worth all it has cost me. Th never has been a time when the Board has not been pretty equally divided between concentrating every- m me in considering leaving the Garden if tempted to Rudolph 17 100 Years of the Missouri Botanical Garden Volume 78, Number 1 1991 о ~ о w w "s o ш S а w m w m do so. А new ruler has arisen in Egypt, who knoweth not Joseph, since Mr. Lackland's death, and a y ago the new President advised me that *we're going to have more garden and less science, now' and that I should have to do a good deal of talking to bring him to agree with my suggestion that I would rather have him say ‘more garden but no less science.’ I did not tender my resignation then, for I felt that it would be wrong to abandon my part on such a prop- osition unless it were strongly enough based as to make further struggle hopeless. There is no question, though, now that I’ve opened a jack pot on queens but the President has four acres; but, as I told him, ээ Trelea (Farlow Library and Herbarium, Harvard University, Farlow paper . "Hunting orchids i in the Andes for St. Louis has its trials,” St. L ' M.B.G. Bull. 52(5): 1-12, 1964; Edgar Anderson, “Learning from the bees: G. H. Pring and water-lily breeding,” M.B.G. Bull. 54(10): 8-10, 1966; and how H. Pring 1885-1974," M.B.G. Bull. 63(5): 1-2, 1974. . “Tropical Life friend, no. as Dr. George T. Moore Tropical Life, Sep. 1931, pp. 170-171; Elder] A[nderson], “Dr. George T. Moore, Director 1913- 1953," M.B.G. Bull. 44: 141-142, 1956; and Car- roll W. Dodge, “George Thomas Moore (1871- ae в = Philos. Soc. Yearbook 1957: 145- 148, Board yes LS Minutes, Mar. 13, 1912 (M.B.G. archives). Board of Trustees Minutes, Jan. 13, 1915 (M.B.C. archives). Board of Trustees Minutes, Feb. 13, 1918; Oct. 17, 1928 ( .G. archives). Missouri Botanical Garden Illustrated Board of our | Са ш, St. а flowers of Missouri," St. Louis Globe- и Арг. 22, 1940; Cora Steyermark, Behind the Scenes, Missouri Во- tanical Garden, St. Louis, 1984, 1 pp; Yatskievych & Lu ther J. Raechal, “In memoriam— Julian Alfred Steyermark, то. М.В.С. Bull. 77(1): 5, 1989; E A. Mor i m 1988), Bull. Torrey. Pis Club 116: 75-76, 1989; and [Julian Steyermark], “Steyermark Recollec- tions," Annals M.B.G. 76: 627-651, 39. Board of Trustees’ Minutes, Mar. 17, 1926 (M. В.С. 4 . Marilyn М. LeDou P . “The Climatron opens to TJ archives); and John D. Dwyer, “The history of bugs collecting in Panama," in William С. Mireya D. Correo A. (editors), The Botany ud Nat. ural History of Panama: La Botanica e Historia Natural de Panama, Monogr. Syst. Bot. Missouri Bot. Gard. 10: т 183, 1985. папі из orchid collection” Amer. Orchid Soc. Bull. 58: 553-558, 1989; and Marilyn M. LeDoux, “Orchids through the years at the Missouri Botanical Garden,” Amer. Orchid Soc. Bull. 58: 696-702, 1989 nday Magazine, Aug. 17, 1930; and Board of Trustees’ Minutes, June 5, 1940 (M.B.G. archives ). . Board of Trustees’ Minutes, Jan. 18, 1933 (M.B.G. archive 8). . М.В.С. Bull. 44: 142, 1956. . "Men of science in St. Louis, world famed botanist Louis 305, 1969; and John J. Finan, “Edgar Anderson 1897-1969," Annals M.B.G. 59: 325-345, 1972. Charles B. Heiser, Jr., "Student days with Edgar .G. 59: 362- Te G. мүм Stebbins, “Edgar Anderson; | f a long friendship," Annals M.B.G. 59: 373-379; and interviews with William L. Brown, M. Donald Duvick, and G. Led- yard Stebbins, 1989 (tapes in M.B.G. archives). archives . Interview with Hugh Cutler, 1988 (tape in M.B.G. ) . Interview with Rolla and Alice Tryon, 1989 (tape in . archives). . *Frits Went, California botanist, named Shaw's Gar- den director," St. Louis Post-Dispatch, May 18, 1958; transcripts from Went's diaries of the period and an interview of 1988 are on file in the M.B.C. archives the public," M.B.G. Bull. 48: 131-133, 1960; “And suddenly last summer," Progressive Architecture, Apr. 1961, pp. 174-178; George McCue, “The ar a architecture, St. Louis Post. Dispatch, Арг. 30, 1961, p. 58; “The Climatron,”” mer.] obi [of] A [rchitects] с Мау 1961, рр. 27-32; Јадић Jenkins, “А g r Shaw 5 Garden, Director Went fut ” St. Louis Globe- Democrat, Nov. 24, 1962; “and Robert Dingwall & Barbara Lawton, “‘The Climatron, Missouri Botanical Garden's о age greenhouse," M.B.G. Bull. 58(4): 12-21, EN Ti Я бе к А music at Shaw's Garden," St. Louis Post-Dispatch, Sep. 23, 1959; “Where to put a cultural center," St. Louis Post-Dispatch, Jan. 8, 1963; "Downtown gets preference for culture cen- ter," St. Louis Post-Dispatch, Jan. 20, 1963; and *"The cultural ger reports," St. Louis Post-Dis- patch, Jan. 27, 1963. ‚ “Missouri Botanical Garden, report Frits W. Went of the director dor 1962," M.B.G. Bull. 51(3): 1- Annals of the Missouri Botanical Garden сл N сл e сл > сл сл сл с 8, 1963; Frits Went, diary Арг. 1, 1963 and other Perry Lawton, **New director of Missouri Botanical diary entries (M.B.G. archives). Garden," M.B.G. Bull. 59(5): 6-9, 1971; and Peter Edgar Anderson, “David M. Gates, the new direc- Raven, “From the director," M.B.G. Bull. 61(6): tor," M.B.G. Bull. 53(9): 1-7, 1965; Robert I 1-3, 19 Stock, * ‘Saving the world, the ecologist’s way," 57. “John S. Lehmann Building, p to meet the fu- York Times Magazine, Oct. 5, 1969; and быш ture," M.B.G. Bull. 75(1): 3-4, interview 1988 (M.B.G. archives). 58. William Barnaby Faherty, S.J., The Missouri Bo- Mark Paddock, “Foreword,” MB. G. Bull. 57(1): tanical Garden through 125 years,’ о ау Нег 1, 1969; and Mark Paddock, “Annual report — паре 5(1): 10-19, 1984; Peter H. Raven, "The 1969, a significant decade," M.B.G. Bull. 58(2): 2- Missouri Botanical Garden in the world's s service,” 4, 1970. Gateway Heritage 5(1): 40-48; The Unseen Gar- | се B. Croat, “The Missouri Botanical Garden den, Research at the Missouri Botanical Garden, n Panama," M.B.G. Bull. 60(1): 19-22, 1972. Missouri Botanical Garden, St. Louis, 1987, 44 pp.; А Paddock 1969; and “The John S. Lehmann Building, and interviews with Tamra Raven, 1988 and Peter a new era for the Missouri Botanical Garden," M.B.G. Raven 1989 (tapes in M.B.G. archives). Bull. 60(3): 6-12, 1972 59. a Strachey, Eminent Victorians, G. P. Put- . David Gates, “Welcome to Dr. Raven from Dr. m’s Sons, New York, 1928, Gates,” M.B.G. Bull. 5965). 3-5, 1971; Вагђага ONE HUNDRED YEARS OF PLANT TAXONOMY, 1889-1989! Barbara G. Briggs? ABSTRACT One hundred years ago taxonomic botany was at the end of the first phase of ps great regional floras and porem uy re especially with the work of George Bentham and Joseph Hooker in Britain, Asa Gray in the U.S.A., Alphonse de Candolle in Switzerland. The vast influx of new botanical discoveries from exploration had highlighted interest in biogeo, ogr Fai interpreted in terms of new evolutionary insights. The subject had been in high regard and major herbaria had recently been founded. A widespread decline of activity and prestige followed, when emphasis shifted to other botanical fields, except in Germany where Adolf Engler further developed encyclopedic coverage of plant families and genera worldwide. Revival of biological taxonomy came slowly through genetics to the “new systematics.” Karyology, inis ac phytochemistry, and comparison of isozymes were in turn seen as panaceas; each fai led to meet thes ntinue f the armory of relevant approaches. Emphasi ade of numerical phenetics. Partly from lack of a satisfactory theoretical basis, phenetics came to be widely d but gave place to cladistics, ranging from the avowedly phylogenetic to abstract pattern analysis. produce о phylogenetic trees without input from botany remains “ап unending synthesis," aspects is gene боб апа from the so that the new developments are a scientific stimulus to the subject a institutions concerned with floristic projects remain within molecular systematics lives up to its early interpretation of morph Macromole promise, the tas sy ology and evolution in terms of highly cian phylogenies—a synthesi onomic levels by selection of appropriate gene systems. Such data may morphology, way the data are analyzed. A challenge is to maintain cohesion in systematics as a whole and so that herbaria and other the mainstream of scientific taxonomy. Insofar as macro- matics may eventually become, in part, the sis on new terms. As well as these more theoretical approaches, po organized information on the diversity of the world's plants nty. through national and regional floras remains a prio A brief paper cannot adequately review a cen- tury of work by taxonomists, who were numbered in tens a century ago and who are now in thousands. Their work filled our herbaria with specimens and our libraries with publications. This whirlwind tour places emphasis on the beginning and end of our time frame, and gives least attention to those fields where ту lack of expertise would otherwise be most conspicuous. Two other limitations will be apparent. The extent of emphasis on work in the English-speaking countries and on seed plants alone is justified by my greater familiarity with those areas. AN UNENDING SYNTHESIS? With this eloquent and much-quoted phrase Lin- coln Constance (1964) characterized the nature of systematics—a term that I use in part inter- changeably with taxonomy (in general agreement with Small [1989] in the use of these terms). Mod- ern macromolecular genetics has contributed such powerful tools to taxonomy that Gottlieb (1988) queried whether this characterization was still ap- propriate, and I will return later to this question. Traditionally, data from many fields have been grist for the mill of taxonomy. So this subject has reason to be responsive to developments in science gen erally, as well as profiting from technological ad- vances that increase its power to assemble an organize relevant data. This responsiveness has en and remains a two-way process. Taxonomy has given the basis for one of the greatest unifying concepts in science—-evolution— since it is tax- onomy that has the main task of making accessible our organized knowledge of the patterns of diversity of living things. A hundred years ago the scientific context for ' I thank my colleagues Lawrie Johnson, Peter Weston, Peter Wilson, and Bob Makinson who read the manuscript and made helpful comments. Preparation of the paper was much aided by Sue McCahon, е typed the manuscript, and by Anna Hallett, Irma Trunov, and Bob Makinson who located and compiled referenc ? National Herbarium of New South Wales, Royal Botanic Gardens, Sydney, Australia 2000. ANN. MISSOURI Bor. GARD. 78: 19-32. 1991. 20 Annals of the Missouri Botanical Garden taxonomy was the result of a century of dramatic progress (e.g., Taton, 1965). Since the start of the nineteenth century there had been: —the acceptance of atomic theory with insight into molecular structure and chemical reac- tivity; —bacteriology with the inception of scientific medicine; —the interpretation of geological structures, with the international geological scale of epochs and periods replacing models of the earth’s past; —the beginning of understanding of the struc- ture of organisms in terms of cells, nuclei, cell "catastrophic" division, and life cycles; — and then, most dramatic in modifying general thought, had been the concept of biological evolution. Sull lacking were: —extension of the limits of knowledge of the universe in its largest extent that would come with modern astronomy and cosmology and in its fine detail extending from modern mi- croscopy to quantum physics — modern philosophy of science with sharper focus on falsifiability of hypotheses; — systems theory and now chaos theory; — sophisticated concepts of energy, matter, and their range of manifestations and transfor- mations in physical and biological systems; — plate tectonics and the history of continental movements and paleoclimatology; —the localization of genetic inheritance largely to chromosomes, with concepts of genes and mechanisms of heredity; — classification and definition of the major cat- egories of organisms beyond simplistic group- ings into plants, animals, and fungi; —the symbiotic theory of the eukaryotic cell; —technical advances in all fields, including chemical analysis, microscopy, computing, and communication; — extension of centers of scientific research and teaching beyond the limits of Europe and North America. END OF AN ERA A century ago Charles Darwin had been dead seven years and the appearance of The Origin of Species was 30 years into the past. What scientist then would have thought that evolution (as distinct from ‘Darwinism’) would still be a contentious issue, still needing to be defended in some circles although not within science, a hundred years into the future? Within science it is evidence of Darwin's impor- tance that his work still attracts critics today, but mostly their concerns are with the nature of the evolutionary process and the way it is described, rather than denying evolution itself. In the previous century, before 1889, exploring parties in many regions had been gathering spec- imens of unfamiliar plants, still bringing them to the seats of learning in the temperate northern hemisphere, rather than expecting or seeking to establish expertise distributed worldwide. This pace of botanical exploration and interest in joining ex- peditions continued to be powerful stimuli in at- tracting active minds to taxonomic biology, as they had been in the time of Charles Darwin and earlier. By 1889 taxonomy was at the end of the first round of great regional floras and encyclopedic treatments of the world's plant taxa. George Ben- tham had died in 1884 and Asa Gray in 1888, although Alphonse de Candolle was still alive and Joseph Hooker would continue publishing for an- other 22 years till 1911. The importance of these botanists and their work is such that we should extend our time frame back a little to consider them all. Indeed, by going back to September 1880 we could have looked in on any of several little dinner gatherings in the director's house at Kew Gardens near London and found all four of them together (J. Gray, 1893). These were all old friends, old in years and who had met many times before, but think of what a prodigious botanical output is represented here. First the host to these gatherings, Joseph Hook- years director of the Royal Botanic Gardens, Kew, London and spent a long active retirement, his last botanical paper be- ing published in the year of his death at the age of 94 (Allan, 1967; Bower, 1919). He had made notable travels in the southern continents, in India, and high into the Himalayas. The Flora of British India and Genera Plantarum were joint with George Bentham, who took the greater share, but major essays on plant geography were his alone. The incisive, wide-ranging phytogeographical assess- ments in his discussion of southern hemisphere plant distributions (Hooker, 1859) were only set aside when modern plate tectonics gave a new framework for interpreting plant geography. Jo- seph Hooker had played a major part in encour- aging Charles Darwin and had been Darwin's clos- est confidant and first supporter. Next there is Bentham, known to his associates as a prodigious worker (Jackson, 1906), who had Моште 78, Митбег 1 1991 Briggs 21 100 Years of Plant Taxonomy published a wealth of critical treatments of partic- ular families as well as the Flora Hongkongensis and the Handbook of the British Flora. His Flora Australiensis had been 16 years in preparation, but during those years he had been pressing on also with the mighty Genera Plantarum, jointly with Hooker. His botanical judgments have stood well, but subsequent authors have too often simply used Bentham's classifications and ignored his com- ments that indicate how often decisions were ten- tative, as seen by himself, and influenced by the limited information available at the time. Alphonse de Candolle brings yet another strand. He had succeeded his father, Augustin-Pyramus de Candolle (A. P. de Candolle), as professor and di- rector of the botanic garden at Geneva, and also inherited responsibility for the Prodromus (A. P. de Candolle, 1824 et seq.). This work attempted a systematic treatment of the world's dicotyledons down to the level of species and variety. The elder de Candolle had turned away from the artificiality of the Linnean system and attempted to apply the modern principles of his time, and his son had reinforced this trend. The Prodromus set standards of format and nomenclature that became increas- ingly widely adopted. After taking over the half- completed work, with only seven of the eventual 17 volumes published, Alphonse made the work more international and better-based through seek- ing external collaborators, by visiting other her- baria, and through extensive loans of specimens. His publication La Phytographie became a stan- dard work on the principles of plant taxonomy, and he had a leading part in the first codification of botanical nomenclature (A. de Candolle, 1867). Then there was Ава Gray. He had collaborated with John Torrey on Тће Flora of North America (Torrey & Gray, 1838-1843) and later had pub- lished widely on the plants and phytogeography of North America and also their relations with the flora of Japan. A notable teacher, he had produced a range of important textbooks, inspired a stream of outstanding botanists, and founded the Harvard herbarium that bears his name. Himself a system- atist, he also encouraged students into other dis- ciplines and was conscious of the value and im- portance of morphology, anatomy, and physiology. А TIME ОЕ CHANGE, А CENTURY АСО BOTANY AS A SCIENTIFIC CAREER The careers and work of these four botanists emphasized changes that were already in evidence and that would set the scene for the future. Botany in the nineteenth century had been both burdened and blessed by being adopted as a socially desirable genteel pursuit for the newly wealthy, and one that combined well with the romantic idealism of the time (Scourse, 1983). This fashion grew as a cor- ollary to increasing industrialization and urbaniza- tion and gave the subject some problem in estab- lishing its scientific credentials. But for our group of botanists there was no doubt: their motivation and careers had been scientific. Hooker and Gray had originally studied medi- cine, thus receiving some formal botanical training; Bentham and de Candolle had taken law. Two had learned much of their botany assisting distinguished fathers, whereas Gray had worked with John Tor- rey at New York. We see here two types of bo- tanical inheritance. Gray received from Torrey inspiration and teacher and student was replacing that of the mas- ter with his assistant. It was now the time of the professional scientific career. Bentham seemed to regard himself as something of an amateur even when he had major achievements to his name, in contrast to Dr. Hooker and others formally trained in botany. Also, Bentham alone of this group had continued the earlier tradition of the unpaid schol- ar, having a modest inheritance and “findin that I need not toil for our support [he observes ], I determined to give up the law and devote myself entirely to botany" (Jackson, 1906). THE GROWTH OF COLLECTIONS AND INSTITUTIONS When William Hooker was appointed as director of Kew Gardens in 1841 it had neither herbarium nor library and was essentially a pleasure garden (Hepper, 1982). He left it as the scientific insti- tution that his son Joseph continued to develop. The original nucleus of the Kew herbarium was William Hooker’s personal collection, purchased on his death by the British government (Blunt, 1978; Desmond, 1975). Its library also started with William Hooker’s own books, much aug- mented when Bentham offered his library of 1,000 volumes; Bentham made only the condition that he be allowed to use it there (King, 1985). That most useful, if simplest, of resources for the taxonomic botanist—the herbarium speci- men—already had a long history, but large her- baria were recent developments. The foundation of London’s Kew herbarium is dated as 1853, and collections were building up in St. Louis, at Har- vard University, and elsewhere. But many other 22 Annals of the Missouri Botanical Garden major herbaria, for example in New York and Chicago, were still in the future. Much of the early contact among botanists arose through requests to collect or identify specimens, as for example Asa Gray's early dealings with John Torrey and later with George Engelmann (J. Gray, 1893; Hum- phrey, 1961). Now that there are large herbaria of worldwide scope widely distributed, it is hard to equal the enthusiasm with which these botanists awaited and received new consignments of speci- mens from distant lands. None of these four great men had successors of comparable standing. Geneva continued as an im- portant botanical center but never again rivaled its place in the time of the de Candolles. Similarly, after the death of Asa Gray, no individual has dominated the American botanical scene as he had (Ewan, 1969; Heiser, 1969), although important successors included Charles Bessey, notable for his attempt at a phylogenetic system of flowering plant classification, and Liberty Hyde Bailey, with his unrivaled works on cultivated plants. Under the two generations of Hookers, Kew developed exchange, especially in plants of poten- tial economic value, with the British colonies, and Kew directors determined the control of colonial botanical gardens. Under its next director, This- tleton-Dyer, with the British Empire at its peak, economic botany came to dominate much of Kew's work and influence (Brockway, 1979). Kew con- tinued as an outstanding center of botanical re- sources, but no longer held the lead in botanical research. DEVELOPMENT OF BOTANICAL PROCEDURES AND COOPERATION Standardized formats and procedures for plant description, specimen citation, and reference had developed early. Works from this time therefore have formats relatively little different from many modern works. Indeed, some standard conventions in botany seem to be vestigial and probably re- grettable survivors from this age when the subject was in the hands of a small group of botanists well known to one another. This applies to some of the conventions for the citation of botanical authorities and brief reference to publications in taxonomic treatments. In the formalization of botanical pre- sentation, both descriptive and nomenclatural, all these leading botanists had set standards that were followed long into the future (Frodin, 1984). Our group of notables had also orchestrated much of the labors of a farflung band of associates who strove to assist, in return for encouragement and approval. Among them had been Ferdinand von Mueller in Melbourne (Willis, 1949), William Harvey in Cape Town (Ducker, 1988), and En- gelmann in St. Louis. All were productive in their own right, but also contributed importantly to joint works or resources available to Bentham, Hooker, or Gray. In the Prodromus, de Candolle had co- ordinated the work of 35 botanists in eight Euro- pean countries (Stafleu, 1966). A comparable de- gree of wide cooperation on international projects is only now being rekindled and is made easier by such organizations as the International Asso- ciation of Plant Taxonomists (Stafleu, 1988), as well as by modern technology. EVOLUTION Each of these botanists had taken up the new evolutionary thinking. Hooker, Gray, and de Can- dolle had used it as a basis for discussions of phy- togeography and had been especially active in pro- moting its acceptance. Yet the system adopted in Bentham and Hooker’s Genera Plantarum was modeled closely on that of the pre-Darwinian Au- gustin-Pyramus de Candolle. As Stevens (1984) has pointed out, the incorporation of the evolu- tionary time-axis and concepts of divergent evo- lution had remarkably little effect on the classifi- cations they produced, which were often reminiscent of earlier ideas of a network of living organisms. THE BERLIN SCHOOL The next major botanical phase was not a direct legacy from any of our four botanists. The focus shifted to Germany, where the German universities had been preeminent for some decades in research on plant structure, developmental morphology, and physiology, as well as providing training to many students from other countries including the United States. This continued with rigorous morphological interpretation, static in some aspects, but with valu- able precision of analysis. Adolf Engler had been appointed in 1889 as professor and director of the botanical garden in Berlin. There, and previously at Breslau (now Wroclaw) and in his general ap- proach, Engler followed in the steps of his mentor August Eichler, but with vastly expanded aims (Stapf, 1930). Engler’s name “‘symbolizes the spec- tacular development of German botanical system- atics in the late nineteenth and early twentieth century” (Stafleu, 1972), and his Syllabus (1892) has been followed by much advanced and very useful further editions up to recent times (Melchior, 1964). The Pflanzenfamilien (Engler & Prantl, 1887-1915) and Pflanzenreich (Engler, 1900- Volume 78, Number 1 1991 Briggs 23 100 Years of Plant Taxonomy 1953) have been the last of the great encyclopedic treatments of the plant kingdom. These projects, envisioned by Engler and assisted by Prantl and many others, continued for decades on much the original plan. y when the current project of Kubitzki and others (Bates et al., 1980) comes to fruition will their status be challenged. Engler's attempt to arrange plant orders and families in what he conceived to be phylogenetically determined sequences was only partly successful since it did not recognize the extent of reduction trends that produced structurally similar forms. DECLINE AND RISE Encyclopedic projects such as Engler's, however vast their scope and achievement, were assembling existing knowledge rather than moving ahead in scientific concepts or methodology. By now the rush of exotic discoveries had lost impact and the physical sciences and biochemistry were on the rise. In both botany and zoology, taxonomy was in decline before the end of the nineteenth century and was seen as a scientific backwater, unattractive to the best minds. Botanical taxonomy was contin- uing the task of cataloging and naming the world's plants, with more emphasis than before on the cryptogamic groups. Áround the world the general rate at which herbaria were founded (Ma, 1988) was increasing almost exponentially, but this was not an indication of the scientific standing of her- baria. Rollins (1965) expressed it well: near the end of last century “herbaria were respected centres of teaching and research. The most eminent men of botany were associated with them.” By contrast, over some decades up to the middle of the present century, “Having established its place in the Uni- versity and in the Museum, the herbarium attracted less venturesome scholars, who narrowed the di- mension of their activities ... the image became one that reflected stultification, diffidence and ri- gidity without sinews to bind it to the main stream of biological teaching and research." Taxonomy would be revitalized a few decades later from a source that was already at hand early in the present century—the new field of genetics. Hull (1988) recounts how several workers around 1900 independently experimenting on character segregation came upon Gregor Mendel's publications. Soon genetics was a flourishing ex- perimental and theoretical field, but at first genetics was seen to be incompatible with and opposed to Darwinian evolution. The reconciliation and fusion of genetics and evolution came when Fisher, Hal- dane, and Wright began to develop concepts of the structure and dynamics of populations, and Darlington (1937) expressed the relevance of ge- netics to plant evolution. The integration of genetics with evolutionary biology at the taxon level came later with the “new systematics," which was led by zoologists Dobzhan- sky, Simpson, Mayr, and Huxley, but which af- fected all systematic biology. NEW DATA SOURCES OR PANACEAS? The rediscovery of Mendel's work was followed by the identification of chromosomes as genetic material and the realization that chromosome mor- phology offered characters for study. oreover, karyology sometimes suggested hy- potheses about processes and directions of evolu- tionary change, although these were often hedged with statements that certain directions were *'un- likely" rather than **excluded from consideration." Early work was summarized in Variation and Evo- lution in Plants by Stebbins (1950), which helped to inspire a generation of research students of my time, as Darlington (1937, 1939) had done for our mentors in this field. Applications of karyology were numerous at all levels, from infraspecific to con- siderations of base numbers and major trends for orders and families and for the angiosperms them- selves (Raven, 1975). The extent and patterns of chromosomal variation led students of phylogeny to mistrust or downgrade the importance of chro- mosomal information for broad considerations (Raven, 1975), but such information has thrown important light on relationships and processes in many groups (Moore, 8) Along with chromosomal studies in the early decades of this century went biosystematics. This field seems to have been variously defined, certainly embracing studies of variation, population struc- ture, and breeding systems, but at other times tak- ing in karyology, cytogenetics, and the whole of genic molecular evolution (Grant, 1984; Vickery, 1984). In line with biosystematic thinking, tax- onomists for decades sought justification of their categories from biological species concepts. Ehrlich & Raven (1969), Raven (1977, 1980), and others have shown that much that has been claimed about species as interbreeding populations does not ac- cord with the situation in nature. They pointed out how limited the gene flow is between the distant local populations of a widespread species, but that members of such a species are held together by common ecological constraints. It is scarcely an elegant solution that the term “species” is now 24 Annals of the Missouri Botanical Garden concurrently applied to various different concepts, t at least some consensus seems to have been reached in botany (e.g., Davis & Heywood, 1963; Raven, 1976, 1986; V. Grant, 1981; W. Grant, 1984; Haufler, 1989) so that semantics no longer dominates over observations on the situation in nature. In general, biosystematics, at least as nar- rowly defined, seems to have contributed less to taxonomy than to understanding evolutionary pro- cesses and part of its field has been claimed by its vigorous offspring, pollination ecology. It remains a useful approach in its own right but somewhat separate from the main thrust of taxonomy. Alston & Turner (1963) characterized the first half of this century as the cytogenetical phase of systematic biology, whereas they designated 1950 onward as the biochemical phase. Twenty years later Harborne & Turner (1984) made a more modest claim, stating that they did “not believe that biochemistry represents a panacea for all sys- tematic problems" and indeed that the role of sys- tematics as a framework for the ordered arrange- ment of observations applies as well for biochemical data as for morphology. Secondary metabolites have proved their use, and compendiums such as Heg- nauer's (1962-1973) have a wealth of data, es- pecially for studies at higher taxonomic levels. In- deed, the strength of Dahlgren's system of the flowering plants (1980) lies mostly in placing great- er emphasis on chemical characters than do other modern systems. So far, the chemical data have generally supplemented morphological data rather than replacing them. The first biochemical phase of taxonomy seems to have largely bypassed the majority of taxonomic practitioners until now. Re- maining oblivious to the macromolecular phase will not be so easy. In respect to systematics, both cytogenetics as envisaged early this century and chemosystematics failed to fulfill their early promises, but they have found their place as valuable aspects of multidis- ciplinary taxonomy. Perhaps one of the surprises has been the resurgence of morphology. The basis of early classifications was exomor- phology, but the range of relevant structural as- pects has been greatly widened as means of ob- servation improved and as concepts of homology have been more strictly applied. Thus that excellent symposium “The Bases of Angiosperm Phylogeny” (Walker et al., 1975) included structural topics ranging from floral or vegetative anatomy and mor- phology to palynology, embryology, and ultrastruc- ture. The new findings have been especially re- warding at the upper taxonomic levels, as with structure of leaf teeth (Hickey & Wolf, 1975) or the ultrastructure of sieve-element plastids (Behnke, 1975). A related advance has been the integration of critical morphology with cladistic and biogeograph- ic approaches into botanical palentology (e. Crepet & Nixon, 1989). A better explored foil record, with more precise morphological observa- tion and comparison, is throwing more light on the interpretation of living groups (e.g., Dahlgren, 1983) EMPHASIS ON ANALYSIS, PHENETIC AND CLADISTIC Many biologists were well pleased to be working with these expanding fields of data but others looked to new possibilities of making systematics more objective through attention to the philosophy and practice of data analysis. The early converts to these numerical methods seemed to find their sat- isfaction more in methodology than in improving knowledge and classifications of organisms. In retrospect, numerical phenetics based on exo- morphology seems something of an aberration and its theoretical basis (or lack of it) was much criti- cized at the time (e.g., Mayr, 1965; Johnson, 1968, 1970). But it held centerstage in both zoological and botanical taxonomy during the 1960s into the 1970s, and the major expositions of this approach were much followed (Davis & Heywood, 1963; Sokal & Sneath, 1965; Sneath & Sokal, 1973). Hull (1988: 130) gives graphic expression to the next turning point: “In the hubbub over numerical taxonomy, hardly anyone noticed scurrying around in the о the ancestors of the next dominant group іп system- atics, as inconspicuous and active as the progenitors of the mammals had been in the age of the Dino- 99 saurs. Cladistics as an established approach dates from the "discovery" by Kiriakoff and others of the work of Willi Hennig (Hennig, 1950, 1966), al- though numerical methods to elucidate phyloge- netic lineages based on synapomorphies were in- dependently developed around that time or soon after by others (e.g., Wagner, 1961; Camin & Sokal, 1965; Johnson & Briggs, 1975). In his comprehensive account of the development of cla- distics, Hull aptly describes Henning's contribution: "Тће methods ће described for reconstructing phy- his principles and the emphasis he placed on them." Volume 78, Number 1 1991 Briggs 100 Years of Plant Taxonomy Since then, cladistics has presented an excep- tionally argumentative scene as it challenged the survivors of the “New Systematics"— great phy- logenists like Mayr and Simpson. Cladistics has sharpened taxonomic thinking and brought em- phasis back to the evolutionary basis of nature's hierarchical patterns (e.g., redge & Cracraft, 1980; Wiley, 1981; Nelson & Platnick, 1981; Farris, 1983; and the greater part of the contents of Systematic Zoology for more than a decade). Even so, many cladists (pattern cladists) sought to make their analyses as theory-free as possible, to separate the analysis from any theoretical expla- nation that might be used to interpret it. Guiding principles have been monophyly or hol. ophyly, use of outgroups or hypothetical ancestors, and parsimony. Similarity in plesiomorphous fea- tures is dismissed as irrelevant, and grouping into clades depends on synapomorphies alone. Some practitioners seem to forget that evolution is not necessarily parsimonious but rather that the choice of the shortest tree, with fewest steps from the mon ancestor to the member taxa, depends on a philosophical approach that excludes uncorrobo- rated suppositions and unnecessary steps not de- creed by the data. The evaluation of character homology has led to much desirable reinterpreta- tion of the morphological and developmental equiv- alence of features. Methods differ in whether they accept trend reversals or work under a restricted parsimony model. Important concerns are falsifi- ability of hypotheses, and the writings of Karl Pop- per and other philosophers are much quoted. Cladists have exaggerated the extent of inno- vation in their approach (Ghiselin, 1984) and have often been overconfident about their results. Mis- leading and ill-based conclusions can come from misinterpretation of homology, lack of commen- surability or balance in the characters selected, extensive parallelism, or from methods that do not show the full extent of competing trees of equal or near-equal length. Such problems are now being addressed by more critical assessments, improved methods (e.g., PAUP, Swofford, 1985), centration on with consensus trees and ““bootstrap” or procedures. Practical difficulties arise from paraphyly (Cron- quist, 1987; Johnson, 1989) and bear considera- tion. When paraphyly is well corroborated it 15 often appropriate to accept a new classification based on cladistic analysis. But, for instance, should one fail to give taxonomic recognition to a well- marked group (one that is itself holophyletic) on the grounds that it leaves a paraphyletic residue and con- “robustness” that is not yet resolved? Perhaps classifications may include both monophyletic and paraphyletic groups, with those that are not monophyletic clear- ly identified as such (as suggested by Wiley, 1981). Cladistics has been justified on the grounds that “truth emerges more readily from error than from confusion. Perhaps [Hennig’s] phylogenetic sys- tematics might prove to be mistaken, but no one was going to be able to accuse it of being confused" (Hull, 1988), and cladistics has been tagged *'cla- distics: being wrong with confidence" (Penny et al., 1990); these judgments carry a valid warning but not a reason to reject cladistics. It has proven its power, and almost any issue of a modern sys- tematic journal will show it as the method of pref- erence for intensive investigations of plant groups at all levels. Noncladistic similarity methods are, however, widely but not unanimously advocated for analyzing macromolecular sequence data, as mentioned below. l attention on another ladistics also problem: if a genealogical tree has been established, how should the categories of the taxonomic hier- archy be assigned to its branches? Major categories tend to be claimed by the divergence of groups at the base of the tree, leaving minor categories to assign to the most diverse and prominent groups. Thus, in the system of Bremer (1985) the flowering plants form subclass Magnoliidae within the Chlo- robionta, whereas the Anthocerotatae are a super- class. This has logic, concentrates awareness on important new understanding, and is probably the way of the future. There is, however, some loss in communication as well as some gain: people are interested in "grade" aspects of organisms and their adaptive syndromes, as well as in their ge- nealogies. Terms like ‘amphibians’ and ‘reptiles’ have conveyed information on levels of organiza- tion, even if we now find that they refer to para- phyletic assemblages. MACROMOLECULAR BIOLOGY Cladistics swept through systematics as some- thing of a revolution little more than ten years ago. Already a new turning point is at hand—the ap- plication of macromolecular genetics to systemat- cs. Isozymes, the first manifestation of this ap- T. proach, made their appearance in taxonomy in the 1960s. Some early workers on isozymes have since led the charge into nucleic acid studies, but iso- zymes continue to have a place, especially in work at the lower taxonomic levels including investiga- Annals of the Missouri Botanical Garden tions of population variation (Crawford, 1983; Giannasi & Crawford, 1986; Ferguson, 1988; Brown, 1990). Isozymes sample only nuclear genes that are expressed as enzymic or structural proteins (Clegg et al., 1984), and exclude genes from other organelles. Next came amino-acid sequencing of proteins, which has been described (Giannasi & Crawford, 1986) as “ап approach whose time has come and gone (with regard to p/ant phylogeny) during the past two decades." Such sequencing has contrib- uted importantly to phylogenetics of flowering plants at higher taxonomic levels (e.g., Boulter, 1973; Martin & Dowd, 1990). But proteins show many parallel base substitutions and this gives rise to distortions in their use to interpret phylogeny (Boul- ter, 1980). Present emphasis is on base-sequencing of ge- nomes. Genomes offer a large number of potential characters for study (Ritland & Clegg, 1987). They also have advantages of universality exceeding any morphological characters and approached only by a few cytological features. Some gene systems have been so conservative that all living things from the level of bacteria include homologous genes (e.g., those coding for hemoglobins, Appleby et al., 1990). Other gene systems have been quite labile and show differences between allied species or between in- dividuals within a population (Clegg et al., 1984; Jorgensen & Cluster, 1988). Noncoding regions may provide data complementary to those from protein-coding regions since they are subject to different functional constraints (Zurawski & Clegg, 1987; Clegg, 1989 The technical difficulties in obtaining data are being overcome, so that information is available from chromosomal, mitochondrial, and chloroplast genomes. Chloroplast DNA analysis provides data with exceptionally low levels of homoplasy, relevant to the time frame of the evolution of significant groups of flowering plants, and that lead to robust trees. Mitochondrial DNA is more variable in struc- ture and organization in plants than is chloroplast DNA, and recombinant types arise in hybrids, whereas chloroplasts are strictly maternally inher- ited in most flowering plant groups. Ribosomal RNA contains both highly conserved and variable regions and so is useful over a wide time range (Baverstock & Johnson, 1990). А notable example of informative restriction site analysis of chloroplast DNA is in the Asteraceae. The distribution of a large inversion suggests that the initial dichotomy was the separation of subtribe Barnadesiinae (Mutisieae) from the rest of Aster- aceae, including from the three other subtribes of the Mutisieae (Jansen & Palmer, 1987, 1988; Bremer, 1987; Palmer et al., 1988) The pace has accelerated in the last few years with important data coming from Fabaceae (Doyle, 1987; Doyle et al., 1990), Onagraceae (Gottlieb, 1988; Sytsma & Smith, 1988), Poaceae (Hamby & Zimmer, 1988), and various groups (Palmer, 1987; Palmer et al., 1988) including Caryophyl- lales, Orchidaceae, Solanaceae, Fabaceae, and studies within the genus Brassica. The living collections of botanical gardens are coming into their own as accessible research ma- terial for macromolecular work, but the technology can now use minute amounts of nucleic acids so that herbarium specimens can serve as sources for enzymes for electrophoresis (Ranker & Werth, 1986) or for DNA (Doyle & Dickson, 1987), and even the partially degraded material of some fossils can give information (Paabo, 1989). he data are voluminous even when only a few gene systems are investigated. If methods of nu- merical analysis were not already accepted pro- cedures in systematics, they would have had to be invented now to cope with this flood of new data. The power and relevance of the macromolecular data are clear, but warnings against excessive con- fidence have been sounded. Penny et al. (1987, 1990) draw attention to problems arising when data are insufficient, not representative of the genome as a whole, or when methods of analysis lead to convergence on an incorrect tree (including cases where a tree is not an appropriate model). Studies of chloroplast or mitochondrial DNA are elucidating the phylogeny of organelles, not directly that of the organisms that contain them. Thus the gene phylogeny may depart from the species phy- logeny. For instance, maternal inheritance of mi- tochondria in higher animals (Neigel & Avise, 1986) produces anomalous results in comparisons of con- specific individuals and recently diverged species. As a result, the phylogenetic distribution of mtDNA may not be concordant with species boundaries. Аз mentioned above, chloroplasts are also strictly maternally inherited in most flowering plant groups. ethods of analyzing molecular data have right- ly generated much discussion. А two-step ргосе- dure with cladistic methods to produce trees and maximum-likelihood similarity methods to evaluate those trees (Templeton, 1983) appears to have advantages. Farris (1985) and Cracraft (1989) have called for cladistic methods to be used more gen- erally with sequence data, but others have pointed to problems that may arise in parsimony analyses (e.g., Felsenstein, 1978; Jansen & Palmer, 1988). Volume 78, Number 1 1991 riggs 100 Years of Plant Taxonomy MAJOR GROUPS AND THEIR RELATIONSHIPS Cladistic studies within orders or large families have tended to show few synapomorphies marking the lower internodes of early-divergent clades, so that some or all of the major groupings are not robust (e.g., Myrtales and Myrtaceae— Johnson & Briggs, 1984; Rosaceae— Кактап, 1988; Caes- alpinioideae, the basal group within the legumes [lack of robust groups is clear from descriptive comment— Polhill & Vidal, 1981—rather than from formal cladistic analysis]. The hypotheses represented by competing trees in such analyses have nevertheless led to marked improvements on earlier views of relationships. Even more improve- ment toward robust phylogenies can be expected where such analysis has available macromolecular markers relevant to subfamilial or tribal groupings. Examples are in Asteraceae (Bremer, 1987, in- corporating data from Jansen & Palmer, 1987) and in Faboideae (Doyle, 1987). Systematics this century has worked through a range of phases, enlarging the spectrum of new approaches and scarcely ever relegating any as irrelevant. As a result, a valuable development has been the symposium that brings many approaches to focus on a particular large group, usually a family or an order. Apiaceae (Heywood, 1971), Asteraceae (Heywood et al., 1977), Solanaceae (Hawkes et al., 1979), Fabaceae (Polhill & Raven, 1981; Stirton, 1987), Myrtales (Raven, 1984), Poaceae (Soderstrom et al., 1987), Hamamelidae (Crane & Blackmore, 1989), and others have ben- efited from this treatment. Often considerable con- sensus has emerged but sometimes such a meeting of minds has emphasized the divergence of views, as with the status and circumscription of genera of Asteraceae (Lane & Turner, 1985). Still wider synthesis has been the task of those who have developed systems of classification for the flowering plants as a whole. Impressive arrays of features contributed to the systems of Thorne (1981), Cronquist (1981), Dahlgren (1980; Dahl- gren et al., 1985; Dahlgren & Bremer, 1985), and Takhtajan (1987). The next round of systems will differ from these by extensive use of macro- molecular data. Work has not been concentrated on the flow- ering plants alone. The algae have been recognized as a heterogeneous assemblage of groups. Fungi and bacteria have been more adequately divorced from green plants, and such discordant concepts as the broadly conceived pteridophytes and the gym- nosperms are tending to drop from scientific usage. Homologies between major groups have been in- vestigated and, with the increased range of mor- phological data available, have been the subject of cladistic studies (e.g., Bremer, 1985). The sym- biotic origin of the eukaryotic cell has been a major unifying theory, now happily confirmed by the gen- ic homology of mitochondria with “purple bacte- ria" and of sags are with cyanobacteria (Pace et al., 1986). her taxonomic levels es- pecially can be bia to repay macromolecular studies if suitably conservative genes are chosen. FLORAS, TENSIONS AND SYNTHESIS This review has looked briefly at the outstanding trends over a century. But what were the taxon- omists actually doing? Fortunately they were most- ly not poised waiting to catch each new bandwagon as it came along. Mostly they were describing taxa, writing floras and monographs, developing the re- sources of collections and published work that we now have. Usually they applied only a sample of the range of available ا‎ to the classifi- cation of each particular plant gro Thirty years ago Stafleu (1959) аманы the state of plant taxonomy аз an “аре of floras and floristic work," noting that this was partly at the expense of synthetic and monographic studies. This characterization applies with greater force today. The increasing recognition of threats to the diversity of the biota, and also recognition of the value of that diversity, have helped taxonomists to obtain funding for large flora projects, regional, national, and international. Floras of the у and Europe have been completed. Projects for Aus- tralia, China, India, North America, and South Africa have joined such longer-established works as Flora Malesiana, Flora Neotropica, Flora of Tropical East Africa (references in Frodin, 1984), and compilations such as Med-Checklist (Greuter et al., 1984-1989) in providing accessible infor- mation on the world's flowering plants. Frodin (1984) observed that so little change in style has taken place in floras over 140 years— since Torrey, Gray, Bentham, and Hooker—that *One who compares some current American state floras with Torrey’s Flora of the State of New York might be forgiven for thinking that the mod- ern works had been written by descendants of Rip van Winkle"—and North America is not the only continent to which such comment applies. Plant taxonomists are used to living in a scientific culture of some tension. Bibliographic search back more than 200 years and Latin diagnoses are still part of their stock-in-trade, although moves to change the former are being considered. Yet for 28 Annals of the Missouri Botanical Garden 50 years intensive studies using varied techniques have competed with descriptive flora and mono- graphic projects. Macromolecular genetics now of- fers involvement with aspects of one of science's fastest-moving fields. There are serious choices to be made in how much of their resources institutions and individuals should devote to their different re- sponsibilities for flora projects, revisions, and using the most powerful tools yet available to elucidate phylogenies and thus improve classifications. A challenge at present is to maintain cohesion in systematics across these streams so that the new developments are a scientific stimulus to the subject as a whole. Academic teachers of biology must have a broad appreciation of these approaches in research and education if the training they provide is to be relevant to the eventual careers of most research students in systematic and evolutionary biology. On the other hand, the herbaria and other institutions, while carrying the major part of most floristic projects, will need to respond positively to new developments if these organizations are to re- main central to the mainstream of scientific tax- onomy. With environmental destruction and change now rampant, it was the taxonomists of these last hun- dred years that sampled the world's biota at the richest development it has offered for study. The immediate future gives the best chance humanity will have to investigate the diverse products and processes of evolution. he concise flora has been an appropriate re- action to the need to record plants in regions under threat and to obtain information on diversity to support the case for conservation. This view has recently been strongly supported by Nooteboom (1988) and by Bramwell (1989). Equally, however, most macromolecular studies require access to live plants. The living collections of botanical gardens are becoming more important as research mate- rials, just as the herbarium collections built up over the last century have become increasingly irre- placeable (Raven, 1988). Macromolecular studies of different gene sys- tems have commonly, but not always, given con- sistent results. There is particular need to inves- tigate cases where there is disparity between results from DNA and from morphological data (Sytsma & Smith, 1988). There are instances of isozymic information on allied species corroborating mor- phological similarities but giving results quite at variance with comparisons of chloroplast DNA in ferns (Yatskievych & Moran, 1989) and with mi- tochondrial DNA in fishes (Dowling & Brown, 1 ). Studies on Lisianthus (Gentianaceae) have given similarly discordant results (Sytsma & Schaal, 1985). It may be relevant that these apparent "failures" of DNA analysis are all applications to groups of closely allied species, whereas there has been greater success at higher taxonomic levels. As information accumulates, it may be possible to select methods appropriately to particular levels and also to interpret such discrepant findings. Patterson (1987) postulates that “molecular ho- mologies are no more secure, and are possibly more precarious than morphological ones," an c- Kenna (1987) writes, “It should also be obvious that molecular studies can suffer from exactly the same ills that beset comparative anatomical ones: a touchstone has not been found." Indeed, the succession of failed panaceas in systematics should be heeded as a warning. Nevertheless, by seeking congruence of data from different gene systems and from morphology, it appears that macromo- lecular methods will often produce convincing phy- logenies at many taxonomic levels, and they offer a vastly expanded sample of characters. "Reliable" taxonomies have long been contrast- ed with speculative phylogenies, even though the assignment of levels in taxonomic hierarchies has been accepted as partly a matter of subjective judgment. The situation is tending to be reversed. Patterson (1987) observed that as little as 30 years ago, when phenetics was at its height, it was thought that “classification by genealogy might be a laudable goal, but not one that was о In The Origin of Species, Darwin (1859: 486) had predicted that **Our classifications ue come to be, as far as they can be so made, genealogies.” To the extent that the promise of macromolecular systematics is realized, phylogenies inferred from sequence data and considered in the light of mor- phological and other data sources may become frameworks to use in asking important questions in other areas of biolo If phylogenies can be — and reliably de ermined at many taxonomic levels from macro- molecular information, Gottlieb (1988) is right to claim that this offers dramatic new opportunities to systematists. Human interest in organisms is not limited to their relationships, but is also in how they work as functional, adaptive systems and how genic change has been expressed in ontogeny and in phenotypes. Botanical systematics will still be an unending and widening synthesis—but on new terms. LITERATURE CITED ALLAN, M. 1967. The Hookers of Kew, 1785-1911. Michael Joseph, London. ALSTON, В. E. & В. L. TURNER. 1963. Biochemical Systematics. Prentice-Hall, Englewood Cliffs. APPLEBY, C. A., E. S. DENNIS & W. J. PEACOCK. 1990. Volume 78, Number 1 1991 Briggs 29 100 Years of Plant Taxonomy A primaeval origin for plant and animal haemoglo- bins? Austral. Syst. Bot. 3: 81-89. Bates, D. M., В. DAHLGREN, P. S. GREEN & К. KUBITZKI. 1980. А prospectus for a proposed new work: the families and genera of vascular plants. Taxon 29: 318-320. BAVERSTOCK, P. К. & А. M. JOHNSON. 1990. ее NA nucleotide sequence: а methods used for its determination, and its use in nope analysis. Austral. Syst. Bot. 3: 101- 110. BEHNKE, H.-D. The bases of angiosperm phy- logeny: ultrastructure. Ann. Missouri Bot. Gard. 62: 647-663. BLUNT, У. 1978. In for a Penny—A m of Kew Gardens. Hamish Hamilton, end BouLrER, D. 1973. The use of amino acid sequence data in E classification of higher em Nobel Symp. 25: 211-216. 1 80. The evaluation of present results and future ке of the use of amino acid sequence is ta in phylogenetic studies и specific ein olent proteins. n F. A. Bisby et al. (editors), Chemospstenatios Principles and Веи ем 919. Joseph Debs Hooker. Macmillan, ew Yor BRAMWELL, D. 1989. Taxonomy: the sands of time. ls mary of green plant phylogeny and classification. Cladistics 1: 369-385. Tribal interrelationships of the Aster- aceae. Cladistics 3: 210-253. Brockway, L. H. 1979. Science and Colonial Expan- sion — the Role of the British Royal Botanic Gardens Academic Press, New York Brown, A. H. D. 1990. The role of isozyme studies in 2 systematics. Austral. Syst. Bot. 3: 39-46. CAMIN, J. R. R. SoKar. 1965. А method for ы а н branching sequences in phylogeny. Evo- lution 19: 311-326. Lied Lois de la Nomenclature Bo- aris. E ое diversity in plant pop- ulations. Pp. 89-115 H. D. Brown et al. WU Population dad Plant Breeding and enetic Conservation. Sinauer Associates, Sunder- lan ^ Treuttel & Wirtz, — А. Н. D. Brown & Р. R. WHITFIELD. 1984. a DNA майы, in wild and cultivated bar- ley: implications for genetic conservation. Genet. Res 43: 339-3 CONSTANCE, L. synthesis. Taxon 13: 2 CRACRAFT, J. 1989. Review of C. Patterson (editor), Molecules and Morphology in Evolution: Conflict or Compromise? Cladistics 5: 202-210. CRANE, P. R. & S. BLACKMORE. 1989. Evolution, Sys- tematics and Fossil History of the Hamamelidae. 2 volumes, Systematics Association Special Volume 40. ss. 43. 1964. Systematic Nc unending 57-273 Phylogenetic and "prc а = from electrophoretic studies. 5 n S. D. Tanksley & T. J. Orton (editors), Isozymes in Plant اا‎ and Breeding, Part A. Elsevier, Amsterda CREPET, W. L. & K. C. NIXON. 1989. Earliest mega- fossil evid of Fagaceae: phylogenetic and biogeo- graphic implications. Amer. J. Bot. 76: 842-855 CRONQUIST, А. J. 1 n Integrated System of Clas- sification of Flowering Plants. Columbia Univ. Press, . 1987. А botanical critique of cladism. Bot. Review 53: 1-52. DAHLGREN, R. M. T. А revised cpg of clas- sification of the angiosperms. Bot. J. Linn. Soc. 80: -124. 1983. General aspects of — evolu- tion ‘and macrosystematics. Nordic J. Bot. 3: 119- 149. EMER. 1985. Major clades of the angiosperms. Cladistics 1: 349-368. Н. T. CLIFFORD & P. Е. Yeo. The Families of the Monocotyledons. Springer- Verlag, Berlin. DaRLINGTON, C. D. 1937. Recent Advances in Cytology, 2nd edition. Churchill, London. e Evolution ae ii Systems. Ca ны ridge Unie: Pul. Cambri DARWIN, C. В. 1859. On the Origin of Species by Means of Natural Pa ue = Murray, London Davis, P. H. & V. H. woop. 1963. Principles of Angiosperm МА лан D. Van Nostrand, New Yor DESMOND, R. . 1975. The Hookers and the да. velopment of the Royal Botanic Gardens, Kew. Biol. J. Linn. Soc. 7: 173-182. DowLiNc, Т. E. & W. M. Brown. 1989. Allozymes, mitochondrial DNA, and levels of plriogenetio res- olution among four minnow species (Notropis: Cy prinidae). Syst. Zool. 38: 126-143. DovLE, J. J. 1987. Variation at the DNA p uses and potential in legume onne S ied nC H. Stirton (editor), Advance Part 3. Royal Vina us. & E. E. Dickson. 1987. Porai of plant samples for ‘restriction endonuclease analysis. Taxon 36: 715-7 d. 1. MA A. Н. D. Brown. 1990. Anal. ysis of a polyploid complex in Glycine we ood plast and nuclear DNA. Austral. Syst. Bot Роскек, S. C. 1988. The Contented Botanist. Letters of W. H. Harve ie i gap and the Pacific. Melbourne Univ. "Pre , Melbou EHRLICH, P. К. & Р.Н. Rav N. 1969. Differentiation of populations. Science 165: 1228-1232. ELDRIDGE, М. & J. CRACRAFT. 1980. Phylogenetic Pat- terns and the Evolutionary Process: Method and The- ory in Comparative Biology. Columbia Univ. Press, New York. ENGLER, Н. С. А. 1892. Syllabus der Vorlesungen über specielle und medicinisch-pharmaceutische Botanik. Gebr. Borntraeger, Berlin. -1953. Das Pflanzenreich. Wilhelm En- gelmann, Berlin. & К. A. E. PRANTL. 1887-1915. Die Natür- lichen Pflanzenfamilien .... Wilhelm Engelmann, Leipzig Ewan, J. 1969. Early history. Pp. 27-48 i snl А Short - ates. Hafner, 19 F Ten ]. 3. 83. "The pem basis of phylogenetic n J. Ewan ang of Botany in the United Annals of the Missouri Botanical Garden m Pp. 7-36 in N. I. Platnick & V. A. Funk (edito Ad in Cladistics, Volume 2. Columbia Univ. {зн New Yor 198 Distance dia revisited. Cladistics 1: 67-85. FELSENSTEIN, Ј. 1978. Cases in which parsimony ог compatibility methods will be positively misleading. Syst. Zool. 27: 401-410. — i 1988. Isozyme studies and their inter- рев n. Рр. 184-201 in D. Г. Hawkesworth (ed- itor), Prospects i in Posts Clarendon, Oxford. кар; D. С. 1984. Guide to Standard Floras of the orld. Cambridge Univ. Press, New York. Скалы. М. T. 19 Narrow approaches to phylog- y: a review of nine books of cladism. Oxford Sur- eys in Evol. Biol. 1: 209-222. GIANNASI, D. E. & D. J. CRAWFORD. 1986. Biochemical Е cs II. A reprise. Pp. 25-248 in М. К Hecht, B. Wallace & G. T. Prance (editors), Evolu- ved Biology, Volume 20. Plenum Press, New ork. GOTTLIEB, L. D. 1984. Isozyme evidence and problem solving in plant systematics. Pp. 343-357 in W. F. Grant (editor), Plant Biosystematics. Academic Press, Canada. . 1988. Towards molecular genetics in Clarkia: gene е and molecular characterization of de genes. Ann. Missouri Bot. Gard. 75: 1169- 79. mu V. 1981. Plant Speciation, 2nd edition. Colum- bia Univ. Press, New York. 1984. Plant Biosystematics. а Press, A Gray, J. L. Let of Asa 2 volumes. Houghton Mifflin, end Им. Franklin, New ork.] Gio. W., Н. M. Вокрет & С. е о" Меа- Checklist. Cons. et Jard. B : Gen HAMBY, К. p & E. A. ZIMMER. 1988. "Ribosomal RNA sequences for inferring TT grass family (Poacea e). Pl. Syst. Evol. 29-37. HARBORNE, J. B. & B. L. TURNER. 1984. Plant Che- а Academic Press, Orlando, Florida. Hawkes, J. G., В. М. LESTER & А. D. SKELDING. 197 The Biology and e of the Solanaceae. i ademic Press, HAUFLER, C. 1989. Species concepts in pterido- a ийише and synthesis. Атег. Fern. J. 79: 90- HeoNavER, 1962-1973. Chemotaxonomie der wi , 6 volumes. Birkhauser Verlag, Basel $ Stu art. HEISER, E B. 1969. Taxonomy. Pp. 110-114 in J. Ewan (editor), A Short sal of Botany in the United States. Hafner, New York. HENNIG, W. Gru ndzüge e einer Theorie der Phy- logenetischen Systematik. Deutscher Zentralverlag, rlin. . 1966. Зак Systematics. Univ. Illinois 1982. Kew — Gardens for Science and Pleasure. H. M. Stationery Office, London. Неузуоор, У. Н. (editor). 1971. The Biology and Chem- istry of the E paid Academic Press, London. . B. BORNE & B. L. TURNER (editors). 1977. The Biology & "а of the Compositae. Academic Press, Londo Hickey, І. J. & J. A. WorrE. 1975. The bases of angiosperm сао vegotative morphology. Ann. : 538-589. : Introductory essay. In: The Bot- any of the Antarctic Voyage of H. M. Discovery Ships ‘Erebus’ and ‘Terror’ in the Years 1839-1843, III. Flora Tasmaniae. Lovell Reeve, London 88. Science as a Process. Univ. Chicago . 1961. Makers of North American Botany. Ronald Press, New York. Jackson, B. D. 1906. George бейш: Dent, London. JANSEN, R. K. & J. D. PALMER. 1987. A chloroplast DNA inversion marks an ancient evolutionary split in the sunflower family (Asteraceae). Proc. Natl. Acad. U.S.A. 84: 5818-5822. & 1988. о implications of chloroplast t DNA restriction site variation in the Mutisieae (Asteraceae). Amer. J. 64. CHAELS. 88. Investi- gations of chloroplast DNA variation in the Astera- ceae Е ewsl. 15: JOHNSON, L 1968, 1970. Rainbow s end: the quest an л optimal taxonomy. Proc. Linn. Soc. New South Wales 93: 7-45 (1968). nia (with Ad- си іп Syst. Zool. 19: 203-239. (1970) . Models and reality: doctrine and prac- eed in у ел Pl. Syst. Evol. 168: 95-108. & B.G 75. On the Proteaceae — the evolution and classification of a southern family. Bot. J. Linn. Soc. 70: 83-182. 84. Myrtales and Myrtaceae — a phylogenetic. analysis. Ann. Missouri Bot. Gard. 71: 700-756. JORGENSEN, В. A. & P. D. CLUSTER. 1988. Modes and tempos in the evolution of nuclear ribosomal DNA: ed | 1988. Тће phylogeny of ‘the Rosaceae. nn. Soc. 98: KING, R. ш Коуа! Kew. > onstable, London LANE, M. А. & B. 1. TURNER (editors). 1985. The generic concep’ in the Compositae: a symposium. Taxon 34: 5-88. Ma, W. L. 1988. The historic current status of herbaria in China. Taxon 37: 870-875. MARTIN, P. С. & J. M. Dowp. 1990. A protein sequence G. study of the dicotyledons and its relevance to the evolution of the legumes and nitrogen fixation. Aus- tral. St Bot. 3: 91-100. 1965. Numerical phenetics and taxonomic . Syst. Zool. 14: 73-97. MCKENNA, М. С. 1987. Ма и and morphological analysis = high- level mammalian interrelationships. Pp. in C. Patterson (editor), Molecules and М dde in Evolution: Conflict or Co Cambridge Univ. Press, Cambridge MELCHIOR, Н. 1964. А. Engler's Syllabus der Pflan- zenfamilien, 12th edition, Volume 2. Angiospermen. Borntraeger, Berlin МООВЕ, D. M. 1978. The chromosomes and plant ta onomy. Pp. 39-56 in H. E. Street (editor), Essays in Plant Taxonomy. Academic Press, Londo NEIGEL, J. E. & J. С. АМЗЕ. 1986. Phylogenetic mi- tochondrial DNA under various demographic models of speciation. Pp. 515-534 in S. Karlin & E. Nevo ompromise 7 Volume 78, Number 1 1991 Briggs 31 100 Years of Plant Taxonomy (editors), Evolutionary Processes and Theory. Aca- demic Press, Orlando, ida. NELSON, G. . I. PLATNICK. 81. Systematics and Biogeography: Cladistics and Vicariance. Columbia Uni 1988. bie should botanists do with their time? Taxon 37: PaaABO, S. 1989. Ancient he extraction, character- ization, molecular cloning, and enzymatic amplifi- cation. Proc. Natl. Acad. U.S.A. 86: 1939- 1943. РАСЕ, М. R., С. J. OLSEN & С. В. Worse. 1986. Ri- bosomal RNA phylogeny and the primary lines of ди а descent. Cell 45: 325-326. PALMER, J. D . Chloroplast DNA evolution and biosystematic uses of chloroplast DNA variation. Amer. Naturalist 130, S6-S29. ————, К. К. Jansen, Н. J. Mich & . MaNHART. 1988. Chloroplast DNA vari- ation „ш plant phylogeny. Ann. Missouri Bot. Gard. 75: 1180-1206. PATTERSON, C. 1987. Introduction. Pp. 1-22 in C. Patterson (editor), Molecules and Morphology in Evo- lution: Conflict or Compromise? Cambridge Univ. dire Cambridge. PENN . Немру & I. M. HENDERSON. 1987. а of evolutionary trees. Cold Spring Harbour Symposium on Aig ntitative Biology 52: 857-862. — ——-, E. A. ZIMMER & R. K. Намву. 1990. Trees from 5 рапасеа or Pandora's box? Austral. Syst. Bot. 3: 21-38. PornuiLL, В. М. € P. Н. RAVEN (editors). 1981. Ad- vances in Legume Systematics, Parts 1 & 2. Royal Botanic Gardens, Kew . E. Vipar.. 1981. Tribe 1. Caesalpinieae. Pp. 81- 95 in R. M. Polhill & P. H. Raven deris Advances in Legume Systematics, Part 1. Royal tanic Gardens, Kew RANKER, T. A. & C. R. WERTH. 1986. Active enzymes бога herbarium specimens: electrophoresis as an af- terthought. Amer. Fern. J. 76: 102-113. RAVEN, Р. Н. 1 The bases of angiosperm vd eny: cytology. Аид. Missouri Bot. Gard. 62: 764. AELS, М. W. CHA —————. 1976. Systematics and plant population bi- ology. cab Bot. 1: 284-316. 1977. The ee ae and evolution of higher plants. Pp. 59-83 . E. Goulden (editor), The Changing Scene in Natu ral Sciences, 1776-1976 Academy of Natural Sciences, Philadelphia, Spec. Publ. 12, Lancaster. . Hybridization and the nature of species 5 higher plants. Canad. Bot. Assoc. Bull., Suppl. to : 3-10. (editor). 1984. The order Myrtales: a sym- a Ann. Missouri Bot. Gard. 71: 631-985. ————. 1986. Modern пи ан of the biological T 5 plants. Pp. 11-29 in K. Iwatsuki, P. H. Rav W. J. Bock (editors), Modern Aspects of "амач Dus b Press, To 1988. The issue of conservation in natural history museums. P. 12 in Abstracts from the 1988 Annual Meeting, Society for Preservation of Natural History Collections. Carnegie Mus. Nat. Hist., Pitts- RITLAND, K. & M. T. . 1987. Evolutionary anal- ysis of plant DNA sequences. Amer. Naturalist 130, 574-5100. ROLLINS, К. С. 1965. The role of the university her- arium in research and teaching. Taxon 14: 115- 0. 12 SCOURSE, М. 1983. ри Victorians and their Flowers. Croom Helm, Lon SMALL, E. 1989. ннн И of biological systematics (or, Taxonomy of taxonomy). Taxon 38: 335-356. SNEATH, P. Н. A. & К. R. Sokar. 1973. Numerical Taxonomy. W. H. pe sat San Francisco. SODERSTROM, T. R., LU, C. S. CAMPBELL & M. E. BARKWORTH (editor з). "1987. Grass Systematics and Evolution. Smithsonian Inst. Press, Washington, D.C. SoKaL, К. R. & Р. Н. A. SNEATH. 1963. The is in ки a Taxonomy. W. H. Freeman, San Fra Stare, F. A. 1959. The present status of plant tax- my. =- Zool. 8: 59- 1966. The great Prodromus. In: A. P. de Ca ndo lle, Prodromus systematis seid regni ve- getabilis, 1823-1873, reprint of volumes 1-8: 5- amer, T КС). 19 2 volumes on cryptogams of “Еп- де und Prantl." Тахо оп 21: 501-511. E E The prehistory and history IAPT. Tax- 0. Cs 3 i s “О. 57). ag Adolf Engler [obituary]. ull. Misc. Inform. 10: -498. STEBBINS, C. L. 1950. ta Si and Evolution in Plants. Columbia Univ. Press, New York. STEVENS, P. F. 1984. Metaphors and typology in the development of botanical systematics 1690-1960, or the art of putting new wine in old bottles. Taxon 33: 169-211. STIRTON, C.H. 1987. Advances in a Systematics, 3. Royal Botanic Gardens, Kew Using Parsimony, Version 2.3. Illinois Natural His- Phylogenetics of the Lisianthius skinneri (Gentianaceae) species complex in Panama utilising DNA restriction frag- ment vu Y Evolution 39: 594-608. & 1988. DNA and morphology: comparisons in the Onagraceae. Ann. Missouri Bot. Gard. 75: 1217-1237. TAKHTAJAN, А. 1987. Sistema c [Systema Малаји Izdatelstvo “Nauka,” Lenin- fae R. (editor). 1965. Science in the vr enin ry. (Translated by A. J. Pomerans.) Tham ondon А 1983. Phylogenetic o без restriction endonukleáse адя reference to the evolution of eal and га es. Evolution 37: 221- Teen R. F. 1981. iic and angiosperm phylogeny: a summary statement. Pp. 233-295 in D. Young & D. Seigler (editors), Phytochemistry and Angiosperm Phylogeny. upon Scientific, New York. Torrey, J. & A. GRAY. 1838-1843. A Flora of North America, 2 volumes. Wile Putnam, New York. VICKERY, В. К. 1984. Biosystematics 1983. Pp. 1-24 т У. К. Grant (editor), Plant Biosystematics. Aca- demic Press, Canada. WAGNER . 1961. Problems in the classifi- паран of feno Pp. 841-844 in D. L. Bailey (editor), 32 Annals of the Missouri Botanical Garden Recent Advances - Botany, Volume 1. Univ. To- ronto Press, d WALKER, J. W. E 1975. The bases of oe phylogeny. res Missouri Bot. Gard. 62: 515-834. WiLEY, E. O. 1981. The Theory and Practice of Phy- logenetic Systematics. Wiley-Interscience, New York. WiLLIS, М. 1949. By Their Fruits — А Life of Ferdinand von Mueller. Angus & Robertson, Sydney YATSKIEVYCH, С. & К. C. Moran. 1989. Primary я ver RONDA and species concepts in ferns. Amer. Fer -45. ZURAWSKI, G. & M. T. Стесс. 1987. Evolution of ies. Annual Rev. Pl. Physiol. 38: 391-418 A HISTORY OF BOTANICAL NOMENCLATURE: Dan Н. Nicolson? ABSTRACT divide botanical nomenclature into three partly о periods: the schismatic ber (1840-1930), the dark ages (1915-1950), and the IAPT renaissance (1950-d or were evolving modified or different Codes from the Candollean, not to mention ly formed Codes from like Saint- Lager (1880. 037, 1881. 04) and Kuntze (1891.10). The 190 nal, Taxon te). The schisms began with the 1843 British Association manifest with the 1867 Paris Congress " Reunification efforts, 2 as those by Dall (1877. 12), failed. la na Congress eliminated all but the nded it. Again a World War n, in which all Code amendment proposals now ur appear, and its serial това Regnum eis in which all subsequent Codes appear at the remorseless six- year pace of the congresses. The Linnaean aphorisms (1737.07) concerning nomenclature basically concern generic nomencla- ture (the first part of the binary system of nomen- clature). Few seem to have realized that the species names Linnaeus discussed are the diagnostic phrase names, not the species epithets, called nomina trivi- alia, that Linnaeus later, for plants (1753.05.01), adopted in his revolutionary binomial (or biverbal) system. The Linnaean nomenclature (binomial) system has two facets: divorcing the name from the di- agnosis and minimizing classification. Systems that totally eliminate classification, such as uninomial systems, are unwieldy. Systems that try to diag- nose, such as phrase names, are unstable. Bergeret devised the worst of all possible systems, called phytonomatotechnie, a 15-letter uninomial system with each letter expressing a descriptive charac- teristic, such as leglyabiajisbey for Belladonna. Du Petit Thouars added a bit of the family name to each generic name (such as Habenorchis to replace Habenaria). Fossil nomenclature is still troubled by mixing generic nomenclature with mor- phology. Priority had no part of the early schemes of nomenclature. Their authors, including Linnaeus, were focused on replacing the past. Linnaeus was the winner of this competition but contemporary and subsequent workers continued to devise new nomenclatural schemes and rules to overturn the past. EARLY RECOGNITION OF MODIFIED PRIORITY (1813 Augustin Pyramus de Candolle (1813), about to initiate the great Prodromus, gave a nice discussion of nomenclature of what might be termed good practices with examples. He (p. 250) favored pri- ority except in the following five cases: (1) if the name was false and contradicted the characters of the plant, such as Lunaria annua (a biennial); (2) if the name was contrary to Linnaean practices such as pre-Linnaean or common names or if named published, be it in the plant or animal kingdom; (4) if the name was a later homonym or a tautonym created in the process of maintaining the epithet of an early name; (5) if the name was a nomen nudum, lacking at least a phrase sufficient to make recognition possible, such as “in a simple catalogue of a garden." De Candolle did not use such modern words as epithet, homonym, tautonym, or nomen nudum but that is what he talked about. This was ! [ thank Krister de (Helsinki) and Ronald Stuckey (Columbus, Ohio), who shared their bibliographies on nomenclature. I also quickly regrowing libra ank Werner Greuter, director, and the staff (Berlin) for access to and all facilities at their ? Department of Boon NHB-166, Smithsonian Institution, Washington, D.C. 20560, U.S.A. ANN. Missouni Вот. GARD. 78: 33-56. 1991. 34 Annals of the Missouri Botanical Garden still a biological code, revealed by the avoidance of homonymy with earlier names in zoology. There undoubtedly are earlier authors who had the idea of trying to maintain usage, unlike re- formers such as Linnaeus. The general idea was to maintain usage from Linnaeus by the principle of priority with certain modifications. It is the mod- ifications that have caused most of the disagree- ments. BEGINNING OF THE BREAK WITH ZOOLOGY (1843) The approval of Strickland's 1843 Code by a committee of the British Association for the Ad- vancement of Science officially began what I call a split. However, they were not thinking in terms of a split, as is clear from the second of the following quotations: The first quote concerns the evil that they wish to address. “Ц consists in this, that when naturalists are agreed as the characters and limits of an in- dividual group or species, they still disagree in the Further on they discuss the sins that create this evil. appellations by which they distinguish it." Finally, so all will know how botanical nomen- clature was perceived by zoologists (including Charles Darwin, a member of the Strickland Com- mittee): “ menclature of the present day stands in much less need of distinct enactment than the zoological. The admirable rules laid down by Linnaeus, Smith, De- candolle, and other botanists . . ., ercised a beneficial influence over their disciples. Hence the language of botany has attained a more perfect and stable condition than zoology an . we conceive that the botanical no- have always ex- its present backward and abnormal state. . . . One of the subsequently liveliest issues was the question of author citation in new combinations. This work recommended (for the first time?) citing only the author first publishing a name for the species (optionally in parentheses), irrespective in what genus the epithet was now used. Peter F Stevens has suggested (in litt.) “that the develop- ment of this convention is associated with changes in the philosophy that species are created, fixed and immutable, and placed in similar genera; what came to be important in priority was who first discovered the unknown species, not who finally put it in the ‘right’ genus. А CANDOLLEAN LAWS OF 1867 On 1 August 1867 Alphonse де Candolle fin- ished the cornerstone work of botanical nomencla- ture for the meeting of the International Botanical Congress of 16 August 1867 in Paris. This 60- page paper has 11 pages of introduction, 19 pages in 68 articles, and 28 pages of com- mentary. The commentary is particularly valuable of “laws” because it comments on past usage with examples. One of the main issues (one-third of the commen- tary)c author citation, favoring the old way (only the name of the transferring author) but noting that others had introduced the new way (only the name of the author first publishing the basionym) in botany. The issue is exemplified by how to cite Robert Brown's Matthiola tristis, pre- viously known as Linnaeus's Cheiranthus tristis; the old way is Matthiola tristis R. Br., and the new way is Matthiola tristis L. According to Bentham (1878.12), it was the nomenclatural objections to J. Müller's treatment of Euphorbiaceae in the Prodromus that occa- sioned de Candolle's producing his Lois. See See- mann (1866.12.01), Gray (1867.01), and A. L. de Candolle (1867.05.01) for contemporary dis- cussion. De Candolle indicated that Karl Koch's proposals (1866) to the London International Hor- ticultural Exhibition and Botanical Congress pro- vided the inspiration. Alphonse de Candolle was president of the London 1866 Congress and might have been stimulated to start thinking about cod- ifying rules. However, I found little in Koch's pro- posals or the proceedings that would have been as inspiring as the nomenclatural novelties that Jean Müller slipped into de Candolle's Prodromus. Mul- ler's major sin was to credit names, such as Croton, to himself when he redefined the taxon. Apparently between 1 and 16 August 1867, a commission, composed of Du Mortier, Weddell, Cosson, J.-E. Planchon, Eichler, Bureau, and de Candolle (representing France, England, Germany, Belgium, and Switzerland), went over the proposed articles and suggested modifications in the form of motions to the Congress. At the Congress (cf. de Candolle, 1867.11) there were the first discussions of vital questions such as Order vs. Family, Cohort as a rank between Class and Order, | about how to name hybrids, more debuts on author citation for new combinations, even a mention of types in discussions. Otto Kuntze, who later will make himself famous in nomenclature, made a proposal about pleonasms and it passed. Kichler argued for replacing -ae- with -i-, as in hederae- folia, but was finessed by Alphonse de Candolle who said this was a matter of latinity and grammar, not nomenclature. It is worth noting that the Code (1867.09.12) was not enforced like the current Code(s). It was "adopted by the assembly as the best guide to follow Volume 78, Number 1 1991 icolson 35 History of Botanical Nomenclature for botanical nomenclature" (A. L. de Candolle, 1867.11: 208). Within 18 months de Candolle (1869.06) felt compelled to address the questions and criticisms raised by “his” Code, including approbation. He gently demanded permission not to speak of the vexing question of author citation, pointing out that he seemed to have involuntarily provoked a kind of polemic and antipathy that rarely contributes to progress in science. He continued to avoid discus- sion of Greek and Latin because it doesn't involve nomenclature. | mention most of the issues that he addressed in his quiet and civil fashion (with my summary): (1) retroactivity of laws (pro); (2) saving the name when a taxon is subdivided (pro); (3) effective publication; (4) citation of unpublished names (ex favored over in); (5) names contrary to fact (con); (6) hybrids and cultivated plants. Dall (1877.12) released an important survey of American zoologists with another Code, including the Candollean laws, for botanists and zoologists. This was not a unified (biological) Code because, when there were real differences (as the different endings for family names) then there were separate but contiguous articles. Otherwise an article per- tained to both botany and zoology. For those in- terested, an excellent discussion of the differences between the two Codes was published in 1944 (see bibliography). Saint-Lager was peeved by the Candollean ef- forts to maintain status quo, characterized as ““con- tinuation of disorder and anarchy." In 1880 he released his reform, aiming to overhaul all the "bad" spelling. This substantial treatment (155 pages, supplemented by 50 more pages the follow- ing year) covered most orthography questions (some that continue to vex us), applying pure classicism or peculiar conventions to change an astounding number of names or epithets. His ideas included (1) replacing Panax (p. 109) with. Panaxus (at least we wouldn't argue about its gender) and (2) replacing genitive (p. 118) substantives (nemorum, segetum, sepium) with corresponding adjectives (nemoralis, segetalis, sepicola). The major effect was to make botanists aware that correctness could overturn usage and to strengthen support for the Candollean laws. In 1883 Alphonse de Candolle published his "New remarks" (commented on by Asa Gray, 1883.12). De Candolle reviewed what happened over the past 16 years. His introduction com- mented on the Codes and Reports of nonbotanical groups. The opening part surveyed ongoing dis- cussions article by article. The second part dealt with new matters, such as nomenclature of organs and fossils. The third part (61-76) included the 1867 Code with an indication of the changes pro- posed. Kew RULE The first mention of what will be known as the "Kew Rule” was by Henry Trimen (1877.06). This was objected to by de Candolle (1877.08). Hiern (1878.03) stoutly defended it. The first formal definition was by Trimen (1878.06: 171). Bentham (1878.12) protested “ creating a new name in order to combine an old specific with a new generic ne.” Jackson (1887.03) discussed his problems with what was to become Index Kewensis. The Kew Rule: “Our practice is to take the name under which any given plant is first placed in its true genus as the name to be kept up, even though the author of it may have ignored the proper rule of retaining the specific name [epithet], when trans- ferring it from its old genus to the new.” The Kew Rule was opposed again by Alphonse de Candolle (1888.10), but supporters perceived it as applying the principle of priority by maintaining the oldest applicable (binomial) name. In January 1888 Asa Gray, the monarch of American botany, died, having supported the Kew Rule in one of his last papers (1887.12). Within months Nathaniel Lord Britton (Joseph Ewan as- sured me that the accent is on Чон) began to show the first inklings of a drastic new" nomen clature, shocking the Old Guard. The Жеты Code was under attack and promised to flood the new Index with unnecessary names. Many new combinations appeared in Poggenberg et al.'s (1888.04) Preliminary Catalogue, hailed by Greene (1888.06) as the “opening of a new era," but attacked by James Britten (1888.09) and AI- phonse de Candolle (1888.10) as '**mischievous," despite a spirited defense by Britton (1888.10). KUNTZE STRIKES (1891) Late October 1891, Kuntze’s new Revisio with 1,074 replacement genera and 30,000 binations appeared, a nomenclatural schism of the first order. Kuntze had a very broad view of what constituted homonymy, treated 1737 as the start- ing point, and that claimed he was only actually applying the Candollean C Kuntze’s work is, like Shot Pagani essentially forgotten, but is full of examples and repays study. It is worth trying to realize what an impact his work had at that point. Geneva (Alphonse de Can- dolle) discontinued the great Prodromus in 1874, new com- 36 Annals of the Missouri Botanical Garden Kew (Bentham and Hooker) finished the Genera Plantarum in 1883 and (Jackson) had just sent Index Kewensis to press, Berlin (Engler) began Die natürlichen Pflanzenfamilien in 1888. Sud- denly thousands of names were thrown into ques- tion, most perhaps challengeable (say by locking on 1753 rather than 1737) bu The major attack came from German botanists. In April 1892 the German Botanical Society ap- pointed a committee to prepare a supplement to the Lois of four propositions (theses). This was mailed to 706 botanists, requesting support and comments on each proposal (called **pamphleteer- ing" by Weatherby in 1949.01). The four were: (1) priority of generic names from 1752; (2) re- jection of nomina nuda and seminuda, including generic names based on simple figures and exsic- cata without description; (3) retention of generic names differing only by the last syllable or inflec- tion, even if they differ only by one letter; and (4) conservation of listed generic names. The results of this poll, with comments received, appeared in Paul Ascherson's report to the Society (1892.07) and again at the September Botanical Congress in Genoa (400th anniversary of Colum- bus). Dall (1877.12) also had used a poll but was more interested in principles than in what to my t many were not. eye appears to be “damage control.” Also included was the first published list of generic names to be conserved, another effort to maintain usage threat- ened by Kuntze who (1895.11, if not before) termed the list an “Index inhonestans.”” There was what in retrospect can be seen as a passing of baton by the nomenclatural leader. Bri- quet, who had included nomenclatural observations in his 1891 treatment of mints for Burnat's Flore des Alpes Maritimes, published (1892.02.04) his comments on what Kuntze had done to the mints. Briquet's evidence was cited by Alphonse de Can- dolle (1892.05) in one of his last nomenclatural papers before his death (1893.04.04). ROCHESTER RESOLUTION OF 1892 n the meantime a botanical club (with N. L. Britton) held a meeting within the American As- sociation for the Advancement of Science (A.A.A.S.) , in August 1892. Their proceedings (Fairchild, 1892.09) included the Rochester Resolutions (later called the Rochester Code), which accepted the Paris Code of 1867 with eight exceptions: (1) priority was fundamental; (2) in Rochester, New York starting point was 1753 for genera and species; (3) the original specific name [epithet] was to be retained unless a tautonym or later homonym; (4) no later homonyms; (5) publication of generic names either by distribution of printed description or by citation of one or more species as examples or types, with or without diagnosis; (6) publication of species names either by printed description or with reference to previously published species as a type; (7) later similar generic names are not to be rejected on account of slight differences, except in spelling; (8) in case of transfer of a species to another genus, the original author must be cited in parentheses, followed by the author of the new binomial. They also approved sending Lucien Underwood to the Genoa Congress to deliver their resolution. The September 1892 Botanical Congress in Genoa was obviously the place where a rousing battle could be expected. Underwood (1892.11) reported on the polyglot sessions. Ascherson presented the substance of his recently published report concerning Kuntze with four proposals. The Rochester platform was pre- sented. After the discussion the first three Berlin propositions were approved with 1753 (a Rochester resolution) for both genera and species. All else (conserved names and the Rochester resolution) was referred to a committee (Penzig 1893. 439), which was to report to the next International Congress (Paris, 1900). MADISON RULES OF UNMODIFIED PRIORITY (1893) In August 1893 a group of American botanists, meeting at Madison, Wisconsin, decided to amend what now are now being called the *'rules of no- menclature" adopted at Rochester (1892) and vot- ed to change “Section Ш” of the Rochester Code by striking out all after the word *'retained." This meant that a species epithet must be retained (re- stored) even if it involved a tautonym or a later homony. They also approved “that precedence in the same volume be regarded as priority," a mod- ification of the first section. Although the Roch- ester / Madison resolution (1893.09) did not men- tion types, the idea of page (and place on the page) priority, i.e., first listed species or specimen, was used for typification by practitioners of the new school until the (1907.04) Brittonian (American) Code provided ways around the first “listed.” In- deed, the major point of these rules, aside from astounding brevity, clarity, consistency, and posi- tiveness, was that they admitted no exceptions. If you found the earliest name, your troubles were over for all time. Also, practitioners had the moral Volume 78, Number 1 1991 icolson History of Botanical Nomenclature high ground in that you recognized the earliest proposer and nothing could be fairer. Note that what Saint-Lager, Kuntze, Britton, and others (as radicals) emphasized are the rules themselves, never mind upsetting usage, which was only a short-term consequence (cost). The Can- dollean (conservative) concept was to try to main- tain the status quo, never mind some complexity of rules, which are really of no importance. Both concepts would agree that stability is the goal of nomenclature, but one wonders if the interaction of these two schools doesn't result in a Code with the worst of both worlds: a complex and constantly changing Code (trying to maintain past usage) and constant conservations (to set aside rules that, de- spite complexity, are not maintaining the past). KUNTZE STRIKES BACK In July 1893 Kuntze published a second dose (first part of the third volume) of his Revisio. It had two features, a detailed comment on each publication (> 50!) that commented on his first dose and his Codex Nomenclature Botanicae Emendatus in German, French, and English in three columns with parallel text. Nothing was too fine to escape his caustic and withering comments, including the Rochester Resolutions, which were far closer to him than any other. John Isaac Briquet, who was to dominate no- menclature for more than 30 years (like Alphonse de Candolle), readdressed (1894.02) the imbroglio with a discussion of eight questions and a conclusion involving Kuntze's impact on the generic names of Labiatae. This work was important for the com- ments on, among other things, nomina nuda and seminuda [subnuda], Patrick Browne, Rumphius, recycling names fallen into synonymy, still-born names, the role of pre-Linnaean documents, and infraspecific nomenclature. He also proposed 10 modifications of the Lois. Kuntze (1894.07) discussed several recent at- tacks, expounding on Thouar's orchid names and other positions of Pfitzer, page priority (against the Madison amendments), an obligatory register for plant names, “once a synonym always a synonym,” and so on. By 1895 Ascherson and Engler were all that were left of the 30-member international commis- sion recognized at the 1892 Genoa Congress. They put forward six propositions (1895.01), of which the last two involved avoiding names that have not been used for 50 years, which would become an element of the Berlin Rules. Americans were showing an increasing disarray concerning their Code(s); the Harvard Code ap- peared on June and August 1895 with five con- servative points. The latter (Anderson et al., 1895.08) was signed by 74 American botanists “of various degrees of repute," a comment made ear- lier about the signers of the Rochester /Madison resolutions. BERLIN RULES (1897, 1902) Engler and his staff signed 14 rules for the Berlin Garden and Museum that appeared in June 1897. This was translated immediately into English (1897.08) and French (1897.09), often with com- ments. The Rules: (1) Priority was usually to be maintained (species from 1753, genera from 1754). (2) Generic names could be dropped if not in gen- eral use for 50 years from establishment unless restored because of the 1867 Lois. (3) Series were to end in -ales, families in -aceae, with exceptions for Coniferae, Cruciferae, etc. (4) For gender one must follow classical designations or, for later names, usage in Nat. Pflanzenfam. Changes ought not be made except for notorious errors in designations from proper names. (5) Generic synonyms were not to be applied in an altered sense for new genera or even sections. (6) Priority was to rule in species names unless objections were raised by monogra- phers. (7) Parenthetic citation of authors was to be used for new combinations, unless the author himself placed the species in another genus. (8) Linnaean capitalization of species names was to be used (for epithets derived from personal names or names of former genera). (9) Rules were given for forming generic names from personal names. (10) One must change -ae- to -i-, as menthifolia, not menthaefolia. (11) One must avoid tautonymy (incl. Elvasia elvasioides) and depart from priority in case of gross geographic errors (North American Asclepias syriaca L.). (12) Hybrids must appear as a formula of both parents (in alphabetic order) with multiplication sign between (binomial nomen- clature unsuitable). (13) Manuscript names, hor- ticultural names, designations in trade catalogs had no rights; a printed diagnosis was required. (14) No changes permitted to a generic or specific name unless moved by weighty reasons such as Rule 11. A supplement to the Berlin Nomenklatur-Regeln was published by Engler et al. (1902.08.29). KUNTZE STRIKES AGAIN In September 1898 Kuntze released the last two parts of the third volume of his Revisio. One sig- 38 Annals of the Missouri Botanical Garden nificant document included is a Supplement to his Codex Nomenclaturae Botanicae Emendatus. PARIS CONGRESS OF 1900 (STYLED THE 1ST INTERNATIONAL BOTANICAL CONGRESS) By October 1900 the stage was set for fireworks but the Congress unanimously decided to put some- one in charge (John Briquet) and revise the Code at the next Congress. This was the beginning of the office of rapporteur général. Hua (in Perrot, 1900: 475-486) made a detailed proposal to es- tablish an international periodical for publication of new names, a forerunner of a proposal referred to a subcommittee for registration of publications. Kuntze struck again, this time (1903.12) with Thomas Erik von Post in the Lexicon Generum Phanerogamarum, including his Codex Brevis Maturus. The work applied Kuntze's Code to pha- nerogam generic names, had a bunch of cryptogam generic names and, in the second part, dealt with names above generic rank, all warranting a look by those interested in suc Kuntze was given little recognition. Stafleu (in TL-2 under Post) excerpted Barnhart's (1904.04) comment about Kuntze, “Possibly the vituperation poured upon all who disagree with him has pre- vented his ideas from receiving as serious consid- eration as they deserve. It is certainly unfortunate that he should regard himself as an infallible referee upon all points in dispute and hurl anathemas at all who refuse to acknowledge his authority, char- acterizing their propositions as | “dishonest,” ‘шех- ecutable,' ‘false’ and ‘lawless.’ BRITTONIAN (AMERICAN) СОРЕ (1907) AND ITs PRECURSOR (1904) The Americans were divided and criticisms were being published, such as that by Fernald (1901.11), pointing out instability (actually inconsistencies) of practitioners of the Rochester resolutions. In early 1903 the Botanical Club of the American Asso- ciation for the Advancement of Science appointed a Nomenclature Commission. They produced a “Code of Botanical Nomenclature" (1904.05) with four principles, 19 canons, and separate sections on orthography and citations. This came simple proposal to the Vienna Congress, “the laws of 1867 be amended by abandonment of all its articles and substitution of the appended Code." This proposal failed at Vienna in 1905, leading to the Philadelphia **American Code of Botanical No- menclature" of 1907.04, which should be called with a Brittonian rather than American. The basic text was mostly unchanged from the originally proposed Code (1904.05), but there were changes in Canon 15 (application of generic names) as well as in- conspicuous additions elsewhere. 1 mention three definitions of now obscure terms (metonym, ty- ponym, hyponym, respectively taxonomic syn- onym, nomenclatural synonym, and name of an unidentifiable taxon). Weatherby (in Sherff, 1949.01) pointed out that this Code had at least one large crack compared to the ancestral Rochester /Madison resolutions’ iron-clad insistence on priority: it gave an elaborate series of directions for choosing types other than the previous implicit recognition of the first listed. Apparently the practitioners of the Rochester/ Madison resolutions were upsetting more usage than they felt necessary and were willing to compromise their simple and brief Code with exceptions that took up more space than the original resolutions in toto. VIENNA CONGRESS ОЕ 1905 (2ND CONGRESS, 1ST CODE) Briquet tracked down all reasonably explicit pro- posals since 1867 to change the code, translated them into French, and devised an organization. Thirty-eight publications were accepted as having specific proposals. This is a remarkable document, a key to the past. After all was compiled and organized (Nov. 1904), he sent it to the 39 mem- bers of the commission with his comments as rap- porteur. They were to vote and return all by 20 January 1905 (31 did). Briquet (1905.03.15) laid out all in four columns: (1) original (1867) text, (2) proposed change(s), (3) the rapporteur's ob- servations to the Commission, and (4) text rec- ommended by the Commission. The Congress convened in June 1905 in Vienna, capital of the flourishing Austro-Hungarian empire. This was the first meeting devoted solely to the vexing “problem of nomenclature." They had an immense amount of business with potential to end or descend deeper into chaos. It is fascinating to read how the proceedings were laid out (order of business), who had votes, and who did what. For example, N. L. Britton nominated Wettstein for president of the Bureau and proposed that French be the official language of nomenclature. Both were approved. Those wanting to speak had to submit a slip with their name and position (nom et qualité) to the president. Speakers were given five minutes (ten by exception) and only could speak twice on Volume 78, Number 1 1991 icolson History of Botanical Nomenclature the same subject. In fact, this Congress provided the model still followed by our nomenclature ses- sions. The proceedings (Briquet, 1906) were also well spelled out, with a concordance of the 1867 Lois with the Rules and Recommendations of the 1905 Congress, the Code (principally for vascular plants) in French, English, and German with the now con- served generic names. Among the many accomplishments of the Vi- enna Congress were the establishment of the first imposed international Code (the 1867 Lois was only recommended as the best guide to follow). It did not satisfy Kuntze, but he died in early 1907. The Brittonian Americans, although they had won some victories, were not happy that names still were not applied according to types and were al- lergic to the new requirement that Latin had to be used for publishing names of new taxa, considered arbitrary. The Americans, at least those supporting N. L. Britton, therefore settled more firmly into the Brittonian (American) Code, a rupture that would continue for twenty-five years. THE BRUSSELS CONGRESS OF 1910 (3RD CONGRESS, 2ND CODE) Aside from the nomenclature section, there was also an important section on bibliography and doc- umentation. The paleontologists, phycologists, and mycologists made many proposals. Harms’s pro- posals on conserving fern names passed. Briquet tendered his resignation but, when everyone was upset, got what he really wanted, help. Harms was elected vice-rapporteur, the beginning of that of- fice. The 1915 London Congress was never held, because of the outbreak of World War I in 1914. Two people were particularly important in the years soon after World War I: Albert Spear Hitch- cock (US) and Thomas Archibald Sprague (K). Each published initiatives that would ultimately mesh. In April 1919 Hitchcock presented some rules for fixing types of generic names that would (1921.04) become the **type-basis" Code. Did you realize that conservation of generic names, passed at the 1905 Vienna Congress, still did not involve listed types? Sprague’s initiative was to publish (1921.06) a юм иы paper on plant nomenclature titled ome suggestions." This appeared in the Journal of Botany with the editor's (James Britten) invi- tation for comments. The paper stimulated a series of pleasant and unpolemic contributions from (mostly) Americans and British botanists that strongly signaled compromise and recognition that unity in nomenclature was better than ongoing strife. In July 1924, Alfred Barton Rendle presided over a discussion at the Imperial Botanical Con- gress of 13 points about the Code published by a committee convened by T. A. Sprague (1924.03). One point still seems current, “The Rules are too long and complicated." The conference resolved several interesting points: that Latin descriptions should not be required, all later homynyms should be rejected, the type method should be adopted, and duplicate binomials (tautonyms) should be re- stored. In Hitchock’s reliquiae at US I found two un- published circulars worth mentioning. One is Hitch- cock’s circular (six pages, not counting cover letter) to American botanists interested in nomenclature dated 15 October 1924 concerning the practica- bility of compromise. The other is by Fernald and Weatherby dated 8 December 1924, apparently also widely circulated in America, which discussed the 12 proposals made by Hitchcock. These un- derlie Hitchcock’s compromises (1926.05). ITHACA CONGRESS OF 1926 (4TH CONGRESS) In August 1926, Sprague, Hitchcock, and Bri- quet showed up at the Ithaca Congress. Sprague’s paper was extremely erudite and offered the basis for a world-wide agreement on nomenclature, com- bining the best features of the International Rules and the Brittonian (American) Code. It is sad that these important proceedings were not published until 1929 (in Duggar). Hitchcock also made concrete proposals, ar- "standard list of accepted generic names, each with the species that shall direct its application." This would begin with con- served names “‘but would be gradually extended to include also all generic names in use." This sounds like a precursor of the “names in current use" effort, which is still divided on whether types should be included. A roundtable discussion was held with papers read by Briquet (review since 1910 with sugges- tions for the upcoming 1930 Cambridge Congress), Marshall Howe (“I don't feel so irreconcilable as I may have"), and Merritt Lyndon Fernald (“Кеер Latin"). A paper from N. L. Britton was read for him (“nothing is constant but change"). The report (in Duggar, 1929: 1782) by Karl McKay Wie- gand, secretary of the taxonomy section, noted the guing vigorously for a 40 Annals of the Missouri Botanical Garden approval of an International Interim Committee with John Briquet as chairman, H. Harms, vice- chairman, with more than 20 named members from all over the world. All nomenclatural resolutions were referred to this committee. CAMBRIDGE CONGRESS OF 1930 (STH CONGRESS, 3RD CODE) The synopsis of proposals cited 20 documents, most typewritten. One of the most important was a 203-page document (1929.08?) by “British bot- anists.”” It included Hitchcock and/or Green's lists of generic names (then 458 currently conserved and 1,159 Linnaean) proposing that the application of these names be controlled by means of the spec- ified ““standard-species.” Sprague's list of proposed conserved names is the first truly documented ap- proach to proposing names for conservation, hith- erto a matter of citing competing names and places of publication (both in proposals and in the Code) and nothing more. Briquet's synopsis was, as before, a model of clarifying confusion. The multicolumn approach continued, but now with only two columns. The first column had the original text of the Brussels (1912) Code and, in italics in the second column, the rapporteur's comments on the proposed new text. The votes of the Commission on the proposals appeared in a second document “Avis préalable" that was handed out at the Congress (apparently rare). I will not attempt to summarize or highlight the debates from proceedings beyond saying that the idea of using types to determine the application of names (rather than the original circumscription) was accepted, and the Latin requirement (regarded as arbitrary by Britton) was moved forward from 1908 to 1932. This ended the American schism. There was an interesting discussion (cf. Brooks & Chipp, 1931: 544-553) about matters that since have come to pass or remain needed, such as Harvey Hall's proposed “International Bureau of Plant Taxonomy," making and distributing pho- tographs of types, and producing an index to col- lections in different herbaria (Ramsbottom). Unfortunately, John Isaac Briquet died in late October 1931, and the task of producing the third edition of the Code (in three languages) fell to Harms, Rendle, Mangin, Hochreutiner, and Sprague. Sprague (1933) published the main pro- visions of the amended Code with examples from the British flora. Rendle (1934.06) published an English text (without appendices). The official Cam- bridge Code appeared about February 1935, only months before the September Amsterdam Con- gress. The Cambridge Code, despite its extremely short life, was the first completely international ode. AMSTERDAM CONGRESS OF 1935 (6TH CONGRESS) Briquet's death had delayed the Cambridge Code which, in turn, shortened the lead time for revisions at the next Congress. In September 1934 an an- nouncement appeared in several journals (J. Bot. and Bull. Misc. Inform.) that proposals to amend the (still unpublished) Code had to be submitted in 100 copies by 1 January 1935. (That should slow things down!) Nonetheless, indefatigable botanists produced 14 submissions and, by July 1935, Sprague sent the synopsis of proposals to the seven members of the Executive Committee of Nomen- clature elected at Cambridge and to seven others, obtaining 10 votes. These were used to compile the Preliminary Opinions distributed at the Con- gress in September. Among the submissions processed by Sprague, I have had most occasion to consult the 43 col- lected proposals *by twelve botanists" that ap- peared in 1935.03. I would draw attention to Wil- mott's /ndex Purgatio, listing works that should be rejected for various cited reasons, and Sprague and Green's provisional list of institutions in 20 geographical areas to get copies to validate names. English now became the official language, re- placing French, which had been official since Brit- ton's proposal to the 1905 Vienna Congress was approved. The preceding Brussels and Cambridge proceedings had been reported in French, but prob- lems arose with the unexpected death of Briquet. Harms, the vice-rapporteur, asked Rendle to pre- pare the English text since most of the proposals had been in English. The French and German texts were generated by translation from Rendle's En- glish text. Sprague thanked Miss Mary Letitia “Manna” Green for her collaboration (in fact, he married her). Sprague asked that the 1935 Cambridge Code be recognized as a faithful record of the 1930 Cambridge decisions. This was wise in view of the great difficulties after Briquet's death, and it would have been even more futile than usual for nomen- claturalists to argue whether or not a proposal had een implemented correctly. He did ask that the Latin requirement date be moved forward again, from 1932 (when the Code had been expected) to 1935 (when the Code came out). He also proposed automatic rejection of proposals with less than 20% support in the preliminary (Commission) vote. Probably the major battle of this Congress con- cerned the application of a combination when mis- Моште 78, Митбег 1 1991 icolson 41 History of Botanical Nomenclature applied, the classic example being Tsuga merten- siana (Bong.) Carr. Carriére based his name on Bongard's Pinus mertensiana but applied it to an excellent illustration that was not Bongard's spe- cies. This is a classic problem of the old circum- scription method tangling with the new type method (just introduced), not to mention the whole question of whether or not parenthetic authors should be cited that had so vexed nomenclaturalists of the preceding century. Hitchcock died on e while returning from the Amsterdam Congre Now World War II а Not only was there no Stockholm Congress in 1940, there never was an official Amsterdam Code. Until the next Congress (Stockholm in 1950) people had to make do with 1935 (Cambridge) Code and what they could find about the Amsterdam Congress (fortu- nately not too drastically changed, but the type method was new to many) mostly from Sprague's summary in 1936 and the Congress proceedings. Approved conserved names were listed by Sprague (1940.06) BRITTONIA CODE (UNOFFICIAL) In the late 1940s steps were taken to revive things, just as after World War I. The word went out in 1946 that there would be a Congress in Stockholm in 1950. In March 1946 the Council of the American Society of Plant Taxonomists ap- pointed a committee (Camp, Rickett, and Weath- erby) to prepare for the Congress. They produced the unofficial ““Brittonia”” Code (1947.04), followed by a symposium on botanical nomenclature in De- cember 1947 at St. Louis (published 1949.01). Weatherby’s paper, “Botanical nomenclature since 1867," said everything that should be said. Meanwhile, in February 1947, Professor Lan- jouw (Utrecht) wrote to several botanists and, hav- ing gotten support, invited taxonomists to a June 1948 conference on preparing a new Code. The results were not published until Пико) 1950 but they had dramatic consequences before. Among the Americans at the Utrecht Conference were Merrill and Rickett, who were already involved in the American initiative. Ray Fosberg was there (and attended the thesis defense of Frans A. Sta- fleu). Lanjouw agreed to be acting rapporteur gé- néral, succeeding Sprague. STOCKHOLM CONGRESS OF 1950 (7TH CONGRESS, 4TH CODE) Lanjouw’s synopsis of proposals was large (255 pages) and was mailed with the first voting form, requesting return by 1 June 1950. I note that A. C. Smith received his copy on 5 June, so he did not return it. There were only 540 proposals to process in 40 hours (4 minutes each). Two hundred mail ballots were received and anything with less than 25% support was automatically rejected. An important event occurred on 18 July 1950 at what Stafleu (1988.08: 795) called “an informal session” (attended by 130 taxonomists). Lanjouw's proposals for an association with an office (bureau) in Utrecht was accepted and became Resolution 10 of the Congress (pp. 67, 68). This was the official beginning that turned into things like the Taxon and Regnum Vegetabile. Taxon, being the official journal of the new International Association for Plant Taxonomy, became the vehicle for pro- posals to amend the Code so that they no longer were scattered through the literature. I would be astonished if any working taxonomist does not con- sult something in Regnum Vegetabile almost every day, be it Index Herbariorum, TL-2, ING, or even the Code. The first issue of Taxon appeared in September 1951. The first issue of Regnum Vegetabile (ac- tually numbered 3) was the Stockholm Code, which appeared in September 1952, barely in time for the Paris Congress. The editing of the Stockholm Code was described by Merrill (1952.01). The Stockholm Congress proceedings came out in 1953. Panis CONGRESS OF 1954 (8TH CONGRESS, 5TH CODE) This was the first Congress to have all proposals published in a single journal. The Geneva Confer- ence of 25-30 January 1954 was a precursor event, as the Utrecht Conference of June 1948 before the Stockholm Congress. Its report appeared in April 1954, soon after the synopsis. Stafleu’s report on the proceedings appeared in August 1955. A feature of the 1956 (Paris) Code was the ap- pearance of a fourth language, Spanish. Another was a concordance of Article and Recommendation numbers among the latest Paris Code and the two preceding Codes (Stockholm and Cambridge). It would be useful to extend this concordance back through the three preceding Codes, Brussels (1912), Vienna (1906), and Paris (1867). Some- how this should then weave in all proposals to the same text (including rejected ones) and the ex- amples proffered. 1 am convinced we would find we are constantly reinventing the wheel, if not repeatedly raising the same examples discussed before. There is neither time nor space to continue since 42 Annals of the Missouri Botanical Garden we are шы half way through and have six more Codes to go, the products of the subsequent Con- gresses: Mod (1959), Edinburgh (1964), Se- attle (1969), Leningrad (1975), Sydney (1981), and Berlin (1987) and 1 was somewhat distracted by being on my honeymoon. My best memory of Edinburgh was the “Bloody Donk” speech. At Seattle I made my first motion on the floor (defeated) and have been told that this was the last ““decent” Code. Leningrad passed my orthography proposals, probably assist- ed by my absence when discussions started, for the first time, with proposals pertaining to the last articles of the This was the first Code to number paragraphs within articles, making it eas- ier to cite. Sydney was the first Congress most of us experienced without Stafleu in the chair and "mice will play when the cat is away." Stafleu rejoined us for Berlin (1987), but said that he will not go to Tokyo (1993 We must end on an upbeat, and І can do no better than repeat Weatherby's (1949.01: 7) translated quotation from Alphonse de Candolle's introduction to his 1867 Lois (with a few minor . Montreal was my first Congress ode. changes since nomenclaturalists can't even quote each other without making ‘“‘improvements’’): “There will come a time when all the plant forms in existence will have been described; when her- baria will contain indubitable material of them; several hundred thousand taxa ranging from classes to simple varieties, and when synonyms will have become much more numerous than accepted taxa. Then science will have need of some great reno- vation of its formulae. This nomenclature which we now strive to improve will then appear like an old scaffolding, laboriously patched together and surrounded and encumbered by the debris of re- jected parts. The edifice of science will have been built, but the rubbish incident to its construction not cleared away. Then perhaps there will arise something wholly different from Linnaean nomen- clature, something so designed as to give certain and definite names to certain and definite taxa. “That is the secret of the future, a future still very far off. “In the meantime, let us perfect the binomial system introduced by Linnaeus. Let us try to adapt it better to the continual, necessary changes in science ... drive out small Pp the little neg- ligences and, if possible, come to agreement on controversial points. Thus we shall prepare the way for the better progress of taxonomy.’ CHRONOLOGIC BIBLIOGRAPHY ON BOTANICAL NOMENCLATURE, WITH SOME ANNOTATIONS The entries in this bibliography are arranged chrono- logically to better reflect the subject, history. The bib- р : as Stafleu and Cowan's Taxonomic Literature 1737.07. LINNAEUS, C. Critica Botanica. 270 pp. Lugduni Batavorum. [Arth M questioned by Heller (1964. 04). Note: As Miss Gre aid, w Linnaeus spoke of specific names, le meant the diagnostic d in the binary сна self, later adopted i in his deii system.] 1753.05.01. LINNAEUS, C. Species Pape .... Hol- ппае. 1798.11. ГАМАКСК, J. В. А. P. MoNNET ре. Nomencla- ture. Pp. 498-499 in Encycl. Meth. Bot. Volume 4. H. Ag , Paris. [Complained of abuses, such a Linnaeus's converting Syringa into Philadelphus, that are and those derived from names o places or пи (the latter were ОК for genera).] 1813. CANDOLLE, А. P. DE. De la nomenclature. Pp. 221- 52 in Théorie Élémentaire de la Botanique .... Paris 1821.01. “SMITH, J. E. Respecting Nomenclature .... Pp. 51-54 in A Grammar of Botany. Longman et al., London. Reprinted by Prakash Publishers, m in 1973. [Objected to corrupt names. “Future gen- eral writers on Botany, of competent ен must 7) ture and Terminology (pp. 454-459) and V Of Sy onyms 460 DC. uses asteriks (*) in syn- onymy = good description. “Mark of admiration" (!) indicates inspection of authentic specimen — thus, "Linn.!. sp. pl. 427."] 1843. зад Н. E. (R с x PHILLIPS, J. RDSO w W. J. BRODERIP, 12th Meeting of the British Association for the Ad- dU ен | е held at Manchester in June 2, pp. 1843. | ыч A A. Notice of some works, recently Volume 78, Number 1 1991 icolson 43 History of Botanical Nomenclature published, on the nomenclature of zoology. Amer. J. + 1-12. ci. Аг! 1860. STIMPSON, W. botanical and Ерани nd clature. Edinburgh New Philos. J. 12: 324. 1864.03. GRA ser. 2, 37: 279-281. ted p Pepe zoological rules (Edinburgh New Philos. J. 1863: —).] 1866.06. BE "Koc OCH, K. Some propositions with respect to systematic botany. J. Bot. 4: 201. [Abstract, trans- lated from German. 1866.?? KocH, K. Einige die Systematik betreffende Vorschlage. Pp. 188-195 in [Secretary?], Int. Hort. Exhib. Bot. Congr. London 22-31 May 1866 Rep. Proc. [total pagination?] Truscott, Son & Simon, London. [English summary appeared on pp. 21-22.] 1866.12.01. [SEEMANN, B.?] [Review of] Prodromus Na- turalis Regni Vegetabilis. Editore Pars. XV., Sectio Posterior, Fasc. II. Sist phorbiaceas. Auctore J. Müller, icu dier dun is. 1866. J. Bot. 4: 387-388. [Objection: he includes forms that had before nd кы, p» displaces the name of the author of the species, and attaches his own to it."'] 1867.01. GRAY, А. Ап innovation in nomenclature in the recently-issued volumes of the **Prodromus." Amer . Sci. Arts, ser. 2, 43: 126-128. [Objected to п Mueller changing the authorship of generic name when ч nging the circumscription.] шы by Seemann: J. Bot. 5: 81-84. 1867.03. 1867.05. 01. DONE: A. L. DE. о соп- cerning criticisms of] Dr. Mueller's Monograph of the Euphorbiaceae. J. Bot. 5: 151-152. 1867.08.01. CANDOLLE, А. L. DE. Lois de la Nomencla- ture botanique redigées et commentées . . . . 60 pp. V. Masson et fils, Paris. Lacs il басне. ] 1867.09.12. Pon: A. L. DE. — € botanique adopteés par le ades Aoút 1867 .... 64 pp. H. ie Genève бя Bale, J. B. Bailliére et fils, Paris. 1867.11. CANDOLLE, A. L. DE (RAPPORTEUR). Discussion des Lois de la Nomenclatiye botanique. Te 208 bx Lois de la Nomenclature botanique. Pp. Ju 255 in E. Fourneri en Actes du Congrés я ае Воѓа .... Рап 1867. H СКЕ Е. La Nom шш роја не au Congrés ei ard de Botanique. 28 pp. [The па citation argument ad nauseam.] 1868. Des MOULINS, C. Lettre а Monsieur Francois Cré- m 10 pp. Bordeaux. [Reprinted from Actes Soc. n. Bordeaux 26(4).] 1868.7 " CAN NDOLLE, А. L. DE. Laws of Botanical Nomen- nd [English translation] by H. A. Weddell. 72 . Г. Re о., 1868. 07. Gray, A. Editorial ik and suggestions [on the laws of botanical nomenclature.] Amer. J. Sci. Arts, ser. 2, 46: 74-77. [Preceded (pp. 66- 74) by a copy of к! English translation. 1869.06. CANDOLLE, А. L. DE. nse à diverses ques- tions et ёзен faites sur le Recueil des Lois de Nomenclature botanique, tel que le Congrés inter- national de 1867 l'a publié. Bull. Soc. Bot. France 16: 64-81. 1870.01? CARUEL, T. & А. Г. DE CANDOLLE. [Corre- spondence.] Una questione di nomenclatura botanica. Nuovo Giorn. Bot. Ital. 2: 146-149. [Names pub- lished with a query. 1874.02-04. MÜLLER, J. Nomenclatorische Fragmente [with six parts.] Flora 57: 89-94; 119-126; 156- 159. [Motion 1 in Briquet (1905.03.15: 7).] I. e die Gültigkeitsbedingungen der systema- schen Namen. P -94. П. Ueber ө Citiren der edo bei generisch neu gestellten Arten. Pp. 119-121 III. Ueber die Autoritàt bei i alo ese Zettelnamen. Pp. 121-123. P. IV. У das Autorschema bei umgeánderten Gat- ngsbegriffen. Pp. 123-126. V. рза - Ableiten der Speciesnamen von Var- ietátsnamen. P. 156. VI. Verliert ein systematischer Name seinem Rang so rt er zugleich sein Prioritatsrecht. Pp. 156-159. 1875.01. CeLakovskY, L. Zwei Fragen der botanischen Nomenclatur. Flora 58: 2-6; 21-31. [Priority right of species epithets and generic names. 1876.12? COGNIAUX, C. A. & A. Г. DE CANDOLLE. [Cor- respondence.] Quelques points de nomenclature bo- tanique. Bull. Soc. Roy. Bot. Belgique 15: 477-485. Fist appearance of the principle “Never attribute n author what he didn't sa y. 1877.06. TRIMEN, H. Some points in botanical nomen- clature. J. Bot. 15: 189-190. [Supports de Can- dolle's position in the published Cogniaux correspon- dence (1876.12?). First mention of what will be known as Kew Rule on p. 190.] 1877.08. CANDOLLE, A. Г. DE. [Letter & response by editor, Trimen.] J. Bot. 15: ud 243. [Opposing the Kew Rule; defended by Trim 1877.09. CARUEL, T. Botanical spans DEN J. Bot. 15 282. [Supported Kew Rule as espoused by Tenants 5 response to de Candolle (1877.08) 1877.10. CARUEL, T. Divisiones plantarum propositae. Nuovo Giorn. Bot. Ital. P: 280-281. [Phaneroga- mae, Schistogamae, Prothallogamae, Bryogamae & Gymnogamae. 1877.12. Batt, J. On some questions of botanical no- menclature. J. Bot. 15: 357-360. [Supported de Candolle (1877. J^ opposing the Kew Rule.] nclature in zoology and ashville Meeting, August 31, 1877. 56 pp. (A.A.A.S., Volume XXVI). [Important реши а of zoologists aud Code with ole ments for botan 1878.03. HIERN, W. P. On a question of botanical no- menclature. J. Bot. 16: 72-74. [Supports the Kew ule. 1878.05. Ваш, J. On MP ma of botanical pica pi J. Bot. 16: 140-142. [Balanced dis- cussion of Kew Rule, basically opposing it. 1878.06. Tous “н n a point in botanical nomencla e. J. Bot. 16: 170-173 [‘‘... the framing of fixed rule ractice is, to employ the earliest older aps na 1878.09. re W. Bot sd HERREN ature. J. Bot. 16: 260-262. adio ne zoological method of author citation. bs и the нус la of the specific + 44 Annals of the Missouri Botanical Garden 1878.11. CANDOLLE, А. L. DE. Botanical borne) J. Bot. 16: 345-346. [Drew attention upco in p (1883.07?), subsequently wines 10). in English Е Britten (1888. 1878.12. Bera M, G. Nomenclature. Pp. 189-198 bs his] Notes o pem eae. J. Linn. Soc., Bot “the rules Lon is О custom amounts to prescription . “ when priority should be set aside, as for Pack o but accepted Adanson. Opposed “correcting” original си Supported what at will be n Kew Rule. Excerpted by J. Bot. 17: .] sa т of Bentham's “Notes on Eu- phorbiaceae." Amer. J. Sci. Arts [Extracted ES Bot. гечи а 4: 158-161. 1879.05; J. Bot. . 1880. Ts 6 с А М. & М. S. COULTER. чш in scientific nomenclature . (€ . Bot. Gaz betw essive (i КЕ (adiectis di Шш: Worshipful of Asa 1879.08 01. MÜLLER, К. von. Einige Worte über die erste Ausgabe von Linné's Species Plantarum in Be- zug auf Vorzugsrecht. Bot. Zeig. (Berlin) 37: 490- 491. 1879.08.01. DRupE, O. Über Nomenclaturfragen. Bot. Zeit. (Berlin) 37: 492-494. 1880.03? SAINT-LAGER, J. B. Réforme de la nomenclature botanique. Ann. Soc. Bot. Lyon 7: 1-154. ; Origines de la Mau ف‎ [incl. Nomenclature des Grecs.]. Pp. 1- . Réforme T eis cine qui ne s’accor- dent pas avec le nom générique. Pp. 38-60. . Réforme du épithètes spécifiques qui m pléonasme avec le n m générique. Pp. 61-66. . Réforme des noms eara d'un radical grec associé à un radical latin. Pp. 67-72 . La Nomenclature botanique est formée de noms 73- id сл — دب‎ ~ == © Y c > с. e. e a] Ф .8 о O a ms génériques. Pp. 111. [With lists of ancient Greek names E without change of inflection, with changes, mod- Greek stems of which the last is masculine, neuter, and feminine. . De la désinence des épithétes spécifiques. Pp. 112-138. . Des épithétes spécifiques barbares [or based on a person's name, or banal, geographic or recalling industrial ог medicinal properties]. Pp. 139- 150. -~J со Conclusions. Pp. 1 Ha 04. SAINT-LAGER, nomenclature botanique. Ann. 3. -154. J. B. Nouvelles remarques sur la Soc. Bot. Lyon 8: — — + . Du genre grammatical des mots ““Lotos et Meli- lotos." Pp. 149-156. e la désinence des noms de genre. Pp. 156- 3. d Remarques sur le genre grammatical w noms génériques d'origine grecque. Pp. 164-169. Des noms de genre tirés d'un nom d’ atas Pp. 170-178. . Orthographe de quelques noms de genre. Pp. 79-183 > сл [ey . Des pléonasmes. Pp. 184-188. 7. Des noms composés. Pp. 188-190. 8. L'adjectif s'accorde avec le substantif auquel il se rapporte. Pp. 191-203. 9. uh cura te ipsum." [Corrections to this and receding work. ] P. 203. 1881. оно В. n some recent tendencies in tanical nomenc catu ure. J. Bot. 19: 14-83. [Ob- erar to parenthetic author citations, “corrections,” faulty capitalizing of species epithets, and mischie- vous author abbreviations. 1882.06. GRAY, A. The citation of botanical e J. Bot. 20: 173-174. [Advocated crediting name in Aiton’s Hort. Kew. to Aiton, not EE R. Brown, etc.] 1883.07? CANDOLLE, А. L. DE. Nouvelles remarques sur la nomenclature botanique ‚... Supplém 1883.10.31. BUBANI, P. Su di alcune divergenze intorno la nomenclatura botanica ed il compilatore delle proteste Leggi Prof. Alf. de Candolle 15 pp. Bologna. [Objected to the tyranny of law 1883.12. Gray, A. Some points in ed nomencla- ture; a review of **Nouvelles Remarques sur la No- menclature Botanique, par. M. Alph. de Candolle,” Geneva. 1883. Amer. J. Sci. Arts, ser. 3, 26: 417- 1887.03. Jackson, B. D. А new “Index of plant-names." J. Bot. 25: 66-71 (Jan.); 150-151 (May). [Kew Rule spelled out on p. 69. Statement from J. Hooker bout the origin of the Index.] 1887.06. Wirrmack, L. Règles à suivre pour la nomen- claturae des plantes en général, et des Orchidées en particulier (5* question au Congrés de 1887). J. Soc. Na n Md . France ser. 3, 9: cvii-cxxiii. [Seen as 16 pp. reprint.] 1887. 10. GREENE E. e permanency of specific names. J. Bot. 25: 301- 303. [Objected to the Kew Rule.] ы 12. GRAY, A. Botanical nomenclature. J. Bot. 25: 53-355. [Attacked e by quoting Bentham “ supported the Kew Rule.] 1888.01. Asa Gray died at тв 1888.04. РОССЕХВЕКС, J. . L. BRITT E. E. STERNS, A. BROWN, T. 23 Po ORTER 4 "i “Hor LLICK (COMMITTEE OF TORREY BOTANICAL CLUB). Preliminary ern enberg (subcommittee), beginning of the Brittonian (American) schism. ] EN L. Botanical no old and new. Pittonia 1: 176-194. [177-183 reviews a Ray 1660 publication with 48 binomials. 184-194 reviews the new preliminary catalogue by Poggenberg et al.] 1888.09. BRITTEN, Recent tendencies in American botanical nomenclature. J. Bot. 26: 257-262. [Ob- ее to Br dc Sterns & Poggenberg overturning nomenclature еа and the 1888.09. umm h : 230- ЈЕ. Е. Тће ow to settle it. Bull. Torrey Bot. Club 1 235. 1888.10. CANDOLLE, A. L. DE. [Letter concerning] B tanical nomenclature. J. Bot. 26: 289. [Agreed zi Britten's (1888.09) protest about Britton et al. and Volume 78, Number 1 icolson 45 History of Botanical Nomenclature — Asa Gray, about whom Americans ought о be proud and follow. hs pposed the Kew Rule.] Ls 10. BRITTEN, = Translation of Art. 48 discussion n A. Candolle's **Nouvelles Remarques. ”J Bot. 26: 290- ete 1888.10. е. . L. [Response to Britten's protest of 1888.09.] J. Ви, 26: 292-295. [Argued that ће only E what Elias Fries, Boissier, Richard Spruce, and a host of others have done (he was wrong about m originalexemplars нен vid priritesfrágor Bot. Not. 1891: 76-8 1891.10. KUNTZE, O. Revisio Generum Plantarum Vo lume I- 2. 101 1 рр. (+ 155 introduction). [A The pela анша eli nebst Motiven botanischen Nomenclatur- Р. li. -cxxii, accepted as Mo- tion 3 in Briquet (1905.03.15: 7) and as Motion 2 in Briquet (1910.04?: 1).] 1891.11.27. Drupe, О. Bemerkungen zu Dr. Otto н s Aenderungen der systematische Nomencla- r. Deutsch. Bot. Ges. 306. 1892. 02. 04. BRIQUET, J. Zur a Nomenclatur der Labiaten. Bot. Centralbl. 49: 106-111. [Bri- quet's maiden nomenclatural paper, what Kuntze did ts. 1892.05. CANDOLLE, А. L. DE. А note on nomenclature. J. Bot. 30: 135. [Noted Briquet’s evidence against Kuntze’s changes of Labiates and avowed great pleas- ure "dans mes vieux jours” with the support of principles he had always espoused.] ToS "a MRA, E. Quatre propositions relatives r un Comité de botanistes Berlin; d'approbation de M. ee se de Candolle. Bull. Soc. Bot. France 39: 13 2 [D e Candolle’s letter was dated 6 Jul "ш. опе La his last published comments on ] olutiane, English J. Bot. 30: 241- 242. 1892.08. Motion 4 in Briquet (1905.03.15: 7).] 1892.08? GREENE, E. L. Dr. Kuntze and his unies Pittonia 2: 263- 281. [Reprinted in Chron. Bot. 249-257. 1951. = Statement by one of E 5 strongest supporter 1892.09. FAIRCHILD, p. 6. SECR.). Proceedings of the 1892. Bull. Torrey Bot. Club 19: 281-297. [Pp. N. L. Britton, J. M. Co Ward Cot on Botanical Nomenclature). Мо. поп 5 in Briquet (1905.03.15: 7).] 1892.11.15. UNDERWOOD, L. M. The nomenclature ques- ion at Genoa. Bull. Torrey Bot. Club 19: 324-330. у ће le Congress t. Gaz. (Crawfordsville) 17: 341-347. 1893.07?- 10? KUNTZE, O. Die Bewegung in Е bota- nischen Nomenclatur von Ende 1891 bis Mai 1893. Bot. Centralbl. 54: 353-361; 385-408 1893.04.04. Alphonse de Candolle died at 86. 1893.04? Ремс, O. (SECR.). Quarta adunanza, Martedì Settembre alle 2 pom. Atti del Congresso Botanico Internazionale di Genova 1892. 81-126 [Interna- tional Commission to consider four pon of re botanists announced on 9 Sep. 440. 117-121 accepted as Motion 7 by Brian (1905.03.15: 7). Holmes’s emendations in botanical terminology (pp. 121-124) accepted as Motion 8 by Briquet (1905.03.15: 7).] 1893. AscHERSON, P. Die Nomenclaturbewegung von 1892. Bot. Jahrb. 15(Beibl. de 20-28. [Also in Ber. Pharmac. Ges. 4: 35-45. 1893.06. Jackson, B. D. Index аини [2nd fascicle in December.] 1893.07. Кин, har Revisio Generum Plantarum 3(1): CIX— ccccexxil datus (pp. neue as ccexvi) е as Motion 9 in Briquet ae 8) and as Motion 3 in Briquet (1910.042: 1).] 1893.09. ла , У. T. А.А lst fascicle (SECR. ). Proceedings of the d as Motion 10, without Art. II, by Briquet (1905.03.15: 8), у Britton's report of same дате. ] 1893.09. Britton, М. Г. Proceedings of the Botanical Club, A.A.A.S., Madison meeting, August 18-22, 1893. Bull. Torrey Bot. Club 20: 360-365. [Pp. 360-361 accepted as Motion 10, without Art. II, by Briquet (1905.03.15: 8), also Swingle's report of same date. 1893? Om en falles Nomenclatur i systematik Botanik for Skandinavien. Fórh. Skand. Naturf. 14 Moede 2 sau 255. [Motion 6 in Briquet (1905.03.15: . Not я 1893. Fraen Botaniki Sällskapets i Stockholm förhand- lingar. Bot. Not. 1893: 158- [Motion 11 in Briquet (1905.03.15: 8). Not see 1894.02. BRIQUET, J. Questions de du CN Bull. erb. Boissier 2: 49-88. [Motion 12 by Briquet (1905.03.15: 8).] 1894.07. KUNTZE, O. Nomenclatur-Studien. Bull. Herb. Boissier 2: 456-498. 1894.12.28. KNoBLAUCH, E. Die Nomenclatur der Gat- en. Bot. Centralbl. 61: 1-6. [Motion 13 in Briquet (1905.03.15: 8).] 1895.01.15. BARNHART, J. H. Family nomenclature. Bull. Torrey Bot. Club 22: 1-24. [Important document (see author's mature comment 1922.09). Motion 14 ENGLER. Erklárung der Gescháftsleitung der vom internationalen botanisch- en Congress zu Genua (1892) eingesetzen Nomen clatur-Commission. Oesterr. Bot. Z. 45(1): 21-35. [Motion 15 in Briquet (1905.03.15: 8 as pp. 327- 335). Also in Verh. Ges. Deutsch. Naturf. 2: 153- 159. 1894? 1895.03. ROBINSON, В. L. On the “List of Pteridophyta ature Committ tanical Club. Bot. Gaz. TUR 20: 97-103. 1895.04. CoviLLE, К. V. A reply to Dr. — 8 эн icism of the “List of Pteridophyta and Sper of y aN America.” Bot. Gaz. айгы ч 20: 1895.04. n EUER E. Beitráge zur Systematik der ideen. I. Zur Nomenclatur. Bot. Jahrb. 19: 1- 1895.05? KUNTZE, О. Bemerkungen zum künftigen bo- 46 Annals of th Missouri Botanical Garden FRE Nomenclatur-Congress. 5 рр. | m Oesterr. Вог. Z. 1895(5): ??-??, not see 1895.06. RoBiNsoN, B. L. On the applicatio n of “once ' to binomials [with a о ature of ra botany.] Bot. Gaz. үз унны. 20: 261-263. 1895 co CoviLLE, F. V. Dr : en and homonyms. t. Gaz. (Crawfordsville) 20: 1895. 07? SMITH, E. Е. The Bud с CheckList: А protest. [Published by ашћог 2] 1 1895.07. Warp, L. К. The idi ran question. Bull. Torrey Bot. Club 22: 308-329. е of history of e ai and for pea 1895.08. ANDERSO O AMERICAN BOTA- NISTS. е ан the nomenclature of systematic botany. Bot. Jahrb. 21 (Beibl. 52): 12- 15. br | school's loyal opposition, called the Harv e. 1895.08. p neos N, B. L. A further discussion of the Madison rules. Bot. Gaz. (Crawfordsville) 20: 370- 371. 1895.09. ORCUTT, С. К. Nomenclature. Out of Doors for Women 3(22): 1-2. [Mam(m)illaria vs. Cactus. ] . Kuntze, O. Les besoins de la nomenclature а “Nomenclator plantarum omnium correctus" y 1905 1895. " нетто» Dyer, W. Nomenclature. Bull. Misc. Inform. 1895: 278- 281. [Р. 280, “То me the bot- anists и waste bei time over priority are like boys who, when sent on an errand, spend their time in pla ying by ii roadside." . J. E. La loi de priorité dans la soc. Fran ncem. Sci р. [Seen as reprint Gave ien in Alyssum and Carex that see i y today’ 8 ол ѕ.] HORS. The nomenclature question. Bot. em (Cranford 21: 82-91. 1. KN E inconsistencies in plant ыз. PP- 2. DAVENPORT, G. E. ae nomenclature. Pp. 85- . KUNTZE, 0. Some remarks on nomenclature. Pp. о . MEEHAN, T. Dates and ico and priority in e ture. Pp. 9 1896. LE Joris, A. Lettre . Bot. France 42: 661-663. 1896.04.12. BUCHENAU, F. С. P. [Untitled.] Bot. Jahrb. 24: 648-668. | gun on nomenclatural € 1896.06. LEVIER, E. La pseudo-priorité et les n "e Bull. Herb. Boissier 4: 369-406. ут n Kuntze's wor 1896. 08. BRIQUET, J. Questions de nomenclature. 14 Рр. {Reprinted from Observations préliminaires. In: E e guy > га e und Bull. Soc. inige (pru Be- kungen. Bot. Jahrb. 22(Beibl. 55): 1 ње 05, а . Die No л. ea letz- e. Bot. Jahrb. 23(Beibl. 56): 1-32. 1897. у ке says 8 Mai]. ENGLER, А., I. URBAN, ж мм, С. HIERON He NGS, M. GÜR , U. DAMMER, Сис, Н. HARMS, P. poem G. VOLKENS & L. DIELS. Nomenclaturregeln für die Beamten des Kón- iglichen artens und Museums zu Berlin. Notizbl. Kónigl. Bot. Gart. Berlin 1: 245-250. [The Berlin Code. Motion 16 in Briquet (1905.03.15: 8). Also in Gartenflora 46: 304-308. 1897.06.01, which Robinson (1897.08) translated into English.] 1897.08. Britton, М. L. [Translation and comments on] Nomenclaturregeln für die Beamten ... zu Berlin. ,B.L. E official оеп не of ni n and Museum of Berlin. Bot 2. (Crawfordsville). 24: 107-110 1897. ов. Бител, Ј. s Berlin Rules for ийнан; 1. . 35: = 1897.08. 25. KUNTZ oO Levier’s Verdrehung von Ar- А 49 дез Pact Codex. Bot. Centralbl. 71: 302- 5. [Protests twisting » the Code.] T ei BRIQUET, J. “Re e Nomenclature pour les botanists ы au jud in et au Musée royaux de Botanique de Berlin" traduites et suives d'observa- tions critiques. Bull. Herb. Boissier 5: sd 779 ENGLER, А. Anacardiaceae africanae. П. Bot. 1898. M BRITTEN, J. The fifty years’ limit in nomencla- Bot. 36: 90-94. [Odina vs. Lannea vs. Calesiam vs. Haberlia. ] 1898.06. ROBINSON, B. L. Some reasons why the Roch- ester Nomenclature cannot be regarded as a consis- tent or stable system. Bot. Gaz. (Crawfordsville) 25: -445. 1898.08.12. Cook, O. F. Stability in generic nomencla- ture. Science n.s. 8: 186-190. [Comments on Rob- inson, 1898.06. 1898.09. KUNTZE, O. Revisio Generum Plantarum. [Last 180- tum" accepted as Motion 17 in Briquet (1905.03.15: 8) and in Briquet (1910.04?: 1 1898.10.29. LE Joris, A. Protestation contre le Reviso en plantarum III". 11 pages. J. Mersch imp., aris. [Reprinted from J. Bot. (Morot) 12.] ps GREENE, E. L. e en difficulties in botany. Catholic Univ. Bull. 4: 62-75. [Presentation of the circumscription method e determining the appli- cation of names. 1898. Levier, E. Le cas du Docteur Otto Kuntze. pp. Florence. [Satirie, pungent, and delightful criti- 5 € and efforts. А test gegen die Schwein- furth'sche Erklárung. e al 77: 259-262. 1899.05.10-17. CELAKOVSKY, L. J. Das Prioritátsgesetz in der botanischen Nomenclatur. Bot. Centralbl. 78: 225-234, 258-268 1899.09.15. ENGLER, А. ET AL. Erklárung. Bot. Jahrb. 7 үө 63): 7. [In reference to Engler et al. (1897.06). Motion 18 in желе (1905.03.15: 8).] 1899. ч 20. KUNTZE, О. Ueber neue nomenclatorische Aeusserungen. Bot. E Ы, 79: 405-412 1899.12. UNDERWOOD, L. M. Review of the genera of ferns proposed prior to 1832. Mem. Torrey Bot. Club 6: 247-283. [Not seen, generic typifications?] 1899. KUNTZE, O. 250 Gattungsnamen aus den Jahren 1737 bis 1763, welche im Kew Index fehlen oder falsch identifiziert sind. Deutsche Bot. Monatsschr. 17: 55-59. 89-91, 107-110. [Seen as unpaged reprint. ] 1899. Voss, A. Nomenklatur und Pflanzenkunde. 8 pp. Volume 78, Number 1 icolson 47 History of Botanical Nomenclature ну. from Gartn. Zentral-Bl. 2. Kuntze sup- r. Attack on 1735 starting d used by Index Kew nsis. English translation by Kuntze 1900.01.] 1900.01 Kuntze , О.Т 1900.02.20. KUNTZE, O. О UE des Botanischen Vereins der Provinz Branden- p. San Remo. 1900. 02 KUNTZE, О. A plea for my 1737 proposal. J. id du 47- и [With J. Britten's editorial com- n рр. ] 1900. 03? i 0. e und Reform internationaler Kongresse. Deutsche Bot. Mon- atsschr. 18: 33-47. [Reprint seen, 4 pe] © 1900.09? WETTSTEIN, R. VON. Der internationale bota- nische Congress in Paris und die Regelung der bota- nischen ear 5 pp. [Reprinted from Oesterr. Bot. 2. 1900 (9): ??-??] 1900.06-09. KUNTZE N Posr. Nomenkla- torische Revision hüherer adis und über men-Register. 39 pp. [Reprinted from Allg. Bot. Syst. 1900: 110-120 6; 1. 150-164 1900.07; l.c.: 179-191. 1900.09. Accepted as Mo- tion 19 In Briquet (1905.03.15: 8).] 1900.08.15. KUNTZE, O. Exposé sur les Congrés pour la Nomenclature Botanique et six Propositions pour le Congrés de Paris en 1900. 15 pp. Genéve. [Proposed putting things off until 1905 in Vienna.] 1900.09.28. Cook, O. F. The method ar Зе in botan- ical nomenclature. Science n.s. 1 481. TS ча aa Krause, E. H. L. Reductio generum plan- . Naturwiss. Wochenschr. 15: 613-614. о eae EMILE (SECR.). Actes du l" Congrés In- ternational de Botanique tenu a Paris à l'occasion de l'Exposition Universelle de 1900. Lons-le-Saunier. 1900. HaLLIER, Н. Das prolifierende persönliche und das sachliche, konservative Prioritatsprinzip in der bo- tanischen Nomenklatur. Jahrb. Hamburg. Wiss. Anst. 17: 55-64. 1900. KUNTZE, О. Additions aux Lois de Nomenclature Botanique (Code Parisien de 1867) d'aprés le Codex mendatus. 15 pp. [Reprinted from J. Bot. (Morot) 14: . 1901.03. FERNALD, M. L. Some recent publications and a nomenclatorial principles r represent. z. (Crawfordsville) 31: 183- ior 1 1. FERNALD, М. L. The ire of the Rochester Nomenclature. Bot. c6 (Crawfordsville) 32: 359- 366. [Literature citation 1901. BELLI, S. uni dins sur la réalité des espéces en nature au point de vue de la о des végétaux. [Motion 20 in Briquet (1905.03.1 8). Not seen. 1902.03. SHE . L. Generic nomenclature. Bot. Gaz. (Crawfordsville) 33: 220-229. 1902.04. Cook, O. F. Types and synonyms. Science n.s. 15: 646-656. Alan .] 1902.08.29. ENGLER, А. ET AL. 215 zu den Berliner Nomenclatur-Regeln. Bot. "Jahrb. 31 Beibl 70): 24 25. [Additions to the 1897.06 “pre a а as motion 21 in Briquet (1905.0 8).] 1902.12. CLEMENTS, F. E. Greek and me tin in Biological Nomenclature. University [of Nebraska] Studies 3(1): 1-85. [Excellent exposition of classical practices in- volving word formation, examples, and corrections.] 1902. Krause, E. Н. yi Die Autornamen. Allg. Bot. Z. Syst. 7/8: ?-?. [Unpaged reprint of two unnumbered pages: advocated “Rubus fruticosus (O. Kuntze Ref. Bromb.).” 1903.05.22. Нітснсоск, А. S. An ote on nomenclature. И С binomials with pre-1753 epithet-bringing synonyms.] 1903. MAIDEN, J. Н. The principles of botanical nomen- clature. Proc. Linn. Soc. New South Wales 27: 683- 720. [Excellent survey of then current issues.] 1903. NoLL, Е. Vorschlag zu einer praktischen Erwei- terung de botanischen Nomenclatur. Bot. Centralbl. 14 (Већеће): 374-380. 1904.01.15. AMANN, J. ET AL. (Un Groupe de Botanistes (1905.03.15: ad e ا‎ i et nomencla- uedam. Ann. Mycol. Мак (1904.07.31), English 4.05)] ie 04. BARNHART, J. H. [Review of Post and Kuntze's] hanerogamarum. Torreya 4: 42- P 1904.05. CLEMENTS, F. E. Saccardo: De diagnostica et nomenclatura mycologica; AUR quaedam. Trans slated. J. Mycol. 10: 109-112. ше: = ARTHUR, J. Jj T [o N. L. AN CLEMENTS, O. F. Cook, J. esi F. V. e VILLE, LE S. EARLE, A. W. Коше, Т. Н Е. , А. Номаск, M. A. Howe, Е. H. KNOWLTON, OROE, L. M W.A MunRILL, H. Н. RUSBY, С. L. SHEAR, W. TRELEASE, R р, D e Nomenclature. Bull. Torrey Bot. Club 31: 249-261 (English): 263-276 re: 277- te (German). [Motion 25 in Briquet Wane М. € V. WITTROCK. Motion au aa international de Botanique Deuxiéme pu sion. Vienne 1905. Nyt Mag. Naturvidensk. 217-220. [Motion 26 in Briquet (1905.03.15) ы Мопоп 71 ha = : n 047: 2).] 1904.06.09. AMES, O. L. Propositions de change- ments aux aen de ан biais de 1867 Torre & Harms numbers for conserved generic names Annals of the Missouri Botanical Garden of flowering plants. Motion 30 in Briquet (1905.03.15: 1904.06.25. MaLINVAUD, E. Motion présentée au Congrés international de Botanique de Vienne. 1 [handwritten] page. [Motion 36 in Briquet (1905.03.15: 9). iuc б. о , P. А. Motions supplémentaires рге- au Congrés International de Botanique de Vienne e. l p. [Mo otion 35 in Briquet (1905.03.15: 9) and Motion 8 in Briquet (1910.04?: 2).] 1904.06.23. ARCANGELI, C. ET AL. Adjonctions au Code de Paris 1867, proposées par quelques botanistes italiens. 12 pp. [Motion 33 in Briquet (1905.03.15: 9). French pp. 1-3; Italian pp. 5-8; German pp. 9- 1904.06? Additions et modifications aux Lois de la No- membres de la Société de France. 12 pp. [Motion 34 in Briquet (1905.03.15: 9). 1904.06. Rouv, G. Additions et modifications aux Lois de la Nomenclature botanique de 1867. 4 tion 37 in Briquet (1905.03.15: 9). Preprinted from longer article (12 pp.) in Rev. Bot. Syst. Géogr. Bot. 2: 1904.07.01.] 1904.06. SACCARDO, r A. Des Diagnoses et de la no- ropositions. 6 pp. [French translation by Pe of Saccardo's Latin (1904.06). Preprinted from Boll. Soc. Bot. Ital. 1904: 281- 286. 1904.07.31. Motion 28 in Briquet (1905.03.15: 9).] 1904.06. JACZEWSKI, А. DE (RAPP.). Projet de la revision des Lois de Nomenclature présenté par la Société Imperiale des Naturalistes de Moscou. 10 pp. Mos kva. [Motion 24 in Briquet (1905.03.15: 8) id Motion 6 in Briquet үө 047: 2).] 1904.06. Hay vo zur Regelung der aera "Nomenklatur 1 servations et propositions présentées au Congrés de Vienne . . par un groupe e botanistes belges et suisses. 8 pp. [Motion 38 in шоо a 03 15: А Not seen . Hrrc ‚ А. 5. menclatural pare specimens of plan ж сїез. каш е 21: 828-8 1905.03. 15. Brig Td: Texte Duobus pn documents ase aux débats d . 16 1905.03.21. BRIQUET, J. Resumé du travail du Rappor- teur général et de la Commission international de den Sie Ө Todos 2 t he did since his appointment Nov. 1 1905. 05. Pau, C., J. ЗЕСАМ. & L. МАМА s. Observaciones APPS nomenclatura оташа dela ser er presentadas al ? Congre menclatura Botánica п Viena. AS D ecl hon "Bol. Soc. Aragonesa 1905. 12. 10. Hanes H. Die Nomenklaturbewegung und der Internationale botanische p dA js Wien 1905. Naturwiss. Wochenschr. 20: 785-7 М, К. VON. Bericht und ое international de 4 and more in R. v іеѕпег А. Zahlbruckner, Е des Internationalen Botanischen Kongresses in Wien 1905. 1906.04.10. GREENE, i L. An unwritten law of n clature. Leafl. Bot. Obs. 1: 201-212. [Otiectad in to using ersonal names as part of generic names. Mo- 9 in Briquet (1910:04?: 2).] 1906. 08? BRIQUET, J. (RAPP. GEN.). Régles internationales a nomenclature botanique adoptées par le Congrés International de Botanique de Vienne 1905. 99 . Fischer, Jena. 1906. NoRDSTEDT, O. The starting-point of the nomen- mids. Bot. Not. 1906: 96-106. [Not Bot. 45: 128. 1907.03.] 1907.01.28. Otto Kuntze died at 54. 1907.02.04. ScHINZ A. THELLUNG. [On still-born e па “Flora der Schweiz" von Schinz und Keller. Bull. Herb. Boissier, ser. 2, 7: 97-112. [Motion 12 in Briquet (1910.04?: 2).] 1907.03. Robinson, B. L. On the rules of botanical no- menclature adopted by the Vienna Congress. Rho dora 9: 30-55. 1907.04. па J. C., J. Н. BARNHART, М. L. BRITTON, NTS, О. Е. Соор, Е. V. CoviLLE, Е. S. EARLE, А. W. EVANS, T. E. HazEN, А. Ноошск, M. A. Howe, Е. Н. KNowLrTOoN, С. Т. MOORE, Н. H. Кизву, С. L. SHEAR, L. M. UNDERWOOD, D. WHITE & W. F. WicHT. American Code of botanical nomencla- ture. Bull. Torrey Club 34: 167-178. [Motion 11 in Briquet (1910.04?: 2).] 1907.1 ‚ Н. & A. THELLUNG. (Circulaire) du 10 ОЕК 1907 relative à la question des noms mort-nés. Au pp. [Motion 13 in Briquet (1910.04?: . Not 1907. 12. Penis A. B. & J. BRITTEN. ict on the "List of British seed-plants" II. J. Bot 433- 445. [Still-born names discussion. о о JANCHEN, E. Zur Nomenklatur der Gattungs- ; rwiss. Vereins. Univ. Wien 6: 65. [Motion 14 in Briquet (1910.04?: 2). Not seen. ] 1909.03.11. ATKINSON, Motions proposant des articles additionels sur la nomenclature des cham- genericorum Pteridop ytorum conservandorum, pré sentée au Congrès international de Botanique de Bruxelles 1910. 3 pp. Berlin. [Motion 35 in Briquet (1910.04?: 4).] Volume 78, Number 1 1991 Nicolson 49 History of Botanical Nomenclature 1909.06. RENDLE, А. B. ET AL. Motion proposing an additional clause to the rules of 1905 concerning the nomenclature of algae, suggested for consideration of the Brussels Congress of 1910 by the botanists of the British Museum and others. [Motion 24 in Briquet ied 04?: 3). Not seen but proposed C. A. Agardh s algal starting point teste TL-2 no. 9067.] 1906. 06? Potonié, Н. Vorschläge zur Regelung der pa- làobotanischen Nomenklature. Jahrb. Preuss. Geol. Landesanst 30: 533- 1909.07. JANCHEN, E. Proposition d'une Amplification de la List de Noms génériques de Phanérogames, qui i Vi ‚ С. C. Proposition. 2 рр. ? [Replaced six names in the conserved names list. Saw only 1 page.] 1909. ScuiNz, H. & A. T [Motion 15 in Briquet (1910.04?: 2). Not see TS JANCHEN, E. Zur Pur der — Namen Wien. [Mo- n 16 in Briquet 1910.04?: 2). Not seni 1909. y reta ,J.C,E.G MENTS, E & A. W. Evans. Propositions relatives à P adc et au complément des Régles inter- nationales de la Nomenclature botanique, adoptées par les Congrés international de Botanique de Vienne en 1905. 9 pp. [Motion 17 in Briquet (1910.04?: . Amendments to the inter- 1909. RENDL DLE, А. B. 3 рр. bera [Motion 18 in Briquet (1910.04?: 2). Not seen.] 1909. Dane NN O. Motion au Congrés international de Botanique. Troisiéme session. Bruxelles 1910. 2 pp- Lund. [Motion 19 in Briquet (1910.04?: 3). Not seen. 1909. STOCKMAYER, S. Motion au Congrés international de Botanique à Bruxelles 1910, relative à la nomen- clature Ae t Acus es. 4 pp. Wien. [Motion 21 in Briquet (1910. 3). Not seen.] 1909. oa . A. Da quale anno debba cominciare la validatà della nomenclatura scientifica delle Crit- togame. 4 pp. [Motion 22 in ae (1910.04?: на Reprinted from Ann. Mycol. 7: 339-342. Мог se Proposed Ay as starting point for Cryptogams oe TL-2 no pee ] 1909. RöLL, ТА zu зета з a e botanischen trage, betr. Aenderungen und Zusátze Regeln in Bezug auf armstadt. 5 ot 25 in Briquet (1910.04?: 3). Not seen.] 1909. ERI Bon, J. Common nommer les formes biolo- otion présentée au ет international de botanique а Мо Bruxelles 1910. Lund tion 26 in Briquet AGNUS. Vorschlage zur No- menklatur der parasitischen Pilze zu Handen des түнчү gue im Kongress von 1910. 1 p. Berne. [Mo- n 27 in Briquet api 042: 3). Not seen.] on proposant un article anglais, francais et italiens. 1 p. [Motion 28 in Briquet (1910.04?: 3). Not seen.] 1909. DURAND, Е.-Ј. A discussion of some әш the principles governing the interpretation of pre- and their bearing on the selection of a starting-point for mycological nomenclature. 13 pp. [Reprinted from Science n.s. 29: 670-676. Not seen. Motion ruxelles und. [Motion 30 in Briquet (1910.04?: 3). Not seen. 1909. РотомЕ, Н. Propositions pour le règlement de la nomenclature paléobotanique. 7 pp. Berlin. [Motion 31 in Briquet (1910.04.?: 3). Not seen.] 1909. Tuzson, J. Vorschlage zur Regelung de Palaeo- Brüsseler Kongress 1910 vorgeschlagen. 6 pp. Bu- BLA [Motion 32 in Briquet (1910.04?: 3). Not 1909. pom F., L. KOLDERUP- -ROSENVINGE & O. m а еї et plécbotanique 59 pp. [Recognized numbered 3 1910.07.22. FARLOW, W. С. & С. Е. ATKINSON. The Botanical Congress at Brussels. Science n.s. 32: 104— 107. [Also in Bot. Gaz. (Crawfordsville) 50: 220- 225. 1910.09.] 1911. COGNIAUX, C. A. Nomenclature horticole. Rapport préliminaire présenté au Congrés international d'Horticulture de Bruxelles (1910). Bull. Soc. Roy. Bot. Belgique 47: 364-424. [First Horticultral Code (pp. 419- d with 16 Articles, criticized by Briquet as contrary to the Brussels Code just passed (1910) but not yet published (1912).] >“ oo J. (SECR.). Compte rendu M БРЕ be de nomenclature botanique. Pp. 43- | пинен е in Ё. Wildeman (SECR. ), Асїез ди m id тене p de Botanique = 1910. Volume 1. Albert de Boeck, Вгихе 1912.09? dic J. (RAPP. GÉN.). Règles bas e la nomenclature botanique . . . deuxième Éditio mise au point d’aprés les décisions du Congrés In- ternational de Botanique de Bruxelles 1910. 110 ages. G. Fischer, Jena. 50 Annals of the Missouri Botanical Garden 1912. СНОАТЕ, HELEN А. The origin s жож ш of the binomial system. Pl. Worl 263. 1914.05. Cook, O. F. Terms ae to acd types. Amer. Naturalist 48: 308-313. e haplo- type, a pseudotype, cited usage of monobasic "d Concerning D cate types. E (1914.05: 309), who cited two pre- s usages in a p du sense. 1919. 04. Hnc sce. A. S. mittee on generic types. Science 49: 333-336. [Rules and recommendations for fixing the types of generic names 1921.04. HITCHCOCK, А. S. Report of the Committee on Nomenclature of the Botanical Society of America. Science 53: 312-314. [The “type- pasis” Code. ] 1921.05. aurea A. S. The type a in systematic botany. Amer. J. Bot. 8: 251- 1921.06. SUE T. A. Plant ini ANN some sug- gestions. J. Bot. 59: 153-160. [Opening of an im- portant discussion: 12 points to change, many ex- amples.] 1921.08. BRITTEN, J. The Vienna Code. J. Bot. 59: 233- 234. [Defended legality of Vienna Code recently attacked by N. Britton, Torreya 1919: 244-246. 1919. 1921.10. "T у [Letter with more suggestions to .] J. Bot. 59: 289-294. . [Comments opposing Sprague 192 61. Bot 59 294-296. 1921. і. Barron, N. L. [Comments opposing Sprague 1921.06.] J. Bot. 59: zd 297. 1921.12. и T. А. Notes on nomenclature. J. Bot. 59: 345-349. оа discussion on orthography (Gluta renghas / benghas mn ae French names ending in -ier and -iére ш ка ОРНАП, T The iS А of plant milies. J. Bot. 60: 69-7 ion: 104. Hic HCOCK, А. S. [Suggestions regarding generic names.] J. Bot. 60: 111-112. [Advoc ee types. ] mu | PENNELL, F. У. [Comments on plant nomen- is Bot. 60: 112- 118. Poder pleasure Spra ue's reopening the subject.] 1922. 05 ра . Plant nomenclature: a reply [to Pennell & Hitchcock]. J. Bot. 60: 129-138. 1922.09. BARNHART, nt nomenclature. J. Bot. 60: 256-263. [Fascinating commentary on the on- . Уимотт, А. J. Some remarks on nomencla- ture. J. Bot. 60: 196-201. [Attack on Schinz & Thellung, пала a “loose” e. 1922.09. BARNHART, J. H. Plant nomenclature. J. Bot. 60: 256-263. о to Sprague 1922.11. SPRAGUE, T. A. [Response to Barnhart. ] J. Bot. : 313- 1922.11. HITCHCOCK, A. S. [Type-basis code.] J. Bot. 60: 316-318. 1923.01. Fritz, E. A. Plea for common sense in changes of botanical nomenclature. J. Forestry PME aD . [Not seen, ref. by г 1943.0 1923. i HITCHCOCK, A. S. Type species E the first 100 genera of Linné, а. Plantarum. Amer. Ј. Bot. 10: 510-514. 1924. SHEAR, C. L. The failure of i principle of priority to secure uniformity and stability in botanical no 4- 258. 1924. 03. BRITTEN, J., J. RAMsBOTTOM, T. А. SPRAGUE (CONVENER), E. M. WAKEFIELD & А. J. WILMOTT 1924.07. JACKSON, B. D. History of the compilation of 2 Index Kewensis. Ј. Roy. Hort. Soc. 49: 22 229. 1925.04. GREEN, M. L. Standard species of the Linnean ER of Tetradynamia. Bull. Misc. Inform. 1925: -58. TUM RENDLE, А. B. (CHAIRMAN). Rules of nomenclature. Pp. 300- 307 and 384 т Е. T. Brooks (Secr.), Im- peria al B | Conference London, 7-16 July 1924. Univ. Press, Cambridge. 1925.07. Hanson, H. E. Codes of nomenclature and botanical congresses. Amer. Botanist 31: 114-120. [Summary history. 1926.03. SPRAGUE, T. A. Standard species. Bull. Misc. Inform. 1926: 96-100. [Residue vs. type method.] 1926.05. O E S. A basis of agreement on no- menclature. . J. Bot. 13: 291-300. [An im- portant un ] 1926.06.30. GREEN, M. L. Standard- "species of nomina , Kew. t deis before and at the Cambridge Congress.”] 2. GREEN, M. L. History of plant nomenclature. Misc. Inform. 1927: OR sonal ideas about the O d. Rijks-Herb. 55: 1- 76. [Published Ма with Part II, Volume 56.] 1923.03.II. SURINGAR, J. V. Personal ideas about the i Nomencla- Ф n et Ф n + р = a £e = at 8 ч e c © = < ~ yal 3 о = — 1929.03.24. ERIKSSON, J. Question présentée au Congrés International de Botanique à Cambridge (England) 1930. 1 P. Stockholm. [Motion 7 in Briquet (1930.07: ш). Capitalize some species epithets or not.] 1929.03. MCKENZIE, К. К. Proposals for modifications of the Vienna-Brussels (so- ou International Rules of Nomenclature. 5 pp. (multigr.) New York. [Motion 12 in Briquet (1930.07?: m "d last gasp in the best style, brief and clear, of the Brittonian school.] 1929.03. REHDER, А. Proposed amendments to the In- ternational Rules of Botanical Nomenclature. J. nold Arbor. 10: 46 | (1930.07 7: ^ Com names of hortic ad. variations. J. Arnold Arbor. 10: 66-68.] 1929.06.07. SHEAR, C. L. Proposed amendments to the International Rules for Botanical Nomenclature. Sci- ence 69: 601-602. [Motion 18 in Briquet (1930.07 ?: iii).] bo] D Volume 78, Number 1 1991 Nicolson 51 History of Botanical Nomenclature 1929.07.05. GUNDERSON, А. An international list of gen- era of vascular plants. oo 70: 15-16. ional nomina conserva a. urne. [Motion 8 M a (1930. 077: ш). 47 names, mostly species 1920.07. "ane J. Suggested modifications of the In- ternational Rules of Botanical Nomenclature. 4 pp е Ottawa. [Motion 1 in Briquet (1930.07: 1929. ‘Duce ‚ В. M. (EDITOR). Proceedings of the In- ка Congress for Plant Sciences, Ithaca, New York, August 16-23, olume 2. Geo orge Ban- ta Publishing, Menasha, esa, [Includ ing:] . 5 PRAGUE, T. A. Principles and problems of plant nomenclature. Pp. 1 2. HITC K, e relation of — P to кашан н Рр. 1434-1439. 3. HITCHCOC | . (CHAIR RMAN). Round-Table dis- cussion: gs cal Nomenclature. Pp. 1556- 1570. Ts ји Briquet, Howe, Fernald & Britton. 1929.08?. BRITISH BOTANISTS. Nomenclature Proposals. 203 pp. Lon (1930.07 2: ii), using titles ME contents. Comprises seven works with these ti AMSBO M, J., E, А. J E. M. WAKEFIELD [Britten died 1924.]. Pro- posals by the sub-committee on nomenclature, appointed by the Imperial Botanical Conference, London, 1924. Pp. 3-45. [See Rendle (1925).] - SPRAGUE, T. A. Proposal for the conservation of 90 additional generic names. Pp. 46-96. . GREEN, M. L. Proposed Ж зада а of Nomina enerica Conservanda. Pp. 97-109. (This is a truncation of the original (Green, 1926.06.30), now omitting the stated reason for each choice.] . HITCHCOCK, A. S. & M. L. GREEN. Standard-species of Linnean genera of Phanerogamae (1753- 1754). Pp. 110-199. 9. WAKEFIELD, E. M. Proposals. P. 200. LMOTT, A. J. Proposal. m ll 202. URK . H. Proposal. P. z^ ` CIFERI, n Proposte di um alle Regole omenclatura Botanica. 1 p. [Motion 2 in Briquet ~ о ~ Рг roposed amendments to the In- m J. 7: 258-259. 1929. 09 Motion 3[A] in Briquet (1930.077. п). Рр. 2-4 Мо- W^ d is t че 1929. s, Ј. W. ым [Motion Committe L. No omina conservanda. 3 Association for the Advancement of Science in 1924." 1929. SEWARD, А. С. ET AL. [Memorandum.] 1 p. (type wri ritten). [Motion Р in Mp cem (1930. 07?: iii) Ки, LT to solve questions of cai ге.) ч FERNAL . L. . À. HERBY. 0076 ed ане to the been Rules of Botanical e 6 pp. Boston. [Motion 9 in Briquet (1930. ب‎ iii). Not seen.] 1929. GUNDERSON, A. Proposed amendments to the In- ene Rules of Nomenclature. 2 pp. (typewrit- ten). Brooklyn. [Motion 10 in Briquet (1930.07?: iii). Completed by: “10A” An international list of genera of vas iiic plants. Science 70: 15-16. 1929.07.05, and “10B” Genera Pteridophytorum Fe Spermatophytorum Variationes Nomenclaturae. рр. (typewritten) 2 dated March 1929.] T Нітснс S. Proposed amendments to the Гаваи Rules of Botanical Nomenclature. 3 pp. ОЕР [Motion 11 іп Briquet (1930.072: 1929. ер w. В. B. Additional articles and amend- 2 pp. Wellington, New Zealand. [Motion 13 in Briquet (1930.07?: iii).] 1929. PENNELL . Some proposals for еи of the "International Rules of Botanical Nomencla- ture.” ои Philadelphia. [Motion 14 in hos (1930. 072: ш).] 1929. PFEIFFER, H. Ueber 70 neue Namencombinationen eines Gattungsnamens in d e conservanda?" 7 pp. Berlin. [Motion 15 in Briquet (1930.07? a 1929. SHEAR, р. (typewritten). Washington. [Motion in Pedes (193 0.07?: ш).] 1929. р J. V. blas on the International Rul rd to the pp. Wageninge Briquet (1930.07?: iii). Completed by “19А” Supplementary remarks. eningen and “19B” Literature to be added in Appendix of propositions on the Intern. IS of Bot. Nomenclature (London Congress 1930). da Wageningen. 60 species names for conserva- 1929. p ne E. T. A plea for e conservation of Shortia as a generic name pore ashington. [Motion 20 in ug (1930.0 iv). 1929. Fue ipn J. V. The America 2 pp. R. Friedlander & Sohn, Berlin. Ве ¡a és te M em de Botanique Cambridge fitis, 1930. R. Friedlánder & Sohn, Berlin. 1931.07. GREEN, M. L. Rules of EPET nomenclature . Empire Forest. J. 10(1): 5 1931, BRIQUET, J. Compte rendu de didi de la sous- ection de nomenclature botanique. Pp. 554-654 in & T. F. Chipp, Report of the Pro- ceedings [of the] Fifth International Botanical Con- gress Cambridge, 16-23 August 1930. University Press, Cambridge. 1931.10.26. John Isaac Briquet died at 62. 1932.03. MOLDENKE, Н. N. A discussion of tau Bull. Torrey Bot. Club 59: 139-156. [Updated Huth 1933. SPRAGUE, T. A. Plant nomenclature. Bot. Soc xch. Club Brit. Isles 1932: 300-313. [Chief pro- Annals of the Missouri Botanical Garden bes as amended illustrated by examples from Brit- ish 1933. 11. pom , D. L. Terminology of t idl. Naturalist 14: 673-668. [Alphabet listing “of 233 3 kinds | types with usage references.] 1934.06. RENDLE, A. B. International Rules of Botanical Nom на adopted by the Fifth International Bo- hips Congress, Cambridge, 1930. J. Bot. 7 2(Suppl.): 1- 1934. "d MOLDENKE, H. N. A supplementary list of taut- onyms and miscellaneous nomenclatural notes. Tor- геуа 34: 5- 1934.12. s Горск, С. У. Proposals for amendment of Art. 20 of the International Rules of Nomenclature. nn. Mi ri Bot. Gard. 21: 709-712. [Proposal 9 in e 2).] 1934.12. SAM А. J. DE. La ig e de types et la а еке analogique cad. Brasil. Sci. 6(4); 173-179. [Proposal 13i in iu 1935.07: 2). Series of examples ru ed in accordance with Rec. X & XI.). Not see 1934. ADAMS, J. Some nd s | the International Rules of Botanical Nomenclature. 8 pp. (typewritten). ман [Proposal 1 in Sprague (1935.07: 1). Not 1934. a B. P. G. Proposals Г PA modi- 1 Art fication of Art. 35, de XXX, an marks concerning Art. 38. 2 pp. бурен) Genève [ a 10 in Sprague (1935.07: 2). Not seen. | 1934. REHD “| to аў тепиб Rules of e bone. 3rd e old Arbo- retum. [Proposal 12 in Suae б 07: 2). Not 19352 ИЧ J. С. Proposed amendments to the In- ropos 1935? AUSTRALIAN BOTANISTS. сан for conservation the die name Haussm ecific ames... А, uds A [Proposal 3in UMEN (1935.07: D Not s 1935. ВЕСНЕКЕК, А. Vorschláge fur die e Kommission des Botaniker Kongresse Amster- dam. 3 pp. ви Genf. мее) 4i in о (1935.07: 2). Мо! з 1935. DANSER, B. Н. Анд concerning the formation, гена and gender of botanical names. 3 pp. (type- written). [Proposal 7 in Sprague (1935.07: 2). See also Danser, 1935.04).] D И Proposals concerning bryological Rev. Bryol. Lichénol. 7: 137-141. [Proposal 8 in Sprague (1935.07: 2). Not seen.] OUP, В. Motions submitted on behalf of various Forestry Institutions and Societies. 3 pp. Oxford: [Proposal 14 in Sprague E 07: 3). Not seen.] 1935.01. Joncmans, W., T. С. HALLE . GOTHAN. Proposed additions to the ueram Rules of Bo- tanical Nomenclature. 15 pp. эң ет. [Proposal 11 in Sprague (1935.07: 2). Not s see 1935.02. “Briquet, J." (deceased). ТЕ ИОК Rules of Botanical Nom clature prepar hn Briquet (deceased). 152 pp. С. Fischer, fin "(English by A. B. Rendle, French by B. P. G. Hochreutiner, German 1935. 03. SPRAGUE, T. A. (EDITOR). [12 botamiste'] Ad- ditions and amendments to the International Rules of Botanical Nomenclature, 3rd edition. Bull. Misc. Inform. 1935: 65-92. [Proposal 5 in Sprague (1935: 07: 2).] 1935.04.11. Looser, С. Motion pour déclarer “nomen conservandum" le genre Cryptocarya R. Br. et Mus .01: 3). Not s adopted at Cambridge in 1930. Blumea 1: 295-304 [Proposal 15 in Sprague (1935.09.01: 2). Orthog- raphy — "leave this question to a committee of bot- anists who are more or less competent ...”; see d Јајни (1935.08).] iii Hun , H. N. Additional notes on taut- >. Bul. "Torrey Bot. Club 59: 139-156. 1935. 04. ое Т. A. Survey of nomenclature (1930- 5). Chron. ae 1: 34-35. 1935.04. RENDLE, A. B. A short history of the Inter- national Botanica na о Сћгоп. Вог. 1: 35- cell m 1864 Brussels Congrés ing issues | sed additions to the In- 13: -113. nar xdi oa зе 07: 2).] 1935.05. 20 ye d е а & "Em meo "e l p. ten). Utrecht. Proposal 16 in ше (1935.09.0 1935.07. SPRAGUE, T. A. (EDITOR). Synopsis of proposals ой submitted to the Six pp. University Pe Cambridge. [Known as the Red Book by its wrapper 1935.08. НА ER, A., C. A. А de R. MANSFELD &M.L.G onyms. Bull. Misc. In form. 1935: 34 ource of conserved homonyms, её ed with doc- umentation. See зеден (1940. 06) for committee dec 1935. 08. PRAGUE T. A. The gender of generic names: a vindication. Bull. Misc. for, wane 45-556. [Proposal 18 in Sprague (1936.0 1: 3). Reb a of Danser, 5.04: important um ла of o thography, dinis PME s, etc em — 01. SPRAGUE, T. ides шй opinions con- imi nom Зуи sals submitted to the Sixth In ternatio а "Botani cal os Amsterdam 1935. : n University іны Cambridge. [Known as the ook by its w 1935. 12 16, Alfred chus Hitchcock died of heart attack 5S. City of Norfolk after remaining in Europe to study grasses following the 6th International Botan- ical Congress in Amsterdam. 1936.04. BECHERER, А. Conservation d'homonymes géné- riques dans les fougères. Candollea 7: 137-139. [Proposal 1940-18 in Lanjouw (1950. 047: xiii).] 1936.05. SPRAGUE, T. A. P nomenclature made by the Sixth ИЕ Во- Ta Congress. Bull. Misc. Inform. 1936: 185- rning T MARTIN, G. W. The application of the ge- ic name Guepinia. Amer. J. Bot. 23: 627-629. [Proposal 1940-24 in Lanjouw (1950.047: хш).] Volume 78, Number 1 1991 Nicolson 53 History of Botanical Nomenclature 1936. SPRAGUE, T. A. 5 GÉN.). Subsection for no- menclature. Pp. 333-383 in M. J. Sirks, Zesde International Botanisch Congres Amsterdam, 2-7 Sept 1935 nee Volume 1. E. J. Brill, eiden. 1936. BurTT-Davy, J. Forestry 10: 166-168. [Proposal 1940-22 ы ы ар (1950.04?: хш). Мог зееп.] 1937.04. Кокт Validity or effectiveness of publication. ‘Chon, Bot. 3: 336-339. [Cf. Hochreu- S odd 1938. 7.06. BOLLE, Е ‘Varietas e ” Notizbl. Bot. Gart. | Dahlem 13: 524-530. [Proposal 1940-13 in Lanjouw (1950.04?: хш).] 1937.10. FURTADO, mmentary on the laws of botanical nomenclature. Gard. Bull. Straits Settlem. 9: 223-284. [Although sometimes dismissed, the dup was ahead of his time (invented terms basi- уш} 1937. 10. Fuss ‚ С. X. The poteri of types Gard. Bull. Straits Settlem. 9: 285-309. [Three kinde of primary types, eight kinds of secondary types, and six kinds of е types. , E. Kakteenkunde 3: ?. [Proposal 1940-19 in Lanjouw (1950.042: xiii). Not seen.] 1938.02.15. PFE R, H. Dichromena oder Rh spora? Zwe Ants an den VII Internationalen Bo- taniker- ко zur ciegos der урегасееп- nomenklatur. Report. Spec egni Veg. 43: 258-262. [Proposal 1940- 23in Lanjouw (1950. 047: xiii). J 1938. 07. 14. REHDER, A., > | pee de L. CROIZAT. Seven binomials propos ambigua Arnold Arbor. 19: 282- es nus 1940- 20 in Lanjouw (1950.04?: xiii).] 1938.09. MELVILLE, R. Is Ulmus campestris L. a nomen . G. Validité des publications а l'occasion d'un article de M. Furtado d n- a Botanica). Candollea 7: 509- 517 1938. бр. К. T. On the citation E aos for botanical names. Science 88: 299-300. 1939.01. FURTADO, C. X. Validating botanical name referring to invalid literature. Gard. Bull. Straits Set. Чет. 10: 162-172. [Objected to recent treatments of Epipogium and Pterocarpus. ] 1939.02.16. WHEELER, L. C. Pedilanthus and Cnidos- culus proposed for conservation. Contrib. Gray Herb. 124: 47-52. [Proposal 1940-21 by Lanjouw (1950.04 ?: xiii). | 1939.02. WHEELER, L. C. Additions and amendments to the International Rules of Botanical Nomenclature. i i . [Proposal R, A. Proposed amendments to the International Rules of Botanical Nomenclature. J. Arnold Arbor. 20: 269- 279. [Proposal 1940-6 т В. Nomenclatural proposals for the 1940 Botanical Congress. Amer. J. Bot. 26: 229- 231. [Proposal 1940-12 in Lanjouw (1950.04?: xiii). 1939. 05. ALM, C. G. . Amendment of Art. 68. 1 p. Góteborg. [Proposal 1940-3 in Lanjouw (1950.042: xiii).] 1939.05. FURTADO, C. X. Amendments proposed to the International Rules of Botanical Nomenclature (1935). Gard. Bull. Straits Зешет. 11: 1-30. [Proposal 1940-9 in Lanjouw (1950.04?: xiii).] : 82-94. [Prbpdedl 1940-10 in Lanjouw (1950. 047: xiii).] 1939.06. Dixon, Н. Н. International bryological nomen- clature. J. Bot. 77: 176-178. [Proposal 1940-17 in Lanjouw (1950.04?: xiii).] 1939.07? ALLAN, H. H. The nomenclature of hybrids. Chron. Bot. 5(2/3): 205-209. [Proposal 1940-4 in Lanjouw (1950.047: xiii).] 1939.07 7. SPRAGUE, T. A. The nomenclature of hybrids. Chron. Bot. 5(2/3): 209-212 1939.10.06. BuLLock, A. A. Act семь Bull. Misc. Inform. 1939: 337- 339. [Proposal 1940-28 in Lan- ITOR). Proposed additions . (ED and Знала вани м the International Rules of Bo- IN of s y written). Arnhem. [Proposal 1940-26 in Lanjouw (1950.04?: xiii).] 1939? Hovrascens, G. ee [to reject Populus can- adensis Moench a n ambiguum]. 1 p. (type- ae Arnhem {Brcposal 1950- 27 in Lanjouw (1950.04?: xiii).] d^ HocHREUTINER, В. P. С. [Typewritten] [торе хш). Not seen.] geln. Repert. Spec. Nov. Regni Veg. Beih. 121: 1940.06. SPRAGUE, T. A. Additional nomina generica names scheduled бе the Кик җе (4th edition) Co 1941. 09? SMITH, А. C. The principle of НЕ. in bio- logical nomenclature. Chron. Bot. 9: 19. 1942.01. Ewan, J. Isotype vs. co-type as bib id for –9. [A definitive 1942?.12. HARLOW, 2. Scientific names and their va- garies. J. Forest. ЖЕКИ 40: [Not seen, ad- dressed by Dayton, 1943.05.] 1943. BLAKE, S. F. Cotype, syntype, and other terms . Should plant менын bé controlled by an open season or otherwise, or alto- 54 Annals of the Missouri Botanical Garden gether suppressed? J. Forest. (Washington) 41: 369- 373. [Comment on Harlow, 1942?.12, and Fritz, 1923. 1943.12. BEETLE, A. A. Specific decapitalization. Chron. 380-381. [In favor.] 1944.12. A discussion on the differences in observance Proc Linn. € London 156: 126- UR zb пне The case for the p. o = w 16 Uu ct un — SS ЈЕ. The case for the zoologists. Pp. 134— WILMO ‚ 1. А criticism. Pp. 138-140. TREWAVAS, E. Comments. Pp. SPRAGUE, T. А. Comments. "à ce 142. ids s. : O, P. summing- up. Pp. 142-146. 1945. 10? une A. C. The ap d per in bio- ogical se ih Chron. Bot 14-119. [Op- posed n ecifica conser i “If it were not for this “ist P about 850 conserved [generic] would id invalidated"] 1945. HYL R, N. Nomenklatorische und system beu Studien über ies che Gefasspflanzen. jt v. Arsskr. 7: 1946. 07. Пи Бе L, 5 Species names with Mu а | ап + epp nomials. Gentes Herbar 74. [For фес ичен а a the com- А binomial and author means **of," opposes infraspecific names. 1947.04.01. Camp, W. H., Н. W. RICKETT & C. A. WEATHERBY. International Rules of Botanical No- menclature . . . revised by the International Congress of Amsterdam, 1935. Brittonia 6: 1-120. [The Brit- SON, . On the preservation of well- known binomials. Phytologia 2: 201-213. [Parthe- nocissus vitacea, not inserta; Nelumbo lutea, not pentapetala; Lathyrus maritimus, not japonicus; Acer saccharum, not saccharophorum. | 1947.10. Airy SHaw, Н. К. T ypification of new names derived from persons or places. Kew Bull. [1]: 35- 39. [How would author deal with Asclepias syriaca?] 1948.03. p В. E., В. Sr. Јонм-Вкоокѕ & В. 5. p. International Bacteriological Code of No- idis uoi J. Bacteriol. 55: 287-306 1948.03. RickETT, Н. W. Citation of author's names in taxonomy. Bull. Torrey Bot. Club 75: 172-174. [In . . apud. 1948.04. LITTLE, E. Г. A proposal to stabilize plant names. Phytologia 2: 451-456. [Nomen extinctum — ип- used for 100 years, upping the ante | the Berlin code (1897.06) of unused for 50 yea 1948.07.15. MERRILL, E. D. Neolitsea (Bentham) Mer- rill, nomen conservandum propositum. J. Arnold Ar- 98-201. [Proposal 1950-2 in Lanjouw (1950.042: xiv.)] 1948.07.15. MERRILL, E. D. Nomenclatural notes on еа ove papers 1804-1840. J. Arnold Arbor. 29: 2 4. [Proposal 1950-3 in Lanjouw (1950.04?: | 1948. LANJOUW, 3 (EDITOR). [Proposals of Dutch bota- nists. 13 pp. НЕ Pies 1950-5 in uui ps (1950.04?: xiv). Lam (p. 1) Introduced the ord “taxon,” hide alent to German “Sippe.”] 1949. 01. ME E. E. (CHAIRMAN). Symposium on bo- tanical nomenclature. Amer. J. Bot. 36: 1-32. Com- prising: l. SHERFF, E. E. Introduction. Pp. 1 . WEATHERBY, C. A. Botanical ted NN since 1867. Pp. 5-7 . BLAKE, S. F. Byways of nomenclature. Pp. 8-9. кошы, Rafinesque, Sain t-Lager, Bubani.] RICKETT, Н. W. Ап editor's point of view. Pp. 10- 1 ~ [2v » сл MERRILL, E. D. Adventures in locating ren pub- lished a unlisted binomials. Pp. 14-19. ENNELL . Toward a simple and clear no- menclature. Pp. 19-22. BAILEY, L. Н. Problems in taxonomy. Pp. . ROSEDAHL, C. E Ме problem of subspecific cat- ка Рр. 24-27. . JUST, T. The ко ш of fossil plants. Pp. 32. d ox © 1949.06.25. Camp, W. H., Н. W. RICKETT & С. WEATHERBY. Proposed changes i in the ата Rules of Botanical Nomenclature. Brittonia 7: 1- [Proposal 1950-62 Sponsored by a group of 55 taxonomists.] 1949.10. Smith, A. C. A legislated nomenclature for species of plants? Amer. J. Bot. 36: 624-626. [Op- posed. 1949.12. FURTADO, C. X. A further commentary on the rules of nomenclature. Gard. Bull. Singapore 12: 311-377. [Proposal 1950-83 in Lanjouw (1950.04?: xvi) Discriminated between binary and binomial 1949.12? MANSFELD, R. Die Technik der wissenschaft- lichen Pflanzenbenennung: Einführung in die Inter- nationalen Regeln der botanischen Nomenklatur. 116 pp. Akademie- verlag, Ве lin 1949. TROUPIN, С. La termi systematique. Bull. Soc. а Вог. Belg. 82: 57-66. 1950.04? LANJOUW, J. (ACTING RAPP. GÉN.). Synopsis of botaniqı 1e cal Congress, Stockholm, 1950. 255 pp. [550 a .] PUE € UNE N. Specific and trivial decapitaliza- . Bull. Torrey Bot. Club 77: 214-221. [Impor- tant statement pro decapitalization. 1950.06. HJELMQVIST, Н. The conservation of species names [invited responses]. Bot. Not. 1950: 328- m l. KETT, H. W. Pp. 328-329. [Opposed.] 2. GRAM, K. Pp. 330-332. [Pro, needed for stabil- y. . CAMP, W. H. Pp. 332-336. [Unrealistic without 3 4. JANCHEN, E. Pp. 336-340. [Yes but must be limite 5. GILMOUR, J. S. L. Pp. 341-343. [Yes but must be lim 6. BAEHNI, C. “С 343-346.[С tion of sp = sea serpent.] Ts one с. Pp. 346-347. [Fix the rules then me changes. 1950.07. RICKETT, H. W. & W. H. САМР. The appli- cation and use of botanical names. Bull. Torrey Bot. Club 27: 245-261. [Residue of circumscription method still in Code vs. type met 1950.07.18. International Association for Plant Taxon- omy born by resolution of the Stockholm Congress. Volume 78, Number 1 1991 Nicolson 55 History of Botanical Nomenclature 1950.08? Lanjouw, J. (EDITOR). Botanical nomenclature and taxonomy, a symposium organized by the In ternational Union of Biological Sciences with support of UNESCO at Utrecht, the Netherlands, June 14- 19, 1948. Chron. Bot. 12(1/2): 1-87. comprising: - SPRAGUE, M. L. Minutes of the Utrecht Confer- m ~ р. ; Ры k Ont eed for an international society of plant taxonomists. Pp. 55-57. NTERNATIONAL COMMISSION FOR NOMENCLATURE. Pp. 59-62 ә > (л |] > UE, T. A. International Rules of Botanical Nomenclature Supplement embodying the al- terations made at the Sixth International Bo- 1950.11. Wir, H. C. D. ре. Changes in the Intermational Rules of Botanical Nomenclature made e by t the Tt h t. Dot. Fl. Males Bull. 1(7): 197-231. 1951.09. Lanjouw, J. The Stockholm 1950 Rules of on botanical nomenclature. Taxon 1: 35-36. [Edi- torial committee. 1952.09. LANJOUW, J. ET AL. (EDITORS). International e of Botanical Nomenclature adopted by the sev enth International Botanical IN Stockholm, July 1950. Regnum Veg. 3: 1-228 1953. LANJOUW, J. (RAPP. GÉN.). ч ы з= report. 457-550 in Н. Osvald & E. Aberg (editors), Proceedings of the Seventh Pian Botanical Congress, Stockholm, July 12-20, 1950. A & Wiksell, Stockholm. A a in Reg- num Veg. 3: 457-550. 1954.01.] 1953.07-09. OD L. On nomenclature: The *'type- method." Taxon 2: 105-107; 124- id [Classic olemic NUM typification and autonym 1953.09. RickETT, H. W. Expedie у А vs. priority in vey nomenclature. Taxon 2: 117-12 with bibliogaphy. 1954.03 Lanjouw, J. Recueil synoptique des propositions concernant le Code International de la Nomenclature K L STAFLEU, F. A. Report [on] The Genève Con- ference on botanical nomenclature and Genera Plan- tarum organized by the Botanical Section of the In- ternational Union of Biologic al Sciences 25-30 January 1954. Regnum Veg. 5 -59. T е Е. A.] Preliminary mail vote. Taxon 162. boi ne ко Е. A.] [Decisions M E VIIIth аа нац Botanical Congress, Paris, 1954, No- 1954. 11. “Stare, F. Ne Nomenclature at the Paris Con- ras s. Taxon 3: 1955. 08. Sr m. F. y: Huitiéme Congrés International de Botanique, Section e Taxon 4: 121- 2 3c STAFLEU, F. A. Nomenclatural conservation in phanerogams. Taxon 5: onu. . [Historical re- about conserved generic 1956. 10? esee J. ET AL. а Code of Во- nical Nomenclature adopted by the Eighth Inter- national M tanical Congress, Paris, July 1954. Reg- num 1957.04. в re С. Fifty clature. ie tonia 1959.01. LAN PP.). Synopsis of proposals con- cerning t is ae Code of Botanical Nomen- clature submitted to the Ninth International Botanical . Regnum Veg. 14: 1 y years of botanical nomen- istory.] Regnum Veg. 20: 1-116. Reprinted from Proc. IX Int. Bot. Congr. Montreal 1959 3: 27-116. 1961.01. SMITH, A. С. The 1960 meeting of the Editorial Committee of the [Montreal] International Code of Botanical Nomenclature. Taxo den 8- 1961.12. LANJOUW, J. ЕТ AL. (EDITORS). International ode of Botanical Nomenclature a e t International Botanical pneus. Montreal, August 1959. Regnum Veg. 1964.02. Lanjouw, J. 4 F. q STAFL EM of proposals concerning the а ан Соде tanical Nomenclature submitted to the Tenth Тиме. national Botanical B ce Edinburgh — 1964. Reg- я 8. [337 proposals.] 1964.04. HELLER, L т, Тће early history of botanical nomen- о Attributed invention to Linnaeus's struggle m book citation.] 1964.06. таи F. A.?] Nomenclature proposals Xth ongress: preliminary vote. Taxon 13: 183-187. 1964.07. STAFLEU, F. А. (SECR. GEN. ona Preliminary report on the ае of names of plants of eco- nomic importance. Regnum -36. 1964.11. STAFLEU, F. A cial e at Edinburgh. Taxon 13: 273-282. 1965.05. Voss, E. G. On citing the names = publishing uthors. Taxon 14: 154-160. [In 1966.11. STAFLEU, F. A. Tenth р Botanical Congress, Edinburgh, 1964: Vae ia Section [proceedings.] Regnum Veg. 44: 1966.12. LANJOUW, J. ET AL. faded y CON ode of Botanical Nomenclature adopted by the Tenth eg вн Botanical с Edinburgh, August 4. Regnum Veg. 46: 1-402. 1967.11. DANDY, J. E. les of generic joe i vascular plants 1753-1774. Regnum Veg. 5 0. 1968.06. МсУлосн, R., AFLEU. Án n glossary of botanical ae Reg- m Veg. 56: 1-31 196 9.02. STAFLEU, Е. A. (RAPP. GÉN.) & E. С. Voss (VICE-RAPP.). Synopsis of proposals on botanical no- menclature, Seattle 1969. па Мег. 60; 1–124 (including appendices pp. 42-124). [284 proposals. ] 1907.02. STAFLEU, К. A. оте at Seattle. Taxon 19: 36-42 1970.02. MoonE, Н. E., К. A. STAFLEU & E. С. Voss XI International Botanical Congress: final mail vote and Congress action on nomenclature proposals. Tax- on 19: 43-51. 1972.02. dee F. А. (RAPP. GÉN.). & E. С. Voss (VICE-RAPP.). Report on botanical UE Se- attle 1960. Regnum Veg. 81: 1-133. Annals of the Missouri Botanical Garden 1972.03. STAFLEU, F. A. (CHAIRMAN) & E. C. Voss (SECR.). International Code of Botanical Nomenclature adopt- ed by the Eleventh International Botanical iuo i attle, Said 69. Regnum Veg. 82: 1-426. 1975.02. STAF F. A. (RAPP. pu & ^ G. Voss пене prs of proposals on botanical no- menclature, Leningrad, 1975, Taxon 24: 201-154 [152 proposals. 1978.04? STAFLEU, К. A. (RAPP. GÉN.). & E. С. Voss (VICE-RAPP.). International Code of Botanical No- menclature adopted by the Twelfth International Bo- tanical Congress, Leningrad, July 1975, Regnum Veg. 97: 1-457. 1979.12? Voss, E. G. Section 1. Nomenclature. Pp. 129- 186 in D. V. Lebedev et al. (editors), Proceedings, XII International Botanical i Leningrad, 3- Я Р. БЕН, © W. GREUTER РР.). aa of proposals on botanical no- о. Sydney, 1981. Taxon 30: 95-293. [210 proposals. 1982.02. Voss, E. C. Nomenclature at Sydney. Taxon 1982.06. GREUTER, W. (vicE-RAPP.) & E. С. Voss (RAPP.-GÉN.). Report on durum nomenclature — Sydney 1981. wore -124. 1983.10. Voss, E (RAPP. rus & W. GREUTER (VICE-RAPP.). اا‎ Code of Botanical No- menclature adopted by the Thirteenth се Botanical Congress, Sydney, August 1981. Regnu Veg. 111: 1-472. GREUTER, W. (RAPP. GÉN.) & J. MCNEILL (VICE- RAPP.). Synopsis of proposals on botanical no- ы | Berlin, 1987. Taxon 36: 174-281. [334 proposals. 1987. 12 McNew, J. XIV International Botanical Con- ongress action on no- 8- RS). International Code of Botanical хш td adopted е Fourteenth International Botanical Congress, Berlin, July-August 1987. Regnum Veg. 118: 1-328. [En glish only.] 1988.08. STAFLEU, Е. A. The prehistory and history of IAPT. Taxon 37: 791 s 1989.04. GREUTER, W., L & D. NICOLSON. Report on bobanical шы M 1987. Englera 9: 1-228 WHO CONQUERED THE NEW WORLD? OR FOUR CENTURIES OF EXPLORATION IN AN INDEHISCENT CAPSULE Joseph Ewan! ABSTRACT We track botanical conquest from о (who met Columbus in 1490 and visited Santa Marta, present Colombia, 1515) to th in 1515) to of the 16th dean the solitary missionary- -naturalist of the century; and the naturalist-collector, then often a member of a government- schemes of classification) in the 18th sponsored expedition in the 19th century. the close of the 19th century. W 17th century; the apologist of systems (who advocates History? We are making history. Plant taxonomy is not a purely descriptive science, it is history, deviously directed toward a conclusion. History rests on archives of two sorts: books published from archives, and plant mummies ticketed or labeled and entombed in herbaria. The conquest of the New World by the explorers who observed and reported begins with Columbus who met Oviedo two years before Columbus set sail. We headline the sixteenth century for its commentators on the plants observed; the seventeenth century by the missionaries who collected seeds and, now and then, specimens; the eighteenth century by apologists of systems, collecting exhibits to support their schemes; the nineteenth century by the field collector-ex- plorer who filled, not cabinets of curiosities, but museums which by the twentieth century would overflow. Floras, faunas, history, everything flows, omnia fleurit. When did the sea bean, Entada gigas, first take the Gulf Stream across the Atlantic? The soft pithlike wood of cuipo or quipo (Cavanillesia pla- tanifolia), a relative of ceiba, for centuries had been washed on the Azores by tempests. Indeed, these waifs were gathering evidence for what in Columbus’s day was still looked upon as shoreless seas.” When he was coasting among the Bahamas on the first of his four voyages, Columbus wrote: “I believe there аге many plants and many trees which are worth a lot in Spain for dyes, and for medicines and spicery; but I do not recognize them, which gives me great grief."* Pineapple as a do- mesticate from the Mosquito Coast was described, for example, by Columbus's son Ferdinand as a "great pine cone; and the plant is grown in large fields [in Veragua, Panama ] from shoots that grow out of the same pina [and] plants last for three or four years, always producing fruit.”* (See Fig. 1.) Pineapple was transplanted by Guaranies of south Brazil and Paraguay—the word Ananas is a Tupi- Guarani name—via other Amerindians probably down the Orinoco avenue of penetration to the Caribbean. The Portuguese had delivered it as far as Goa on the Indian Ocean during the sixteenth century. But pineapple was not the staple food that maize was, “the most important of all foods for men and their livestock.” By the late sixteenth century the peoples of central Africa were raising maize, manioc, peanuts, and squashes, all Ameri- can domesticates, but we do not know exactly who those phyto-conquestadors were, who delivered the foodstuffs, or how early they arrived. So dispro- portionate was the west-to-east movement that Vavilov listed 640 New World plant domesticates but only 50 of African origins.^ Weeds from tropical or warmer climates had arrived in the Mediterranean from even before the days of Marco Polo, and these Old World emigrants arrived in Hispaniola in 1494 on the 17 ships of Columbus's second voyage. (Do you remember the Santa María, Pinta, and Nina from your school- book history days? The Мепа returned to Hispan- iola in 1494.) On board were 1,200 men (no wom- en) taken abroad to build the settlement near Isabela.’ Imagine the weed seeds that arrived when rootcrowns of the grape vines and sugar cane from Spain and her possessions were planted. Bixa, the body paint of jungle Indians, in mod- ! Missouri Botanical Garden, Р.О. Box 299, St. Louis, Missouri 63166, U.S.A. ANN. Missouni Вот. GARD. 78: 57-64. 1991. Annals of the Missouri Botanical Garden FIGURE 1. at the botanic garden in Amsterdam. This Pineapple grown and illustrated in 1697 drawing served a basionym in the description of the ee by ene in the Species Plantarum in 175 ern times has been sublimated into coloring for cheese and margarine. In 1526 Oviedo said the "paint their bodies [and] make themselves hideous [with] various figures on their faces and other parts of the body [thinking] they look good in it, and besides they like the appearance of the paint. Gaspard Bauhin cited Oviedo, and from the Caribs dehiscent capsule of annatto placed it in the ge- neric group, Castanea.? When Linnaeus named Bixa orellana he commemorated Francisco Orella- na, taking his flag from Plukenet who had pictured Bixa by its other name achiote. Alexander Smith, employed in the Economic Museum at Kew, con- tributed to the Treasury of Botany of Lindley and Moore. Smith wrote that Bixa paint [was] almost their only article of pne n Perhaps Bixa gave the designation **Red Man Tidings were ышы in Pizarro's time that a sort of cinnamon grew on the east slope of the northern Andes. “The Land of Cinnamon" was placed as east of Quito, and Orellana was dis- patched to accompany Francisco Pizarro's brother Gonzalo on an expedition to a province called Qui- 5 jos to find this “Canella alba" or “Ispingo. са Pizarro chronicled that Mies they were “on the skirts of a volcano," evidently Cayambe, it did not cease to rain for a single day for two months. He said the Spanish “received great in- jury, and much of their clothing became rotten.” In the vicinity of Sumaco the Indians went naked, Pizarro concluding, “they go naked because the country is so hot, and it rains so much that clothes No specimens from this Spanish Pizarro period exist. It was probably Joseph de Jussieu who in 1747 first collected a sterile spec- imen that Lamarck labeled Laurus quixos and preserved in the Paris herbarium.'* Here at the Missouri Botanical Garden, Henk van de Werff is piecing together this complex history four centuries after Pizarro. Ithough Santa Catalina Island was discovered off the southern California coast in October 1542 by Cabrillo, there was no botanist aboard the San Salvador or its consort, La Victoria, and it was 300 years before William Gambel botanized on the island in 1847. The University of Padua, not Ca- brillo's Spain, opened the first botanic garden in 7912 become rotten. Europe nearly a half century after Columbus.'* The next botanic gardens, Leyden (1590) and Mon- tepellier (1592), featured materia medica exhibits or plants involved in the economy of the nation. American plants were grown at Oxford, which opened its garden in 1621, Paris in 1635, and Uppsala in 1655. The great surge of living collec- tions advanced with Kew in 1759 and her 13 satellite gardens throughout the far-flung British Empire. In due course there was a particular one outside the British crown—Shaw’s Garden, created in the image of The growth of botany, systematic botany in par- ticular, depends on the comparative forms of flow- ers afforded in a botanic garden. Just as vegetation types are mapped by so-called indicator species, botanic gardens stake the growth of horticulture and advance taxonomy. Notable are the Chinese studies nurtured by Arnold Arboretum's Chinese plantings. We recall that Mexico City and Lima opened universities in 1553 and 1555, and that the New World's first herbal, though not published for four centuries, dates from 1552. It is the Badianus Manuscript. This Aztec herbal records such mod- ern concepts as the relation of hummingbirds to the tubular scarlet flowers of, for example, Loeselia coccinea (Polemoniaceae). There are puzzles: the aroid huacalxochitel, meaning “basket flower,” was illustrated for its value in curing angina. Tom Croat tells me it is “presumably a fanciful rendition Volume 78, Number 1 1991 Ewan 59 Who Conquered the New World? of some Philodendron." The tendrils in the draw- ing are unnatural; a few aroids do have lobed leaves, but none has blades close to those pictured. Many narrow-endemic philodendrons occur in Mexico, and so, in twentieth century thinking, “perhaps it represents an extinct species." Among other engaging illustrations in the Badianus Manu- script are the very earliest illustrations of American cacti, a cereoid and an opuntia.'* Remarkably this 1552 herbal offers notes on Old World immigrants: Rumex, Oxalis, and Cynara cardunculus. Early records of New World Parthenium hysterophorus, and such fruits as mombin (Spondias purpurea L.) are in this Mexican herbal. The Franciscan friar André Thevet was on the coast of Brazil in November 1555." Later he sailed along the Florida coast and reached the Gulf of St. Lawrence. Thevet's Canadian journal was cited by Linnaeus in his Bibliotheca botanica (1747), where it was the first of what Linnaeus listed as ” per- egrinationes Americani." Thevet antedated Jacob Cornut's Canadian account by 78 years, but Cornut offered about 30 plant descriptions. The indehis- cent capsules of the attractive genus Thevetia bring him to mind. (The Flora of Panama describes these distinctive fruits as **drupes.")'* A century before Linnaeus “Ahovay” was included by Gaspard Bauhin in his Pinax under “Pomiferae Arbores" together with genipat and durion. Under a six-word polynomial Bauhin cited Jean de Lery, Protestant pastor who visited Brazil the year after Thevet, and who noted the fruit of Ahovay (Thevetia) as a crepitaculum, Latin for rattle. Domestic uses of plants are the first chronicled by the sixteenth century herbalists Clusius and L'Obel. The Tupi- nambas of Brazil wore Thevetia “shells” as anklets in their dances. When the barber-surgeon John Gerard produced his Herball or Generall Historie of Plantes for British readers, the year Shakes- peare's **Henry the IV" was on the stage, Gerard figured a woodcut of Thevetia, the “Indian Morrice bells," in an allusion to the exotic Morocco dance theme. He said their Шы. sound doth much delight" the Indians.’ Only one of the several herbalists of the sixteenth century included American discoveries, so slow was the spread of plant lore. Brunfels included no Amer- ican headlines, not even holy wood, Guaiacum officinale or sanctum, the prized antisphyilitic, which by 1530 was selling briskly. Fuchs alone of the herbalists illustrated two New World squashes and maize." By 1583 Clusius of Antwerp, who profited from his residence in a port city on the North Sea, mentioned the “White beans of Gua- temala" and “дата de noche.” Mulli LE поп s erratis 22943 FIGURE 2. Pepper Tree of Chile illustrated by Feuillee in 1714, introduced into California during the Mission Period. Before we leave the foreign correspondents of the sixteenth century, a word about Old World preconceptions that have left us with some every- day misnomers. For centuries maize was known as Turkey wheat, the American bison is still called buffalo, and the (spotted) j jaguar, the Tupi Guarani he- d World association, el tigre, is the popular Latin American name today. When the American agave was named it became the **Amer- ican aloe.” David Quinn called these “products of wishful thinking." The Greek word for a spiny plant is kaktos, appropriated for the cactus. John Lindley knew our avocado as Alligator pear, ““pear” going back to Oviedo from its shape.” Who were the solitary missionary-naturalists of the seventeenth century? First let us mention John Banister of the York River region of Virginia.” d the Reverend Banister lived beyond his 38 years, North American natural history would have taken a very different course indeed. Banister took his M.A. degree at Oxford in 1674 and became clerk and chaplain at Magdalene College. He stud- ied and collected especially the surviving plants brought back from the York River area that had 60 Annals of the Missouri Botanical Garden earlier been noticed by John Tradescant between 1632 and 1654. From 1678 to 1692, when Ban- ister was accidentally shot while collecting on the Roanoke River, two-thirds to three-quarters of the plants described by Banister from Virginia were new. Leonard Plukenet had 89 of Banister's draw- ings finely engraved and reproduced in his Phy- tographia, with the Latin descriptions. However, he failed to credit Banister with 11 of these, mostly those received in England after Banister's death. Plukenet's Banister records were an important con- tribution to the writings of Linnaeus 50 years later. When Linnaeus quoted Gronovius or the works of Ray, Morison, Plukenet, Petiver, and Dillenius for records of Virginia plants, he was quoting Banister and was basing his newly described species on one of Banister's drawings or descriptions, although Linnaeus rarely attributed these to him. The monk Louis Feuillée (1660-1732) explored Martinique, Caracas, and elsewhere before visiting the coasts of Chile and Peru between 1707 and 1712, ob- serving especially the ports of Concepción, Tal- cahuano, Coquimbo, and Callao. Linnaeus cited plants from the illustrated journal of Feuillée, but no specimens are known. The pepper tree, so well known in southern California, was cited from Feuil- lée as Schinus molle where the trivial epithet is without meaning, a mistake for mulli, the native vernacular name in Chile.” (See Fig. 2. The Catholic missionary Charles Plumier (1646— 1704) visited Martinique and Haiti in 1689 with Joseph Donat Surian, physician-pharmacist.” Sur- ian returned to France the next year with a plant collection, which is preserved in ten books in the natural history museum in Paris. Instead of col- lecting plants Plumier made about 6,000 line draw- ings and prepared early descriptions of the plants observed. These were the bases of four botanical works published between 1693 and 1760. They were completed by Johannes Burman who wrote the later descriptions. The eighteenth century may be marked by the advocates or apologists for the two historic systems of plant classification launched in that century. The Sexual System espoused by Linnaeus, and its im- pact, has been skillfully told by Frans Stafleu in his Linnaeus and the Linnaeans (1971), with some references to the succeeding scheme, the emerging Natural System of Jussieu. | summarized in my "Plant Collectors in America: Backgrounds for Lin- naeus” in 1970,28 but we await a capsulated review of the American collectors and sources that con- tributed to Jussieu's natural arrangement. Linnaeus is well known for his Systema Naturae published in 1735. The same year a clockmaker in England, John Harrison, invented the first chro- nometer to enable ships to determine longitude at sea. The use of the chronometer was opposed by the astronomer royal, Nevil Maskelyne, and so it was not used on the Endeavour by Captain Cook. It became crucial in later voyages.” The chronom- eter and the sextant enabled expeditions to reach and to map charted shores accurately. The harvest of those expeditions would still be sorted by the numbers of their stamens and pistils into new gen- era and species according to Linnaean classes. The description and classification of new plants in the New World followed in the wake of such applied botanical activities as the culturing of silk- worms on leaves of mulberry trees. The founding of the Colony of Georgia is said to have rested on three needs: an asylum for debtors and persecuted Protestants; a trial garden for the growing of silk, wine grapes, and drug plants; and a relief of pop- ulation pressures in the Old World.* Old Ebenezer, a sterile and unattractive spot about 21 miles from Savannah and about four miles below the present town of Springfield, Effingham County, was settled in the spring of 1734 by 78 Salzbergers. George Genzmer described them as “simple folk, pious, industrious, uneducated, impoverished by confis- cation and exile, and bewildered by their situation on the Georgia frontier." Baron von Reck and Reverend Johann Bolzius led a ““tour of observation into the adjoining country [to Savannah] sponsored by Governor Oglethorpe.* Old Ebenezer had proved unsatisfactory—in 1736 the settlers moved to a high ridge and founded New Ebenezer. Central to my mentioning Ebenezer is the appearance in 1756 of an almost entirely overlooked botanical paper published in Hamburgisches Magazin enumer- ating 87 plants, mostly economic or potentially useful species: the report of the Reverend Bolzius.* Dr. William Houstoun botanized in the West Indies, Vera Cruz, and the Spanish Main on behalf of Sir Hans Sloane and the Chelsea Physick Gar- den.* His premature death in Jamaica, August 14, 1733, brought an end to the plan of Oglethorpe and the South Sea Company, by whom Houstoun was retained as a physician, for shipping plants and seeds from tropical America to West Indian ports, to Charleston, and to the “Colony of Georgia in America." Some West Indian plants reached the Savannah plantations from Houstoun’s trav- els. His energies in other regions and the many new species he discovered and made known by Philip Miller indicate that the history of Georgia botany would have been very different had he lived. Men of the nineteenth century at watersheds around the world who directed the flow of knowl- Volume 78, Number 1 1991 Ewan 61 Who Conquered the New World? edge from explorations included Barton, Humboldt, the Hookers, the de Candolles, Martius, Engler, Torrey, and Gray. Humboldt's enquiry into many fields is “unlikely ever to be surpassed, for thor- oughness, perseverance, curiosity, courage, and sheer force of intellect." Humboldt’s was the last moment in history when an individual could master all branches of science. William Jackson Hooker, as did S. F. Baird in this country, implemented the field naturalists, often out of their own pockets, to assemble at Kew the records of global exploration. Witness, for example, that George Gardner, M.D., aided by Hooker, from 1836 to 1841 wrote 36 letters, some 18 and 24 pages long. Gardner's comments in his Brazilian Travels (1846) are often to the point: the Rio botanic garden, Gardner wrote, is "more a promenade than a Botanic Garden." Joseph Dalton Hooker, who cultivated Kew after his father's seeding, saw the international values of colonial floras, still of first-line usefulness, for the West Indies, Hongkong, British India, and so forth. The elder de Candolle, Augustus Pyramus de Candolle, first used **endemique" in 1820.” His Prodromus, published from 1823 to 1873, with his son, Alphonse, involved British and continental authors. "Prodromus literally means the northeast wind which blows for eight days before the rising of the dogstar," wrote Willis Linn Jepson. “The work was intended as a forerunner of that natural system which would sometime appear in effulgent glory. . . . The Prodromus blew for fifty years but the dogstar has not yet risen." Geopolitics dictates exploration. Banks and So- lander were not welcomed when they visited Rio de Janeiro in 1769, and later Humboldt was for- bidden to enter Brazil, then in tight control of Portugal. But Napoleon's invasion of Spain in 1807 set off independence movements in America. The tables turned in Portugal. Many left Lisbon for an emergent Brazil. The capital was moved from Bahia to Rio de Janeiro in 1808; Brazil was opened to English commerce, and the botanic garden was founded in that same year. For the succeeding decades a veritable parade of field-collecting ex- plorers from Europe went to Brazil. Included were Langsdorff, Sellow, Saint-Hilaire, Pohl, Claussen, d'Urville, Mawe, Bowie, Burchell, Cunningham, and many more, a parade unmatched perhaps in the annals of botanical exploration.* The sixt German naturalist in the decade was Karl Friedrich Philip von Martius, born in Erlangen, northern Bavaria. Martius was joined by the zoologist Johann Baptist von Spix.* They traveled 1,400 miles from 1817 until 1820, on an expedition *'that certainly ranks among the most important enterprises un- dertaken for scientific purposes in the [nineteenth ] century." As many as 6,500 plant specimens, as well as living plants and seeds were sent to Munich. Zoological collections included 85 mammals, 350 birds, 116 fishes, and 2,700 insects. Seeded by the plant collections was the epic Flora Bras- iliensis. Martius, then of age 46, prompted by the ill-fated Endlicher, in 1840 launched the first of 130 fascicles, which in 20,733 pages made 15 volumes. Martius completed 46 fascicles by his death in 1868. A. W. Eichler and Ignatz Urban continued to edit the Flora to its completion in 1906. Sixty five botanists of nine European nations contributed their accounts to Flora Brasiliensis.” t may surprise us to read Аза Gray's letter to George Engelmann in 1863 that **Martius " че a very remarkable botanist, but good." Gray added Martius “is a genial, philosophical id = of Plato, etc.), a good explorer, has worked up the Palms, etc., and is a wonderful man for the amount he knows on a vast number of different subjects— philology, antiquities, philosophy, et id genus omne.”* Gray had first met Martius on his Eu- ropean tour in 1839—Gray was 28, Martius, 45, “а small man, not so tall as Г” says Gray, “quite thin, but very good looking."* They met at the Royal Cabinet in Munich, once a Jesuit College, then containing the immense library, which ac- cording to Gray was one of the two largest in the world, though the collections except for the Bra- zilian punte ienna." After his visit to the botanic garden, Gray dined at the “house of Martius," met his “very intelligent and pleasant” wife, and their four chil- dren. The party then retired to the royal chapel, listened to the chapel musician on the piano “‘play to perfection," then, says Gray, to crown all, Mar- tius seized his fiddle, quite to my surprise, and played with great spirit." For the Brazil of Spix and Martius that will never return, listen to their “first night in the open air" "make no great show after that of not far from Rio: At Retiro, a niis — lying sideways fro mpy s esed first night in the open air. The u a had ceased his strangely sounding notes, the swarms of g oppers com en ight set in, their monotonous chirp, at beetles fluttered in luminous circles through the Annals of the Missouri Botanical Garden hedges, till at length a heavy rain veiled all in dark- s. 5: The next mountains over which we passed are lower, and rise at intervals. The road is sometimes cut very deep in the soil, which consists of red clay, confining all travellers to one narrow path, prevents it from being quickly overgrown, as would otherwise еп.» No single botanist in this country accomplished so much in the Nineteenth Century to create the image of American botany as did Asa Gray. John Torrey, botanist, chemist, and mineralogist, dif- fered from Gray, though they had characteristics in common. Both had a deeply religious turn of mind, although I have always suspected that Asa Gray attended Church as a proper Bostonian, whereas Torrey, a more retiring Christian, seldom mentioned his religious associations. Both took medical degrees, and both preferred the magnifying lens to the stethoscope.“ John Torrey wrote to Professor A. P. de Candolle at Geneva on January 6, 1837, “whenever I can find time I continue to work on my Florae Amer- icae Septentrionalis on which I have been engaged for several years past. We are now doing much in N. America to perfect the Flora of our country." At this time he was expecting Gray to join the Wilkes Exploring Expedition. “Dr. Gray who has worked by my side for nearly three years past, is to be the botanist. He is well qualified for the situation. . . . The United States can well afford to send out this expedition as we have 50 millions of dollars in our treasury and have paid all our debts. I hope we shall yet have a national botanic gar- деп. ...* Gray, however, declined, and together they proceeded with the Flora. he Flora of North America was planned, fos- tered, and stands today as Torrey's monument, which Gray helped to raise. Gray worked hand in glove with Torrey, as Bentham and Hooker worked across the Atlantic. Gray's trip to Europe in 1838 was the first trip of an American botanist to visit European museums expressly to study type spec- imens essential to the writing of a flora. By 1867 Asa Gray had many loyal followers with an almost idolatrous devotion. Frequent ref- erences in letters between botanists to “the Good Doctor Gray" attest to the fraternity of spirit. Gray had created a garden, a garden of living and dried plants, and by this time a garden of memories. It may fairly be said that he had even built an empire. To summarize Torrey's influence: he contributed to a sound taxonomy of the Cyperaceae; he fostered the work of many botanists including one, Rafin- esque, with whom he disagreed; he wrote an ex- emplary pioneer state flora; and he prepared many descriptions of new species for the government- sponsored Mexican Boundary Survey and the Pa- cific Railway Reports. Torrey was 55 when they were Берип.“ Asa Gray was a man of sound scholarship, broad culture, and of compelling personality, yet he was jealous of his preeminence. Alphonso Wood, with neither the talents nor the advantages of Gray, competed successfully in the textbook field. The first edition of Wood's Class Book of Botany (1845) promptly sold out, the second edition and subse- quent printings from stereotyped plates held their popularity. All the while the author was traveling and learning new floras. He made a major revision in 1861, and others followed throughout his life- time. Altogether between 800,000 and one million copies of Wood's botanies were sold. Henry Nich- olas Bolander, author of a catalogue of limited popularity on the plants of San Francisco area (1870), wrote enviously to Аза Gray in 1866, “What is to be done with Professor Wood?" Columbus Ohio bryologist Leo Lesquereux wrote to Professor Bolander in 1864 about an endemic feature of California: “We were greatly concerned about your shocks of earthquake and happy to know that you had escaped a living burial or some- thing of that sorte.” Who conquered the New World? For the nat- uralist there was no real conquest, only a quest. And the search continues. In 1841 George Ben- tham named the genus of plants Heliamphora based on Schomburgk's collection from Roraima. Today nine heliamphoras are known from the Gua- yana Shield.? The pigeonholes of this garden's herbarium hold verified undescribed taxa. What of those “‘indehiscent capsules”? I taught plant taxonomy at Tulane, disseminating the def- inition that a capsule was dehiscent except when it was indehiscent, for example, in the Brazil nut. The American naturalist Herbert Huntington Smith wrote of the “‘nut-capsules ... five inches in di- ameter and weighing two or three pounds; falling a hundred feet or more, they come crashing through the branches like cannon balls, and bury themselves often six inches deep in the ground. You can imag- ine," said Herbert Smith, **that a man's skull would be small proof against such a missile.” He had gone to Brazil at age 19 with his Cornell professor, C. К. Hartt, in 1870 and later collected off and on until 1902 in South America, Mexico, and the West Indies. In 1910 Smith became curator of the Alabama Museum of Natural History. Deafness Volume 78, Number 1 1991 Ewan Who Conquered the New World? overtook him in time, and as he was walking on March 22, 1919 along a railroad track he failed to hear an approaching locomotive.** To get back to one of those indehiscent capsules, the poet Ca- tullus in the first century BC wrote about what he called a “little chest”; he called it a capsula with no mention of indehiscence. John Lindley, master lexicographer, stressed the тм іп 1830, just as Leeuweuhoek in rote of the “capsula breaking upon the ripening ot the seed.” ho really knows who conquered America? Can we encapsulate the story indehiscently? Definitions flow. Omnia fleurit. The pictures drawn of Barro do Rio Negro by Wallace and Spruce are different from the Manaus I saw in 1985. How long will the New World remain “new”? Capsules generally crack open, but what about the lecyths? They rot. Time is short and the water rises. Unfortunately one lifetime seems hardly enough to allow a scholar full mastery of every aspect of huge subjects like these. Peter Gayl, Dutch historian said: ‘History is an endless debate and he who delights in debates with historians enjoys the spice of historical writ- ing." But, as Samuel Johnson said, “I may freely, without shame, leave some obscurities to happier industry, or future information." NOTES 2. Samuel Eliot Morison, Admiral of the Ocean Sea: a Life of Christopher Columbus (Boston, Little Brown, 1942) 2 volumes. 1: 60. One volume edition (1942) lacks the chapter notes. . Ibid. 1: 325. . Quoted in Carl О. Sauer, Early Spanish Main (Berkeley, Univ. California, 1966), 133. Morison, = о tanical ideas with special reference to the Sixteenth Century" in First Images of America. The Impact of the New World on the Old, Fredi о editor (Berkeley, Univ. California, 1976), 8 history, origin and taxonomy e of the pineapple," /nterciencia 14 (1989), 235- 241. сл . Alfred W. Crosby, Jr., Columbian Exchange. Bio- logical and Cultural Consequences of 1492 (West- port, и Greenwood, 1972), 171, 192. Ibid., 5. А з. 2: 102- . Sterling А. Stou — editor, Oviedo's Natural History of the West Во 1526 (Chapel Hill, Univ. North Carolina Studies Romance Languages and Lit- eratures, 19 . Amerindians used three body paints: red, fro m bija or urucü or achiote with vari- ant spellings, de fruit of Bixa; bluish black, from jagua, or genip, genipapi, or jenipapeiro, Genipa со че americana L.; yellow, from bark of Bixa. See Hand- book of South American Indians, Julian H. Steward, editor, (Bur. Amer. Ethnol. Bull. 143, 1948). 7 volumes. 9. Diagnostic phrase-names of Bauhin: Arbor Mexi осапа fructu castaneae coccifera. Pinax (1671), 10. к m y ci botanicum (London, 96), 272. pl. 209, 11. ко S(mith) in john Lindley and Thomas oore, Treasury of. و‎ revised edition (London, Longmans, Green, 1889), 148. 12. Clements R. Markham, Expeditions into the Valley of the dice: 1539, 1540, 1639 (London, Hak- luyt Soc., 1859), 6-7 12a. For Joseph de Jussieu in Ecuador see Arthur B. Steele, Flowers for the King (Durham, North Car- dm, Duke TM. 1964), 20- 22; Frans Stafleu, BIA Li Intern. Assoc. nt Taxonomists, 1971), 278- 279. For Jussieu's deserun with Lamarck see Ludwig Diels, “Bei- ü Vegetation und Flora ' Bibliotheca Botanica 116, (Stuttgart, ), 48. 13. J. Ewan, “Early History" in J. Ewan, editor, Short Galvezia speciosa (Nutt.) A Condor 18 (1916): 3-14. pp. 11-12. T "Notes on persons whose names appear i menclature of California birds" Condor 30 (1928): 261-307, р. 278. Howard Campbell, “William Gam- bel, Naturalist" New Mexico Wildlife 26(6) (Nov.- Dec. 1978): 2-9, mentions Gambel's passing through St. Louis in April-May 1841, enroute to Indepen- dence, Missouri, but his diary (if any) is lost. 14. J. Ewan, * *Notable living тонар т and botanic gar- dens" in Ian MacPhail, Hortus Botanicus, the Bo- tanic Garden and the n (Lisle Illinois, Morton Arboretum, 1972), 118—1 15. Emily W. Emmart, Badianus Manuscript (Balti- more, Johns Hopkins Univ., 40). Martin de la Cruz, author of the Aztec text, latinized by Juannes Datura meteloides DC., yageurs naturalistes du clerge francis En 5 Revolution (Paris, Lechevalier, 1932), 11-21. 18. Annals CAR Bot. Garden 57 (1970): 92. was one of th a in Linnaeus, op. PL 0759), 208- 209. The — New World ovate-leaved **yoyote" or “yoyotl” is now Т. ahouai (L.) A. DC. Linnaeus adopted the Indian name given Annals of the Missouri Botanical Garden ~ © ~ pat N N N о N с -~J © ~ ыы > د wh‏ in Thevet’s Singularitez de la antarctique la France (1588), 66. The third, Old World Cerbera manghas according to Herbert Huber in M and F. R. Fosberg, Flora of Ceylon (Washington and New Delhi, 1983) 4 histories of Linnaean pim are not unusual. . John Gerard, Herball VERE 1633, Johnson ed.), pp. 15 А Crosby, 177. See note 4. — also J. 7. Finan, “Maize іп the great herbals," Annals Missouri Bot. Garden 35 (1948) 149-191. . Joseph Ewan, “ТҺе Columbian discoveries and the growth of botanical ideas with special reference to the sixteenth century” in First Images of America, edited by Fredi Chiappelli = California, Univ. California Press, 1976), 2: 807-812, n. 4. . David Beers Quinn, “New geographical | Lit- erature" in Chiappelli, ea Pp. 637-638 . Stoudemire, 86-87. See . J. & N. Ewan, John Ba nister and his Natural с of Virginia, 1678-1692 (Urbana, Illinois, 1970). : 53, 54. Such involved . Feuillée's diagnostic phrase name, Mulli foliis non serratis, appears under Schinus areira L., pl. 389, characterized by entire leaflets, the usual vari- anical Magazine with serrated leaflets, is the Linnaean S. molle. Feuil- lée's works were noticed by J. Ewan in Herbaris 25 (1959), 60-64, and in Amer. Fern Jour. (1960): 26-32. . J. Ewan, “Plant collectors in America: backgrounds Ји Linnaeus" in Essays in Biohistory (Utrecht, Regnum Veg. 71. 1970) 22, 50. Fournier, 53-59 (see note 16). Typification problems reviewed by T. Gillis and W. T. Stearn, Taxon 23 (1974): 188- 190. I. Urban, Symbolae antillanae (Berlin, 1898) 62. . See note 26. . J. C. Beaglehole, Exploration of the Pacific (London, A. & C. Black, 1934, 3rd edition. 1966) and Life of Captain James Cook (Stanford Univ. Press, 1974), passim. . J. Ewan, "Silk culture in the colonies, ал particular reference to the Ebenezer colony and the first local flora of Georgia" Agric. Hist. 43 (1969) 129-141. . Kristian Hvidt, Von Reck's Voyage. Drawings and Journal of Philip Georg Friederich von Reck (Sa- 4. Pastor Ha Me Ichior Мане nberg, mentioned as a visitor, not to be confused with botanist Rev. Heus Ernest Muhlenberg, who did not visit Georgia. А "a а 140-141. See note 29. „1. andy, Sloane Herbarium (London, British Ma Nat Hist.) 1958), 139-140. Urban, 3: 62. See note 26. ~ . Herbert H. 02. Bra сл w . Ewan (1969), 44-47. See note 29 5 Robin Furneaux, 7he Amazon (London, H. Hamil- ` ton, 1971), 97. . Hooker correspondence, volume 68, Royal оо Garden, Kew. С. Gardner, Travels in the Interi of Brazil ... 1836-1841 (London, eue 1846), 34. Portuguese edition, (Sào Paulo, ‚3 . Dict. des Sci. Naturelles (1820) 1 У. L. Jepson, High School Flora Г California 9. К. W. Pennell, “Historical sketch: n Verdoorn, Plants and Plant Sciences in Latin America (Wal- tham, Massachusetts, Chronica Botanica, 1945), 35- 48. . Nelson Papavero, Essays on the History of es tropical Dipterology (Sào Paulo, Univ., 1971) 65-69. . I. Urban т С. Е. P. von 2. 005 Flora brasiliensis (1906) 1 (Pt. I): 210-2 ; vig Loring Gray, ы Letters ig Gray (Bos- ње Vom 1893) 2 ibid. | 44. Ibid. ТВ. von Spix & С. F. P. von Martius, Travels in Brazil in the ears 1817-1820. (London, Long- man, etc., 1824) 1: 284-285. чи iie n, tse ten feet less” in J. E. Gunckel, Current Top D ар ience (New York, Аса- demic Te 1969), 1 0. . J. Torrey c P. de Candolle, Ar- , Jan. 6, 18 aa Conservatoire Jardin Botanique Genève. ; T M. Johnston, Jour. Arnold Arbor. 24 (1943): ; Publications of Alphonso Wood, 1810-1881, listed in Agric. Library Notes (1927) 2: 95-100. Charles ]. Lyon, “А 50 Edition Best Seller" Dartmouth Alumni Mag. 31 (March dk p» 18, 81-83. . Leo Lesquereux, Columbus, Ohic, 13 April 1864, to N. olander. Bolander papers, Bancroft Library, Univ. e ES Berkeley ire Missouri Bot. Garden zil. The Amazon and the Coast (New York, сеппе, 1879), 311. Јеап де Laet described the tree, called totocke in Novus Orbis (1640), 632. ae boldt praised de Laet's **remark- able description,” see Ew 3. Bertholletia ait are “functionally indehiscent" fall to the ground, the seeds held inside by the small- ness of the opercular opening or "plug" tually germinate. See Sc г Chapter 6 in R. E. Dickinson, editor, Geophysiology of Amazonia (New York, Wiley, 1987), 69-89. . T. S. Palmer, Auk (1920): 637-638. Anton von Leeuwenhoek, Phil. Trans. 17 (1693): 706. BOTANICAL EXPLORATIONS IN THE AMERICAN WEST— 1889-1989: AN ESSAY ON THE LAST CENTURY OF A FLORISTIC FRONTIER James L. Reveal! ABSTRACT . 4, > апу Атепсапз the term “frontier” evokes images of cowboys and Indians, but for historians the “frontier were lone generalists; they were followed by institutionally sponsored collectors in organized expeditionary groups. These naturalists typically presented their discoveries With the settlement of the Initially the first floras w evolved into ез illustrated regional floras with а monographic format. The fr ard movement of institutional power caused by i there should be a westw to urban-situated * frontier came local naturalists who published their own discoveries. Published eoa evolved ver time as each region viec settled, with more details being included as the native plants a “professionals,” who published the results. s became nown. e expansive and general. These were followed by concise, descriptiv e local ma uals s, whic ontier hypothesis i that ncreased conservatism in the established eastern institutions о with а movement of talented people closer to the frontier. The ани ауа also states that . The key to the future of systematic botany is to fight the tendency to лана of the lolas] frontier in the continental United States, the systematic community must decide bos it will on in their numbers, ater ir of ideas ess new ideas. With h and power to consolidate its institutions, what its missions will be, and how it will attempt to remain creative. For many Americans the term “frontier” evokes the Wild West images of cowboys and Indians. Generations of pulp magazine readers, radio lis- teners, moviegoers, and now television viewers have grown up with the exploits of Buffalo Bill and Billy the Kid. To the younger generation who missed the weekly adventures of Roy Rogers and Trigger, characters or the setting, it is the idea of a frontier that is important here—the concept of one or more adventuresome individuals searching for some un- known in a region that is itself largely unexplored. The long-romanticized American frontier is dis- tinct in character from the staid colonies of the eastern seaboard, and vastly different from that of Europe, where “frontier” alludes to a boundary between populous nations. The American character was shaped by the frontier experience and this in turn molded a unique civilization. No single event or feature can be credited with this; our European heritage of customs and institutions contributed but were changed by a mingling of peoples. All even- md were transformed by the industrial revolution d th and military growth of the United @ 5 States into a world power. Taking a step back from these larger trends one finds, as the American historian Frederick Jackson Turner found nearly a century ago, that along the American frontier people and institutions underwent repeated rebirths of civilization at the raw edge of settlement during the four centuries required to cross the continent. Turner formulated the “frontier hypothesis.” He believed that one of the factors that drove the frontier westward from the fall line along the East Coast was the availability of free land on the edge of advancing settlement. Into that unoccupied wil- derness came new immigrants who hoped to better themselves. With the institutions. and traditions of politics, economics, and religious and social customs. These were adapted to fit the more primitive conditions of the frontier, and as the m came ' Department of Botany, University of Maryland, College Park, Maryland 20742-5815, U.S.A. ANN. Missouni Bor. Ganp. 78: 65-80. 1991. Annals of the Missouri Botanical Garden frontier moved even further west these modified institutions and traditions changed again. he primitiveness of the frontier's edge did much to force the reshaping. Highly developed institu- tions and customs gave way to more simple asso- ciations and more pastoral pursuits. After the third century of westward expansion, reaching, as the great arch in St. Louis now symbolizes, the begin- ning of the trans-Mississippi West, the people, in- stitutions, and customs were already significantly different. Each generation on the advancing fron- tier had transformed its antecedent with an evo- lutionary selective process as significant as any seen in the biological world. Adaptation and change were common to all aspects of the frontier existence. Laws and customs were based on those found just to the east on what previously had been a line of frontier settlement. As each frontier community became more settled, the structure of the institu- tions became more complex and local customs more sophisticated. Governments and institutions be- came rigid and conservative with the economic structure specialized. Religious and social customs became stratified into specific beliefs and social classes. In turn, these would be contributed to the adjacent community forming on the edge of the wilderness in a freer, less-sophisticated, and more rebellious form oupled with this was the imposition of the ever- changing local climate and environment. Just as there were continuous rebirths of civilization on the edges of the American frontier, so too did new ways and customs evolve and adapt to the changing environment. To Turner this was significant for it resulted in the **Americanization" of people, cus- toms, and institutions. The American character, he felt, was the result of the evolution and adap- tation of society to fit the ever-changing physical environment. Into this grand equation we now place the plant explorer. By examining this minor character we might better understand: the field of American systematic botany; the historical significance of the large eastern institutions and why it is that the power base is moving westward; the causes leading to an increase in conservative thought and the narrowing of the power base to a few; the increasing failure of institutions with an historical mission of supporting systematics; the concentration of fund- ing and opportunities to do research; and the fleet- ing importance of the methodologies promoted by the different factions in systematics. Perhaps then we might also better contemplate the future, for as Turner dated the end of the American historical frontier as 1890, so too must we now date the end of the American floristic frontier as 1990. What lies beyond is our future. Much as Butch Cassidy and the Sundance Kid left Utah to find a more exciting frontier, so too are many of our systematists leaving temperate North America for the wilds of South America. This is not unexpected because the frontier hypothesis predicts it should be so. In the continental United States we have a known flora and an ever-dwindling number of nov- elties. With the disappearance of our floristic fron- tier, so too are those who long for the excitement of the unknown and the thrill of discovering new frontiers. BOTANICAL EXPLORATIONS AND THE FRONTIER HYPOTHESIS To a significant degree, botanical exploration in the American West is as much a tale of romance and adventure as that of any fur trapper or cowboy. The tales of those who risked life and limb to find a plant simply for its uniqueness and beauty have long been the stuff of novels and television movies. The first botanical explorers in temperate North America were lone individuals who were nearly always transients. Many were ship's surgeons or ministers. Others were members of the military, surveyors, physicians, or wives. Some collected to win a favor from one of Europe's rich and powerful. But there were real collectors, too, who sailed to America and probed the rich forests along the great rivers and onto the backbone of the Appalachian Mountains. through these mountains than naturalists were along them, collecting whatever they could find. Both the Atlantic and Pacific coasts had early explorers. In the 1600s, plants from the colonies were taken to Europe. A century later, ships from several nations roamed the inlets of the West Coast No sooner had paths been pushe gathering all kinds of plants, not just trees and shrubs as had been the trend prior to the 1720s. For most Europeans, interest in the flora of eastern North America waned after 1700 with the discov- ery of the horticultural treasures found in China and Japan, and even the West Coast of North America attracted little interest until the diversity of the inland forests became known in the 1820s. Still, naturalists came. For the new nation of the United States, science was largely neglected until the presidency of Thom- as Jefferson. Jefferson sought and obtained con- gressional support for a party of explorers to ven- ture up the Missouri Hiver and to cross the Volume 78, Number 1 1991 Reveal Botanical Explorations in the American West mountains to the Pacific Ocean. They were to survey the natural history and to observe the native peoples and their customs. The land was to be mapped, the features named, and the geology un- derstood. Thus, Lewis and Clark explored the wil- derness beyond the frontier at the mouth of the Missouri River. The plants they found were des- tined to be taken to Europe, described by foreign- ers, and essentially lost for nearly a century. Other expeditions followed. One went up the Red River of the Arkansas and another followed the Arkansas itself. Neither was successful scien- tifically. William Bradbury and Thomas Nuttall, each associated with competing trading companies, traveled up the Missouri River in 1811 to the Mandan villages of North Dakota, but the War of 1812, more thievery, and other unforeseen fates befell the men and their collections, rendering them less Yun than they might have been. e West Coast, after half a century of coastal не by the French, English, and Russians, David Douglas explored the interior and collected hundreds of new species, many of which made their way into European gardens. Like the fur trapper's beaver pelts that were taken to foreign lands where others profited, so too were North American plants sent to Europe where they were described and classified. The exploitation of our natural resources is predicted by the frontier hy- pothesis and is expected at this stage of develop- ment on any new frontier. European naturalists who sought plants in the North American wilderness after 1800 were often supported by scientific organizations, companies, groups of wealthy men, or governments. For the new nation, there were few formal organizations able to support natural history expeditions, and with almost no wealthy class the task fell mainly to the federal government. Only after the cream of the botanical novelties had been gathered by others did American scientific expeditions begin to move across the continent. Previous generations had found the obvious; this one collected critically. The year 1850 was significant for by then the frontier was west of the Mississippi River. The gold rush to California was at its height. Texas had been added to the nation, the Oregon question settled, and the land from the crest of the Rocky Mountains to the California coast acquired as a result of a war with Mexico. No sooner had the frontier come within sight of the western boundary of the United States than a vast new wilderness was added to the nation's map. Still, there were scattered settlements on the fringes. The Oregon, California, Mormon, Santa Fe, and Spanish trails were in constant use. Farming communities had been established in the Central Valley of California and the Willamette Valley of Oregon, and the Mormons had arrived in Utah. The small mission villages along the Cal. ifornia coast from San Diego to San Francisco had been augmented by a series of boom towns dotting the foothills of the Sierra Nevada, and gold and silver would soon be found elsewhere. To the north, the old forts of the Hudson Bay Company were giving way to new fishing, lumbering, and farming communities. s before, naturalists came to the frontier, crossed it into the botanical wilderness, but in doing so came to a new frontier moving eastward. The unexplored regions of the American Southwest and the Great Basin attracted greater attention, with the two massive cordillera, the Rocky Mountains, and the Sierra Nevada-Cascade Range bringing many to search for their novelties. Government expeditions to map the interior of the American West provided numerous opportunities for natu- ralists to explore remote regions. On both edges of the frontier, institutions were being established, indicating а higher generational order near the frontier. In St. Louis, George Engelmann and Неп- ry Shaw were joining forces and the forerunners of the Missouri Botanical Garden were taking shape, while in San Francisco, a group of naturalists led by Albert Kellogg were coming together to establish the California Academy of Sciences. The Civil War years of the 1860s, and then a decade of Indian wars in the 18708, took their toll on scientific efforts. Many naturalists who might otherwise have gone west in search of scientific curiosities were delayed or hesitant. When the tur- moil ended, many went west, some to become local amateur naturalists, some to be professional col- lectors, and a few to serve in the new scientific military or to join the few academic institutions offering botany. In time, as Congress reduced the role of the military, more and more of the scientific research done in the wilderness was carried out by civilian agencies By the 18708, the wilderness beyond the frontier was restricted to the Rocky Mountains, Great Ba- sin, and desert regions of the Southwest. Into these areas, as predicted by the frontier hypothesis, came lone naturalists in this rebirth of civilization on the botanical frontier. This generation was more edu- cated and critical than previous ones, reflecting the increased sophistication of their immediate ances- tors. Some were supported by the federal govern- ment, but many were associated with colleges or 68 Annals of the Missouri Botanical Garden with scientific institutions. To a lesser extent, some were private collectors interested in selling their specimens. Most came to the frontier and entered the wilderness only for the growing season, but by the 1880s, some devoted all of their energy to settling the American, as well as the botanical, frontiers. In its simplest form, the closure of the American frontier in 1890 was signified by the settlement of the land by people. The botanical frontier lagged behind for there were still remote regions to be explored for their botanical treasures. The lure of new species drove naturalists into the wilderness where the discovery of novelties was still routine. For the botanical frontier to close, those remote regions had to be explored and their botanical nov- elties discovered. THE CLOSING OF THE BOTANICAL FRONTIER The last century of botanical explorations and discoveries, the writing of floras and florulas, and the development of today’s botanical institutions did not occur in a vacuum. There was a past, and that past molded what is in the present. Numerous collectors came to temperate North America before and during the Linnaean era, and these collections were used by Carl Linnaeus to describe nearly 90 species from temperate North America in his 1753 two-volume work, Species Plantarum; the major- ity were based on specimens obtained by just two men, John Clayton and Pehr Kalm. Even the ma- jority of North American plants named by Linnaeus after 1753 were based on the collections of John Bartram alone. The exploits of post-Linnaean naturalists in east- ern temperate North America did much to settle that floristic frontier by 1825 with Thomas Walter, John Fraser, André Michaux, Frederick Pursh, Thomas Nuttall, Stephen Elliott, Jacob Bigelow, John Torrey, and Amos Easton being major con- tributors to the effort. While Nuttall continued to explore on the edge of the frontier, Torrey remained in the floristically settled East where he rose to a position of influence and became the first to dominate systematics in temperate North America. That mantle was soon passed on to Asa Gray at Harvard University, who became the single most influential man in our fron- tier history. Together, Torrey and Gray imposed their will up to To settle in a wilderness is not to conquer it. Although Torrey and Gray did much to describe the flora of temperate North America, much like Linnaeus before them, they too relied upon others to discover the novelties. It was those people who closed many of the botanical frontiers across the continent. For the Northeast, Torrey and especially Gray made substantial contributions. They were followed at Harvard by Sereno Watson, Benjamin L. Rob- inson, and Merritt L. Fernald, all of our century," 1889-1989. These men had a competitor in the form of Nathaniel Lord Britton and later Henry A. Gleason at the New York Botanical Garden. Each man championed a series of flora projects, students, and institutions as shall be noted below. The honor of closing the floristic frontier in the Northeast, however, must be granted to Fernald. In 1950, following nearly half a century of devoted fieldwork on his part and a careful study of type material, Fernald published a fitting tribute to all who had followed Asa Gray. The richness of the southeastern United States was discovered initially by the likes of Mark Cates- by, Thomas Walter, and the father and son teams of Bartrams and Michauxs. Overlapping our cen- tury were Alvan W. Chapman, Samuel B. Buckley, and John L. Riddell. John K. Small did much to reduce the botanical frontier in the South during the first half of our century, but only now, thanks to the efforts of Robert W. Long, Olga K. Lakela, Robert K. Godfrey, Daniel B. Ward, and Walter 5. Judd in Florida, Albert E. Radford, C. Ritchie Bell, and their colleagues in the Carolinas, and Delzie Demaree, Robert Kral, M. Murray Evans, and R. Dale Thomas elsewhere in the South, has its botanical frontier closed. The flora of the Old Northwest was almost al- ways considered synonymous with that of the Northeast, a fact reflected by the regional floras of the area. The university community was active in the search for new species, but the individuals were largely dominated by those at Harvard and New York. State floras were written and are now being written, but without a definable region where novelties are to be expected, one must conclude that the botanical frontier ended in 1950, when it closed in the Northeast. Early plant explorers on the Great Plains were mainly transients who hurriedly crossed the region to get to the mountains of the West. The Great Plains were the deeded home of many American Indians, and travel off established routes was rarely done. With the gradual demise of the Plains In- dians, and in the 1860s and 18705 the loss of their homeland, plant explorers were able to penetrate beyond the limits of the frontier trails established by the cattlemen, miners, and cavalry. Even the Volume 78, Number 1 1991 Reveal Botanical Explorations in the American West resident naturalists were often transient. Albert S. Hitchcock left Kansas for Washington in 1901, and Per Axel Rydberg left Nebraska in 1895, eventually accepting a position at the New York Botanical Garden in Unlike in the Old Northwest, specimens col- lected by resident plant explorers on the Great Plains tended to remain in local university herbaria. This was, in part, because the Great Plains were sufficiently far enough removed from the eastern centers to evoke a degree of independence, but also the men and institutions themselves had come from a more isolated, former frontier. Persons like Charles E. Bessey in Nebraska and Theodore D. Cockerell in Colorado collected for years, trained numerous students, and built large herbaria. To be sure, outsiders continued to come to the Great Plains, and there were other resident naturalists who collected throughout much of the area, yet credit for closing that botanical frontier must be granted to the late Homer A. “Steve” who made some 90,000 collections before he re- tired in 1975 Botanical explorations in Texas began in 1828 with Jean Louis Berlandier, who was followed by Thomas Drummond in 1833, by Ferdinand J. Lind- heimer, Auguste Trécul, and Charles Wright in the 1840s, and ES ge Reverchon, who collected from 1870 to . The remarkable diversity across the 2 to say nothing of its size, meant that the botanical wilderness in Texas was more a series of pockets than a single definable line. The work of Victor L. Cory, Lloyd H. Shinners, Barton H. Warnock, Frank W. Gould, Donovan S. Correll, Billie L. Turner, and Marshall C. Johnston led ultimately to a 1970 flora of Texas and brought its floristic frontier to an en The West Coast floristic frontier differed dra- matically from its eastern counterpart in many aspects besides the great diversity in its flora. Here Stephens, naturalists were often far in advance of others in the exploration of the West, and the naturalists demonstrated their independence and adventure- someness by their very presence. Most significant of all, they were well isolated from the eastern establishment so that they were able to make crit- ical decisions on the future of their profession. At first Collignon collected the first plants near Monterey, California, in 1786—some 280 years after the first plants were taken to Europe from the East— and Jean Baptiste Lamarck described the first trans- Mississippi species, Abronia umbellata, in 1791. the discoveries were slow. Jean-Nicolas The Spanish, English, Russian, and American gov- ernments sent other botanical explorers to the West Coast so that the roster of collectors grew rapidly, but it was David Douglas who was destined to be the most formidable, for he forced the botanical frontier away from the Pacific Ocean. Others en- tered the Pacific Northwest by land, following the Oregon Trail with Nathaniel J. Wyeth, Thomas Nuttall, Karl A. Geyer, and John C. Frémont fol- lowing that same route. Since 1889, many have devoted much of their professional life to closing its botanical frontier. In Washington, Wilhelm M. Suksdorf, Charles Vancouver Piper, R. Kent Beattie, F. Marion Own- bey, and Harold St. John explored many parts of the state. Fieldwork was done in Oregon by Thomas Jefferson Howell, William C. Cusick, and Morton E. Peck, with lesser contributions made by Louis F. Henderson and LeRoy E. Detling. More recently, Helen M. Gilkey and Kenton L. Chambers and their students and associates have been actively collecting in the state. In adjacent Idaho, natural- ists tended to pass through the region rather than explore it for its novelties. Still, the team of John H. Sandberg, John B. Leiberg, Daniel T. Mac- Dougal, George Vasey, and Amos A. Heller col- lected extensively in northern Idaho around the turn of the century. Louis F. Henderson of the University of Idaho concentrated his efforts in east- ern Idaho, and Aven Nelson of the University of Wyoming and several of his students, including James F. Macbride and Edwin B. Payson, collected in Idaho during the first two decades of this century. John H. Christ of the Soil Conservation Service collected nearly 10,000 numbers in the 1930s when Ray J. Davis, who started the herbarium at Idaho State University in 1930, was encouraging others to collect. Among Davis’s students were Arthur Cronquist and my own father, Jack L. Re- veal, who gained a lifelong appreciation of grasses from Davis. The team of C. Leo Hitchcock and Clarence V. Muhlick actively collected throughout the Pacific Northwest. It is to them and to 1973, when Hitch- cock retired, that one must credit the closing of the floristic frontier in the Pacific Northwest. The diverse flora of California has long attracted botanists. Its botanical frontiers moved away from the coast into the mountains and deserts early in its history so that California had its own published flora by the mid 1870s. The cast of post-1889 characters was, and even today is, remarkable. Mary Katherine (Curran) Brandegee and Alice Eastwood are two of the more notable characters. Both collected actively throughout much of the American West. Their lives were closely interwo- ven as Eastwood succeeded Brandegee, in 1894, Annals of the MEE Botanical Garden as the curator of botany at the California Academy of Sciences. Both had ardent followers and sup- porters, and each in turn promoted botany as a science. The university scene was equally gifted with the likes of Willis L. Jepson, who replaced Edward L. Greene in 1895, and like Eastwood, remained active in botanical explorations for half a century. At Stanford University, the reign of William В. Dudley began in 1892, with LeRoy Abrams joining the staff in 1902. Berkeley and Stanford graduated many students who continued explorations in California. Jepson students included Harvey Monroe Hall, who graduated in 1906; Ivan M. Johnston, John Thomas Howell, David D. Keck, and Conrad V. Morton, all of whom graduated in the 1920s; and Lincoln Constance, who finished his degree in 1934. Herbert L. Mason went from Stanford to Berkeley, receiving his doctoral degree in the early 1930s whereupon he joined the Berke- ley staff. Also from Stanford were such active col- lectors as Elmer I. Applegate, Ira L. Wiggins, Charles Piper Smith, and Roxana J. Ferris. Numerous collectors were associated, in some fashion, with each of the three institutions in the San Francisco Bay region. At the Academy, where John Thomas Howell held the curator's chair after Eastwood, numerous knowledgeable “amateurs” came for guidance. The list is long, but one cannot ignore the likes of Ernest C. Twisselmann, Gordon True, or Gladys L. Smith. Samuel B. Parish and Charles Russell Orcutt, both active in southern California from the 1880s until the 1920s, con- tributed their personal herbaria to Stanford. Adolph D. Elmer collected widely in California, with his specimens going to Stanford. Numerous individuals were attracted to Berke- ley. Ethel K. Crum, Annetta M. Carter, Helen K. Sharsmith, and Lauramay significant collections. Contributing their herbaria to Berkeley were such collectors as John СШ Lem- mon, Joseph P. Tracy, and Ira W. Clokey. In addition to the large collections made by Jepson, Dempster each made Mason, Constance, and more recently Robert Orn- duff, who were on the Berkeley faculty, the efforts of Jepson Herbarium staff, notably Rimo Bacigalupi and Lawrence R. Heckard, made great inroads into many rarely collected portions of California. In southern California, Pomona College came to be the focal point of plant explorations and systematics. Philip A. Munz, who became director of the nearby Rancho Santa Ana Botanic Garden, was deeply involved in floristics. From 1917, when he joined the college's faculty, until his death in 1974, Munz was devoted to the California flora and made numerous collections. His students read like a “who's who" in plant explorations both in California and beyond: F. Raymond Fosberg, C. Leo Hitchcock, George J. Goodman, David Keck, and Louis C. Wheeler are just a few. Munz con- centrated his efforts in the mountains and deserts of southern California, but made occasional ex- cursions to the North. His student and eventual colleague Lyman D. Benson was also a formidable collector, albeit more specialized. The basis of the Pomona College Herbarium was the private herbarium of Charles F. Baker, who amassed a collection of some 100,000 specimens. То this, Munz added the herbarium of Marcus E. Jones in 1923; it too numbered approximately 100,000 sheets. The collections of Frank W. Peir- son, M. French Gilman, Percy C. Everett, and Carl B. Wolf may now be examined at the Rancho Santa Ana Botanic Garden. In the years since Munz died in 1974, many new species have been found by persons willing to explore the most remote portions of California. ary C. DeDecker discovered Dedeckera, which was described in 1976, and James R. Shevock is still uncovering novelties. Yet, the finds are few and generally scattered so that it seems appropriate to close California's botanical frontier in 1976, a century after the state's first flora was published. The Rocky Mountains have always attracted the botanically curious. When Joseph Dalton Hooker came to America in 1877, Asa Gray took him to the Rockies. Numerous naturalists came by train to spend the summers in the coolness of their height and to regale in their splendor. In 1889, no other region of the United States was attracting such a diversity of botanists. Per Axel Rydberg, Marcus Jones, and Joseph W. Blankinship did much to explore previously unstudied portions of Montana, with Charles Bessey sometimes joining Rydberg. C. Leo Hitchcock collected in western Montana, especially during his tenure at the University of Montana from 1932 until 1937. Several univer- sity-based professors and their students have climbed “ат- ateur" Klaus Lackschweitz among the forefront of recent collectors. ven Nelson is synonymous with floristics and W yoming. This remarkable man, a botanist, teach- er, and president of the University of Wyoming, was productive throughout his 97 years. He сој- lected widely in the state, expanding its floristic the mountains in search of plants, with the frontier away from the narrow Oregon Trail and Central Pacific Railroad route where nearly all before him had concentrated their efforts. He en- couraged others to collect, and the students he influenced to continue in systematic botany are Volume 78, Number 1 1991 Reveal 71 Botanical Explorations in the American West now legendary: Elias E. Nelson, Leslie N. Good- ding, James Francis Macbride, Edwin and Lois E. Payson, George J. Goodman, Louis O. Williams, and Reed C. Rollins. C. L. Porter, who came to the University of Wyoming in 1929, continued the tradition of exploring the wilds, concentrating in the 1930s on bryophytes and then vascular plants. The massive field efforts by Ronald Hartman in the 1980s, although still demonstrating the need for additional fieldwork, have conquered the last vestiges of a botanical frontier in Wyoming. The Colorado Rockies have been a beacon to naturalists since the first specimens were taken by Thomas Say and Edwin James in 1820. The work of Cockerell, Rydberg, Charles Penland, George E. Osterhout, Harold D. Harrington, and a host of others who have explored the Rockies since 1889 have allowed William A. Weber of the University of Colorado, long a persistent and diligent collector, to be the last to collect on its botanical frontier. Across the American Southwest, the floristic frontier moved slowly, being pinched in from both the East and the West. The first to collect in New Mexico was William Gambel, who traveled the Old Spanish Trail from Santa Fe to Los Angeles. During the war with Mexico, Major William H. Emory, Lieutenant James W. Abert, and Charles Wright entered New Mexico, collecting specimens when- ever possible. Augustus Fendler and Friedrich A. Wislizenus arrived in Santa Fe with Emory and his "Army of the West"; Fendler remained in New Mexico collecting in the mountains to the north while Wislizenus turned southward into Mexico. Thomas Coulter was the first to explore Arizona botanically, visiting Yuma in 1832. Naturalists as- sociated with other military expeditions, and es- pecially the Corps of Topographical Engineers, dis- covered numerous botanical novelties. The most important collector of our **century" was Edward Palmer, who gathered plants in the Southwest for nearly 40 years. Others have searched for plants, with many finding a surprising number of novelties in the more remote regions since 1930. The floristic frontier in the Southwest was gradually reduced through the efforts of Forrest Shreve, Thomas H Kearney, and Robert H. Peebles. The botanical wilderness finally disappeared along the border of central Arizona and New Mexico in the 1980s through the efforts of Donald Pinkava in Arizona, Richard Spellenberg in New Mexico, and their many colleagues and students. The floristic frontier swept into the Intermoun- tain West from all directions. The historical Cali- fornia and Spanish trails to the Pacific Coast had been rapidly botanized prior to 1889. Like in the Southwest, many military-sponsored expeditions crossed Utah and Nevada, with the geological sur- veys led by Clarence R. King, and especially by John Wesley Powell and George M. Wheeler, put- ting their botanists into exceedingly remote areas. The university communities in Utah and Nevada encouraged botanical explorations by their faculty, and the faculty in turn encouraged numerous stu- dents. Among the more significant were Amos А. Heller and Patrick B. Kennedy of the University of Nevada, and Walter P. Cottam and Seville Flow- ers of the University of Utah. The post-1889 history of botanical explorations in the Intermountain West cannot be told without a brief mention of Marcus Eugene Jones. He set out to explore the whole of the **Great Plateau," as he termed the Intermountain West, and suc- ceeded in traveling throughout most of it. He col- lected wherever he went and described many nar- rowly endemic species as he amassed his herbarium. While Jones lived in Salt Lake City (1880 to 1923), he carefully guarded the frontier with his biting pen and his quick printing press. Those who ven- tured upon his territory were all too often destined to be the subject of his scorn. His desire to write a “Flora of the Great Plateau" eluded him, how- ever, and the manuscript was still unfinished when he died in a car accident in 1934. Many botanical teams came to the cold deserts in search of plants. Eastwood and Howell were active in the 1930s and 1940s, with H. Dwight Ripley and Rupert C. Barneby, Bassett Maguire and Arthur H. Holmgren, and Annie Alexander and Louise Kellogg continuing the trend into the 1950s. Noel Holmgren and I collected together in the 1960s, with N. Duane Atwood and Larry C. Higgins replacing us in the 1970s. Arnold ““]еггу” Tiehm and Margaret Williams, working mainly in Nevada, were active in the 1980s. Stanley L. Welsh and his students and associates at Brigham Young University moved it for Utah in 1987. In Nevada, Tiehm, who dis- covered a wealth of new species in isolated portions of Nevada, completed the task and may be con- sidered the last to have explored on the botanical frontier in the continental United States. THE EVOLUTION OF FLORISTIC STYLES ON THE AMERICAN FRONTIER— 1889-1989 Many of the early floras treating the known plants of the world included species that had been collected on the edge of whatever frontier existed at the time. Among the early settlers were men 72 Annals of the Missouri Botanical Garden and women willing to collect natural objects for others to study. Occasionally, some desired to pub- lish their own findings, but events and others usu- ally conspired to prevent this, as happened to John Clayton whose manuscript of the flora of Virginia was published by Gronovius, with the help of Lin- naeus, in 1739 and 1742. Forty years later, Thom- as Walter's Flora Caroliniana was edited and published in London by John Fraser. The French naturalist, André Michaux, who explored on the edge of the frontier in eastern Canada and the United States, did write his own 1803 Flora Bo- reali- Americana, but many still believe it was ac- tually the work of L. C. Richard. One of the first American books devoted to the plants of a region was the little-known 1808 Flora Carolinaeensis published in Charleston, South Carolina, by John Shecut. This work was designed for the physician. It contains excellent descriptions and surprisingly well-informed distribution statements. Only one of its two volumes was published. Frederick Pursh, after spending several years in the United States, published his 1814 Flora Boreali-Septentrionalis in London. Most floristic manuals treating the plants of tem- perate North America prior to 1814 were at best general approximations of the diversity. The style was Linnaean— short descriptions and a brief state- ment regarding distribution—as was the classifi- cation (Shecut being an exception in all respects). Both were dictated by strong historical ties with England and the influence of James Edward Smith and others of the "Linnaeans" then in London. The transition in styles and formats may be seen in the 1814 Florula Bostoniensis by Jacob Big- elow, who maintained the concept of the Linnaean diagnosis, but added a more detailed description for those plants he knew from the field. He also provided a common name for each species. In А Synopsis of the Genera of American Plants pub- lished in Georgetown by J. M. Carter, also in 1814, one finds a key to the classes and orders, but the descriptions themselves are classically Linnaean. In Stephen Elliott's 4 Sketch of the Botany of South-Carolina and Georgia (1816-1824) the diagnoses are expanded and given in both Latin and English. A full description of each species was given with the characters arranged in a manner not unlike a modern flora. Elliott provided infor- mation on abundance and the time of flowering, and cited previous places of publication and illus- trations. The use of “natural orders," term “families,” or what we would in floras was highly debated at this time. Most works, such as William Barton's Compendium Florae Philadelphicae and Thomas Nuttall's Genera of North American Plants, both published in Philadelphia in 1818, followed the Linnaean system with the genera artificially ar- ranged in nearly 60 classes and orders. In 1789, A. L. de Jussieu had proposed 100 natural orders, and these were coming into use, especially in Ger- many. The first North American work to use nat- ural orders was Florula Ludoviciana, published in New York by Constantine Rafinesque in 1817. Two years earlier, Rafinesque had proposed more than 300 natural orders. Amos Eaton was more cautious in his 1817 Manual of Botany. He arranged his plants following Linnaean principles, but in his in- troduction he presented an outline of the Jussieu system, complete with descriptions, representative genera, and medical uses. He also provided a glos- sary. Unique to this work, the genera were de- scribed in one section with the species accounted for in another. A habitat statement was given, but where іп “‘the northern and middle states of Amer- ica" the plant was found was not provided. Torrey's A Flora of the Northern and Middle Section of the United States, dedicated to Thomas Nuttall, appeared in 1823 and 1824. True to his heritage, he arranged plants in the Linnaean fash- ion. The genera were briefly described, with ref- erences given to previously published descriptions and synonymy. He also provided a common name. However, he noted the natural order that each should be assigned. The professional union of John Torrey and Asa Gray was to alter immediately the style of American floras. The Americans were no less grandiose in their plans for continental floras than their Euro- pean counterparts. The first part of William Jack- son Hooker's Flora Boreali-Americana appeared in 1833, and in 1838 Torrey and Gray began their own never-finished А Flora of North America. Like other efforts of its kind, the project suffered from an overwhelming amount of new material that could not be readily summarized. For the evolution of floristic models, however, this was to be a significant work. The usefulness of natural orders in classifying plants had been amply demonstrated by John Lindley in 1830, with the publication of his /ntroduction to the Natural System of Botany, and he had swept away the last influences of the Linnaean sexual system on plant classification. Gray was an ardent follower and in- corporated the concept in the Flora, going so far as to provide dichotomous keys to the families based on those in Lindley's 1835 А Key to Struc- tural, Physiological, and Systematic Botany, and even to genera as seen in S. F. Gray's 4 Natural Volume 78, Number 1 1991 Reveal 73 Botanical Explorations in the American West Arrangement of British Plants (1821). There were no keys to the species; rather, Torrey and Gray provided diagnostic headings to groups of species following the format designed by A. P. de Candolle for the Prodromus. More practical regional floras began to dominate the floristic scene, with the majority written for the settled portions of the United States. John Darby published 4 Manual of Botany Adapted to the Productions of the Southern States in 1841. Gray responded in 1848 with the first edition of his Manual, which was a comparable treatment for the plants of the northern states. The first edition of Chapman's Flora of the Southern United States was published in 1860. In all of these works, the plants were arranged in families, keys to the fam- ilies were provided, and some means were provided which allowed one to discriminate among the var- ious groups of species. Works for students began to appear as well. Gray published the first edition of his Elements of Botany in 1836. Alphonso W ood's Class-book of Botany, initially published in 1845 and variously titled, underwent many re- visions and editions, selling perhaps a million copies over the next 50 years. For those at female sem- inaries there was Mrs. Lincoln’s Familiar Lectures in Botany. Gray and Wood followed Lindley. W ood's key to the families, however, was bracketed and only partially dichotomous; Mrs. Lincoln tend- ed to retain the Linnaean mode In addition to floristic studies, a growing number of naturalists began to specialize on groups of plants. Mosses, bryophytes, ferns, grasses, and even cer- tain genera of flowering plants became the focus of intense study. As Gray's career evolved during his lifetime, he too became an authority on a mul- titude of families (notably Asteraceae) and such diverse genera as Епоропит and Penstemon. For those who were on the western frontier, however, the aging Gray eventually became an impediment as he seemed to become increasingly more rigid and demanding that his authority not be questioned. To the independent westerner this was intolerable, and a revolt began in the form of defiance by Edward L. Greene, Marcus E. Jones, and others who lived and worked on the floristic frontier. [t was traditional for most new species to be sent to Gray for evaluation and approval. After the mid 1870s, evaluation was left to Sereno Wat- son who, like Gray, was at Harvard University. Approval was not guaranteed. Furthermore, editors often submitted articles to Gray or Watson for comment. Some westerners alleged that their new species were thus delayed awaiting Gray or Watson to describe the same plant from material in their own holdings. Later, Per Axel Rydberg required some western authors to submit vouchers with their manuscripts so that he could make his own eval- uation of their new species. И was this pretense of eastern superiority that many on the floristic fron- tier resented and rebelled against. The result was sectionalism and its expression was in the form of a host of privately published journals. Books, manuals, and regional floras were difficult to write and expensive to publish on the frontier. Most westerners had limited resources and thus it was simpler to publish new species. Botanical jour- nals are, even today, the bread and butter of the systematic community, but they were not common prior to 1900 and most were restrictive in some fashion. The Transactions of the American Phil- osophical Society, one of the first American jour- nals, began in 1771, and the American Journal of Science started in 1818. The Proceedings of the American Academy of Arts and Sciences, long the journal used by Gray and Watson, began in 1846. Institutional journals first appeared in 1817 when the Journal of the Academy of Natural Sciences of Philadelphia was established; it was followed by Smithsonian Contributions to Knowl- edge in 1848. Some local and regional western American academies had their own publications. The Proceedings of the California Academy of Sciences published its initial volume in 1854. The Transactions of the Academy of Science of St. Louis had its origin in 1860. Current journals, such as the Bulletin of the Torrey Botanical Club, began in 1870; this was soon joined by the Bo- tanical Gazette in 1875 and Rhodora in 1899. These three eastern-dominated and establishment- sponsored journals were soon the primary outlets for individuals who were not associated with an institution to propose new species. With the advent of open journals, the institutional ones often be- came largely even more restricted. Examples in- clude the Smithsonian's Contributions from the United States National Herbarium (1890), Har- vard's Contributions from the Gray Herbarium (1891), Chicago's Publications of the Field Co- lumbian Museum (1895), and the Bulletin of the New York Botanical Garden (1896). The ancestor to the Annals of the Missouri Botanical Garden, which began in 1914, was the Annual Report of the Missouri Botanical Garden (1890). To isolated, independent westerners, these jour- nals were either unavailable or the editorial policies made it difficult for them to publish their findings. To a substantial degree each journal was edited by a strong personality who had political, economic, or institutional reasons for excluding the works of 74 Annals of the Missouri Botanical Garden outsiders. Most of the dissidents had earned their reputation as a “botanical pest” (to quote one of them speaking of another) so that very often there were sound scientific reasons why an editor should be cautious. Аз a group, the westerners were a quarrelsome lot—if one made an error, another would immediately write the editor to complain. Another factor for the independent western jour- nals, and one predicted by the frontier hypothesis, was the need to establish an independent identity. What better way than by a scientific journal? The proliferation was immediate: Greene began Pitton- ia in 1887, Townshend S. Brandegee and Mary Katherine Brandegee started Zoe in 1890, and Jepson released the first issue of Erythea in 1893. Heller started Muhlenbergia in 1900, and Marcus Jones began his Contributions to Western Botany series in 1902. With Gray's death in 1888, one year before our "century" began, the domination of the eastern establishment weakened and Watson, Robinson, and Fernald, who followed Gray at Harvard, never gained his stature and respect. Britton and Rydberg at New York, and John Merle Coulter in Wash- ington, further weakened Harvard's importance by imposing their own forms of domination. Still, western floras were written. The initial ef- forts came in the form of a cooperative effort between a university and a state agency. William H. Brewer of the California Geologic Survey and Watson of Harvard University published the first volume of their flora for the state in 1876. Volney Rattan followed with his privately published А Pop- ular California Flora in 1879. Greene, who be- came professor of botany at the University of Cal- ifornia in 1885, privately published the first part of his Flora Franciscana in 1891. At Wabash College, Coulter compiled a Manual of Botany of the Rocky Mountain Region, which was published in 1885. These floras followed the format devel. oped by Gray for his Manual. There were no innovations except by Brewer and Watson, who had sections devoted to discussion much as seen in Torrey and Gray’s Flora, but heretofore not found in a local flora. Prior to 1900, keys to species were not common in floras and only rarely seen in revisionary studies. One of the innovations in the three-volume Ап Illustrated Flora of the Northern United States, Canada and the British Possessions by Britton and Brown was the use of indented, dichotomous keys to the species. The idea was probably adopted from the Berlin-based taxonomists who routinely provided keys to species in their works. The in- novation was immediately adopted by western au- thors. For example, keys are lacking in Coulter's Flora of Western Texas, published in 1891, in Frederick V. Coville's 1893 report on the Botany of the Death Valley Expedition, and in Thomas Jefferson Howell's Flora ој Northwest America (1897). But Jepson, who replaced Greene at Berke- ley in 1895, included keys in his suite of studies on the plants of that state. In the 1901 А Flora of Western Middle California, Jepson indicated that his keys to family, genera, species, and in some instances even varieties, were designed to "apply only to the species listed” and that many can appreciate today —he had spent a “very great amount of time and effort" on their con- struction. He maintained the same degree of care in his other efforts: 4 California Flora (1909) and his condensed treatment А Manual of the Flow- ering Plants of California (1923). Jepson was as aware of the student market as Gray, Woods, or Britton. In 1902 he published 4 School Flora for the Pacific Coast. My first taxonomic work was Jepson's 4 High School Flora for California, which I used as a freshman at Sonora Union High School in 1955, 30 years after its initial publication. One of the major evolutionary changes in flo- ristic design that may be attributed to Jepson was as the monographic style he developed for his А Cal- ifornia Flora series. Like Gray before him, and his contemporaries Britton and Fernald, Jepson examined type material and wrote extended com- ments about each species. While on Jepson, it should be mentioned that he started the University of California Publications in Botany in 1902, which was an institutional journal, and was the first editor of Madrorio, the journal of the California Botanical Club; it was first published in 1916. LeRoy Abrams, who arrived at Stanford Uni- versity in 1902, published a Flora of Los Angeles and Vicinity in 1910 in cooperation with the New York Botanical Garden, and in competition with Jepson, published the first volume of his Ilustrated Flora of the Pacific Statesin 1923. Abrams adopt- ed the illustrated flora format of Britton and Brown. Like them he included synonymy and provided each species with а common name. Unlike his eastern counterparts in this regard, Abrams fab- ricated the names; Britton and Brown had evidence for their use. The Smithsonian Institution began its western campaign by publishing a series of local and state floras in the Contributions. The Flora of the State of Washington by Piper appeared in 1906, and Elmer O. Wooton and Paul C. Standley published their 4 Flora of New Mexico in 1915. A decade later, Ivar Tidestrom's Flora of Nevada and Utah Volume 78, Number 1 1991 Reveal 75 Botanical Explorations in the American West was published. The format of each book was a series of descriptive keys with distribution and eco- logical data added in an adjacent section. The de- tailed descriptive format of the Harvard- and New York-based floras was deleted. By the 19308, this was the trend in most federally supported flora projects such as Kearney and Peebles's 1942 Flowering Plants and Ferns of Arizona. The New York Botanical Garden was also ex- tending its influence into the West. Rydberg’s Flo- ra. of Colorado was traditional in format, but fol- lowing the lead of his colleague John Kunkel Small, who was engaged in floristic studies in the south- eastern United States, Rydberg later adopted a concise floristic style in his 1917 Flora of the Rocky Mountains and Adjacent Plains and his 1932 posthumously published Flora of the Prai- ries and Plains of Central North America. These works, with their precise keys and sparse but de- tailed descriptions, became the model that many others have copied. The first modern manual was Fernald's Gray's Manual of Botany, published in 1950. It was carefully written, with detailed descriptions and excellent distribution and habitat data given for each species. In the descriptions, Fernald high- lighted the diagnostic features of each species in italics. 4 California Flora, published by Philip A. Munz in 1959, also had excellent descriptions and distribution data. While Fernald illustrated an oc- casional species, Munz illustrated each family. The model developed by Munz was adopted by Correll and Marshall C. Johnston in their Manual of the Vascular Plants of Texas, published in 1970. When the Pacific Northwest and Intermountain West floras were in their early planning stages in the 1950s under the influence of the New York Botanical Garden, Arthur Cronquist and C. Leo Hitchcock decided to merge various flora designs into two publications. The first was to be a multi- volume, large-format, illustrated, monographic work with room for ample discussion. The second was to be a concise one-volume manual. The five-vol- ume Vascular Plants of the Pacific Northwest was published from 1955 to 1969, with the one- volume Flora of the Pacific Northwest appearing in 1973. The latter work followed the expanded key-format of the Smithsonian-based series, but added illustrations along the margin of each page. The first volume of the /ntermountain Flora was published in 1972 with an expanded intro- duction. This format is being followed in the new multivolumed flora of North America north of Me- xico. As might be expected, as newer models for floras evolved in the West, they were later adopted by writers in the East. Strausbaugh and Core began publishing their illustrated Flora of West Virginia in 1952. The Manual of the Vascular Plants of the Carolinas, by Radford, Ahles, and Bell ap- peared in 1968; it included distribution maps. The two-volume work on the vascular plants of Mary- land by Russell G. Brown and Melvin L. Brown used color photographs and line drawings. They also annotated their illustrations so that in many instances it is possible to key out a plant merely by examining the figures. The Flora of the Great Plains is a two-volume work, with one dedicated to distribution maps and the second a descriptive manual modeled after those written in California and Texas. Two excellent Midwest floras are being published currently: The Illustrated Flora of Illi- nois by Robert Mohlenbrock and Michigan Flora by Edward Voss. These are monographic in style, but the books are of a smaller size so that it will require several volumes to complete the work. Specialized manuals rapidly came into being as well as the more general flora. While illustrated silvas were common, dating back to those by F. A. Michaux (1841-1842) and Nuttall (1842- 1849), Charles Sprague Sargent’s The Silva of North America, published from 1890 to 1902, is the best known. Other specialized manuals were published as well: the Manual of the Grasses of the United States by А. 5. Hitchcock first appeared in 1935, with a revision by Agnes Chase released in 1952. A Manual of Aquatic Plants was pub- lished by Norman C. Fassett in 1940 and in a revised edition by Eugene C. Ogden in 1957. This illustrated manual led to Mason’s А Flora of the Marshes of California in 1957, which led to Aquatic and Wetland Plants of Southwestern United States by Donovan S. Correll and Helen Correll in 1972. The latter was followed by a com- panion work for the southeastern United States by Robert K. Godfrey and Jean W. Wooten in 1981. Kenneth K. Mackenzie’s 1940 North American Cariceae is reminiscent of the kind of illustrated works published in the 1700s in Europe. The increased popularity of wildflower books with fine illustrations or photographs evolved into informative works on American trees, mosses and bryophytes, fungi, lichens, and algae. One of the first western books was The Wild Flowers of Cal- ifornia written by William Doxey and published in 1897. Harold W. Rickett’s series of books pub- lished from 1966 to 1973 on the wildflowers of the United States with its elegant colored photo- graphs has become a classic of its kind. The same may be said of Carlyle A. Luer’s two volumes on 76 Annals of the Missouri Botanical Garden the orchids of the United States and Canada pub- lished in 1972 and 1975. The scholarly А Field Manual of the Ferns & Fern-allies of the United States, published in 1985 by David Lellinger, is illustrated by a series of excellent color photographs taken by A. Murray Evans. Numerous checklists and local florulas have been published for portions of many states, especially in western America. John Thomas Howell in Califor- nia and Stanley L. Welsh in Utah have authored or assisted in many such publications. Recent works have been published for Louisiana and other states in the Deep South. Two trends have evolved over the past 30 years. First, many manuals are so expensive that some— notably the 1980 4 Flora of New Mexico by William C. Martin and Charles R. Hutchins, and Robert D. Dorn's 1977 Manual of the Vascular Plants of Wyoming—are hardly used. Second, regional floras now have multiple authors, and even those for states or portions of states often have several contributors. While multiauthored works often result in highly variable taxonomic treatments between groups, there is often a greater consistency between publications because the same specialist is invited to contribute treatments for several works. The Manual of the Vascular Plants of Texas was one of the first multiauthored volumes in the mod- ern era, although the unfinished Contributions To- ward a Flora of Nevada, begun in the 1940s, was earlier. This trend will reach its pinnacle in the revision of Jepson's Manual for California and in the new national flora for North America north of Mexico. 'THE EvoLuTION ОЕ BOTANICAL INSTITUTIONS AND THE FRONTIER HYPOTHESIS The evolution of institutions on the frontier was, like that for explorations, time-related and influ- enced by developments in adjacent regions of set- tlement. The first scientific societies began in 1660, but such modern institutions as the British Museum were not established until 1753. Governmentally supported research in Britain began in the 1 and gradually replaced that which had been pri- vately funded in the past. In the eastern United States, individuals with modest libraries and col- lections were not uncommon in the formative years of the nation, but few could afford to send out major expeditions. Thus, the federal government, rather than a group of individuals, a society, or an institution, sponsored the Lewis and Clark expe- dition. Few specimens gathered in temperate eastern North America remained in the New World. The majority went to England and France where they were incorporated into private and public collec- tions. When specimens did remain, they were too often lost or destroyed. The first set of John Clay- ton's Virginia plants, his Ehret drawings, and his library and manuscripts were burned by the British during the Revolutionary War. Still, the herbarium of Stephen Elliott at The Charleston Museum, es- tablished in 1773, has managed to survive. The privately held collections of Henry Muhl- enberg, William Barton, and others in the Phila- delphia area formed the nucleus of the Academy of Natural Sciences, whic ablished in 1812. Many private collections were eventually housed at universities along the Atlantic seaboard, with Harvard, Yale, Columbia, and Princeton having major herbaria early in their histories. The Gray Herbarium at Harvard University, formally estab- lished in 1864, had a long prior history thanks to the efforts of Nuttall, Gray, and others. Even the herbarium at the Smithsonian Institution was active before its formal date of establishment in 1868. e tradition eastern institutions inherited from Europe was one dominated by influential men in positions of power. Prior to 1776, most of those in power in Europe were also wealthy. In the United States, a monied class did not exist so that insti- tutional power came from a combination of position and education. By the 1830s and 1840 Torrey and Gray dominated systematic botany, European institutions had evolved to the point that talent too was more important than wealth. Brit- ain's Bentham and Hooker were not poor in any sense of the term, but both were sufficiently tal- ented to dominate taxonomy regardless of their individual monetary wort n the United States, as systematic institutions s, when began to form nearer the western frontier, Torrey and Gray dominated the field and virtually con- trolled all who collected and published. They chose who was appointed to government-sponsored ex- peditions, and not surprisingly, saw to it that all of the collections generally went to them for identi- fication and publication. When Kellogg and his associates established the California Academy of Sciences in 1853, and Englemann and Shaw joined forces to form the Missouri Botanical Garden in 1859, their activities were blessed by Gray. Systematic institutions in the United States grad- ually became fragmented with the establishment of competing government, museum, and university herbaria. Because museums were found primarily on the East Coast, university herbaria (with the exception of Harvard) there generally failed to Volume 78, Number 1 91 Reveal Botanical Explorations in the American West become major centers of power. Even when the systematic community began to fragment region- ally, no university herbaria in the East was able to dominate floristic and systematic research to the extent that the Smithsonian, Harvard University, and the New York Botanical Garden had done. In the Midwest, university and museum her- baria, more isolated but not totally divorced from their eastern counterparts, began to expand their circles of influence so that they could produce significant regional floras, but these efforts were minimal until the n the Pacific Coast, university herbaria competed directly with local museums and botanical gardens for support, ad- justed with them to the changing requirements of systematics, and yet often failed to cooperate. Still, one survival strategy was for university herbaria to concentrate on monographic research, with mu- seums and botanical gardens devoting their re- sources to floristics. The regional sectionalism that resulted after Gray's death in 1888 was exacerbated by the in- stitutional infighting between Harvard and the Smithsonian Institution over the demand by the Smithsonian for the return of collections made on federally supported western expeditions. With the increased independence of western botanists, the sudden growth and the attempted domination of the eastern elite by the New York Botanical Garden caused additional bitterness. This came in the form of New York's rapid domination of floristic regions previously considered to belong to Harvard or the Smithsonian. To protect its interests, each insti- tution made further demands upon universities, taxonomists, and editors for allegiance and support for their particular points of view The development of class and NT section- alism are natural stages in the evolution of insti- tutions. Like the individuals at them, there was a concomitant increase in conservative views, with occasional attempts by the major institutions to dominate, or at least control, the ideas and concepts that were evolving on or near the frontier. The small western institutions argued that the eastern establishment was attempting to control floristic studies at the state and regional level. Yet these institutions had the collections, libraries, and type material to do the work. They also had the talented scientists with the time and funding to accomplish the tasks. While local florulas were left to the state e eanga most western floras were worked on the permission of or in concert with a major eastern institution. y 9 the so-called eastern establishment had expanded to include the California Academy of Sciences and the Missouri Botanical Garden, with a few universities, such as the University of Cali- fornia at Berkeley, assuming major leadership roles in systematic botany. Smaller institutions in the West were often under the direct influence of one of these larger institutions, and good working re- lationships between them were common. For in- dividual western collectors the truce was anything but smooth, yet most became associated with one of the establishment institutions. Greene, a fester- ing sore in Gray’s side, went to the California Academy, then Berkeley, and finally to Washing- ton, D.C., where he used the facilities of the Smith- sonian. Jones worked as a collector for the federal government and identified his collections at the Smithsonian. He also published initially in journals controlled by the Academy and by Jepson. Heller, isolated at the University of Nevada, collected for the federal government, then for the Academy, and finally for the New York Botanical Garden. In each instance, the lure of a larger institution was overwhelming even to the most argumentative of the independent naturalists. Specimen exchange programs soon developed among most of the institutions, with the majority of specimens moving from the smaller herbaria to the larger ones. As the frontier moved westward, those near the edge gradually became the favored repository of the local collectors because by then they had enough of a library and herbarium to serve their needs. Large herbaria soon developed at most of the land-grant universities. Herbaria at the University of Washington (1880), the Rocky Mountain Herbarium at the University of Wyo- ming (1889), the University of Arizona Herbarium (1891), and the herbarium at the University of Texas (1900) were quick to grow and dominate floristic and monographic research efforts in their respective regions. Unlike the general success of the major insti- tutions to attract large collections and talented scientists, the success of the smaller ones was usu- ally dependent upon the personality and ability of a single taxonomist. The growth and success of university herbaria such as the Intermountain Her- barium at Utah State University was due largely to the efforts of Maguire and Holmgren, its first two curators. Likewise, the enormous growth of the herbarium at Brigham Young University since 1965 can be credited to Welsh. Throughout the West there was a step-like evo- lution of people and institutions as the floristic wilderness disappeared. General collectors were re- placed by more regional collectors who were then replaced by monographers. As taxonomy became Annals of the Missouri Botanical Garden more sophisticated, the monographers were re- placed by experimentalists so that now, at least in many universities, few systematists are knowl- edgeable of even the local flora. Even the teaching of systematics has changed. In the past students were first introduced to the regional flora, often writing a master's thesis of a particular area. Once having gained first-hand ex- perience with several groups of plants and the cognate literature, the student went on for a doc- toral degree specializing on a genus. This changed in the 1960s and 1970s, so that now few students are familiar with any local flora, and many have little experience beyond the species level because most do their doctoral research only on a portion of a genus. Likewise, many students today have families, the botanical literature, and almost no sense of the past. Nonetheless, modern innovations in systematics over the past 50 years have tended to come from the university community with its young and ag- gressive faculty and innovative students. Several have taken a leadership role in certain areas. On Ф мапе and fall from рготіпепсе, with influence and power shifting to more dynamic institutions. The Беле Ке: tat th kha “> + il 11 Чио» WHI Е 4 always be closer to the frontier. Universities with large and diverse faculty in systematics tend to be successful in changing em- phases d intaini ful [ g tul programs simply because of their size. Whether or not they retain their position in systematics depends to what extent they abandon their traditionally strong areas of research. Interestingly, where the systematics staff is small, the institution tends not to rid itself of the service aspects of its taxonomist; rather, it allows the other necessary academic disciplines, such as anatomy, cytology, morphology, and paleobotany, to fall aside, thereby preventing continuation of a viable graduate research program. The national disappearance of botany depart- ments and the demise of organismal biology on university campuses may seem to discount the above assertions. Still; many administrators realize that for an institution to become ““modern,”” it must be at the expense of the classical fields, and thus it becomes a philosophical question whether their tra- ditional mission should be fulfilled. To avoid this uncomfortable question, administrators tend to es- tablish isolated "institutes" so as to divorce their researchers from the day-to-day routine of dealing with academic matters. Classical scientists are therefore retained in teaching and service positions with limited in-house funding—well isolated from the activities of those who are considered to be doing real science. This results in internal section- alism, which is often coupled with profound рег- sonal bitterness among the individuals. Mature museums and botanical gardens, in an effort to be on the cutting edge in research, often fail in their community responsibility of being а primary resource center. When they attempt to follow fads, and available research dollars, by hiring narrowly trained specialists in peripheral, often ex- otic fields, it is typically done at the expense o those people who are familiar with plants and whole floristic regions of the world. The newly hired cu- rators, to remain competitive, cannot afford to be interested in collections, exchange programs, OF large-scale fieldwork. As a result, the primary mis- sion of the institution is not fulfilled, to the distress of others who depend upon its accumulated m sources. For those who remain, there is a lessenmg of available funding as the institution fails to attract the research dollars needed to operate its exotic programs, and with fewer broadly trained staff, 1t can no longer maintain its position of importance. This often results in internal classism with accu sations and condemnations that can only haste? the demise of the institution. The underlying force behind all of these pre dictable trends, according to the frontier hypoth- esis, is the lack of a near frontier. Researchers once interested in the wilderness flora of temperate orth America are now working in the trop! where there is still a substantial floristic frontier: Institutional interests at all levels, but particularly those in the larger national and regional museums, botanical gardens, and universities, have alway‘ been on the frontier, and today that means the tropics. Funding agencies, and those few system atists who control the dispersal of research funding: —A ÁS bue dit. memo Volume 78, Number 1 1991 are often guilty of concentrating available funding on the latest fad, arguing that this is the only wa they can justify their budgets to the administration and to Congress. The disquieting trend, and the one that must be fought, is the tendency with maturity to assume that the work done on the old frontier should end simply because the excitement and challenge of the frontier are gone. BEYOND THE AMERICAN FLORISTIC FRONTIER What is the future of systematic botany in tem- perate North America, especially in the continental United States, now that there is no longer an un- known wilderness beyond a floristic frontier? Ac- cording to the frontier hypothesis, we can expect а consolidation of institutions and a reduction in their numbers. We can expect a greater conser- vativeness to envelop the field, especially as it re- lates to nomenclature and other aspects of classical taxonomy, including floristics. While the natural evolution of technology will continue to be exploit- ed, its application will be restricted to fewer indi- viduals and fewer institutions. Furthermore, the half-life of any new fad will shrink, with most con- tributing little to our broad understanding of evo- lution and systematics. These trends are predictable and will come to pass. Policies to prevent them will not succeed and generally will not be attempted. The problem to be addressed by those who work on the temperate flora is how do we proceed? Predictably, we must consolidate our research efforts, reduce our ex- pectations, and sacrifice knowledge on the biology and natural history of all our plants in favor of being permitted to evaluate a few. п short, to work on the plants of this newly settled region we must do precisely what the fron- пег hypothesis says we will, even with the predicted adverse consequences. Sectionalism will increase, exacerbated by the haves and have-nots. Instead of the kind of sec- tionalism now found, it will evolve into one involving tonal versus international interests. Large, mul- ‘institutional international projects will dominate "узета св at the expense of individual scientists *Xamining a single group of plants. Even mono- Braphic work will change, with funding going pri- id to those who put together a team of scientists ccomplish what one, in the past, could have Опе just as well at far less expense. Still, with the “mise of the classical “herbarium taxonomist, " TT be such a profound loss of knowledge Will take years to recover that traditionally Reveal Botanical Explorations in the American West inherited "institutional" knowledge of the plant kingdom so many have been able to learn simply by having a willingness to work and listen. The frontier hypothesis predicts many of the best and brightest will continue to work on and beyond the floristic frontier wherever it exists. Funding in the future, as in the past, will continue United States no longer has a floristic frontier, other nations in the Americas do. We must be willing to address the concerns and needs of those still on the frontier. Much as we resented the ex- ploitation of our wilderness flora by the Europeans in the past, so too do those on the frontier resent exploitation by American systematists today. We have done much to address the concerns of others; yet, we must do more and to do so will be costly. That means an effective transfer of funds, people, and knowledge must occur. We must also be aware that the amount of time for each evolutionary stage to occur on the re- maining frontiers will be much faster in the future. It took some thr ituri the floristic frontier to move to the base of the Rocky Mountains and the Sierra Nevada, but only one to eliminate the remainder of the wilderness across the Great Basin of the American West. With the rapid, exploitative destruction of the tropics, we have little time to make even the most basic observations on the native vegetation before it will be so altered that its significance will be lost forever. Therefore, a degree of urgency does exist—which did not exist in 1889—and if we are to gain any useful knowl- edge of the plants found in many of the world’s last remaining wildernesses, we must act as any adventuresome taxonomist would: head for the frontier! That is where the challenge is; that is where the research dollar is. CONCLUSION In this symposium we have commemorated a century of botany, 1889-1989. The changes that have occurred over the century since Turner pro- claimed the close of the American frontier have been great. A maturation of people, institutions, customs, and traditions is predicted by the frontier hypothesis. We should continue to improve well into the future as long as systematists and system- atic institutions do not become so rigid that they cannot change with the times. Annals of the Missouri Botanical Garden The past century of botanical exploration has brought us to the point in our evolution that, like this nation a century ago, we too are in a position to expand and mature. We do not yet need to fear rigidity, but we do need to be cautious. Indeed, we will experience advances similar to those of the nation since 1890, but in a much shorter period of time. If our future is not to be dim, we must be alert. How we direct our future, and how much we are willing to invest in the maturation of our na- tion's floristic and systematic community, must be something we decide. To leave it to a few, or worse yet, to outsiders, means that we are forgetting one of the most important lessons of the frontier: to be free is to have liberty, and to have liberty means we must participate in the processes that give it a voice. THE NIGER AND THE NILE: BOTANICAL EXPLORATION AROUND TWO AFRICAN RIVERS F. Nigel Hepper! ABSTRACT owledge of e beets of Africa has been obtained by exploration and taxonomic study. While west Africa is Kno dominated by the Niles. er Niger, northeast Africa has the famous Nile, with its major branches of the White and Blue This paper ње бе more than two centuries of botanical exploration by individuals and organizations in these regions, relating them to the colonial and ош history of each period, and links the value of the knowledge da gained to the environmental concerns of to The Niger and the Nile—the two great rivers of Africa north of the equator—lie to the west and east of the continent. For centuries even geogra- phers thought that the Niger flowed out through Senegal, owing to the hidden nature of the Niger's debouchment with the Atlantic Ocean via a mul- titude of outlets in Nigeria; they also thought the Niger and the Nile had a common source in the Mountains of the Moon (Senex, 1721). The real source of the Nile, however, was a mid-eighteenth century objective as much sought after as a moon landing was in the mid-twentieth century. If we look back on the past century we see that this time span encloses most of Africa's colonial period. Before the 1880s only a few, scattered European trading posts located along rivers, such as the Niger and Senegal, could be reached by ship. This situation was altered by the “зсгатЫе for Africa," starting with the Berlin West Africa Conference of 1884-1885, when European im- perial powers scrambled to gain colonial posses- sions. By 1898 all of Africa except for Liberia and Abyssinia was carved up into colonies (see Fig. 1). Less than three-score-years-and-ten later, howev- er, the colonies were becoming independent nations. Gold Coast became Ghana in 1957, and in quick succession other British and French possessions gained nationhood, with only Portuguese territories resisting until the 1970s. Botanical exploration adapted to these changing conditions, and it is no accident that the main African herbaria were established in the imperial cities of London, Paris, Berlin, Brussels, and Lis- bon, which were the capitals of Britain, France, Germany and Portugal, the states that controlled their African colonies. PRELIMINARY EXPLORATIONS Ever since the time of Henry the Navigator (1345-1460), when the Portuguese in the fifteenth century sent their caravels to explore coastal regions ever further around Africa, European nations sought the spices of the Far East, rather than conquest of unhealthy tropical Africa itself. Not until the eighteenth century were herbarium collections and serious botanical observations made. The first were the French, with Adanson in Senegal (in 1749- 1754), and Palisot de Beauvois (in 1790) in Oware and Benin, which are now parts of Nigeria (Aubré- ville, 1962). The British Duchess of Portland sent the Dane Henry Smeathman to Sierra Leone to settle released slaves in 1771; William Brass was sent by Kew’s Sir Joseph Banks to the Gold Coast in 1780 {although Keay [1962] stated that Brass actually collected more gold than plants). The Danes were early seafarers who established forts on Af- rica’s west coast. It was from Christiansborg Castle in present-day Ghana that P. E. Isert and Peter Thonning made their important plant collections between 1783 and 1803. They set out to find medicinal herbs and to help the Africans create better farms to alleviate their poverty (Hepper, 1976, 1979a). Virtually nothing was known of the natural his- tory of northeast Africa until the second half of the eighteenth century (see Fig. 2). Two notable expeditions dramatically opened up the flora and ' The Herbarium, Royal Botanic Gardens, Kew, Richmond, Surrey TW9 ЗАЕ. ANN. Missouni Вот. GARD. 78: 81-86. 1991. 82 Annals of the Missouri Botanical Garden 15 10 5 0 GURE 1. The Niger and West African colonization. (Based on Ј. Каре, Edward Arnold, London.) fauna by approaching the region along the Egyp- tian Nile and the Red Sea. The first was a royal Danish expedition with the botanist Peter Forsskal, who died in the Yemen in 1763. (I include this because many of the plants [and animals] he col- lected have an African distribution and affinity, rather than an Arabian one.) The sole survivor of this expedition, Carsten Niebuhr, sent back Forss- kal's natural history collections to Copenhagen University. I am at present revising that important botanical material (Hepper, in prep.). The other expedition was led by James Bruce to Abyssinia (now Ethiopia), which he reached in 1770 and about which he published his five-volume Travels (1790). This work included illustrations of 15 species of plants (the second edition of 1805 increased the number to 27), as well as engravings of some birds, mammals, and other creatures. Much has been written about this expedition to discover the source of the [Blue] Nile (Head, 1836; Moore- head, 1962), but little has been known about his companion Luigi Balugani. This Italian artist, who died in Gondar in 1772, had drawn more than 150 plants and many animals during this African ex- French a есе 1999 | | pas Sa TS 903) -4 5 10 15 1978, An Atlas of African History, pedition, some as superb finished water colors. Most had been overlooked until I worked on them at the Yale Center for British Art and realized that when Balugani drew them they were new species that remained unknown until later explorers collected them. At last Balugani takes his place, according to Paul Hulton, formerly of the British Museum, as a major eighteenth-century artist alongside John Parkinson of James Cook's voyage (Hulton et al., 1990). THE First HALF OF THE NINETEENTH CENTURY Napoleon’s savants in Egypt included the bot- anist Delile (1809); further up the Nile, Bruce and Balugani were followed by Henry Salt (1832-1833) and more importantly by others including Quartin- Dillion and Petit in the 1830s, and the German W. P. Schimper, who lived in Abyssinia from 1837- 1878 and who made large and important collec- tions of plants. The Sudan was visited mainly by French (F. Caillaud in 1819-1822) and Italian (C. Raddi in 1827-1829) explorers (Cufodontis, 1962). In West Africa, the early nineteenth century saw increasing numbers of expeditions, partly for Volume 78, Number 1 Hepper 83 Botanical Exploration Around 1991 the Niger and the Nile 25 30 - 1 ~ | Es ! 1 _1 ANGLO- EGYPTIAN ie) р $орам (1699) 7 = SUDAN FIGURE 2. The Nile and East Africa. See text for explanation. trading purposes and partly to resist the Portuguese slave trade; along on the ships went naturalists. In the 1820s and 1830s the French sent the botanists Perrottet, Leprieur, Richard, and Heudelot to Sen- egal, and their herbarium material was received at Paris. The British explorers, such as Mungo Park, who perished in the Niger in 1805, and Clapperton, who reached the Niger from the coast in 1826, collected few plants. An ill-fated expedition to the Niger in 1841 lost most of its crew because of disease. The botanist Theodor Vogel also suc- cumbed, although he had collected many plants that found their way to Kew. Fortunately, the next expedition under Dr. W. B. Baikie in 1854 was equipped with quinine, which successfully con- trolled malaria. During Baikie’s return visit to the Niger in 1857, however, his ship “Dayspring” wrecked at the Jebba rapids, and Charles Barter's herbarium collection was lost. Undaunted, Barter made friends with local chiefs and collected another 2,000 specimens in the vicinity while waiting for a replacement ship. He did not survive the two years required, however, and was in turn replaced by another Kew botanist, Gustav Mann (Hepper & Neate, 1971) THE “HEROIC PERIOD” OF AFRICAN EXPLORATION According to Cillett (1962), the **heroic period" extended in East Africa between 1860 and 1888, with famous expeditions such as those by Speke and Grant and the German explorations to Mt. Kilimanjaro led by Decken. Here I overextend my terms of *'the Nile area" in East Africa. The 1886 Anglo-German agreement decided the spheres of influence of the trading companies, and the direct imperial administrations created the period of co- lonial establishment between 1888 and 1918 (Gil- lett, 1962). The British areas of Uganda and Kenya included the headwaters of the Nile at Lake Albert, while Lake Victoria impinged on German Tangan- yika (see Fig. 2). Meanwhile in the north, following the British occupation of Egypt in 1882 and the Mahdist revolt in Sudan 1881-1898 (during which occurred the death of General C. C. Gordon in 1885), the joint Anglo-Egyptian Sudan was set up in 1899. This opened up the entire Nile to explo- ration, although in 1863 Samuel Baker had almost reached Unyoro (now Bunyoro in Uganda), and in the opposite direction Speke and Grant arrived at 84 Annals of th Missouri Botanical Garden Khartoum in 1863. Georg Schweinfurth's epic journey from Congo to Egypt (1866-1872) yielded a large collection for Berlin (Cufodontis, 1962; Gillett, 1962). In between the Nile and the Niger, the French conquest of the Lake Chad region and the Shari and Ubangi river basins by Lamy and Archambault in 1899 linked equatorial Africa with French West Africa, but the Germans took Kamerun (Cameroon) eastward of British Nigeria. C. Ledermann was the botanist on the German Adamaua expedition through Cameroon from 1908 to 1909, during which he collected 6,492 plants. Unfortunately, few duplicates were distributed and most were later destroyed in Berlin during World War II (Hepper, 1974). Fortunately, many of G. W. J. Mildbraed's Cameroon collections are represented in other ma- jor herbaria and were included in Letouzey's (1963) Flore du Cameroun, which itself has provided an incentive for further biological exploration. Nigeria had seen an amalgamation of trading companies in 1879; by 1897 Nigeria was taken somewhat reluctantly under the British crown, and military expeditions pushed north to Sokoto de- claring it a Protectorate. In the meantime (1879— 1893), the French military had made rapid ad- vances from Senegal eastward through the Sudan country under the asssumption that the most ad- vanced and productive parts of West Africa lay there. Thus, the British and French forces met, and defining territorial limits was necessary to avoid disputes. It was during boundary commissions, such as that between Sierra Leone and French Guinea (1891-1892) on which С. Е. Scott-Elliot was able to collect plants for Kew, that serious botanical investigations took place. the 1860s the director of the Royal Botanic Gardens at Kew, Sir William Hooker, had initiated the plan for the Flora of Tropical Africa as one of the proposed series of colonial floras detailing the plant life and potential products of the British Empire. With hindsight one cannot help but be amazed that such an undertaking was started when so little was known about the interior of Africa, yet it was a significant work that continued until 1937. The preparation of this 11-volume work must have stimulated botanical exploration and the collection of herbarium material, as did the first edition of the Flora of West Tropical Africa by J. Hutchinson & J. M. Dalziel during the 1920s and 1930s First HALF OF THE TWENTIETH CENTURY Ву the twentieth century in both West and East Africa botanical studies began to be made by res- idents based in departments or offices. In 1908, Н. М. “Timber” Thompson founded the Gold Coast Forest Department and soon afterward became su- perintendent of forests at Lagos (W. о where he designated forest reserves. In 1892, mans established a botanical garden on da Cam- eroon coast at Victoria where Dr. P. R. Preuss studied the flora; in Tanganyika in 1902 the Ger- mans founded a herbarium at Amani (which much later was transferred to Nairobi's East African Her- barium). Botanical exploration of the Anglo-Egyp- tian Sudan for British Museum (Natural History) collections resulted in Andrew's Flora (1950- 1956). In Egypt, the Swede Vivi Tackholm and her Egyptian colleague Mohamed Drar published four volumes of their Flora of Egypt (1941-1969), which was never completed. From 1899 to the 1930s the notable and en- ergetic Auguste Chevalier sponsored West African botany in the French territories, and he amassed a huge collection at the Paris herbarium. He was followed by another of similar stature, André Aubréville, who published much including an illus- trated, three-volume La Flore forestiére de la Cóte d'Ivoire т 1936. SECOND HALF OF THE TWENTIETH CENTURY Since the Second World War, increasingly more African nationals have become involved in botan- ical exploration. Following the war, institutes of higher learning and research stations with herbaria were established and were linked by motorable roads. Roads had a counterproductive effect on plant collecting as well, as it became increasingly difficult to travel off the roads! Porters were ex- pensive or unprocurable, and the scatter of rest houses about a day’s walk apart fell into disuse in favor of hotels or lodges at wide intervals along the roads. On the positive side, resident government officers were able to settle down to extended study of large or small areas. Some notable collections have been made by amateurs, including wives of officials posted to out-of-the-way places—the best known is Mary Richards, who lived in Zambia outside our area of consideration. Others such as Mrs. Faulkner in Tanzania and Mrs. Tweedie in Kenya were good correspondents of Kew. Con- trolling governments, and later aid donors, sent missions to regions to report on the forestry, pas- turage, geology, and so forth, which resulted in large collections of often, regrettably, inadequate specimens. Local herbaria were active in identi- fying material, usually in cooperation with Kew or Paris. Thus in Sierra Leone, Е. С. Deighton’s study of pathogenetic fungi resulted in the formation of Volume 78, Number 1 1991 Hepper 85 Botanical Exploration Around the Niger and the Nile an excellent agricultural herbarium at Njala. In Ghana the Forestry Herbarium at Kumasi was under the erudite, self-taught A. A. Enti, who trans- ferred to the University of Ghana and there built up а much larger and active herbarium under С. D. Adams, J. K. Morton, and J. B. Hall. In Nigeria the largest herbarium is still at the Department of Forest Research in Ibadan, which was very active under К. W. J. Keay, who helped prepare Nigerian Trees (1960-1964). As well as collections by res- ident botanists, exploration by visitors continued for general collections like my own to remote parts of northern Nigeria (1957-1958) and from Sen- egal to Lake Chad by hovercraft along the River Niger (1969) for the revised edition of the Flora of West Tropical Africa. Specialist collecting grad- ually increased again with fine work by F. White (Ebenaceae); D. J. Mabberley of Oxford (mountain flora); O. Hedberg of Uppsala (mountain flora); R. M. Polhill of Kew (Loranthaceae); C. Grey- Wilson of Kew (Impatiens); B. L. Burtt of Edinburgh (Gesneriaceae); and many others. The Association of African Taxonomists (AETFAT) has done much to channel and coordinate such studies and explo- rations, and they have collaborated on publications such as their Index (Léonard, 1954—) and vege- tation maps (Keay, 1959; White, 1983). The tremendous burst of activity after indepen- dence from colonial rule was badly hit by the rise of oil prices in the 1970s, the slump in oil demand in the 19808, and inflation generally. Foreign cur- rency was unavailable for botanical journals and equipment, fuel, and even food. For example, in Nigeria and Ghana most activities were suspended for many years, yet J. B. Hall & M. Swaine (1981), with D. Abbiw's help, managed to make a study of Ghana's tropical forests that is probably the most comprehensive of any in the world. An active cen- ter in the Ivory Coast has continued under Pro- fessor L. АКе Assi. In Kenya, Christine Kabuye, following J. B. Gillett and colleagues, collaborated with Kew on the Flora of Tropical East Africa (Turrill et al., 1952-). Exploration of Ethiopia con- tinued sporadically (Cufodontis, 1962), but the re- cent incentive by the founding of a flora project (Hedberg & Edwards, 1989) has meant concerted efforts by many international expeditions, unfor- tunately at a time when the vegetation is already decimated. FUTURE OF THE ENVIRONMENT The present international awareness of and con- cern for the environment at last involves govern- ments and donor organizations. They now begin to realize that tropical rainforests are crucial not only to the economy of African countries, but also to the well-being of the biosphere. Overseas aid is partly being directed toward conservation instead of development in the old sense—that of cutting down “the useless bush” in order to establish a plantation. Since the ancestors of the plantation crops occur in the bush, developers now realize that it is not useless after all. This gives added impetus to exploration of the vegetated habitats and the investigation of their constituent species. Joint projects with zoologists, the World Wide Fund for Nature, and other donors in cooperation with national governments are concentrating ex- ploration and research on special areas, for ex- ample, Gola Forest (Sierra Leone), Sapo National Park (Liberia), Tai Forest (Ivory Coast), Ankasa Forest (Ghana), Oban Forest (Nigeria), and Korup National Park (Cameroon). One project with which I have been concerned since its inception in 1986 is the Limbe Botanic Garden project for the con- servation of genetic resources of associated rain- forests оп Cameroon Mountain. This project is supported financially by the British Overseas De- velopment Administration and the Cameroon gov- ernment and is linked with the Royal Botanic Gar- dens, Kew. It is refurbishing the old German garden at Victoria (now Limbe) as the administrative and research base for the species-rich forested areas (Hepper, 1989). Another project concerns the Mis- souri Botanical Garden. For more than a decade significant discoveries have been made by station- ing plant collectors in strategic places, such as Korup in Cameroon where Duncan Thomas gained unrivalled knowledge of the flora and made many additional records. Other Missouri collectors else- where in Africa are making this herbarium Amer- ica's major one for tropical African study. This may be the pattern for the future—international cooperation at various levels. CONCLUDING REMARKS Published maps (Léonard, 1965; Неррег, 1979b) show large gaps in the cover of botanical explo- ration in some African countries, such as Cameroon and Gabon. In Gabon, Breteler (1990) estimated that only 40,000 herbarium specimens have been collected thus far in a tropical forest covering a quarter of a million square hectares, and probably only 75% of the species are described. (Owing to general deforestation in Africa one may ask wheth- er they will ever be known before they are de- stroyed.) By contrast, Ghana is now so well known that few additional species are likely to be found. Little is still known of the ecology or biology of most tropical species, and this is where future field 86 Annals of the Missouri Botanical Garden studies must concentrate (Hawthorne, 1990). Ex- ploration of, say, the tree canopy to study epiphytes or the study of insect pollination may not be as hazardous a quest as the old Niger and Nile ex- peditions, yet we must find out what sustains plant ife, or human life will be unsustainable on this planet LITERATURE CITED 2 hie I W. 1950-1956. Flora су the [Anglo- Egyptian] Sudan. rod & Co. ath. Ambar A. 19 Flore а де la Cóte d'Ivoire. [2nd edition. 1959] L e Chevalier, Paris. ore forestière таала Те Chevalee, Paris 1962. L'exploration deg de l'Afrique occidentale Frangaise. 4 in А. Fernandes (editor), mie a rendus de la IV réunion pléniére de PA.E.T.F.A.T. Junta de Investigagoes do Ultramar, Lisboa. imi ue Е. J. 1990. Gabon’s evergreen forest. Mitt. t. Allg. pe Hamburg 23a: 219-224. mua. Jae 1790. Travels to Discover the Source Nile. J. A rar Edinburgh. Pe preliminary contribution to the dde of the botanical exploration of north- eastern Ponte Africa. Pp. 233-248 in A. Fer- nandes (editor), Comptes rendus de la IV réunio pléniére de ГА.ЕЛ.Е.АТ. Junta de Investigacoes до 09. Description de l'Egypte 2: pls. 1-62, Paris. [Reprinted 1988.] башт. 1 B. 1962. The history of the botanical ex- ploration of the area of The Flora of e iam East Дена Gem, Kenya, Tanganyika and Zanzibar). 229 in A. Fernandes (editor), Comptes dae de la IV réunion pléniére de l'A.E.T.F.A.T. Junta de Pino conie do Ultramar, Lisboa. Hair, J. B. & M. Swaine. Distribution and Ecology of Vascular Plants in a Tropical Rain Forest Vegetation: Ghana. Junk, Th HAWTHORNE, У. D. 1990. Knowledge oi йзге species in the forest zone of Ghana . Allg. Bot. Hamburg 23a: 177-185. Heap, F. B. 1836. The Life of Bruce. Murray, London. n I. & S. Epwarps. 1989. Flora of Ethiopia. Addis Ababa University, Addis A HEPPER, F. N. 1962. Botanical collectors i in West Af. rica, except French territories, since 1860. Pp. 69- 75 in А. Fernandes (editor), Comptes rendus de 1 IV réunion pléniére de l'A.E E.T.F.A.T. Junta de In- vestigacoes do Ultramar, 1963-1972. Flora of West d و‎ SE xevited edition, Volumes 2 & 3. Crown Agents don. 1974. C. Ledermann's botanical posers localities in Kamerun 1908-09. Kew Bull. 29: 381. 1976. The West African aran of Isert and Thonning. Bentham-Moxon Trus — ———. 1979a. Africa: present stage of ا‎ ex- ploration. In: 1. Hedberg (editor), Saan Botany, Plant Utilization and Biosphere Con n. Alm- qvist Maii ges Stoc 979b. n of rog map showing the 2 of floristic ee estat A: se south of the Sahara, published by A.E.T.F.A n: G. Kunkel (editor), Taxonomic piba of Feds Economic tany. Las Pa a ew. 88. Taxonomic analysis of the plants drawn from James Bruce an Luigi Balugani's Ethiopian travels. In: Modern шо Studies in African Botany. Monogr. Syst. Bot. Missouri Bot. Gard. 25: 575-58 . Limbe Botanic Garden, S.W. Camer- oon. cR 11-85 in International Association of Bo- tanic eun Proceedings 1987. Palmengarten, Frankfur i Plant collections of P. Forsskal made on the Royal Danish Е to the кеги 1961-1963. Kew Bull., Additional Ser. (in prep.). К МЕ ЕАТЕ. 1971. Plant аады in West Africa. International Bureau of Plant Taxonomy and Nomenclature (Regnum Vegetabile Volume 74) Utrecht. Hutton, P., Е. №. Hepper & I. Frius. 1990. Luigi Balugani's Drawings of African Plants. Yale Center for British Art, New Haven, and A. Balkema, Rotterdam. HUTCHINSON, J. & J. M. DALZIEL. 1927- IS Flora of West Tropical Africa. Crown Agents, London Keay, В. W. J. 1954-1958. Flora of West Tropical Africa, revised edition, Volume Crown Agents, London. 59. ел Map of Africa South of the A.E.T.F.A.T. € U.N.E.S.C.0., Oxford & Yun . 1962. Botanical collectors in Wes prior to 1860. Pp. oi 68 in i Fernandes i mptes rendus а IV réunion plénière des l'A.E.T.F.A.T. e n latii eng М do Ultramar, isboa. 960, ‚ С. F. A. ONOCHIE & D. P. STANFIELD. l 1964. Nigerian e Federal Govt. Printer, Lagos LÊONARD, J. 1954-. A.E.T адн Index of Papers on Systematic Phaner у and of New Таха n- rog cerning Africa South of the Sahara and Madagascar. A.E.T.F.A.T., Bruxelles 1965. Map of dei extent of floristic ара in Afris south of Sahara. Webbia 19: 911- -914. e R. 1963-. Flore du Cameroun. Museum National d'Histoire Naturelle, Paris. к. REDHEAD, E., В. M. Рогнил. & W. B. T 1952-. Flora t Tropical East Africa. А. А. kema, Rotterdam and Kew ! ne шегу» 1962. The Blue Nile. Hamish Н London. dr T 1721. A New Map of Africa. London TunRILL, W. B., E. MILNE-REDHEAD . D. Рони. 1952-. Flora of Tropical East Africa. Crows Agents, London, now A. A. Balkema, Rotter WHITE, F. 1983. The „бе ge. Africa. U.N.E.S.C.O., Paris. TICODENDRACEAE: A NEW FAMILY OF FLOWERING PLANTS Jorge Gómez-Laurito' and Luis D. Gómez P.? ABSTRACT A new family is described to accommodate the genus Ticodendron. The unusual characters of this monotypic genus, with only 7. incognitum, ћ n discussed by us elsewhere (Gómez-Laurito & Gómez P., 1989), analyzed ave bee by several experts in their special fields, and reported elsewhere in this issue of the Annals. The discovery and description of new families of flowering plants is uncommon these days; onl about 16 of a total 372 recognized by Cronquist (1981) were described since 1950. Many of the most recent new families, e.g., Alzateaceae (Gra- ham, 1984) and Cecropiaceae (Berg, 1978), are taxonomic segregates of long-known entities. Oth- ers such as Saccifoliaceae (Maguire & Pires, 1978) and the most recent new family, Lacandoniaceae (Martinez & Ramos, 1989) were known almost from the beginnin be very close to long-estab- lished families (Gentianaceae and Triuridaceae, re- spectively). Such familiae novae as Tepuianthaceae (Maguire & Steyermark, 1981) and the new family described below for Ticodendron are rarer beasts, whose placement, even as to order, relies on de- tailed studies by numerous specialists looking at a variety of characters from wood anatomy to pollen ultrastructure. Results of these studies of Ticoden- dron are reported in a series of papers in this issue of the Annals. Ticodendraceae Gómez-Laurito & Gómez P. fam. nov. Magnoliophytarum. TYPUS: Ticodendron Gomez-Laurito & Gómez P. cati 3 . Agar ‹ Us. Foliae alternae; lamina elliptico-ovata, subcoriacea, 13 in parte superiore serrata, nervis secundaris 8- нм. Inflorescentia mascula 1.5-4 cm longa, simple ramificata, aliquan s a coronata. St on dehiscen E basifixae; thecae geminatae, longitudin. linis tri mus, conectivo subulato producto. Granulla po- ея ach ‚ minute spinulata, plerunque suboblata. Flo Bracteae S тюк dede. Оп surcolorum dispositae. ‚ mox deciduae. Ovarium inferum in репап- пао CUR EMEN RUNE E thum tubus inclusum, bicarpellatum, quadriloculare utro- que loculo 1; placentatio axillis; ovula hemitropa, apicallis, pendula, crassinucellata, unitegumentata. Fructus dru paceus monospermus. Embryo oleosus, cotyledonibus rec- t Dioecious or less commonly polygamodioecious trees. Twigs terete with the scar of stipules evident. Stipules subulate, encircling the stem, caducous. Petioles grooved above. Leaves alternate, blades elliptic-ovate, subcoriaceous, the margins serrate on the upper part, with 8-13 secondary veins. Male inflorescence 1.5-4 cm long, simple or branched, sometimes crowned by a solitary female flower. Stamens many, arranged in 2-4 verticels surrounded by 3 deciduous bracts, Filaments 2-3 mm long; anthers oblong, 2 mm long, geminate, attached by the base, the connective with an apic- ulate appendage, longitudinally dehiscent. Pollen triporate, minute spinulate, mostly suboblate. Fe- male flowers solitary on pedicels, surrounded by 3 early-deciduous bracts. Ovary inferior, bicarpel- late, included in the perianth tube, 4-locular with 1 ovule in each locule; ovules hemitropous on axile placenta, apical, pendulous, crassinucellate, uniteg- ic. Fruit drupelike, 1-seeded. Embryo oily with straight cotyledons. LITERATURE CITED BERG, C. C. 1978. Cecropiaceae, a new family of Urtica- les. Taxon 27: 39-44. CRONQUIST, А. 1981. Ап Integrated System of Classi- fication of Flowering Plants. Columbia Univ. Press, & L. D. Gomez P. 1989. Ticoden- . Mis- ‚ a new tree from Central America. Ann d 4 1984. Alzateaceae, a new family of Myr- GRAHAM, S. e Шу of tales in the American tropics. Ann. Missouri Bot. Gard. 71: 757-779. ' Herbario Nacional de Costa Rica, Apartado 749-1000 San José, Costa Rica. Wi San "В. & С. ilson Botanic Garden, Vito, Coto Brus, ANN. MI Costa Rica. ssouRI Bor. GARD. 18: 87-88. 1991. 88 Annals of the Missouri Botanical Garden MAGUIRE, B. € J. M. Pires. 1978. Saccifoliaceae, a MARTÍNEZ, E. & C. H. Ramos. 1989. po пи es new monotypic family of the Gentianales. Mem. New (Triuridales una nueva familia de México. Ann ork Bot. Gard. 29: 230-245. Missouri Bot. Gard. 76: 128-134. & J. STEYERMARK. 1981. Tepuianthaceae, Sapindales. Mem. New York Bot. Gard. 32: 4-21. NEITHER OAK NOR ALDER, BUT NEARLY: THE HISTORY OF TICODENDRACEAE! Barry Hammel? and W. С. Burger? ABSTRACT The newly described monotypic family Ticodendraceae is endemic to Mesoamerica, where it ranges from southern Mexico to central Panama. Although locally comm on in some areas very close to old collecting sites (at least in Costa Rica), few collections from prior to 1085 are known. Recent funds for exploration and concentration on the arborescent flora may account for the "rediscovery" of Ticodendron during the 1980s. It may also have been overlooked by botanists because it looks so familiar, like an oak or an alder. Most evidence suggests that it can be placed comfortably often grows in association with endemic taxa of Alfaroa nnoa, Nyssa, remnants of the Tertiary Laurasian flora. Further study and survey of this area will surely provide insight into the composition of that ancient flora surviving in Mesoamerica. With the formal description of Ticodendraceae (Gomez- Laurito & Gómez P., 1991) in hand, this paper will present discussions of the history of discovery, distribution and habitat, biogeography, and a summary of relationships based on the fol- lowing five papers, which present results of detailed studies of the wood anatomy (Carlquist), leaf anat- omy (Hickey & Taylor), sieve-element characters (Behnke), floral anatomy and morphology (Tobe), and pollen morphology (Feuer). HISTORY OF DISCOVERY In one sense, when Ticodendron incognitum Gomez-Laurito & Gomez P. (1989) was described, its epithet could be taken as ironic. The recent history of this plant attests to the fact that many people knew a lot about it even before the species was formally published; correspondence among specialists during the years immediately preceding this compendium is measured in reams. We now know Ticodendron to have a rather wide distri- bution; as recently as 1984 it was known only from Costa Rica, but now is known to be locally common, from southern Mexico to central Panama. The plant is also quite well known to woodsmen; in Costa Rica alone it has at least four common name alder (Gómez-Laurito & Gómez P., 1989). Yet, although Псодепагоп was far from unknown al the time of publication, the Latin ““incognitum”” can have much the same meaning as the modern English usage of “incognito,” and the fact that the family placement of this species has been such a puzzle makes the epithet appropriate. The plant has been famous in many ways, but its true identity has remained hidden (Fig. 1). In spite of the fact that Ticodendron grows in relatively hospitable and accessible areas of moist to wet mid-elevation forests, some of which have been frequented by botanists for nearly a hundred years, we can cite only eight collections made prior to the 1980s. Four of those eight came to light recently, after the publication of the genus; Tico- dendron has been hiding not only in the woods, but in the herbarium as well. Now that we know the tree we find it obvious and abundant, sometimes easily accessible at pasture edges and, at the right time of year, loaded with more or less conspicuous fruits (Fig. 2A) or littering the ground with the decaying fruits t first viste especially when vegetative, this tree and its fallen leaves are easily mistaken for the common Alnus of the region, which often grows in the same areas as Ticodendron (Fig. 2B). One of the earliest known collections (Lankester s.n., 1925), found recently by Aaron Goldberg at US, had been identified by an Alnus specialist аз А. ' Supported in part by National Science Foundation Grant BSR-8700068. ng all of the critical material for nama from the ational en | Е | А, * Field Museum of Natural History, Roosevelt Road at Lake Shore Drive, Chicago, Illinois 60605, U.S.A. ANN. Missouni Bor. GARD. 78: 89-95. 1991. Annals of the Missouri Botanical Garden FIGURE 1. acuminata subsp. arguta. Herbaria have scores of collections of Alnus from Central America, but why so few of Ticodendron? Why didn’t indus- trious collectors like Brenes, Otón Jimenéz, Ma- tuda, Pittier, Standley, Tonduz, or Williams collect it? Actually some of them may have; a rec ently rediscovered specimen at F with an early anno- tation “not Flacourt. " made us take Steyermark off the above list at the last minute. Alternatively, why was this species not described by one of the early botanists who did collect it? It i 18 conceivable that the species has been described, e. g., from Mature tree of Ticodendron incognitum in forest at Monteverde, Costa Rica (Gentry et al. 71474). sterile material, in a large genus, where no one thought to look. The most likely reason that Ticodendron has gone so little collected and so long incognito is that it is most common in wet primary forest where intensive and persistent sampling of trees has begun only recently. Certainly trees are difficult to collect, and Ticodendron may have been overlooked for that reason. Also, the flowering period is apparently rather short and the flowers are inconspicuous. Many places accessible now by road were un ا‎ completely isolated 40 or 50 years ag? | Volume 78, Number 1 Hammel & Bur | 1991 i d | History of Ticodendraceae ч "o 1 2.—A. Fruiting specimen of Ticodendron incognitum (Gentry et al. endron incognitum (Hammel & Rivera 17834). E one might be conservativism. If good ros bos nna fallen fruits beneath a tree with Кой» : p s mo t “fit” the fruits—i.e., because the Su. ттн до по! look like those of TR ied may assume the fruits came from many 5 adii and not pursue the matter. How itin ot Prin their time to collect a difficult So do "i : Inus in primary forest when it could "fam 5 : seis along the road? Likewise, even they иге а it with fruits or flowers attached, don’t es euge ely to describe it right away if they ow in which family to place it. The tree 71474).—B. Young sapling of looks so familiar and is so common, surely it has already been described! This concept is suggested by field notes on the Lankester specimen: *W ood- land tree ... having a distinctly European ap- pearance.” According to label information, the first botanist to collect Ticodendron (see below) thought it might be a genus of Polygonaceae close to Coccoloba. А more recent sterile collection, the only one from Nicaragua, was also first identified tentatively as a species of Ru prechtia (Polygonaceae). This family is implicated by a conspicuous amlexicaule scar Annals of the Missouri Botanical Garden left on the twigs of Ticodendron by the large, paired stipules. Thus, the belated discovery of Ticoden- dron by botanists is curious and yet perhaps ex- plainable by the fact that although it looks very much like something well known, e.g., Polygona- ceae, oak or alder, it really is something different. The earliest known collection of Ticodendron, recognized at US by R. Foster, was made in Costa Rica by Carlos Werckle (probably between 1900 and 1910). Eight other collections (all from Costa Rica, except as indicated) prior to the 1980s have been found: C. H. Lankester (1925), Austin Smith (1938), Steyermark (1942, Guatemala), W. H. Hatheway (1965, 1966), Luis Poveda (1974), and John Pipoly (1979, Nicaragua). For nearly a decade before specimens from Cos- ta Rica and Panama (1986) began to arrive and arouse interest at the Missouri Botanical Garden, botanists at the Herbario Nacional de Costa Rica had regarded the plant as an unknown of consid- erable interest, possibly related to Hamamelidales. However, few, if any, others were aware of the existence and significance of this taxon. By then, fruits and male flowers were known; the species was assumed to be dioecious, but female flowers were still needed. When the existence of a fairly common and conspicuous Costa Rican tree that defied identification by the usual route became more widely known, a three-year-long stream of correspondence ensued. The goal was to motivate collectors in Costa Rica and Panama and specialists around the world, from paleobotanists to phyto- chemists, to gather and analyze the appropriate material in order to place this new tree among its proper relatives. DISTRIBUTION AND HABITAT The single species of Ticodendraceae is known from southern Mexico to central Panama at ele- vations from 500 to 2,400 m. A number of col- lections, most notably those from Mexico, have come to light since the genus was described (Gómez- Laurito & Gómez P., 1989 collections are cited. We now have seen specimens rom Mexico, Guatemala, Nicaragua, Costa Rica, and Panama, and are confident that other speci- mens or the plant itself will turn up from the remaining countries of Mesoamerica, where the appropriate habitat occurs. . Below, all known Specimens examined. MEXICO. OAXACA: е de San Miguel Сћипајара, o a 1,850 m Aug. 1986 (fr), Wendt et al. 5380 (MO); Municipio + Santiago Comaltepec, La Esperanza, 1,600 m, 30 Sep. 1987 (fl & fr), Luna 49 (MO); Municipio Tontontepec, 2 km SW de Tontontepec, 1,900 m, 17 June 1986 (fr), Tore & Téllez 8620. GUATEMALA. HUEHUETENANGO: rro Cananá, 2,500- 2,800 m, 18 July 1942 (st), Stey- а 49111 (Е). NICARAGUA. ZELAYA: Cerro Га Pi- mienta numbers 1 and 2, 900-1,600 m, 17 Apr. 1979 (st), Pipoly 5233 (MO). Costa RICA. ALAJUELA: Cachi, m, 27 Sep. 1925 (st), Lankester s.n. (US); Reserva Biologica Monteverde, valle del Rio Penas Blancas, 800 m, 30 Mar. 1989 (fl), Bello 773, 774 (MO); 1 Mar. 1990 (immature fr), Bello 1971 (MO); 850 m, 28 Mar. 1987 (fl), Haber & Bello 6840, 6841, 6842 (MO); 29 Mar. 1987 (fl), 6851, 6852 (MO); 900 m, 30 Mar. 1987 (fl & fr), 6868 (MO); 9 Mar. 1987 (bud), 6909 (МО); 0 m, 24 Маг. 1987 (fl), Haber & Cruz 7071, 7072 oe 850 m, 5 Apr. 1987 (fl & fr), 7286, 7604 (MO); 750 m, 12 June 1986 (st), Hammel et al. 15385 (MO), (fr) 15396 (MO); бойлы Forestal d Bosque Eterno e los Ninos, 1,100 m, 6 Mar. 1990 (fr), Bello 2154 (MO); Rio Aguas Gatas, 1 000 m, 2 Jan. 1990 (8), Bello 1877 (MO); Rio Cario Negro, 1,200 m, 17 Feb. 1990 e He 1885 (MO); 800 m, 19 Feb. 1990 (f), Bello 987 (st), Gómez-L. 11403 (CR); 30 7 (8), Herrera et al. 515, 516, 517 (МО); Upala, Bijagua, 1,000 m, 9 July 1988 (fr), Herrera et al. 2032 (MO); Zarcero, La Brisa, 1,800 m, 19 July 1938 (fr), Austin Smith H970 (F); Zarcero, Volcán Viejo, 1,800 m, 11 Feb. 1986 (fr), Gómez-L. & Carvajal 11100 (CR). CARTAGO: Reserva Forestal Tapanti, 1,500-2,000 m, 8 July 1986 (st), 1. Chacón 1985 (CR). GUANACASTE: Tilarán, Río Chiquito, 1,100 m, 7 June 1987 (fr), Haber & Bello 7442 (MO). HEREDIA: Cerro Chompipe, 1,800 m, 17 June 1990 (st), Hammel et al. 17815 (CR); Sac- ramento, Montana La Isla, 2,300-2,500 m, 13 Jan. 1987 (fr), Gómez-L. 11306, 11307 (CR); 3 Apr. 1987 (8), Gómez-L. et al. 11472, 11473 (CR); 17 July 1990 (fr), Hammel & Rivera 17834 (CR); Volcán Barva, above San José de la Montana, 1,950 m, 19 Sep. 1965 (fr), Hatheway 1681 (05); 1 May 1966 (st), Hatheway 1712 (US). LIMON: Canton de Talamanca, Bratsi, 700 m, 15 July 1989 (st), Herrera 3339 (CR); Cantón de Guapiles, Rio Blanco, 700 m, 24 Feb. 1990 (st), Herrera & Schik 3791 (CR); Rio Blanquito, 1,100 m, 26 Feb. 1990 (fr), Herrera 3846 (CR); Parque Internacional La Amistad, 650 m, 17 July 1989 (st), 4. Chacón 190 (CR); Reserva Biológica d m 1,024 m, 3 Sep. 1989 (fr), Ham- mel et al. 17685 (CR). РЏ NTARENAS: Cantón de Golfito, Quebrada LS 500 > 23 July 1990 (st), Herrera 4020, Monteverde, Quebrada Máquina, 1,400 m, 12 Nov. 1988 (fr), Bello O 17 Mar. 1990 (immature fr), Haber 9809 (MO); 14 July 1990 (fr), Gentry et al. 71474 (МО); Parque Nacional Corcorado, Cerro Brujo, 969; Reserva о , 1,600 m, 21 July 1985 (fr), Haber & Bello 2058 (MO). SAN JOSE: Bajo La Hondura, 1,300 m, 21 May 1974 (st), Poveda 864 (CR) San Cristóbal, Wercklé s.n. (US Dam area fruits, McPherson & Hammel 8401 (MO). c ~ à ortuna Dam area, 1,100- 1, m, 7 Sep. 1987 McPherson 11659 (МО); 26 Apr. 1988 (fl), McPherson 12 (МО); М slope of Volcán Вага, 1,750- (st) й of El Copé, 750 m, 22 June 1988 (st), McPherson 12615 MO). Volume 78, Number 1 1991 Hammel & Burger 93 History of Ticodendraceae Ticodendron grows in mid-elevation cloud for- ests on both sides of the continental divide through- out Mesoamerica. It is most often found in very diverse, evergreen forests wetter and somewhat lower (2,400 m at most) than those dominated by oak, and is typically found on deep, humic soils. At its lowest it descends to 500 m in the Osa peninsula of Costa Rica. Otherwise, in the southern part of its range, where the isthmus is divided sharply by the central mountain chain into wet (Atlantic) and dry (Pacific) sides, the species is more common and descends lower (650 m) on the At- lantic slopes. However, near San Jose, Costa Rica, numerous populations are known from the very wet southern (Pacific) slopes of the Cordillera Cen- tral. [n this part of its range Ticodendron flowers from January through April, and the fruits mature several months later. Germination has not been observed, but from the numerous young saplings observed we surmise that germination is hypogeal. RELATIONSHIPS Even with detailed analyses, which follow, Ti- codendraceae finds its position in the Hamameli- dae sensu Cronquist (but see Hickey & Taylor, 1991) and Fagales almost by default. Outside Ham- amelidae, Euphorbiaceae were considered (because of some similarity to Picodendron) but rejected because Ticodendron lacks their characteristic ob- turators (Tobe, 1991). The very reduced unisexual flowers lead by overall similarity to the “higher” Hamamelidae, where many of the old Amentiferae lie. Pollen data also lead there but by way of an array of features unique among dicots (Feuer, 1991; Zavada & Dilcher, 1986). In the end, Ticodendraceae has been described as a distinct family because it has been shown neither to share any rigorously supported synapo- morphy or set of synapomorphies with any taxon already described nor does it fit comfortably in overall similarity within any described family. By a process of elimination, its ordinal placement seemed most likely among Juglandales, Myricales, or Fagales. Urticales were rejected early because Ticodendron lacks the foliar trichomes character- istic of that order (Tobe, pers. comm.). Ticoden- dron also does not coincide with Urticales in pis- tillate flower (Tobe, 1991) and sieve-element characters (Behnke, 1991). Ticodendron is elim- inated from Juglandales and Myricales because of floral and trichome considerations. Taken together, the details of wood and bark anatomy (Carlquist, 1991), leaf anatomy (Hickey & Taylor, 1991), sieve-element characters (Behnke, 1991), and flo- ral anatomy (Tobe, 1991) suggest that Ticoden- dron cannot be placed exactly within any of the compared families, but that it does not conflict seriously with placement in Fagales. Hickey & Taylor (1991) place Ticodendron closest to a fossil, the fagalean genus Fagopsis. Although the pollen data show the highest sim- ilarity with Myricaceae (Feuer, 1991), other char- acters argue strongly against placing Псодепагоп nearer to Myricales than Fagales. Even the pollen data show considerable similarity with the coryloid Betulaceae (Corylaceae sensu Carlquist). This agrees with the data from bark and wood anatomy (Саг!- quist, 1991) and sieve-element characters (Behnke, 1991), as well as the observation that the nutlets of Carpinus are remarkably similar in shape and sculpturing to (though much smaller than) the en- docarp of Ticodendron. Although the floral anat- omy of Псодепагоп has a superficial resemblance to Nothofagus (Tobe, 1991)—which itself com- prises a distinct family, sister to Betulaceae rather than to Fagaceae (cf. Jones, 1986; Nixon, 1989)— among Betulaceae, Ticodendron perianth char- acteristics are more like those of the coryloid gen- era (Tobe, 1991) (see Table 1). In Cronquist’s (1981) Hamamelidae, Ticoden- dron would find a home in Fagales simply by mod- ifying the key to orders to include plants with hemitropous ovules. On the basis of floral anatomy (Tobe, 1991) and wood and bark anatomy (Carl- quist, 1991), Ticodendraceae would be more prim- шуе within Fagales than other families. With our present knowledge, the best estimate of relation- ships of Ticodendron is represented by placing it in its own family within the order Fagales, where it is distinctive for its drupaceous fruits, four-loc- uled ovary, and hemitropous ovules. BIOGEOGRAPHY Ticodendraceae is endemic to Mesoamerica, and by virtue of its alliance to the higher Hamamelidae is probably a remnant of the Tertiary Laurasian flora. As such, its discovery and distribution are significant at a level with the trigonobalanoid Fa- gaceae of Colombia, Alfaroa and Oreomunnea (Ju- glandaceae), Molinadendron and Matudaea (Ha- mamelidaceae), and Nyssa talamancana Hammel & Zamora (Cornaceae) (see Hammel & Zamora, 1990; Nixon & Crepet, 1989; van der Hammen & Cleef, 1983). The fossil histories of these or related taxa tell us that they are remnants of a once widespread northern forest. Indeed, according to leaf characteristics (Hickey & Taylor, 1991) Ticodendron appears to be most closely related to a fossil from Oligocene deposits in Colorado. The rather obvious extrapolation is that fossils of Ti. Annals of the 94 Missouri Botanical Garden “oye тудо ue шош sno¡npuad sama() > "рај шавала St j[9S1t 91215 лојовлецо 991 JUSTUBAUOD элэЧм *ure11o2un 10 UMOUXUN ¢ ‘3urpn [our ing 3[qEUEA + *jua1ogtp X ‘se aures = :uo1puapooiy 0} ралефшоз зле S[OQUIAS q 'овзовимепзео) 'аваоетдолого) “иолригролла 'suvjnf 'sngvfow10A| ‘эвэоерцешешен *eeeoeur(] DAW *njnjag 'sisdoXajs() *uoupuapooi] :exe jo гиоцетлоладе jo uorsuedx . 9 у 6 6 6 OI OT 11 11 £I LI (+ 10 =) ЧЗЧУН$ х = = х = = = х = = quasaid S31n4LLS h - x х = = = = = = a¡duns SIAVIT = = = = = x = = = = 1 54335 ¿× = = = X = = = = = e[ndno ou jnu + = = mu sdeo B = jnu jnu ednip LINYA 1q iq iq = = Iq q = = = этш8әнип ф= очзло ¿ 04310 вив вие вив 04310 вив вив зподолпшоц é [2524 5 [254 = = aide [254 = = опхе uomnejuooe[q 12919 4 = 12919 = = = 19919 = = зпојприод 5 = r4 = 5 = = = = © зрпоој/ 1 TINAO с 1 © 1 $ 05 ГА X I 6 r4 Y S310201 = 1 = = + = == = = = с лзануЭ = 1 = = E = = = == = с SATALS x é ¿ х х X х = х = juasqe эшхэриЗ = é x = x x x = = = ојелоф N3TIOd x = х = = = = x x - juasaid ојешој = х = x x x x = х «= juasqe ALW H.INVIHd Asva H040 мога ТЭП НІОМ VWVH УП IVAW Lad YLSO ODIL (1861) 1зтаБиоло) pue umrpueduroo siy} ш siaded шолу Кцѕош әзе exe 'раледшоо ueoq зец и чом чим ехе] aui jo 9uios pue UOIPUIPOIIT jo зэцзмазовтеЧ”) °1 319%], Volume 78, Number 1 1991 Hammel & Burger 95 History of Ticodendraceae codendron and its associated Laurasian taxa must exist and ought to be sought in North America, Europe, and Asia. Both the high mountains and lowlands of south- ern Central America have floras dominated by rel- atively recently (ca. Pliocene) arrived taxa; the rising mountains were invaded by northern species and the lowlands by South American species across the isthmus (Gentry, 1982b; Grayum & Churchill, 1987; Hammel, 1986; Rich & Rich, 1983; Stan- dley, 1937). However, Ticodendron and all of the other above-mentioned Laurasian taxa, endemic to the region between southern Mexico and northern Columbia, are found for the most part in wet to moist mid-elevation forests. These areas, more than anywhere else in the tropics, are characterized by the kind of mild, humid and aseasonal climate that was prevalent across the Northern Hemisphere in Paleogene (65-25 mya) times. Further exploration and careful phytogeographical analysis in the areas where Ticodendron grows, especially along the iso- lated, wet Caribbean slopes is likely to reveal ad- ditional archaic forms that may, in some cases, give us direct, living insight into the kinds of species widespread in the northern forests tens of millions of years earlier. LITERATURE CITED PU + "i 1991. Sieve- seu characters of Tico- . Ann. Missouri Bot. . 78: 131-134. ет S 1991. Wood an but anatomy of Ti- codendron: comments on relationships. Ann. Mis- souri Bot. Gard. 78: 96-104. CRONQUIST, А. An Integrated System of Classi- = on of Flowering Plants. Columbia Univ. Press, or FEUER, "s. 199]. Pollen morphology and the systematic relationships of Tico ыя incognitum. Ànn. Mis- souri Bot. Gard. 78: 143-151. GENTRY, А. Н. 1982a. cd patterns as 1 a Chocó refuge. Pp. 112-136 in С. T. Prance (editor), Dini uen Diversification in the Trop- ics. Columbia Univ. Press, New York. —————. 1982b. Neotropical floristic diversity: phy- t hical connections bet Central and South America, Pleistocene climatic fluctuations, or an ac cident of the Andean orogeny? Ann. Missouri Bot. Gard. 69: 557-593. GOMEz-LauriTo, J. & L. D. Gómez P. 1989. Mug. cM dron: a new tree fro Central America. Апп. souri Bot. Gard. 76: 1148-1151. 1991. Ticodendraceae: a new family of flowering plants. Ann. Missouri Bot. Gard. 78: GRAYUM, M. H. & H. W. CHURCHILL. T au Ап in- a La Selva Costa Rica. Amer. Fern J. 7 Иш. B. E. ls of ii ical analysis jey a subset of the flora of La Selva (Costa 1990. Nyssa talamancana (Cornaceae), an addition to the remnant Laurasian Tertiary flora of southern Central America. Brittonia 42: 165-170. Hickey, І. Н. & р. У. Taytor. 1991. The leaf ar- chitecture of Ticodendron and the application of foliar characters in discerning its relationships. Ann. Missouri Bot. Gard. 78: 105-130. Jones, J. Н. Evolution of the Fagaceae: the | of foliar features. Ann. Missouri Bot. Gard. NIXON, K. с 1989 Origins of Fagaceae. Pp. 23-43 in P. R. Crane & S. Blackmore (editors), Evolution, Systematics, pii Fossil History of the Hamamelidae. Clarendon, Oxford. — & W. L. СВЕРЕТ. 1989. Trigonobalanus (Fa- gaceae): taxonomic status and phylogenetic relation- ships. Amer. J. Bot. 76: 826-841. H. 1983. The Central American dispersal route: biotic history and paleogeography. 12-34 in D. H. Janzen (editor), Costa Rican Natural History. Univ. Chicago Press, Chicago, Il- lin енн Р. 1937. Flora of Costa Кіса. Field Mus. Nat. Hist., Bot. Ser. 18: 5-63. Тове, H. 1991. Reproductive morphology, anatomy, and relationships of Ticodendron. Ann. Missouri Bot. Gard. 78: 135-142. VAN DER HAMMEN, T. & A. CLEEF. 1983. Trigonobala- e tropical amphi-pacific apa in the North Andean forest. J. Biogeogr. 1 ~ M. & D. DILCHER. 1986. Со E pollen morphology a and its relationships to phylogeny of pollen in Hamamelidae. Ann. Missouri Bot. Gard. 73: 348-377. WOOD AND BARK ANATOMY OF TICODENDRON: COMMENTS ON RELATIONSHIPS Sherwin Carlquist! ABSTRACT single known species of Ticodendron has notably primitive wood (scalariform perforation plates with numerous e bars; primary wall remnants in perforations; scalariform lateral wall pitting in vessels; imperforate tracheary elements all tracheids; axial parenchyma diffuse; rays Heterogeneous Type I). Although wood of Betulaceae and Corylaceae (Betulaceae subfamily Coryleae of most authors) is more specialized, primarily in tracheary element morphology and ray histology, ray parenchyma; wood features in these families are consonant with relationship to Ticodendron. Presence of crystals in rays is alike in Псодепагоп, Corylaceae, and Fagaceae. Bark an fusiform-shaped nests of sclereids containing large rhomboidal pee! druses in co and subdivision of ray cells that contain druses. Placement of Ticodendron in Fagales as a separate atomy features are shared by the fagalean families: rtex, phloem parenchyma, an family, more primitive than others in the order, is indicated by the data reviewed. Wood anatomy is of interest in phylogenetic studies because it contains characters that suggest relationship, as well as characters that are sensitive indicators of ecology and of degree of primitiveness and specialization. Bark anatomy, if not as rich in characters, does contain certain character assem- blages indicative of relationship. The discovery of the distinctive tree Ticodendron (Gómez-Laurito $ Gomez P., 1989) provides an opportunity for the use of wood and bark anatomy in order to deter- mine the relationships of this interesting genus. e samples of Псодепагоп, provided me by Peter H. Raven and Barry Hammel of the Missouri Botanical Garden, proved excellent material for analysis of anatomical features. Liquid-preserved stems 37 mm diam. with thin bark were collected by Gordon McPherson (11659, MO) at Fortuna Dam, Chiriqui, Panama. A dried section of a large log (about 40 cm diam.) with bark 6 mm thick was collected by William A. Haber (7071, MO) at Penas Blancas, east of Monte Verde, Costa Rica. Viewed together, these samples permit study of a broad range of histologic and ontogenetic aspects of wood an k Comparison of wood and bark anatomy of Ti- codendron with that of other dicotyledons is the main focus of this study because of the distinctive features and apparent isolation of the genus. For this purpose, the wood slide collection at the Ran- cho Santa Ana Botanic Garden proved unusually valuable because of its size and because some of the wood sections in this collection have bark at- tached to them. The data below show that bark seems unusually decisive in determination of re- lationships of Ticodendron. Regardless of systematic position, Ticodendron possesses a constellation of characters that must be judged unusually primitive for a vessel-bearing dicotyledon. These wood features offer an unusual dimension, because one might not have guessed that the moderately specialized floral features of the genus are coupled with such primitive wood. Wood of this type is related to occupancy of mesic sites (Carlquist, 1975), and thus wood of Ticoden- dron is of interest with relationship to ecology as well. MATERIALS AND METHODS A wood segment from near the cambium and a bark portion were taken from the large dried trunk section Haber 7071. These were boiled in water and stored in 50% aqueous ethyl alcohol. Similar segments of McPherson 11659, a liquid-preserved collection, were transferred to 50% aqueous ethyl alcohol. Wood and bark of both collections were sectioned on a sliding microtome without softening agents. Some sections of McPherson 11659 were dried between glass slides and examined with a scanning electron microscope. Sections for light microscope study were stained in safranin and light- ly counterstained with fast green to improve con- Rancho Santa Ana Botanic Garden Pomona College, Claremont, California 91711. 4412, , Claremont, eid 91711-3101, U.S.A., U. and Department of Biology, ANN. Missouni Вот. GARD. 78: 96-104. 1991. Volume 78, Number 1 1991 Carlquist 97 Ticodendron Wood and Bark Anatomy trast of pit details. Macerations were prepared with Jeffrey's fluid and stained with safranin. Mean number of vessels per group is calculated on the basis that a solitary vessel is rated 1, a pair of vessels in contact is rated 2, and so forth, and 25 such figures are averaged. Vessel diameter is based on lumen diameter at widest point. Means for quantitative features are based on 25 mea- surements except for vessel wall thickness, tracheid diameter, and tracheid wall thickness; for these features, a few typical conditions were selected and averaged. ANATOMICAL DESCRIPTIONS WOOD A full description is given here for the sample McPherson 11659. Data are then given for Haber 7071, but the latter description is limited to fea- tures in which this wood sample differs from Mc- Pherson 11659. Ticodendron incognitum Gómez-Laurito & Go- mez P., McPherson 11659 (Figs. 1-15). Growth rings essentially absent; a small fluctuation in vessel diameter observable (narrower vessels one-third of the way from bottom of photo, Fig. 1). Vessels solitary, rarely in pairs (Fig. 1). Mean number of vessels per group, 1.04. Mean vessel diameter, 73 um. Mean number of vessels per mm?, 23. Mean vessel element length, 1,725 um. Mean vessel wall thickness, 2.6 um. Perforation plates scalariform (Figs. 6, 7), all bars bordered, borders wider at ends of perforations (Fig. 7). Mean number of bars per perforation, 62. Perforation plates frequently with meshlike pattern in this sample (Figs. 9-11). Primary wall remnants in perforation plate exten- sive (Figs. 6, 10, 11) to moderate (Figs. 7, 8), mostly the latter, but with some vestiges of pit membranes present in all perforation plates. Mod- erate remnants take the form of webs and strands (Figs. 7, 8), or have a clotted appearance (Fig. 11), the latter possibly the result of inhibition of enzymatic pit membrane dissolution by deposition of some secondary plant products. Primary wall remnants in perforations may show various sizes of micropores (Fig. 8). Lateral wall pitting of vessels scalariform, less commonly transitional or opposite, on vessel-axial parenchyma (Fig. 13) and vessel- ray contacts. Vessel-tracheid contacts bear cir- cular to oval pits alternate in arrangement but often rather sparse (Fig. 12). Vessels rounded in tran- sectional outline (Fig. 1). Imperforate tracheary elements are all tracheids with fully bordered pits, pit cavities about 4 um diam. (Fig. 5). Mean tra- cheid length, 2,473 ит. Mean tracheid diameter at widest point, 29 um. Mean tracheid wall thick- ness, 6 um. Pits are somewhat more abundant on tangential walls of tracheids than on radial walls, as determined from study of transections. Axial parenchyma moderately abundant and chiefly dif- fuse, with tendencies toward diffuse-in-aggregates and abaxial occurrence (Fig. 1). Axial parenchyma strands often about 14 cells long, the mean length of each component cell about 150 um. Rays both multiseriate and uniseriate, about equal in fre- quency (Fig. 2). Mean height of multiseriate rays, 1,193 ит. Mean height of uniseriate rays, 594 um. Mean width of multiseriate rays at widest point, 3.2 cells. Multiseriate rays mostly with a uniseriate wing at one or both tips. Uniseriate rays and uni- seriate wings of multiseriate rays composed of up- right cells. Upright cells also present adjacent to wings of multiseriate rays and occasional as sheath cells in multiseriate rays (Fig. 2). Procumbent cells present in multiseriate portions of multiseriate rays, “ч much longer horizontally than vertically (Fig. homboidal crystals present singly in a scat- “> of ray cells (Fig. 4), the crystals more abun- dant in the transition from one year's wood to the next. Starch grains spherical (Fig. 14) to slightly angular (Fig. 15), abundant in axial parenchyma (Figs. 3, 14) and in ray cells (Figs. 4, 15). Dark- staining “‘resinlike’’ compounds present in man (but not all) ray cells (Figs. 2-4). Unidentified smaller spheroidal bodies present in ray cells (Fig. 15). Ticodendron incognitum, Haber 7071. Mean number of vessels per group, 1.13. Mean vessel diameter, 116 um. Mean number of vessels per mm”, 22. Mean vessel element length, 1,782 um. Mean vessel wall thickness, 2.6 um. Perforation plates scalariform (no meshworklike patterns ob- served). Mean number of bars per perforation plate, 32. Primary wall remnants in perforations (as ob- served by light microscopy) minimal. Mean tra- cheid length, 2,537 um. Mean tracheid diameter at widest point, 31 um. Mean tracheid wall thick- ness, 7.2 um. Axial parenchyma diffuse (sparser than in McPherson 11659), with a slight tendency toward diffuse-in-aggregates and abaxial arrange- ments. Mean height of multiseriate rays, 2,312 ит. Mean height of uniseriate rays, 525 um. Mean width of multiseriate rays at widest point, 6.0 cells. Uniseriate rays and uniseriate wings of multiseriate rays composed of upright cells. Multiseriate por- tions of multiseriate rays composed of procumbent cells, with only a few upright sheathing cells pres- ent. Procumbent cells radially elongate (radial di- mension 1.5-8X the vertical dimension). Crystals Annals of the Missouri Botanical Garden 2 etre Set ges eae KET KD < - Ф.Ф. am 4 - .- 1 Я TEE EEEL PEE = > ec. Vara e vt "m Џ . ee . р = а > E i " ‚о фе с EA LEI оса as» - ae Y TM oe hos |, ee... аг „А дк» » ~" D. ` =" ÉL uec andis da^ V) Vo -e am = a. ® ae 1 4 ES 1-4. Wood sections bus Ticodendron incognitum, McPherson 11659. — 1. Transection; vessels are ноћу (pair of vessels in contact, angential section; rays conform to Kribs's TS ype 3. ila = dark dots) and dark-staining compounds: cell shapes ease Ee scale above Figur 1 (divisions = 10 џ m). Figure 3, scale above Figure 3 (divisions = 10 um). Figure 4, scale above Figure 4 (divisions = 10 Volume 78, Number 1 C 99 1991 arlquist Ticodendron Wood and Bark Anatomy FIGURES 5-8. Wood sections of Ticodendron incognitum, McPherson 11659. — 5. Radial section, showing pits on tracheid in face view (center) and sectional view (left). 6-8. Perforation plates from radial sections, SEM fibrillar webs of pit membranes in perforations. Figure 5, magnification scale above Figure 4. Figures 6-8, scale indicating magnification (bracket = 10 um at bottom of respective figures). 100 Annals of the Missouri Botanical Garden iw 14 2 А Fıcures 9-15. SEM photographs of wood sections of Ticodendron incognitum, McPherson 11659. 9-11. Perforation plates from radial sections, showing meshlike patterns. — 9. Perforation plate with minimal remnants of pit membranes. — 10. Perforation plate portion with nearly intact pit membranes in perforations. — 11. Perforation plate portion with irregularly dislodged pit membrane portions. — 12. Vessel—tracheid pitting from radial section (vessel side shown). — 13. Vessel— axial parenchyma pitting from radial section (vessel side shown). — 14. Starch grains in axial parenchyma cell, from radial section. — 15. Starch grains and smaller spheroids in ray cells, from radial section. Figure 9, magnification scale at bottom of Figure 7. Figures 10-15, magnification scale at bottom of Figure 6. Volume 78, Number 1 1991 Carlqu 101 riqui зае Wood and Bark Anatomy occasional in ray cells (more abundant than in McPherson 11659) The ways in which Haber 7071 differs from McPherson 11659 are to be attributed most to different sizes and therefore ages of the samples. Although the wood of McPherson 11659 does not represent a juvenile pattern, its features can be described as more mature than Haber 7071 with respect to vessel diameter, multiseriate ray height and width, and degree of radial elongation of pro- cumbent ray cells. The meshlike perforation plate pattern, so abundant in McPherson 11659, may be a juvenile feature. This pattern is perhaps less an anomaly than one might think. If a vessel in which bars of a perforation plate follow an oblique or diagonal pattern rather than a strictly horizontal one, a meshlike pattern occurs where the perfo- ration plates (which are combinations of secondary walls of two adjacent cells) occur. The two collec- tions differ slightly in crystal abundance, a feature that sometimes is related to age in dicotyledons. BARK The collection McPherson 11659 (Figs. 16- 19) is used as the basis for description; features by which Haber 7071 differs are given in a later paragraph. Druses are present in cortical cells (Fig. 17, right), phloem ray cells (center strip two cells wide, Fig. 18), and axial phloem parenchyma cells (Fig. 18, upper left and lower right). Only a fraction of the axial phloem parenchyma cells contain druses. hus, when axial phloem parenchyma cells develop thick lignified walls (Fig. 17), only a few of these sclereids contain druses. Druses occur in “сһат- bered" fashion in ray cells that are subdivided horizontally, each cell with a single druse. n older parts of the secondary phloem (Fig. 17), sclereids derived from sclerification of axial phloem parenchyma cells occur in patches. Anal- ysis of longisections shows that these sclereids are not fibers, because they are subdivided into strands as are xylem axial parenchyma; they are also wider than fibers typically are. Nests of sclereids form in the cortex (Fig. 19), as well as in the old secondary phloem. In the phloem, the sclereid nests evidently are derived from axial phloem parenchyma, but develop into cells much larger and more nearly isodiametric in shape than the sclereid strands mentioned in the above paragraph. In three dimensions, the sclereid ests are often fusiform in shape, the long axis paralleling the stem axis. Many of these sclereids contain one large rhomboidal crystal each (Fig. 19 The surface of McPherson 11695 (Fig. 16, top) demonstrates a periderm with about 10 layers of highly compressed phellem cells; this periderm is a dark band in this photograph. Dark-staining compounds are evident in many cells of the secondary phloem (Figs. 16-19). ‘Cells that contain druses usually do not also contain the dark-staining compounds The thick bark of Hou; 7071 reveals several differences. Periderm was not observed in this bark; evidently it had sloughed off and no new peri- derms originate at levels deep in the bark. Haber 7071 contains a large number of patches of the sclerified phloem parenchyma strands, but nests of large sclereids, some of which cells contain rhom- boidal crystals, are also present. These differences are probably related to age rather than to genetic distinctness. RELATIONSHIPS OF TICODENDRON Wood of Ticodendron is exceptionally rich in primitive features. This suggests unbroken occu- pancy by this phylad of mesic areas, according to earlier considerations (Carlquist, 1975). Features of Ticodendron wood that are regarded as primitive and the authorities for those interpretations in di- cotyledons at large are as follows: long vessel el- ements (Bailey & Tupper, 1918); perforation plates scalariform with numerous bordered bars (Frost, 1930); perforation plates with remnants of pi membranes (Carlquist, 1988: 56-60); tracheids present as the imperforate tracheary element type (Metcalfe & Chalk, 1950: xlv, “fibers with dis- tinctly bordered pits"); axial parenchyma diffuse (Kribs, 1937); rays Heterogeneous Type I (Kribs, 1935). Primitive wood features are not by them- selves evidence of relationship, for they may be retained in many phylads of dicotyledons indepen- dently. The presence of crystals in rays of Ticodendron, a feature that is neither primitive nor specialized, is not evidence of relationships because the number of families that have rhomboidal hela in rays is quite large (Carlquist, 1988: 225-228). Crystal presence in wood also must ing interpreted in a larger context: crystals are more often present in leaves and in primary stems than in wood, sug- gesting that absence of crystals in wood should be considered significant only if crystals are also ab- sent elsewhere in the plant. In searching for woods that may resemble that of Ticodendron, difference in evolutionary level (especially if moderate) does not rule out relation- ship provided that there are no features with dis- tinctive systematic distribution that would contra- Annals of the Missouri Botanical Garden FIGURES 16-19. Ticodendron incognitum, McPherson 11659, views of bark transections. — 16. Transection; small portion of secondary xylem, below, periderm at top (dark); sclereid nests, center (gray). — 17. Sclerified axial phloem parenchyma cells. —18. Druses in secondary phloem ray (center strip two cells wide) and in adjacent phloem parenchyma cells (upper left, lower right). — 19. Portion of sclereid nest with prominent rhomboidal crystals. Figure 16, magnification scale above Figure 1. Figures 17-19, magnification scale above Figure 4. Volume 78, Number 1 1991 Carlquist 103 Ticodendron Wood and Bark Anatomy dict relationship (e.g., presence of vestured pits, presence of intraxylary phloem). Such features have distinctive systematic distributions but are not found in Ticodendron. The wood of Betulaceae differs moderately from that of Ticodendron in degree of phyletic advancement (data from Tippo, 1938, and Metcalfe & Chalk, 1950). In Betulaceae, perfo- ration plates are scalariform with numerous bars and axial parenchyma is diffuse as in Tiocoden- dron. Somewhat more specialized in Betulaceae than the corresponding feature in Ticodendron is the presence of the fiber-tracheids (rather than tracheids). The rays of Betulaceae (Homogeneous Type I and III) are markedly more specialized than those of Ticodendron (Heterogeneous Type 1), as is the alternate pitting of vessels (as compared to scalariform to лека Rhomboidal crystals are not present in f Betulaceae, but they are present in т parts of the plant. Woods of Betulaceae may contain dark-staining deposits like those of Ticodendron. Wood of Corylaceae (Betulaceae subfamily Cor- yleae of many authors) also shows resemblances to Ticodendron, but about the same degree of diver- gence in level of specialization (data from Tippo, 1938, and Metcalfe & Chalk, 1950). In Coryla- ceae, perforation plates range from scalariform to simple and lateral wall pitting is alternate; fiber- tracheids are present; axial parenchyma is diffuse; rays are Homogeneous Type II, or have a small degree of heterogeneity (Heterogeneous Type IIB) in Carpinus and Ostrya. Wood of Corylaceae often contains dark-staining deposits like those of Tico- dendron. The f grouped vessels in Betulaceae and Corylaceae is not a specialized character in- dependent of others. Vessels may be grouped in woody dicotyledons with fiber-tracheids or libriform fibers (degree of grouping depends on ecology), but vessels are solita т so in woods with | 1984), as in Ticodendron. omboidal crystals, like cds of Ticodendron, do occur in p = согун еае. harartar f systematic И тања in woods of Betulaceae or Corylaceae that are not also found in Ticodendron. Fagaceae has many of the same features as Betulaceae and Corylaceae (e.g., crystals in ray cells), but wood is rather more specialized. Thus, comparisons to 7i- codendron need not involve Fagaceae. Bark anatomy, however, more clearly shows relationship between Ticodendron and the fagalean families (observations on these families original, based on material in the Rancho Santa Ana Botanic Garden wood slide collection). Fusiform nests of sclereids, cells of which contain large rhomboidal sclereids, are found in bark of Ticodendron, Сог- ylus cornuta Marsh. (= C. rostrata Ait.), Betula pendula Roth, and, in Fagaceae, Lithocarpus den- siflora (H. & A.) Rehder. These highly distinctive sclereid nests alone might be evidence of relation- ship. In addition, druses in cortex, axial secondary phloem parenchyma, and phloem ray parenchyma characterize Ticodendron, Betula pendula, and Corylus cornuta (druses in cortex but rhomboidal crystals in axial phloem parenchyma occur in Lith- ocarpus densiflora). Subdivided phloem ray cells with a druse in each of the cells occur in Псодеп- dron and Corylus cornuta. Ticodendron does not have phloem fibers, but it has a possibly allied phenomenon, conversion of strands of axial phloem parenchyma to strands of sclereids. Phloem fibers do occur in Betula pendula and Corylus cornuta. Phloem fibers are absent in bark of some species of Betula and A Él pone m an external periderm, but evidently does not develop successive deep- seated sio : deep-seated orm are likewise absent in Alnus, Betula, and at least some Fagaceae. Thus, there are no features in wood that rule out a relationship between Ticodendron and the families of Fagales (Betulaceae, Corylaceae, Fa- in bark between Ticoden- lationship between Псодепагоп and Fagales, one must note the appreciably more primitive wood of Ticodendron. 'This accords with the rather primi- tive (compared to other Fagales) floral morphology of Ticodendron (Gómez-Laurito & Gomez P., 1989). Ticodendron can be considered a very primitive element in Fagales. It could not be in- cluded in any of the existing fagalean families with- out markedly altering the family description, with Therefore, wood and bark anatomy support place- ment of Псодепагоп in a monogeneric family re- ferred to Fagales. Wood and bark of other families considered earlier in discussions among the authors of other papers on Ticodendron were considered in comparisons: Brunelliaceae, Dilleniaceae, Ju- glandaceae, Myricaceae, Picrodendraceae, and Nothofagus (which several authors now segregate from Fagaceae). The anatomical resemblances be- tween Ticodendron and these families are appre- ciably fewer than the resemblances between Ti- codendron and Betulaceae and Corylaceae (or Coryleae of Betulaceae), and so such comparisons have been omitted. 104 Annals of the Missouri Botanical Garden LITERATURE CITED BaiLEY, I. W. & W. W. Tupper. 1918. Size variation in rng cells. I. A comparison between the sec- ondary xylems of vascular cryptogams, gym ae and а Proc. Amer. Acad. ren 54: -204 im oed a 1975. Ec ological Strategies of Xylem n. Univ. California Press, Be Ек 4. Vessel grouping in dico woods: significance and на to icu. tracheary elements. Aliso 10: 505-525. —— 1988 к. ы Anatomy. Sprin- ger-Verlag, Berlin & Heid Екозт, F. H. 1930. Sora in secondary xylem in dicotyledons. I. Origin of vessel. Bot. Gaz. (Craw- fordsville) 89: 67-94 GÓMEZ-LAURITO, J. & L. D. Gómez P. 1989. Ticoden- dron: a new tree from ce = Ann. Mis- souri Bot. Gard. 76: 1148- 1935. Salient Mn A structural special- wood rays of dicotyledons. Bot. Gaz (Crawfordsville) 96: 547-557. — 37. Salient lines of structural specialization in the wood rays of dicotyledons. Bull. Torrey Bot. Club 64: 177-186. МЕТСАТЕЕ, C. В. & L. CHALK. 1950. Anatomy of the Dicotyledons. Clarendon Press, Oxford. TiPPo, O. 1938. The comparative anatomy of the ondary xylem of Moraceae and T presumed allies, -99. Bot. Gaz. (Crawfordsville) 100: THE LEAF ARCHITECTURE OF TICODENDRON AND THE APPLICATION OF FOLIAR CHARACTERS IN DISCERNING ITS RELATIONSHIPS! Leo J. Hickey? and David Winship Taylor? ABSTRACT Leaves of the recently described an Ticodendron are simple, alternate, stipulate, symmetrical, elliptical with as ‚а sing le o omous н) venation, percurrent tertiaries, we diio veloped, random areolation, anomocytic stomates, and T-shaped hairs. other chara Phenetic and cladistic analyses of -— their demonstrated potential as syst of 70 characters 48 leaf an Ticodendron belongs to a group o cters of Ticod Several new characters, including t‏ ا were added to the suite of standard leaf are њен descriptors because of matic indicators. After filtering out nonapplicable and symplesiomorphous сћаг- d ; к за mentiferous taxa, with the extinct genus Fagop dendron were inp in order to determine eins joining the midrib of a followed by Brunellia, Castanopsis, "Ca astanea, pe and Nothofag us. Our bur manes show Ticodendron grouping with the Fagales and occasionally wit Cunoniaceae. These data suppor and with closest affinities to Fagaceae and Betulaceae. e Myricaceae in a rosalean t recognition of he. new genus as a member of a a deir family in the order Fagales se base is embedded in the The recent discovery of a relatively common tree of unknown affinities in middle elevation, wet primary forests of Panama and Costa Rica is vivid evidence of what is being lost as the deforestation of the tropics continues. Determining the systematic affinities of Ticoden- dron Gómez-Laurito & Gómez P. is especially chal- lenging because it has a mosaic of characters that defy ready placement. Its general affinity is with the more derived members of the highly unnatural group of dicotyledons with reduced flowers (Behnke, 1989; Thorne, 1989; Wolfe, 1989) known as the Amentiferae. Thus, during the exchanges by the working group, evidence has been cited for rela- tionships to families as disparate as the Euphor- iaceae, Dilleniaceae, Juglandaceae, Betulaceae, and Ulmaceae. However, detailed anatomical and morphological investigations by members of the group are beginning to provide reliable characters. А consensus has developed that the affinities of Ticodendron lie near a group of amentiferous fam- ilies that comprise the Hamamelidae of Cronquist and that include the Fagaceae, Betulaceae, Myr- icaceae, and possibly Juglandaceae. The main objectives of this paper are to provide a concise and rigorous description of the leaves of Ticodendron and an assessment of its affinities, with an emphasis on foliar characters. We will also examine its phylogenetic relationships based on leaf and other characters assisted by cladistic tech- niques. Throughout, we will attempt to show the concordance between inferences of relationship that arise from an analysis of leaf characters with those of other organs. Our focus on leaves is based on the fact that they are morphologically complex organs, with a diversity of characters such as teeth and detailed venation patterns that often indicate phylogenetic affinity (Hickey & Wolfe, 1975). However, precisely because it now appears that certain elements of leaf architecture are congruent ! The assistance of Barr acknowledged Haven, Connecticut 06511. U. S.A y Hammel in coordinating this project and of MO in supplying specimens is gratefully ew York Botanical Garden for allowing preparation U for preparing the slides, and members of the research group studying this pai Peabody Museum, Yale University, P.O. Box 6666, 170 Whitney Avenue, New ANN. Missouni Bor. GARD. 78: 105-130. 1991. 106 Annals of the Missouri Botanical Garden with systematic affinity at higher taxonomic levels, it is essential that more rigorous and explicit tech- niques of phylogenetic analysis be applied to lower taxonomic levels of the angiosperms in order to demonstrate the utility of leaf characters. A number of such test cases already exist or are being developed. Levin (1986a, b, c) explicitly studied the correlation between the relationships based on leaf characters with phylogenies based on traditional characters within the Euphorbiaceae, a notoriously difficult family. The phenetic and cladistic results not only show the similarity be- tween leaf and other data sets, but also suggest that the leaf data may be superior in some cases. Levin now plans to extend this survey to the Fla- courtiaceae (pers. comm.). More recently, Todzia & Keating (in press) used foliar characters to dis- tinguish clades in the Chloranthaceae where re- productive characters had provided ambiguous re- sults. So has Hershkovitz (1988), who reported that such leaf architectural characters as vein course and leaf rank prove useful in the systematic in- terpretation of the Centrospermae, a group whose predominantly herbaceous character might lead to an inference of a high degree of leaf architectural plasticity and attendant homoplasy. Finally, Miller & Nowicke (1990) have used leaf architectural characters, in conjunction with a suite of other characters, to show the relationships of two genera in Boraginaceae. t the present time, this and other evidence suggest that leaves may well provide information on phylogenetic relationships that is at least equal to that derived from any other organ or data set in the plant. Thus, although the major purpose of this study is to describe and elucidate the relation- ships of Ticodendron, we also hope to show that cladograms based on leaf characters are similar to those derived from other organs. To optimize our use of leaf characters, we refined and extended many of their definitions, such as tooth type, so that they would be more useful in distinguishing entities and in pointing to relationships at lower taxonomic levels. This increased the general ro- bustness and utility of our leaf characters as well as elucidated the relationships of Ticodendron it- self. MATERIALS AND METHODS Cleared leaves of Псодепагоп were prepared for this study from two specimens supplied by MO. These were prepared by the methods outlined in Hickey (1979) and Hickey & Wolfe (1975). Com- parisons were made using the cleared leaves found in the National Cleared Leaf Collection (NCLC) housed at Yale Peabody Museum (YPM). Addi- tional vegetative characters as well as the repro- ductive characters used were obtained by study of herbarium specimens housed at YU and NY or by recourse to the literature of systematic and com- parative botany. A list of taxa examined, both from cleared leaves and herbarium specimens, appears in Table 1 After a broad search of the National Cleared Leaf Collection at Yale University, 18 operational taxonomic units (OTUs) at the generic level from eight families (Table 1) were selected for detailed comparison to Ticodendron. This group included representatives from five amentiferous families, namely species of Castanea, Castanopsis, Noth- ofagus, the extinct genus Fagopsis (Fagaceae), Alnus (Betulaceae), Pterocarya, о (Ju- glandaceae), Canacom a (Myrica- ceae), and Rhoiptelea ыл In addi- tion, several clades from three putatively primitive families of Rosales were examined. These included Brunellia (Brunelliaceae); Davidsonia (Davidson- iaceae); and Acsmithia, Ceratopetalum, and Cu- nonia (as representatives of Cunoniaceae). The yrica, genera of Cunoniaceae were chosen due to their apparently primitive phylogenetic position within the family ba reproductive, and wo characters (Dickison, 1980b) and general similarity to Ticodendron. sed on leaf, Formulation of a list of characters and character states that adequately described the leaf architec- ture of Ticodendron and allowed its comparison with other taxa was a crucial aspect of this analysis. Our starting point was the outline of leaf architec- tural characters established by Hickey (1979). As we collected data we found other overlooked char- acters that proved to have potential value in de- termining relationships. This led to an expansion in the terminological base, especially for characters that deal with tooth morphology and marginal ve- nation (see Results below). Initially, 71 vegetative characters were collected from cleared leaves and herbarium specimens. These included leaf architectural as well as node, stipule, cuticular, and twig features. In addition, seven wood characters and 17 floral and pollen characters were obtained from the working group and the literature (Gómez-Laurito & Gómez P., 1989; Carlquist, 1991; Feuer, 1991; Tobe, 1991). The high degree of similarity that characterizes the pollen of many of these OTUs (Batten, 1989) has been suggested to be a shared derived character some groups of families in the Hamamelidae for (Muller, 1984). Volume 78, Number 1 1991 Hickey & Taylor y Ticodendron Leaf Architecture 107 E l. List of species, OTU abbreviation, and collection information for the specimens used to generate the leaf du and other vegetational data u sed in this study. for Fossil Plants; YPM-PU — Peabody Coll. ex Princeton Univ. ات‎ Coll. ssil collections: YPM — Yale Peabody Catalogue Source of leaf architectural characters Source of other vegetative characters T. incognitum (Hammel, Bello, Haber, Kins- man & Lierheimer 15285, YU), (NCLC T. incognitum (McPherson 8401, YU ex MO), F. longifolia YPM 25404, 25467; УРМ-РЏ 9809 B. colombiana (Cuatrecacas & Llano 27314, C. gummiferum (Wilkes s.n., YU), (NCLC 6713) D. puriens (Cronquist 11622, NY), (NCLC 6766) P. stenocarpa (Hickey s.n., YU), (NCLC 3220) . strobilacea (Chen 3706, GH), (NCLC 7001) P P. strobilacea (1757053, US), (NCLC 410) R. chiliantha (1577640, US), (NCLC 804) 781) C. concolor (Schoh 101, US), (NCLC 627) C. pumila (Kearney, Jr. 892, US), (NCLC N. discoidea (Baumann-Bodenheim 15111, N. fusca (Walker 4696, US), (NCLC 1765) A. nepalensis (Stibolt s.n., US), (NCLC 5387) Genus (family) OTU Ticodendron (un- TICO known) 6383) (NCLC 6765) Fagopsis (fossil) (Fa- FAGO gaceae) Brunellia (Brunellia- BRUN ceae) US), (NCLC 3186) B. еме ed (Barclay et al. 3413, US), (NCLC 3190) Acsmithia (Cunonia- ACSM A. о (Hickey 5274, YU), (NCLC ceae 6545) Ceratopetalum (Cu- CERA noniaceae) Cunonia (Cunonia- CUNO C. capensis (Cummings 300, US), (NCLC ceae 3165) Davidsonia (David- DAVI soniaceae Pterocarya (Juglan- PTER aceae) Platycarya (Juglan- PLAT daceae) Rhoiptelea (Rhoipte- RHOI leaceae) Canacomyrica (Myr- CANA C. monticola (Hickey 5210, YU), (NCLC icaceae) 6768 Myrica (Myricaceae) MYRI М. cerifera (358864, US), (NCLC 143) M. californica (E. Hall 468, YU), (NCLC Castanopsis (Faga- CAOP ceae) Castanea (Fagaceae) CAST 6136) Nothofagus (Faga- NOHI ceae) US), (NCLC 221 Nothofagus (Faga- NOH2 N. moorei (ex Herb. Melbourne 5.п., US), (NCLC 55) Alnus (Betulaceae) ALNI Alnus (Betulaceae) ALN2 A. rhombifolia (Plaskett s.n., US), (NCLC 5402) same same B. comocladifolia (Wright same same C. cf. deplanchei (Hickey 252, YU P. caucasia (Brewer s.n., Englehardtia spicata (Mer- rill 253, YU one same M. cerifera (Curtiss 2606, YU) C. chrysophylla (E.W. Hammond 346, YU) C. dentata (Godfrey s.n., YU) N. balansea (Hickey s.n., YU) N. balansea (Hickey s.n., YU) A. crispa (A. Hill 2541, A. crispa (A. Hill 2541, YU) Several operational difficulties were tooth characters. Problems included the existence of two different types of teeth (cu- Each t in coding the tered 1975) on the basic teeth types (e.g., in Juglandaceae, Rhoip- telea, Davidsonia, and Myricaceae). oth type can be described by a suite of 2 nonioid and rosid; Hickey & Wolfe, same leaf, and in rare cases transitional forms, both in the same species and in closely related taxa, that lay morphologically between the common cee (see below, Fig. have two sets of descriptive tooth characters, one for each type. Thus tooth type 1 (cunonioid) is 108 Annals of the Missouri Botanical Garden coded by characters 6-11 and type 2 (rosid) by characters 12-17 and 20. This created a problem when an OTU had only one tooth type. Normally, if a character is missing, it is either coded as un- known, which allows computerized cladistic algo- rithms to assign an optimal state, or coded as a single character state for the absence of the char- acter. Yet if an entire tooth is missing, in the second case each state of the entire suite of characters is coded as absent, and these states could act as shared derived characters for a group of taxa. Yet the absence of these characters could be due to parallel loss, the ancestral condition (thus are sym- plesiomorphies), or actual shared derived states. his presents difficulties for our analysis because the lack of rosid teeth is also ancestral, based on outgroup comparison. Our solution was to code each of the characters of the missing tooth type as a nonapplicable state (NA) in the cases where only one toot state for each family where at least one OTU was lacking a tooth type. We suggest that this method results in underweighting of these characters (they type existed. We created ап МА are either autapomorphies or synapomorphies at the top of the tree), but at least avoids grouping taxa based on symplesiomorphies or on less secure character loss states. In order for our analysis to deal with OTUs having transitional teeth, we chose to use the most fully developed extreme of each of the two tooth types. In fact, in a given OTU these extremes tended to be the most common. Once the matrix was formed, we found a number of characters, particularly of the leaf, that proved to be inappropriate at the level we were using them. They were removed if they either did not vary at the taxonomic level being examined or if they were polymorphic at the generic level and thus only useful for elucidating specific relationships. In ad- dition, we removed all autapomorphic characters, characters in which many of the OTUs were miss- ing data, and finally, a few characters that further study indicated were not structurally homologous. he remaining 70 characters (48 leaf and 22 reproductive and wood) were used in our phylo- genetic analysis (Appendix). Some OTUs had mul- tiple character states for specific characters (the states not used in the analysis are provided at the end of the description of each character in the Appendix). Since the version of PAUP we used to implement our cladistic analysis can only use one state for each OTU, we had to devise an objective protocol in order to decide which character states to include for a given OTU. For an OTU, if a character was polymorphic and shared a state with Ticodendron, we chose that state because the prin- cipal objective of this analysis was to ascertain the affinities of Ticodendron. If none of the states were the same as those of Ticodendron, the state that was most common in the OTU was chosen. If all states were equally common, the state most com- mon for all of the OTUs was used. If none of the states were similar to the most common state, we used the state most similar to that of a closely related taxon. Finally, if two closely related taxa (same family or genus) shared two character states, each OTU was assigned a different state. The resulting data matrix was arranged to form two sets (Tables 2, 3), with the second set divided into two additional subsets (characters 1-48 and 49-70) o compare vegetative and other data sets it was necessary to reduce the number of OTUs because data on reproductive and wood morphology were missing for several of the OTUs, or were the same for Nothofagus2 and Alnus2. Thus, our second data set (Table 3) had 14 OTUs and 70 characters. Two types of analyses were performed on each of the two data sets. First was a simple similarity comparison, whose similar- ity coefficient was calculated by counting the num- ber of character states shared by Ticodendron and each OTU and dividing it by the number of known characters. The remaining analyses were based on synapomorphy and parsimony. Our phylogenetic analysis used the program PAUP (Swofford, 1985). he major options of this program that we used were global branch swapping, mulpars, and rooting to a hypothetical ancestor to find the shortest trees. Variable options were ordering of the characters with two or rarely three character states, the option to treat the characters as completely unordered (those characters which are ordered are indicated as such in the Appendix), and the initial addition of sequence either closest or simple. Lastly, if mul- tiple parsimonious trees were found, the trees were combined to form strict consensus trees (Swofford, 1985). The specific choice of outgroups, construc- tion of the hypothetical ancestor, the method of ordering, and the resulting trees are discussed in the Results section below. RESULTS In the following description of the leaves of Ticodendron and analysis of its affinities, the terms used are from Hickey (1979) with the addition of the following new terms or clarification of previous ones: Volume 78, Number 1 Hickey & Taylor 109 1991 Ticodendron Leaf Architecture TABLE 2. Input data matrix for data set 1. OTUs are identified in Table 1 and characters are described in the Appendix. OTU Characters 111111111 14444444 23 156789 2345678 1 HYPO 2 TICO A. 1102122333 10010111121103113021203311 3 FAGO 1101044954513544922142 4100 121 3271 112 92? 221212904 4 BRUN 01020000000343 211001??00220000110010103321 5 ACSM 1001 6 CERA 124 7 CUNO 3000000000 8 DAVI 011301011003335351162100120113200200011000002200 9 PTER 011000010003336354161102220200121001112110012210 10 PLAT i ие IN 11 RHOI 2100110100110201104010000110 12 CANA nn 1102302302110001125100101100 13 MYRI A. A 1104000010034364601 125220201100 14 CAOP | 1001556668333635206220010001 eee 101221202312 15 CAST , 16 МОН] pan ОНЕ ee 1630200 17 NOH2 11010001023436350152201001112221112101021215000 18 ALNI 19 ALN2 ооо MM о nee (1) sinus—an indentation in the margin of the closest tooth (Fig. 1). This is done by dropping а leaf associated with a tooth or a lobe. Their iden- perpendicular (cd in Fig. 1) to the nadir of the tification is straightforward if they are angular, but indentation from a line (ab) connecting the apices shallow, rounded sinuses may grade imperceptibly of the adjacent teeth and comparing the length of into undulations of an entire margin. In such cases, this line with the distance (in this case ac) to the sinuses may be recognized by measuring the max- closest tooth. If the depth (cd) is greater than the imum depth of the marginal indentation and com- distance to the closest tooth (ac or bc) the marginal paring this with the distance to the apex of the indentation is a sinus, otherwise not. TABLE 3. Data matrix for data set 2. OTUs are identified in Table 1 and characters are described in the Appendix. OTU Characters 1234567890123456789012345678901 1 НУРО 2 TICO 1102122333 100101111211031130212033112141200001011212301111 3 FAGO ава ou Mu л |" 4 ВЕОМ 01020000000 01??002200001 30200 5 ACSM 10021 6 CERA 1124 7 CUNO 140020000 8 DAVI 0113010110033353511621001 2011320020001 1000002200100?2220001 20413120210 9 PTER ie pe E РАТЕ 10 PLAT 0112000000033363611 21100002101141242101001351201211 11 RHOI 0112000001034363521621001 10100110201 1040100001102141 2000000001 13000210 12 MYRI 11 1252202011001141202001001151211211 13 САЗТ | 3220001101221003312003013101?000033020211 14 NOHI 1 221 e о 1103301?0111?2401211 15 ALNI 111 001 4110102?011112401211 Annals of the Missouri Botanical Garden --e-e-- --- a ~ О •---- Determination of the presence of a sinus between teeth in three different marginal configurations (labeled 1, 2, and 3) where the nadir of the indentation is rounded. Tooth apices lie at points a and b, with d indicating the nadir of the sinus (in the case of d,) or of the marginal indentation (d, and d,). Where cd > ac or cb the indentation is a sinus; where cd « ac or cb the indentation is not In the course of this work we found it necessary to elaborate upon the terminology available to de- scribe the types of venation and vein courses at the margin and especially within the teeth of the leaf. The complete list of character states for these appear in the Appendix, but we give definitions for several of the most important characters below because of their importance in our analysis. The first set of these are the veins associated with the teeth. These begin with: (2) the principal vein—the thickest vein found within the confines of the tooth (Fig. 2, P). This is followed by: (3) the admedial vein—the first branch from the principal vein below the tooth apex on the admedial side of the tooth (if from a tooth located at the leaf apex, then the first branch from either side of the principal vein that has the subsequent characters) that is the same, or one vein order lower than the principal vein, is distal to any ex- medial vein of similar strength, and has > 60% of its vascular tissue at its junction with the prin- cipal directed away from the tooth apex (Fig. 2, (4) accessory veins— collectively, all the veins between the tooth apex and the admedial vein that either branch from or merge with the principal vein including the conjunctal veins (defined below) (Fig. 2, Ac). (5) conjunctal veins—accessory veins of the tooth that converge upon or merge with the prin- cipal vein and contribute vascular tissue to the tooth apex. Conjunctal veins must have at least 60% of their vascular tissue directed toward the tooth apex at their point of convergence or fusion with the principal vein. They may occur in pairs, arising either oppositely or a age! to one an- other, or they may be single (Fig. (6) comb vein—an admedial branch of a sec- ondary vein, arising near the margin and running along it, from which originate several tertiary veins that run ere to the superadjacent secondary vein (Fig. 5 as an inner und an outer comb vein. C). They occur singly or sometimes In an unpublished survey of angiosperm vena- tion, we have noted that the number and config- uration of the different classes of veins arising from the midvein of the leaf have proven to be of con- siderable value in determining affinity. Because the categories of veins used in this analysis carry an inference of evolutionary homology, we felt that a new set of terms was necessary to distinguish them from venation terms such as secondary, tertiary, and intersecondary in the system of Hickey (1979) where categories are recognized strictly on the basis of similarity in course and size where they attach to the midvein. These new terms are as follows: (7) deuteral veins— consisting of the secondary veins of a leaf, the midveins of lateral leaflets, lateral primaries, and all those secondaries of the E 2. An ideal tooth with the major types FiGUR venation labeled. A с — accessory veins, Ad — admedial vein, C— conjunctal veins, P— principal vein Volume 78, Number 1 1991 Hickey & Taylor 111 y Ticodendron Leaf Architecture FIGURE 3. Venation of a simple leaf showing the deuteral (D), opadial (O), and trinal (T) veins. same size-class or whose size is part of a smooth gradation of sizes and whose behavior does not differ markedly from the majority of secondaries (Fig. 3, D). (8) opadial veins—comprising the intersec- ies of abruptly different ondaries and th size and behavior from the majority of the sec- ondaries (Fig. 3, O). (9) trinal veins—the next thinnest vein set con- sisting of those tertiary veins that connect to the midrib and which arise either from secondary veins or from intersecondaries (Fig. 3, T). The sum of these three vein sets will be designated as the D-O-T veins collectively. A final addition to the original set of leaf ar- chitectural terms involves a new term for the con- figuration of the marginal ultimate venation, as follows: (10) spiked —margin with short, outward- pointing ultimate veins that are either unbranched or with a single pair of short branches (Fig. 4). FIGURE 4. Spiked marginal ultimate venation. LEAF ARCHITECTURAL DESCRIPTION OF TICODENDRON INCOGNITUM Leaves alternate with encircling bases, stipulate with deciduous stipules attached to a ridge of pet- iolar tissue that extends transversely from the sides of the enlarged and raised site of leaf attachment (Fig. 7). Stipules completely encircling the apical and axillary buds. Leaf abscission scars with three traces, lacunae 3. Leaf simple, symmetrical, except slightly asymmetrical at the base; shape elliptical with a length to width (Мм) ratio of from 1.8 to 2.8 (mean 2.1, standard deviation 0.2, N = 15). Apex attenuate, base acute, texture chartaceous (Fig. 5). The leaf margin is serrate in a single order with from 5 to 12 teeth (mean = 11) per side; generally there is one tooth per secondary vein with an average spacing in mature leaves of 1.3 cm. Ser- rations concave on their apical and basal sides (type C3) and have elongate, abaxially directed, soon deciduous, glandular tylate processes (a capitate pad of clear glandular tissue at the tooth apex; Fig. 8). The principal vein of the serration is medial and represents the continuation of a secondary vein that enters the serration straight and is then de- flected slightly downward (Fig. 5). Principal vein termination tapered to abrupt, the principal ac- companied by an admedial vein and at least one admedial conjunctal vein with accessory branches on its inner margin (Fig. 18). Conjunctal veins in opposite pairs when both are present. The presence of a medial principal vein and conjunctals on both sides of the principal in a tooth of this morphology implies that Ticodendron possesses modified rosid teeth in the classification of Hickey & Wolfe (1975). The sinuses between the teeth are rounded and unbraced but are approached by a conjunctal vein or a branch from the conjunctal. Venation pinnate, craspedodromous (Fig. 5). Secondaries diverge at a moderate acute angle (45—60°) and have a uniform spacing and angle of divergence. Secondaries slightly recurved and un- branched until they reach the base of the serration where they give off a comb vein, that is, a branch that runs close to the margin while giving off several tertiary veins to the superadjacent secondary (see 112 Annals of the Missouri Botanical Garden FIGURES 5-11. Details of leaf characters. 5-8. Ticodendron incognitum (Hammel et al. 15285, YU).— 5. Whole cleared leaf. c— comb vein (NCLC 6383). Scale bar = 10 mm.—6. Lower cuticle showing hair base and stomatal anomocytic complexes (NCLC 6383). Scale bar — 100 um (0.1 mm).— 7. Apical о, of woody twig. Note the large stipules that encircle the axes, and in the node below, the circular scar that remains on the axis. Scale bar — 10 mm. = Scanning electron microscope (SEM) view of а tooth of an unexpanded leaf; note the elongate rocess at its apex. Scale bar = 100 um (0.1 mm). —9. Davidsonia puriens, tooth of a young leaf with an elongate, eed iba process near its apex (Cronquist 11622, NY). Scale bar = 100 um (0.1 mm). — 10. Castanea den ung tooth with apical, capitate process (Nichols s.n., YU). Scale bar = 10 um (0.01 mm).—11. Corylus ano Ке tooth with a short, apical, deciduous process (Nichols s.n., YU). Scale bar = 10 um (0.01 mm). Volume 78, Number 1 Hickey & Taylor 113 Ticodendron Leaf Architecture 16 17 Ficures 12-18. Camera lucida drawings of teeth from cleared leaves.— 12. A typical dillenioid tooth from Doliocarpus dentatus (Dilleniaceae) (Bartlett 11367, US) (NCLC 857).— 13. А tooth of Tetracera volubilis (Dil- leniaceae) with the typical dillenioid-type architecture but with a deciduous process (Gentle 3363, US) (NCLC 828).— 14. A typical theoid tooth from Hartia sinensis (239672, US) (NCLC 5).—15. Tooth of Ceratopetalum gummiferum (Cunoniaceae), which has many similarities to the theoid tooth in 14 (Wilkes s.n., YU) (NCLC 6713).— 16. Two dimorphic teeth of Canacomyrica monticola antag The upper tooth is a typical cunonioid tooth while the lower is a rosid tooth. Note the asymmetrical vena the lower tooth (Hickey 5210, YU) (NCLC 5210).— 17. A typical rosid tooth from Alnus nepalensis. Note voi ee venation in pa tooth an ли eciduous apical tip (Stibolt s.n., US) (NCLC 5387). — 18. A tooth of Ticodendron incog he similarities to the Т. іп 16 and 17 indicating that и is а rosid tooth d pie 8401, YU ex MO) ARES 6165). All scale bars 1 . Figu and re 16 same magnification as 14, 17 same as 15 8 same as 114 Annals of the Missouri Botanical Garden TABLE 4. Table of the similarity of Ticodendron to other OTUs in each data set. These figures are calculated from the number of shared characters divided by the number of characters known for each OTU. OTUs are identified in Table 1 Data set 1 Data set 2 b c (all (leaf (reproduc- (leaf ^ characters) characters) ti ) characters wood) FAGO 0.537 0.491 0.537 0.500 BRUN 0.500 0.373 0.500 0.095 ACSM 0.319 0.304 0.319 0.273 CERA 0.250 0.243 0.250 0.227 CUNO 0.277 0.304 0.277 0.364 DAVI 0.312 0.290 0.312 0.238 PTER 0.312 0.371 0.312 0.500 PLAT 0.333 0.400 0.333 0.545 RHOI 0.417 0.471 0.417 0.590 CANA 0.250 NA NA МА MYRI 0.354 0.429 0.354 0.591 CAOP 0.521 NA NA NA CAST 0.458 0.377 0.458 0.190 NOHI 0.354 0.426 0.354 0.600 NOH2 0.458 NA NA NA ALNI 0.375 0.478 0.375 0.714 ALN2 0.458 NA NA NA C on Fig. 5). Secondaries in 5-11 pairs; intersec- ondary veins lacking. Tertiary veins percurrent, opposite, closely spaced (> 0.5 mm); sigmoidally curved; exmedial angle of iom slightly acute to perpendicular (mean 85°, mode 90°); admedial angl pl the high- ly cau ed of being aldo obtuse (mean 95%, m 97°). Fourth-order veins strongly im- e Scent fifth order strongly impressed, generaly random; highest vein order fifth; marginal ultimate venation looped. Freely ending veinlets present, curved, mostly twice-branched. Areolation well developed, randomly oriented, mostly quad- rangular, of medium size (mean diameter 0.6 mm); leaf rank (following Hickey, 1977) 3r'. Upper epidermis with straight-sided cells and many radial arrays of cells as from a hair base. Lower epidermis with anomocytic (rarely ap- proaching cyclocytic) stomates, guard cells with thickened polar walls and a highly distinctive an- nular thickening on their inner ledges (Fig. 6). Both epidermal surfaces, but most commonly the lower one, with what appear to be overthrust glands in pits formed by the projecting edges of the sur- rounding cells, which are sometimes aligned in ra- dial arrays away from the pit. Trichomes very rare in the mature leaf, common in immature specimens (Fig. 8); T-shaped, no scales noted. GROSS PHENETIC AND PHYLOGENETIC AFFINITIES The similarity of the leaves of Ticodendron to our OTUs is given in Table 4. The combination of simple, craspedodromous leaves with a single tooth per secondary, angle of basal secondary origin the same as for the upper secondaries, an excess of 60 D-O-T veins; strongly percurrent, sinuous ter- tiary veins; with teeth having a tapered or abrupt, but not bulbous, termination; one or sometimes two pairs of conjunctal veins with inward-branching accessories; and the presence of marginal comb veins strongly suggests a group of taxa in the so- called Amentiferae, consisting of the Fagaceae in- cluding Nothofagus, Betulaceae, the fossil genus Fagopsis, and also the family Brunelliaceae (Ro- sales). In addition, the venation and marginal con- figuration of some pinnately compound leaves in the Juglandaceae and the Rhoipteleaceae war- ranted a more careful comparison to those families. Emphasis on the amentiferous group of genera is reinforced as well by the reduced flowers (Tobe, 1991) and the pollen (Feuer, 1991) found in Tico- dendron. Leaves of the extinct genus Fagopsis, with a similarity coefficient of 0.537 (Table 4), are most like those of Ticodendron even to the tapered terminal portion of the principal veins of its teeth. Fagopsis has convex-convex teeth (character 4) lacking conjunctal veins (16), angular sinuses (5), uniformly symmetrical leaf bases (47), well-developed areolation (31), and a deuteral-opa- dial-trinal (D-O-T) vein total of between 41 and 60 (35) as opposed to > 60 for Ticodendron. Among living genera, Brunellia, Castanopsis, and Castanea have similarity coefficients (s) of 0.50, 0.521, and 0.458, respectively, and also share the character of having tapered terminations of the principal veins of their teeth. Furthermore, both Castanea and Castanopsis have rounded si- nuses similar to those of Ticodendron. However, Brunellia has pinnately compound leaves (char- acter 1) with angular sinuses and semicraspedodro- However, mous secondaries (29), while the two fagaceous genera have teeth with tylate processes (22), non- deciduous stipules (25) that are not attached to the expanded leaf base (26), more than 18 deuteral veins (33), and a total of deuteral and opadial veins that ranges between 18 and 28 (34). АП three genera have well-developed areoles (31) and glan- dular peltate and simple trichomes 6). Leaves of Alnus (s = 0.458 and 0.375) and Volume 78, Number 1 1991 Hickey & Taylor 115 Ticodendron Leaf Architecture Nothofagus (s — 0.458 and 0.354) are similar to those of Ticodendron in their general shape and vein configuration. However, both genera lack the comb veins (character 44) found in Псодепагоп, and most other genera of Rosidae examined have well-developed areoles (31), glandular peltate tri- chomes (Nothofagus) or both glandular peltate and simple ones (Alnus) (36), D-O-T vein numbers of between 41 and 60 (35), and splayed (Alnus) (Fig. 17) or truncate and splayed (Nothofagus) termi- nations to the principal veins of the teeth (18). Alnus differs further from Ticodendron in having two orders of convex-convex marginal teeth (4) with clear glandular (21), tylate (22) apices and splayed principal vein terminations (18) (Fig. 17). Additional differences between Nothofagus and Ti- codendron include teeth of different shape (4), leaf bases that are always symmetrical (47), basal sec- ondaries at a lower angle than those above (3), and petioles that have both wings and stipules attached (24) (see Tables 2, 3). Among the remaining hamamelid genera ex- amined in this study Rhoiptelea (s = 0.417), Platycarya (s = 0 , and Pterocarya (s — .312) were dissimilar because of their pinnately compound leaves. Two genera of the Myricaceae, Myrica (s = 0.354) and Canacomyrica (s = 0.250), have simple leaves but these show nu- merous differences such as teeth of the wrong shape (character 4) (Fig. 16), angular rather than round- ed sinuses (5), a lack of stipules (24), symmetrical leaf bases (47), glandular, peltate, and simple tri- chomes (36), and a D-O-T vein number of from 19 to 38 (34). These data suggest that the affinities of Псодеп- dron are with amentiferous families with rosid teeth (Hickey & Wolfe, 1975) which, despite their in- clusion in the subclass Hamamelidae in the clas- sifications of both Takhtajan (1969) and Cronquist (1988), have leaf architecture that is indistinguish- able from that of the subclass Rosidae (Hickey & Wolfe, 1975; Wolfe, 1989). These two groups also share other characters (Dickison, 1989). For these reasons we also examined several other fam- ilies of Rosales that are regarded as representing the most primitive families in Rosidae Acsmithia (s = 0.319) is notable because it is the member of the Cunoniaceae with the highest level of similarity to Ticodendron. The genus has simple leaves with a single order of teeth (character 3) of the proper shape (4) and only of the rosid type. However, these teeth appear to be a less derived form of the rosid type than in Ticodendron because they possess a full set of accessory vein (8) and have splayed terminations. Other differ. ences include opposite leaves (2) with semicras- pedodromous venation (24) and paracytic stomates 37). The above families form the basis of our cladistic analysis described below. In this, we used Cunonia- ceae to polarize our characters. In Takhtajan’s 1969 and 1980 classifications this family is thought to be the most primitive of his Saxifragales, the primitive order of Rosidae. Studies by Dickison (1975a, b, 1980a, b, 1984, 1989) have further strengthened the position of the Cunoniaceae. Cronquist (1988) also regarded the family as very primitive in the Rosales, although not the most ancestral. What is apparent for this family, like so many other truly primitive taxa, is the high degree of morphological variability to be found within it as well as the somewhat capricious expression of this variability (Dickison, 1989). Thus the family has, in some cases, tricolporoidate pollen (Hideux & Ferguson, 1976), extremely primitive wood (Dickison, 1977), and low- to moderate-rank leaves. Extreme variation is apparent in characters in- ~ cluding the perianth, which varies from present to nearly absent, ovary position from hypogynous to epigynous, the frequent occurrence of a nectarifer- ous disk surrounding the carpels (Dickison, 1984), and the presence of both simple and compound leaves within the same genus, such as Weinman- nia. Cunoniaceae are especially variable in their carpellary morphology, which ranges from follicles to capsules and even drupes. Superficially the fam- ily runs nearly the whole gamut of diverse fruit types seen in the Rosidae and the с even to configurations seen їп Liquidambar (Pancheria), Staphelia (Gillbeea), and Pterocar- ya (Ceratopetalum) (Dickison, 1984). In addition, a number of fused, bicarpellate fol- licles with two stigmas or stigmas that approach fusion into a single bifid structure and whose flowers have highly reduced perianth are found in the genus Acsmithia and in the satellite family David- soniaceae (Dickison, 1984; Cronquist, 1981). Cronquist (1988) has commented on the diffi- culty of distinguishing the Cunoniaceae and other Rosales from Dilleniidae. We uncovered a striking instance of this transitional position when we found marginal teeth in the leaves of the Australian gen- moderately derived Dilleniidae (Fig. 15 and com- pare to 14). Thus, the transformation to the cu- nonioid tooth type found in the Rosales and the amentiferous taxa discussed here must occur within the Cunoniaceae because both occur there. Nothing closely resembling the rosid tooth type 116 Annals of the Missouri Botanical Garden ABLE 5. Comparison between data sets and different options. The options are addition of sequence — closest (closest), addition of sequence — simple (simple), unor- dered in part (part), and unordered all (all). Data set 1 included characters 1-48 and 18 OTUs, and data set 2 included all or parts of characters 1-70 and 14 OTUs. Num- Length Consis- ber o о tency Data set options trees trees index Data set 1 Closest part 2 260 0.527 Simple part 2 260 0.527 Closest all 1 249 0.550 Simple all 15 250 0.548 Data set 2 All characters Closest part 3 305 0.587 imple part 3 305 0.587 Closest all 1 295 0.607 Simple all 2 294 0.609 Characters 1-48 Closest part 1 213 0.601 Simple part 1 213 0.601 Closest all 9 204 0.627 Simple all 9 204 0.627 Characters 49-70 Closest part 3 75 0.680 Simple part 3 75 0.680 Closest all 3 74 0.689 Simple all 3 74 0.689 with its frayed termination of the principal vein of the family edid including Dillenia and Doliocarpus, resemble Ticodendron in form, sec- ondary venation, and str d percurrent tertiaries (Wolfe, 1989). Their teeth (Figs. 12, 13), in con- trast to Ticodendron, have the slightly bulbous termination of their principal veins that charac- terizes the dillenioid tooth type (Hickey & Wolfe, 975), and lack conjunctals with inner branches and the comb veins that characterize many mem- bers of Rosidae. For this reason our search was limited to the Rosidae and Amentiferae. EVOLUTIONARY ANALYSIS Our evolutionary analysis of the relationships of Ticodendron progressed through a number o methodological stages, roughly along the lines sug- gested by Eldredge (1979) and Neff (1986). The first and potentially most important of these stages was character analysis (Kaplan, 1984; Neff, 1986; Stevens, 1987; Steele et al., features are described, 1988). At this stage characters and character states are defined, and useful characters are se- lected. This is followed by character polarization n which ancestral states and structural transition series are identified. These characters and their polarized states are then used in the next stage. For this we used pattern sequence analysis to iden- tify the optimal branching patterns. The last stage is systematic analysis, which uses a variety of data to identify the evolutionary relationships of the taxa. We paid particular attention to character anal- ysis during this research. Such an analysis pro- uced a matrix showing the distribution of char- acter states that are hypothesized to be structurally homologous. During this stage we made no attempt to equate structural homologies with evolutionary homologies. We considered that if the same char- acter state is found in different taxa it is funda- mentally similar in structure. We feel that our explicit terminology and rigorously defined char- acter states facilitated our recognition of structur- ally similar features. In the Appendix, these char- acters are defined by topological, anatomical, or morphological similarity, while the states are dis- tinguished by more subtle differences. When assigning polarities to character states we hypothesized the ancestral states and the transi- tions between them using outgroup comparison (Maddison et al., 1984). Because we consider Tico- dendron to be a member of Rosidae, for the reasons given above, we chose as our outgroup the family Cunoniaceae, which we consider to be primitive in the subclass. The occurrence of theoid-type teeth in some Cunoniaceae (Fig. 15) permitted us to form a sec- ondary outgroup lying outside of the Rosidae. Theoid teeth are found in at least two clades, the thealean clade and the flacourtialean clade (Hickey & Wolfe, 1975), and at this time it is neither clear that the origin of this tooth type is monophyletic nor which of the above clades is more closely related to the Rosidae. Based on this equivocal placement, we chose the Dilleniaceae as our secondary outgroup because of its inferred sister group relationship to the Rosidae, as well as the thealean and flacour- tialean clades. The dillenioid tooth type (Figs. 12, 13) is clearly ancestral to the theoid-type tooth (Hickey & Wolfe, 1975) (Figs. 14, 15) and pro- vides an unequivocal secondary outgroup. A hypothetical ancestor was then formed for polarizing the character states and to root the trees. is was necessary when using PAUP because the program uses one taxon when several outgroups Volume 78, Number 1 Hickey & Taylor 117 199 Ticodendron Leaf Architecture О "S = — e a < = NY O = т д = = О = I = = © б > Ut «ва 5ш o BH 3 x i py о 55 << < = о < © 1T T 152 * 57197 f 131111 1 10 + 194] 3341 18+4 6*1 7+3 16*5 21-0 5*1 9-0 6+2 23*1 34*1 21+1 8-0 9x1 8+4 20+3 30*3 23-0 10*1 7+2 26*1 3542 29-0 17*6 15%5 9+2 29-0 43-0 45*2 21+1 5-0 843 31-0 40*1 31-1 45*0 30%3 10+4 30*2 27+2 444 14*4 9+3 32-2 45%3 33%1 32 11*5 43-0 4+ 30*2 13+4 15+4 10+3 38*1 46* 34-0 36-0 10*1 14*4 45-0 6+5 31-1 16+4 16*5 11-4 39* 44*1 37-0 31-1 15+3 46-0 7+5 32*1 17*6 17+2 19-0 41= 32%1 34*1 8+6 33- 20*4 31-1 24*2 47+0 38-0 23*1 35*1 9+6 40*1 35-0 25+2 39+4 39-0 44-1 10+6 41-0 44-0 28+2 46*] 43*1 4843 11+8 43*1 3*1 47+1 39*2 45*3 п 18*2 4*0 48+1 40*1 7+1 a 19-0 11*2 iise i 25*1 2242 8+1 12+1 0 28-0 24*2 24*1 6+4 13*3 is AN 33-0 2543 27* 7+4 18+1 1442. Ml 48+2 28+3 30*2 8+5 26+2 jer An 29-0 46-0 9+4 29%1 iés 32 38+2 10+5 45*2 17+1 39+5 13+5 18*3 42-0 18*2 3*1 19-0 2242 45+1 23*1 25x] 2042 23*1 46*1 26*1 34*2 24*1 47+1 36+3 36*1 27*1 38-1 18*3 39*3 19-0 48+4 20+2 24*1 1 5*1 27*1 13*3 11+7 5%] 36*1 2045 6*1 4042 34*] 9*1 44-0 35*1 11+1 1 -— 34*2 1-0 9+1 37-0 18*2 29*1 8+2 26*1 30+1 9+5 41+1 38-0 11*2 46 43+2 18*3 45*3 30*2 т Ј 36+4 4+1 33*1 3642 4241 44*1 47+0 I J 2+1 12+3 1344 14+3 15+6 16+3 17+5 19*1 20+6 28+1 31+2 32-2 38*1 39+1 41+2 46*3 J 1 1+1 4+2 21+2 22+1 47+1 t 3 31+1 37+1 HYPO——À FIGURE 19. Cladogram of data set 1, which includes 48 leaf characters and 18 OTUs (Table 3). This is the shortest tree with 249 steps and a consistency index of 0.550. —, reversal; =, multiple reversal; *, synapomorphy; llelism. +, paralle 118 Annals of the Missouri Botanical Garden О < > 2 - = = - = 2 oc a © (o > O т ТЕ. дм т = е Q 5 © О CC < 2 I > = 3 О = + О < e T tr > T а. 2 < ш Е 1*1 4-0 | A 245 je 8*1 14 +4 *1 10+6 br 17-5 18*1 15*4 +2 11 +8 fx 21%1 29%1 1)+1 +2 19*1 17+4 вж Jul!) 33- 20+4 +3 23* 18* 19-0 ge 4 ја“ 7 39*1 +3 26* 33 21=0 15* 11+ 44*] 10+3 31- 34* 2242 18* 12+ 1 32* 356 23*1 26% 13+ 1 38* 40* 28-0 32- 14+ 1 39% 45 38- заж 15 + 3 49* 46* 40* 36- 6+ 4 51* 50* 42* 37- 17+ 61* 54 46* &1- 18* 62* 56* 4B+ 45* 19- 66* 58* 51+ i. 20+ 68- 63* 53* 49* 24+ 54* 9+ 27+ 55* 60+ 6* 57, 2+ 9* 63+ 65*1 44* 70*1 47- 51+ 54* 3*1 55* 10+1 57 13-4 26%1 31-1 32*1 314*2 38- 41*] 23*1 26*1 а + | 41*1 16+5 43+1 20+5 55*3 6+5 25=0 56+3 8+6 42*1 66*3 9+6 46*3 29*1 53-0 30+3 60*1 1*1 39*1 r *1 55*2 1*1 6*1 6 18-0 9*1 2543 11+1 33*1 18-0 39,2 34*2 58*1 37-0 61*1 51+6 6442 65+4 67+1 70*1 mod Y 8+2 45-0 9+5 46+1 11•2 49*2 30*2 50+1 36+4 51*4 6344 52+1 6 54-0 62%1 13*3 25+1 36+1 39+4 5342 63*1 65-0 69+1 Y 2+ 12+ 13+ 14+ 15+ 16+ 17+ 19* 20+ 28, 31 + 32* 38 41+ 45* 46+ 54* 64+ 4+2 21+2 22+1 47+1 65*1 — 18,2 31+1 37+1 68+2 HYPO —— Volume 78, Number 1 1991 Hickey & Taylor 119 y Ticodendron Leaf Architecture are specified. Also, since there is no cladistic anal- ysis of Cunoniaceae, the hypothetical ancestor had to be determined from the most common character states found in the three OTUs of Cunoniaceae. If states were equally common, the state shared with or found in the secondary outgroup was used. Some character states were ordered (see the Appendix). Those characters with only two states were ordered because we were able to specify by outgroup com- parison which states were ancestral and which were derived. However, we polarized only a few char- acters with multiple states, based on structural transformations (see discussion of specific char- acters in the Appendix). All of our character po- larizations are explicit hypotheses of the ancestral character states as well as the structural transfor- mation in the characters that are ordered. We performed a cladistic analysis as part of our pattern sequence analysis in order to identify the optimal branching pattern. PAUP uses parsimony to find trees with the minimum number of steps. Such an analysis gives hypotheses as to which structural homologies are evolutionarily homolo- gous and which are evolutionarily homoplasous. In addition, it identifies the shared derived characters for eac the clades and taxa. Our analyses included two data sets, one with 48 leaf characters and 18 OTUs (Table 2) and a second that included 70 characters and 14 OTUs (Table 3). The latter data set was subdivided and resulted in three sub- sets: the first group has all the characters (desig- nated as 2a), the second group (2b) has characters 1 to 48 (the same characters as data set 1) and the third (2c) has characters 49 through 70. Each analysis is summarized in Table 5 and discussed separately below. he use of global branch swapping for data set 1 resulted in a single shortest tree (Fig. 19) of 249 steps and a consistency index of 0.550 (Table 5). The general topology of this tree has the cunonioid taxa Ceratopetalum, Cunonia, and Acsmithia at the base and a monophyletic origin for the re- maining taxa, which are split into two clades. One clade includes the two juglandaceous taxa Ptero- carya and Platycarya, and the OTUs Davidsonia and Brunellia. The second clade is essentially a fagalean clade with the addition of Myrica and Canacomyrica (Myricaceae). In this tree, Псодеп- dron is the sister OTU to the fossil genus Fagopsis. The set of trees one step longer (Table 5) is similar to the shortest trees in their stability and variability of the topology. Thus it is significant that through- out both sets of trees, the relationships of the taxa of the Cunoniaceae never change, with Acsmithia remaining the sister group to the remaining taxa. In addition, these trees always had two clades, though in some trees Brunellia was placed as the sister OTU to both clades. Lastly, Ticodendron remained well-embedded in the fagoid clade but as a sister OTU to Castanea and Castanopsis or to Fagopsis and Alnus2. We have separated the synapomorphies into four categories based on the characters in our matrices and their distribution on the trees. The first are unistate synapomorphies, in which only one state exists in the entire clade above the node. out lean Е: he next in which two states exist above the node, but one is considerably more common than the other. The third are asymmetrical tristate synapomorphies, which have three states above the node, with one much more common. The last are polystate syn- apomorphies, which have more than three states above the node or have additional states that are nearly as common. Monophyly of the non-Cunoniaceae taxa is sup- ported by 16 character states. Seven of these char- acters relate to the origin of what we term the type 2 (rosid) tooth but this tooth type is found in Cuno- niaceae and other Rosidae, and it is not clear if its origin is monophyletic. Most of the synapomorphies for the clade are polystate. The exceptions include alternate leaf arrangement (character 2), which is a unistate synapomorphy; medial position of the principal vein in a type 2 tooth (12), an asym- metrical bistate synapomorphy; and stipule base that is wider than its attachment to the axis (28)— an asymmetrical tristate synapomorphy. The jug- landoid clade is supported by six character states of which two are asymmetrical bistate synapomor- phies. These are branching of the secondary veins common (46) and slightly zig-zag branching of the twigs (41). The fagoid clade is supported by six character states and all of the synapomorphies are polystate. The analysis of data set 2a using global branch e FIGURE 20. Cladogram of data set 2a, which includes 70 vegetative and reproductive characters and 14 OTUs. This is one of the two shortest trees with 294 steps and a consistency index of 0.609. The other shortest tree differs by having Davidsonia as the sister OTU to all the higher taxa including Castanea, and Castanea becomes the sister taxon to the remaining OTUs. —, reversal; =, multiple reversal; *, synapomorphy; +, parallelism. 120 Annals of the Missouri Botanical Garden HYPO m О = = =. = = q= H- = О = d D. 5 6 > ff < I 2 o t€ © z 2 Ш O (Г I < = О a < = < S О O < co та O а. o > < О = ш | Jb в. 4-0 4+5 4+2 23*] 33%] 1844 6*5 5x] 26*1 34*1 &+3 21%1 8-0 3*1 7+5 14 +4 6+2 31-0 35+2 6* 29-0 17*6 15*5 8+6 15+4 7+2 32*2 40*1 10+1 9* 31 45*0 16% 9+ 16*5 843 38*1 45*3 31-1 15%5 33%1 18-0 10+6 1742 943 3943 46*3 32*1 30*3 34-0 20+3 11+8 20+4 10+3 38-0 32-0 44*1 26-1 17+4 35+1 11+4 39+4 36-0 30*2 19-0 44*1 19-0 46*1 37-0 24*2 32*1 21-0 31-1 41-0 2542 39*2 25%] 47-1 47-0 2842 45*0 28-0 &8+1 39*2 46= 29*1 40*1 74 33-0 11+6 4842 4+4 6+4 12+1 16+4 7+4 8+ 13+2 17*6 8+5 13*3 1442 18-0 9+4 18+1 15+1 23-0 10+5 1*1 26+2 16+2 24*2 1147 3*1 29*1 17+1 2543 13+5 pu 23*1 45+2 18*3 13*3 28+3 20+5 9+1. 334 19-0 2141 38+2 26-1 11+1 45*3 2042 2232 42-0 29%1 18-0 24+1 38-0 45+1 34*1 game 274 47-1 46*1 36+3 37-0 3642 i 44*1 о. 30+1 39+5 4342 8+2 3*1 9*1 9+5 25*1 23*1 11+2 34*2 30*3 30*2 40+2 36+4 45*3 | 1*1 4+1 26-0 33*1 41+2 42+1 46*3 47-0 ] Y 36+1 39+1 T J 2+1 12+3 13+4 14+3 15+6 16+3 17+5 19*1 20+6 26*1 28+1 31+2 32*2 38*1 41+1 46+2 Y 4+2 21+2 22+1 47%1 18+2 31+ 37+1 Volume 78, Number 1 Hickey & Taylor 121 y Ticodendron Leaf Architecture CUNO CERA ACSM BRUN RHOI DAVI PTER PLAT ALN1 CAST МҮН! ЕАСО TICO — МОНЇ HYPO Nn FIGURE 22. swapping (Fig. 20) resulted in two trees with 294 steps and a consistency index of 0.609 (Table 5). The general layout of the trees has the cunonioid taxa at the base and the remaining taxa forming a monophyletic group derived from Acsmithia. The other Rosales, Brunellia and Davidsonia, are at the base of this clade, followed by Rhoiptelea and the remaining hamamelid taxa, except that Cas- tanea is placed as a sister group to Davidsonia. Again, Ticodendron is well-embedded at the top of the hamamelid clade near members of Fagales as a sister OTU to Fagopsis. This topology remains quite consistent in the other most parsimonious trees, with the only difference being that instead of having Davidsonia and Castanea as sister taxa, Davidsonia becomes the sister OTU to all the higher taxa including Castanea, which in turn be- comes the sister taxon to the remaining members. The next most parsimonious tree, with one addi- tional step, also has the same topology The monophyly of the taxa derived from the cunonioid Acsmithia is supported by 18 character states. Again, seven refer to character states of the rosid (type 2) tooth. А unistate synapomorphy is alternate leaf arrangement (character 2), while the medial position of the principal vein (12) in a Strict consensus tree of the nine trees with the shortest length from data set 2b. rosid tooth is a bistate synapomorphy. The pres- ence of both admedial and accessory veins (14) is a tristate synapomorphy. Relationship of Ticoden- dron to the upper hamamelid clade is given by A fused bract-bracteole , while an eight character states. complex (60) is a unistate synapomorph acuminate/acuminate (4) tooth shape is a bistate synapomorphy. he cladistic analysis of data set 2b, performed with global branch swapping (Fig. 21), resulted in nine trees with a length of 204 steps and a con- sistency index of 0.627 (Table 5). The general topology of the tree has the OTUs of the Cunonia- ceae at the base, and the remaining taxa forming a monophyletic group derived from Acsmithia. Brunellia appears next as a sister OTU to the remaining taxa that are again divided into two clades. One is the juglandoid clade with Pterocar- ya, Platycarya, and Rhoiptelea, while the second is the fagoid clade with the remaining taxa. Псо- dendron is well-embedded in the fagoid clade and is the sister OTU to Fagopsis. A strict consensus tree of all nine most parsimonious trees (Fig. 22) shows the major differences from the tree described above (Fig. 21). Here the upper hamamelid clade is unresolved, as are the three clades leading to — E 21. Cladogram of data set 2b, which includes 48 leaf characters and 14 OTUs. This is one of the nine shortest trees (consensus tree in Fig. 22) with 204 steps and a consistency index of 0.627. —, reversal; reversal; *, synapomorphy; +, parallelism. —, multiple is one of the three shortest trees (consensus tree in reversal; —, multiple reversal; *, uc b. н Fig. 2 Brunellia, the Juglandaceae, and the upper hama- melids. In data set 2b, the monophyly of the taxa de- rived from Acsmithia is supported by 16 character states of which seven refer to character states of the type 2 (rosid) tooth. The only unistate syna- pomorphy is alternate leaf arrangement (character 2), while the medial position of the principal vein (12) in a type 2 (rosid) tooth is a bistate synapo- morphy. The presence of both admedial and ac- cessory veins (14) is a tristate synapomorphy. The juglandoid clade is supported by three states while the fagoid clade is supported by eight states. The latter has one unistate synapomorphy—zig-zag twig branching (41)—and one bistate synapomorphy — axillary buds out of the plane of the leaf (42). Our analysis of data set 2c, again using global branch swapping, produced three trees (Fig. 23) 122 Annals of the Missouri Botanical Garden z E = О = - CC == I T 7 т 2 d > o 2 о gs ut Z 3 8 g O 5 O O E o E & | | 2 E 51+3 67-0 66+1 53%1 54%3 62+2 55%1 65+3 50+2 5643 57*1 49%] 54+4 68*1 56+2 63-0 63+43 5941 61-0 56+1 58*1 5-0 60+2 62+3 70*1 62+4 49*1 53*2 55*2 55*2 60*1 63+4 645 66*2 67*1 66*2 64+1 60*1 69*1 6841 68-0 65+2 T 55*3 58*1 51+6 64*3 64+2 65*1 65*1 65+4 П 67*1 70*1 50+1 52+1 53*2 64*3 54+2 69*1 66+3 В J 49+2 51+4 J T 51+5 61+1 62+1 || Џ 63+1 I 68+2 HYPO ——4 FIGURE 23. Cladogram of data set 2c, which includes 2 reproductive and wood characters and 14 OTUs. e with 74 steps and a consistency index of 0.689. pacar ie that had a length of 74 steps and a consistency index of 0.689 (Table 5). The general topology has Acsmithia at the base with the remaining taxa forming a monophyletic group. This is split in turn into two groups, the first with Davidsonia, Cas- tanea, Brunellia, and Ceratopetalum, and the sec- ond with the remaining taxa. Note that the second clade includes Cunonia (Cunoniaceae). Ticoden dron is well-embedded in the second clade near some of the fagalean taxa. There are two other equally parsimonious trees. The only variation from this pattern is the relationships among Ticoden- dron, Alnusl, and d as shown in the strict consensus tree (Fig. In this tree (Fig. 23) the a S of the taxa derived from Acsmithia is supported by a single polystate synapomorphy. The Ceratopetalum/ Brunellia clade is supported by one bistate synapo- Volume 78, Number 1 199 Hickey & Taylor 123 y Ticodendron Leaf Architecture Don D 2 $9.5 T РО $5 а 5 = — == ч О a O о ix О c = Ц. 2 < m мо — FIGURE 24. Strict consensus tree of the three trees with the shortest length from data set 2c. morphy — vessels whose perforation plates are sim- ple or scalariform with only a few bars (54). The second clade includes the remaining taxa and is supported by three characters of which two are unistate synapomorphies. These states include ver- rucate pollen sculpturing (51) and a simple, un- branched inflorescence (61). n addition to these data sets we tried several experimental sets (Table 6). In these trials we at- tempted to use partially a priori methods to reduce the number of characters that were inappropriate for this hierarchy. We did this in order to resolve the situation that arises as the scope of a phylo- genetic analysis is broadened to include more di- vergent groups, which causes certain characters to become homoplasous. Thus, in any randomly chosen set of characters, some characters will be more appropriate for comparisons at low hierar- chical levels and some at higher. We thus devised a filtering method that was based on our two out- groups, Cunoniaceae and Vipera to select characters appropriate to the hierarchy we ex- amined. We suggest that if there is | variability of characters between outgroups it may indicate a potentially homoplasous character for the ingroup. Our method was simply to remove any character that did not have any states shared in both out- groups. In this way, 26 characters were removed, 25 from the leaf character set (1—48) and 1 from the list of other characters (49—70; Table 6). Although the topologies subsequently derived using these sets are generally similar to those found in the previous analyses, the cunonioid taxa switched position, with Acsmithia at the base and Ceratopetalum and Cunonia as sister taxa that together form the sister group to the remaining taxa. This is a particularly interesting result in light of previous suggestions that Acsmithia may be the most primitive genus in the Cunoniaceae (Dickison, 1989). The resulting tree lengths and consistency indexes are shown in able 6. As one would expect, a smaller character base decreased the tree lengths and increased the consistency index; however, the resolution de- creased because there were more trees of the same length. Note that, although the leaf data set (2b) had nearly the same number of characters as the nonleaf set (2c), the consistency index of the former was considerably higher. This suggests that these leaf characters may be better for resolving these relationships at the familial level among the so- called Amentiferae than the combination of repro- ductive and wood characters making up set 2c. The sum of our cladistic analyses shows several general and recurrent similarities. Ticodendron is consistently placed with members of the Fagales я 124 Annals of the Missouri Botanical Garden TABLE 6. Experimental data sets with characters re- moved. The characters in parentheses are the number of characters. The major options are addition of sequence — closest (closest), addition of sequence — simple (simple), unordered in part (part), and unordered all (all). Data set 1 included characters 1-48 and 18 OTUs, and data set 2 included all or parts of characters 1-70 and 14 OTUs. Num Length Consis- ber of of tency Data set options trees trees index Data set 1 (23) Closest part 4T 110 0.682 Simple part 4T 110 0.682 Closest all 47 110 0.682 Simple all 47 110 0.682 Data set 2 а All characters (54) Closest part 8 177 0.667 Simple part 8 177 0.667 Closest all 8 176 0.670 Simple all 8 176 0.670 b Characters 1-48 (23) Closest part 10 94 0.745 Simple part 10 94 0.745 Closest all 10 94 0.745 Simple all 10 94 0.745 с Characters 49-70 (21) Closest part 18 70 0.686 Simple part 18 70 0.686 Closest all 18 70 0.686 Simple all 18 70 0.686 and occasionally Myricaceae. The base of the clade is always found to lie with members of Cunoniaceae and other taxa considered to be members of the Rosidae ry Takhtajan (1969, 1980) and Cronquist (1981, 1988). From our analyses it appears that Cunoniaceae may not be a monophyletic group, but we have not sampled the family sufficiently to show this. A major exception to the consistency of our analyses is the placement of Fagopsis, which switched from the top of the tree to near Cunonia near the base in data set 2c. This variability is probably due to the number of unknown character states in this fossil taxon. Examination of the similarity matrices also shows comparable results (Table 4). Ticodendron is most similar to Fagales in data sets 1 and 2b, based on leaf data, though it is also similar to Brunellia and Rhoiptelea. The nonleaf data set (2c) also shows Ticodendron to be similar to Fagales and Myrica but with strong similarities as well to the two jug- landaceous taxa and К hoiptelea. Overall, Ticoden- dron is similar to the two fagalean taxa Fagopsis (0.491) and Alnusl (0.478), as well as to Rhoip- telea (0.471), Мупса (0.429), Nothofagusl (0.426), and Platycarya (0.400) when all char- acters are included, as in data set 2a of Table 4. Both the similarity matrices and the cladistic analyses show that the leaf and the nonleaf data sets have higher internal similarities ог consisten- cies than does the combined data set. We suggest that this is partly due to the lack of evolutionary homology in some of the characters. In particular, the Normapolles-type pollen that is found in many members of Hamamelidae (e.g., Pterocarya, Platycarya, Rhoiptelea, Myrica, Alnus, and Ti- codendron) and thought to be a natural group marked by synapomorphies (Muller, 1984), may actually have evolved more than once (Batten & Christopher, 1981; Batten, 1986; Kedves, 1989). Thus the pollen characters of this group may act to force these taxa together artificially. Another suite of characters that may have evolved inde- pendently is that of reduced flowers. Since reduced flowers of similar morphology are found in other Thorne, 1989), these are probably homoplasous characters. Similarly, some leaf characters may be evolu- tionarily homoplasous at this familial level of anal- nonrosid hamamelids (e.g., Urticales; ysis. These suspicions were confirmed when we ran experimental data sets in which about half of the characters were dropped, resulting in a dramatic increase in the consistency index. The fact that some of our leaf characters are evolutionarily ho- moplasous is not surprising at this early stage in the development of leaf architectural methods. As future studies re our efine the use of these characters edge of their variability at various tax- onomic levels will be established. CONCLUSIONS Based on these analyses we suggest the following relationships. Ticodendron is most similar to an extinct taxon of Oligocene age known as Fagopsis (Manchester & Crane, 1983), and both of these taxa are most similar to members of the families Fagaceae and Betulaceae (in the sense of Cron- quist, 1981). Manchester & Crane (1983) placed Fagopsis in the Fagaceae, but we do not feel that the structure of the female inflorescence or details of leaf architecture support assignment to that fam- ily. Rhoiptelea is a centrally placed taxon with relationships to both the juglandoid clade and the fagoid clade. Davidsonia is probably a sister taxon to Juglandaceae. Cunoniaceae is the sister group to all these taxa and has characters that are tran- sitional to members of Dilleniidae as well. The Моште 78, Митбег 1 1991 Hickey 4 Taylor 125 Ticodendron Leaf Architecture Cunoniaceae also appear to be paraphyletic but this needs further study. From the standpoint of its systematic placement, therefore, our analysis demonstrates that the affin- ities of Ticodendron lie with the order Fagales (sensu Cronquist, 1981) but that it is clearly distinct from any of the families that are iced under- stood as belonging there. We sd recommend that Ticodendron be reco closely related, the status of the fossil genus Fa- gopsis in relation to the new family is unclear because of a number of missing characters as well as the rather specialized nature of the fruiting struc- tures in Fagopsis In addition to indicating the relationships of 7i- codendron, this study demonstrates the utility of leaf characteristics. Comparisons of the consistency index of data sets 2b (leaf) and 2c (nonleaf) show very little difference. This indicates that these char- acteristics are as useful as floral, pollen morpho- logical, and most anatomical characters that have traditionally been used in plant systematics. Inter- estingly, when we attempted to filter inappropriate characters from the experimental sets, we foun that the consistency index of the leaf data set dramatically improved (Table 6). The leaf set had 3 characters and a consistency index of 0.745 while the nonleaf set had 21 characters and a consistency index of 0.686. We believe that the best approach to such anal- yses is to select characteristics based on their in- formation content at the appropriate hierarchical level. Our experimental data suggest that, for this level of the taxonomic hierarchy, the leaf data set reduced, largely wind-pollinated inflorescences of the Amentiferae (Thorne, 1989). Thus the leaf data set may actually be superior to the nonleaf set at this hierarchical level in this group of plants. Finally, we feel that the analysis presented here has provided a plausible and consistent resolution for the surprisingly difficult systematic problems presented by this unusual new taxon. We also hope that this effort has shed some new light on the evolution of that group of rosid/hamamelid taxa oosely known as the Amentiferae and has eluci- dated some of the methods and characters whereby that analysis could be extended still further. LITERATURE CITED BATTEN, D. J. 1986. Possible functional implications of exine sculpture and architecture in some Late Cre taceous Normapolles pollen. Linn. Soc. Sym. Ser. 12: 219-232 — € J. A. WOLFE. 1989. Systematic Tips n between Nor- mapolles pollen and the Hamamelidae. Pp. 9-22 in . Crane & S. ra ricum (editors), Evolution, Systematics, and Fossil History of the Hamamelidae, Volume 2. ‘Higher’ Hamamelidae. Clarendon Press, Oxford. . А. CHRISTOPHER. | 1981. recognition of N similar pollen genera. Rev. Palaeobot. Palynol. 35: 3. Key to the BEHNKE, H.-D. 1989. Sieve-element plastids, phloem proteins, and the evolution of flowering dix) Hamamelidae. Pp. 105-128 in P. R. Crane & S. Blackmore (editors), Evolution, Systematics, and Fos- sil History of the Hamamelidae, Volu Intro- duction and ‘Lower’ ебали ван: n on Press, xford. CARLQUIST, 5. 1991. Wood and bark anatomy of Tico- dendron: comments on relationships. Ann. Missouri Bot. Gard. 78: 96-104. CRONQUIST, A. 1981. An Integrated System of Classi- fication of Flowering Plants. Columbia Univ. Press, New Yor 1988. The Evolution and Classification of Flowering Plants, 2nd edition. The New York Bo- tanical Garden, Bronx, New York. ни У. С. 1975a. Studies on the floral anatomy f the "cda ie Amer. J. Bot. 62: 433-44 E 975b. Leaf anatomy of Cunoniaceae. J. Lin dos. pe 71: 275- з 77. Wooda of Wi г. Bull. Тотеу. Bot. Club 104: 12- 23. —— —. 1980 iverse nodal anatomy of the Cuno- niaceae. Amer. J. Bot. 67: 975- pues . Comparative wood a y and e шой n the Cunoniaceae. епова. ok 581- 321. 84. Fruits and s of the Cunoniaceae. (Cuno J. sel Arbor. 65: 149-190. 9. Comparisons of primitive Rosidae and a Pp. 47-74 in P. R. Crane & S. Black- more (editors), Evolution, Systematics, and Fossil History of the Hamamelidae, Vol and ‘Lower’ Hamamelidae. Clarendon Press; baat ELDREDGE, 3 а Cladism and common 165-198 in J. Cracraft & N. Eldredge facis сени Analysis and Paleontology. Columbia Univ. Press, New Yor FEUER, S. 1991. Pollen morphology and the pro relationships of | incognitum. Ann. souri Bot. Gard. 78: l. GÓMEZ-LAURITO, J. & L. p. "ous P. 1989. Ticoden- ni a new tree from е America. Апп. Mis- uri Bot. Gard. 76: 1148-1151. 1988. ^r architectural studies Amer. J. Bot. 72(6, part 2): 179. 1977. ` Stratigraphy and Paleobotany of he Golden Valley Formation (Early езин Л A Ta North Dakota. Geol. Soc. Amer. Mem. 1 1979. A revised classification of the an tecture of dicotyledonous leaves. Pp. 25-39 R. Metcalfe & L. Chalk (editors), Anatomy of Di- cotyledons, 2nd edition, Volume 1. Clarendon Press, rd. 1975. The bases of angio- а phylogeny: Va cbe ari morphology. Ann. Mis- uri Bot. Gard. 62: 538-589. ineo: M. J. & J. K. Ta The stereo- structure of the exine and its evolutionary signifi- 126 Annals of the Missouri Botanical Garden cance in Saxifragaceae sensu lato. Linn. Soc. Sym. Ser. 1: 327-378. KAPLAN, D. К. 1984. The concept of homology and its Stuessy (editors), Cladistics: Perspectives on the Re construction of Evolutionary History. Columbia Univ. Press, New York. 1989. Evolution of the Normapolles com- plex. Pp. 1-8 in P. R. Crane & S. Blackmore (ed- itors), Evolution, Systematics, and Fossil History of the Hamamelidae, Volume 2. *Higher' Hamamelidae. Clarendon Press, Oxford. Levin, С. 1986a. Systematic foliar morphology of Phyl- lanthoideae (Euphorbiaceae). I. Conspectus. Ann. Missouri Bot. Gard. 73: 29-85. Systematic foliar morphology of Phyl- lanthoideae (Euphorbiaceae ae). II. Phenetic analysis. Missouri Bot. Gard. 73: 86-98. 6c. Systematic foliar morphology of Phyl- PE EE (Euphorbiaceae). Ш. Cladistic analysis. ot. 11: 515-530. MapDISON, W. P., М. J. DONOGHUE & Р. R. MADDISON. 1984. Outgroup analysis and parsimony. Syst. Zool. 33: 83-103. MANCHESTER, S. R. & P. R. CRANE. 1983. Attached Florissant flora of Colorado, U.S.A. Amer. J. Bot 70: 1147-1164. MILLER, J. S. & J. W. Момлске. 1990. Dioecy and a reevaluation of Lepidocordia and сле (Во- raginaceae: Ehretioideae). Amer. J. Bot. 77: 543- 551. MULLER, J. 1984. Significance of fossil pollen for angio- sperm history. Ann. Missouri Bot. Gard. 71: 419 443 NEFF, N. A. 1986. A rational basis for a priori character IR Syst. Zool. 35: lr . E. HOLSINGE . ЈАМЗЕМ & D. W. 1988. Phylogenetic relationships i in green plants —a comment on the uses of 5S ribosomal RNA sequences by Bremer et al. Тахоп 37: 135-138. STEVENS, Р. К. 1987. Pattern and process: P RC P 155-179 i ner (editors), Biological Metaphor and Cladistic Classification. Univ. Penn sylvania Press, Philadelphia. oo D. L. 198 PAUP version 2.4. atural History ~ Champaign, Illinois. TAKHTAJAN, A. 1969 owering Plants: Origin and Dispersal. Smithsonian Institution Press, Washing- t Illinois n, 80. Outline of the classification of flowering plants (Magnoliophyta). Bot. Rev. (Lancaster) 46: 225-359. TAYLOR, D. W. 3 елик of оуше char- acters in angiosperms a and their out-groups: impli- cations for origins of the иен Amer. J. Вог. 75(6(2)): 211- Seg күр rm ovules and c carpels: their charac- arities, distribution in basal clades, and THORNE, К. melididae': a commentar Pp. 9-1 6 n P. R. сЕ E S. Blackmore (editors). Evolution, PDA and Fossil History of the Hamamelidae, Volume 1. Introduction and ‘Lower’ Hamamelidae. Clarendon Press, Oxford. ToBE, H. 1991. Reproductive morphology, anatomy, гиз relationships of Ticodendron. Ann. Missouri Вог. 5-142. fi їй. "С. А. & К. C. KEATING. 1991. Leaf architec- ture of the Chloranthaceae. Ann. Missouri Bot. Gard. in press). Wo re, J. А. 1989. Leaf-architectural analysis of the Hamamelididae. Pp. 75-104 in P. R. Crane : Blackmore (editors), Evolution, Systematics, and Fos- sil History of the Hamamelidae, Volume 1. Intro- duction and ‘Lower’ Hamamelidae. Clarendon Press, Oxford. ENDIX. Characters and character states for the data matrices (Tables 2, 3) used to produce the cladograms tgrou * indicates that the character states are ordered; NA indicates not applicable 1. *Leaf organization: О pinnate compound, 1 simple inna nd state, even though it is actually pinnatisect, with wings between the pinnae on the leaf axis 2. *Leaf arrangement: О opposite, 1 alternate. Opposite leaves are not common in this group of OTUs nor in the primitive Dilleniidae or Rosidae. Nothofagus has a few species that are opposite. Leaf margin: 3. *Margin: 0 toothed in one order, 1 toothed in more than one order. This character refers to a situation where 2 is a second order of teeth originating on the margins of the first. Two orders of teeth are widely аа but not found in closely related clades. 4. ee е iia and basal margins): О concave, convex, 1 convex, convex, 2 concave, concave, 3 con- cave, dor nate 4 acuminate, acuminate, 5 concave, straig The 2 rst term refers to the apical side and the second to the basal side of the toot ‚ There i is considerable vari- both character state 1 and teeth that are x, acuminate. In Alnus the small teeth are concave, concave. 5. *Sinus shape: O angular, 1 ннн Angular sinuses are the most common type in sperm p. Castanea iub DLE ine both па ки states. 1 +} = = 1 Toot ype 1: th typ Г F о t © from close to the sinus and has no accessory vein along the admedial side. This basic tooth type has previously been called cu- nonioid (Hickey & Wolfe, 1975) and our recent studies show that it is similar to the third tooth found in the Volume 78, Number 1 1991 Hickey & Taylor 127 Ticodendron Leaf Architecture primitive Dilleniidae. All the ancestral character states are found in theoid teeth. 6. Position of Alay aid vein Ка tooth type 1: 0 lateral 2 ое МА, 3 is МА, 5 Faga Platycarya most commonly medial Gy: € also lateral 0). 7. Course of principal vein of tooth type 1: 0 apically deflected, 1 nothofagoid, principal displaced to аон 2 Ticodendron NA, З Betulaceae NA, 4 Еарорз МА, 5 Fagaceae МА. Some ВгипеШа are slightly basally = Acsmi- thia are usually straight and only sometime 8. The presence of veins associated with the principal vein of tooth type 1: 0 both admedial and accessory, 1 only admedial, 2 only accessory, 3 Ticodendron NA, Betulaceae NA, 5 Fago psis NA, 6 Fagaceae n Platycarya the veins are usually missing and only sometimes are accessories found (2). 9. Strength of admedial vein of tooth type 1: 0 same vein order, 1 lower vein order, 2 Betulaceae NA, 3 Tico- dendron NA, 4 Fagopsis NA, 5 Cunoniaceae NA, 6 Fagaceae NA. 10. Course of conjunctal veins close to tooth apex of tooth type none, 1 single, 2 in alternate pairs, 3 Ticodendron NA, 4 Betulaceae NA, 5 Fagopsis NA, 6 Fagaceae Rhoiptelea may be either none (0) or single (1). 11. Major ан closest to sinus of tooth type 1: 0 admedial, 1 accessory or branch, 2 principal, 3 admedial and accessory, 4 Ticodendron NA, 5 Betulaceae NA, 6 nothofagoid: branch from principal, 7 Fagopsis NA, 8 ien e N oth type E this type has the principal vein origi- th with an The transitions between teeth in taxa with both types of teeth, transitions between taxa with only one type of tooth, and outgroup кагы suggest that type 2 teeth are derived from t . Note that in leaves with both types, type 2 teeth are n with strong secondaries. 12. Position of к vein of tooth type 2: 0 Cunonia- ceae NA, 1 Nothofagus NA, 2 Brunelliaceae NA, 3 al. edi Although the teeth of the Magnoliidae and Dilleniaceae are medial, so are some type 1 teeth and theoid teeth. 13. Course of qim vein of tooth type 2: 0 Cunoniace- ae NA, 1 Brunelliaceae NA, 2 Nothofagus NA, 3 straight, 4 apically defle ced. 5 basally deflected. The principal vein in Fagopsis is өү straight a but occasionally basally deflected (5), and in Platycary mostly straight (3) and sometimes њи а Рэнд а). 14. Presence of veins associated with principal vein of tooth type 2: 0 Cunoniaceae NA, 1 Brunelliaceae NA, 2 Nothofagus NA, 3 both admedial and accessory, 4 only accessory 15. Strength of admedial vein of tooth type 2: 0 Cunonia- ceae NA, 1 Nothofagus NA, 2 Brunelliaceae NA, 3 Betulaceae NA, 4 Fagopsis NA, 5 lower vein order, 6 rder same vein or 16. С f conj l veins of tooth type 2: 0 Cunonia- ceae NA, 1 Brunelliaceae NA, 2 d re o da NA, 3 in opposite pairs, 4 in alternate pairs, 5 n Ticodendron mostly has only an admedial but when there are а di aen tals, they are in opposite pairs (3). Myri- ca. has mostly alternate conjunctals (4) but occasionally opposite (3). 17. Fusion of principal and [dee veins of tooth type 2: 0 Cunoniaceae NA, 1 Nothofagus NA, 2 Fagopsis NA, 3 Brunelliaceae NA, 4 Fagaceae NA, 5 separate, 6 d. fuse 18. Termination of vein at apex of tooth: 0 splayed, 1 attenuate, 2 tapered, 3 Shee 4 bulbous Nothofagus (2) has both splayed and bulbous termi- nations. Brunellia may be splayed (0) or tapered (2), Canacomyrica is tapered (2) but sometimes truncate od 3), and A ( splayed (0). 19. *Sinus bracing of teeth: 0 not braced, 1 braced by convergent thickened veins. Not braced (0) is the state found in most teeth of angiosperm leaves. Brunellia may be not braced (0) ог braced (1 20. Major vein(s) closest to sinus of tooth type 2: 0 Cunoniaceae NA, 1 Brunelliaceae NA, 2 Nothofagus NA, 3 admedial, 4 accessory or branch, 5 conjunctal or branch, 6 admedial and accessory. Nothofagus usually has a branch from the admedial and rarely just the admedial (3). 21. Glandularity of tooth apex: 0 anos dark, 1 glandular, trans lucent, 2 nongla ndu Glandular, dark a di the theoid tooth type; presumably homoplasous clear, glandular termina- tions are found in ancestral forms such as the Dilleniales. Castanopsis usually has character state 2 but occasionally has character state 0 2. Apical учен ааны of boi in cleared leaf prepara- tions of mature leaves: 0 setaceous, 1 simple, 2 tylate. Setaceous di е occur in theoid teeth and in presumed DEUS transitions to that t Castanopsis is usually simple but sometimes tylate. 23. *Characteristics of the leaf base or petiole: O does not encircle the axis, stipules discrete or missing, 1 en- circles the axis (> 50% at apical nodes or if opposite, > n axis apparently first ap- pear in Rosales and the primitive, palmate Dilleniidae. 24. Presence of wings on the petiole or stipules associated with the base of the petiole: 0 has stipules, 1 has both wings and stipules, 2 no wings or sti "a es. ings are stipules associated with encircling leaf bases and are common in the Magnoliidae and primitive Dil- leniidae 25. Stipules deciduous below the second node with ex- 128 Annals of the Missouri Botanical Garden panded leaves: 0 yes, always deciduous, 1 no, not usually deciduous, 2 Juglandaceae NA, 3 Myricaceae NA. Cena] n ly А dise Р © J “Prva | : 5 Р Castanopsis and Castanea both may have twigs that are deciduous (0) while others are nondeciduous (1). 26. Petiole attachment: 0 enlarged base without stipules attached, 1 enlarged base with portion of stipule attach- without enlarged base. The distribution of this character is unclear. Brunellia and Castanea may be enlarged (1) or not enlarged (2) 27. Petiole scar: 0 3(4) trace/3 lacunar, 1 arc of more than 4 traces, 2 no scars visible. Castanopsis and Castanea both have character states and 1. 28. Stipule base: 0 not wider than attachment, 1 wider than attachment, 2 Juglandaceae NA, 3 Myricaceae NA. Distribution of this character is unclear but from our observations it appears that stipules with narrow attach ments (1) do not appear until the level of Flacourtiaceae and Cunoniaceae. 29. *Principal venation: 0 pinnate semicraspedodro- mous, 1 craspedodromous Semicraspedodromous (0) is the general case in the primitive Dilleniidae and Rosidae. Craspe edodromy i is rare but fou ndi in the isolated clade originating in Dilleniaceae (e.g., However, because of the difference i in tooth type in their clade, the craspedodromous state here represents a homoplasy with the craspedodromous veins (1) of the Rosidae. Davidsonia usually has character state 1 but occasionally also 0, and Myrica is both brochidodromous and semicraspedodro- 0). ctinidiaceae 30. Angle of origin of lower basal veins: 0 with two pairs at a lower angle and subsequent at a higher angle though d be romewhi irregular і in and spacing, 3 basal ۷ t veins at lower angles, i ac s rom Distr oon "ot this е is variable, although it appears that those with two pairs at a lower d (0) are more common in the nin im angiospe rms 31. Areoles: 0 incomplete, 1 imperfect, 2 well developed. In ncomplete areoles (0) are found throughout the prim- itive angiosperms and those with low-rank leaves 32. Marginal venation: О incomplete, spiked, 1 incom- pletely looped, 2 looped Distribution of this character is unclear. 33. *Number of deuterals: О greater than 18, 1 less than or equal to 1 н P. this у is incompletely е but in primitive angiosperm groups the number of deu ira is usually less than 10. 34. Number of deuterals + d 0 19-28, 1 less than or equal to 18, 2 greater than Distribution of this character is ле Р known, but in primitive angiosperm groups the number of deu terals is usually less than 10, plus opadials is usually less than 15. 35. Number of deuterals + opadials + ips О greater than 60, 1 41-60, 2 less than or equal to Distribution of this character is ини еН known, but in ancestral angiosperm groups the D-O-T number is usually less than 35 in unreduced leaves. 36. Trichome type: O simple, 1 glandular /peltate and simple, 2 glandular /peltate, 3 T-shaped, 4 Cunoniaceae Diao of this character is unclear. 37. *Stomate type: O paracytic, 1 anomocytic, 2 vari- able. Both paracytic (0) and anomocytic (1) stomates have a variable distribution in primitive angiosperms. Dehision of tooth tip in young expanding leaf: 0 oss a eae 1 deciduous, 2 попе Nondeciduous = шша are the general case among primitive angiosper 39. Shape of tooth tip in young expanding leaf: 0 short, stout, 1 elongate, capitate, 2 short, capitate, 3 elongate, simple, 4 Rhoiptelea NA, 5 Мупса NA. Distribution of this character is unclear. 40. Direction of apex of tooth tip in young expanding leaf: 0 pointing abaxially, 1 pointing adaxially, 2 in plane of lamina. Distribution of this character is unclear. kn ОЕ of twigs: O straight, 1 slightly zig-zag, 2 "ns ribution of this character is unclear Castanopsis and Castanea are zig-zag (2): to slightly zig- zag (1). 42. *Axillary buds: 0 in plane of leaf base(s), 1 out of plane of leaf base(s). Axillary buds in plane of leaf base (0) is the expected state for primitive angiosperms, but our observations of this character are incomplete 43. *Uniformity of course of basal secon daries compared to di stal portion of leaf. O no, 1 nearly so, 2 yes. of uniformity in secondaries (0) is the general case in primitive angiosperms. Uniform, or nearly uniform (2, 1) secondaries have a similar distribution to character 29. 44. *Existence of comb vein: 0 present, 1 abse Distribution of this character is unclear, tough ‘comb veins are often found in craspedodromous leav 45. *The existence of intersecondaries: 0 common (great- er than 40%), 1 few (3 to 5), 2 rare (1 or 2), 3 none (never an Leis Анев are frequently found in the magnoliids and are lost in numerous groups of more derived dicots. Branching of secondaries: О rare (1 or 2), 1 few (3 o 5), 2 common (greater than 40%), 3 none (never any). ranched secondaries are common among the Ma noliidae Volume 78, Number 1 1991 Hickey & Taylor 129 Ticodendron Leaf Architecture 47. "Вазе of a number of leaves or terminal leaflets asymmetrical: О none (never any), 1 common Distribution of this character is unclear 48. In type 1 teeth: O principal vein originates from near sinus, 1 Ticodendron NA, 2 Fagaceae NA, 3 Betulaceae NA, 4 Fagopsis МА. 49. Pollen a 0 prolate, spherical, 1 suboblate, spher- ical, 2 oblat The pr asd to spherical iia ш appears throughout the primitive triaperturate grou 50. Pollen aperture: O colporate, 1 porate, 2 colpate. ceae. The porate condition found in group probably had multiple origins. Cunonia is both colporate (0) and colpate (2). 51. Pollen sculpturing and density: 0 reticulate, 1 spi- nulose, moderate, 2 spinulose, few, 3 striate, 4 spinulose, many, 5 verrucate, 6 psilate/ perforate. thro ughout the prim- itive triaperturate groups. Cunonia is both reticulate (0) and spinulose, many 52. *Exine туре tectate/perforate: О columellate, 1 сој- umellate-gran Cab n. (0) appear throughout DORUM. npe sperms. The coding of TEM sections. 53. Exine de жарын 0 sexine greater than ог equal to nexine, 1 sexine greater than nexine, 2 sexine much greater than nexine. The distribution of this character is unclear. The coding of Cunoniaceae is uncertain and needs TEM sections. 54. Vessels: О scalariform with many bars, 1 scalariform with few bars, 2 simple and scalariform with moderate to many bars, 3 simple and scalariform with few bars, 4 simple. Scalariform perforation plates with many bars (0) ap- pear throughout primitive angiosperms. 55. Imperforate type: O tracheids and/or fiber-tracheids with good bordered pits, 1 tracheids and fibers with bor- dered pits, 2 fibers with bordered pits, 3 fibers without bordered pits bud first state (0) appears throughout primitive angio- sperm 56. Axile parenchyma: 0 diffuse, 1 banded, 2 both, 3 other Distribution of this character is unclear. Ceratopetalum is banded (1) and sometimes diffuse (0). 57. шт 0 е 1-3+ з к; 1 heteroge- пеоиз, 1-25 ‚ 2 homogeneous, 1-2 зепа!е. The first state e (0) appears throughout ficken angio- sperms. 58. *Ray crystals: 0 no, 1 yes Distribution of this character is unclear. The coding of Acsmithia, Cunonia, and Pterocarya for this character is not certain. 59. *Bract- мен complex off main axes: 0 уез, 1 only bract, Distribution of this character is unclear. Ticodendron is variable, either 0 ог 1, and dependent on the sex of the inflorescence. The coding of Acsmithia and Cunonia is not certain. 60. *Bract-bracteole complex fused: 0 по, 1 yes, 2 Bru- nelliaceae NA. The first state (0) appears throughout eg n angio- sperms. Псодепагоп is variable, either 0 or 1, and de- pendent on the sex of the inflorescence. The coding of Acsmithia and Cunonia is not certain. 61. *Main inflorescence axis: 0 branched (panicle, cy- mose), 1 es unbranched (unbranched catkin, thyrse, ceme). The first state (0) appears throughout primitive angio- sperms. Сипота is probably variable between the two es. m. £e 62. Structure in T of bracteoles: O inflorescence axis (complex cyme), 1 flower, 2 Ticodendron variable, de- pendent on sex of inflorescence, 3 nothing, 4 Davidsonia МА. € ў state 0 apean throughout pe рая iable inflorescence (0) a and. Ca stanea is d. with an пре ни 0) апа ae a flower (1). 63. Gynoecium: 0 apocarpous, lateral-admedial, 1 2 syncarpous and ascidate, carpel and attachment lateral- admedial, 2 2 syncarpous and a scidate, carpel and at dmedial, 5 2 syncarpous and ascidate, carpel and at tachment basal. The apocarpous state of the gynoecium appears throughout primitive angiosperms. The syncarpous state has a variable occurrence and appears to have multiple origins. (Terminology and distribution from Taylor, in P Ovule integuments: О apobitegmic, 1 apounitegmic, à eniunitegnis, 3 hemibitegmic. primitive. angiospe erms. (Terminology and distribution fun Taylor, 1988, in press.) Coding of Brunellia, Acsmithia, Cunonia, Davidsonia, Nothofagus, and Alnus is not сег- tain. 65. Ovule morphology: 0 ا‎ hemianatropous, 1 bitegmic, anatropous, 2 unitegmic, orthotropous, 3 uni- терт (0) is common minology and dis- tribution from Taylor, 1988.) Coding of Acsmithia, Cu- nonia, Castanea, Nothofagus, and Alnus is not certain. 66. Stamen attachment and SEE hs basal with apical ex- tension, 1 basal without apical extension, 2 dorsi iventrally fixed without apical extension, 3 ede fixed with apical extension. The first state (0) appears throughout primitive angio- perms. Rhoiptelea is either 0 or 1, and coding of Dav- Does and Platycarya is uncertain. 130 Annals of the Missouri Botanical Garden 67. *Flowers: О bisexual or reduced bisexual, 1 unisex- ual. The first state (0) appears throughout primitive angio- sperms. 68. depen other than bract-bracteoles: whorls, , 2 yes, one whorl. TT of this character is nn E is variable with character state 0 o Fa agopsis is uncertain. We consider the e ине in Myri- a to be perianth. 0 yes, two 69. *Endosperm: 0 yes, 1 no or scant dui first state (0) appears throughout primitive angio- sperm 70. *Ovary: 0 superior, 1 inferior or semi-inferior e first state (0) appears полне primitive angio- sperms. Cu 8 Р for Erie coding. SIEVE-ELEMENT CHARACTERS OF TICODENDRON' H.-Dietmar Behnke? ABSTRACT ong the families of Hamamelidae, Ticodendron incognitum Gómez-Laurito & Gómez P. contains S-type sieve- element plastids, the specifics of which (diameter and starch grains) would place it within or Corylaceae. The absence of persistent, nondispersive P-protein close to Betulaceae/ e sieve elements of Ticodendron excludes bodies in th the families of the Urticales and the Fagaceae, all regularly containing persistent P-protein bodies, from the list of its closest relatives Sieve-element characters, i.e., information ob- tained with the transmission electron microscope on sieve-element plastids, phloem proteins, and oth- er morphological features, have largely been used to delimit higher taxa within angiosperms (Behnke, 1981a), but in a few cases were also helpful to assign single genera to a family (Lophiocarpus: Behnke, 1974; Halophytum: Hunziker et al., 1974; Hectorella: 1975; Swartzia: Behnke, 1981b; Geocarpon: "Behnke, 1982) The hitherto unknown taxonomic position of the newly described genus Ticodendron (Gomez-Lauri- to & Gómez P., 1989a, b) and its probable rela- tionships to various families of the Hamamelidae prompted an investigation of its sieve-element char- acters. Moreover, a survey of the sieve-element characters of all of the hamamelidaean families (Behnke, 1989) both facilitates a comparison with and asks for a complementation by data from 7i- codendron. MATERIAL AND METHODS Shoot parts of a sapling of Ticodendron incog- nitum were collected in Costa Rica (Bello & Haber 9809), divided with a razor blade into longitudinal sections, and immediately immersed into a fixative containing qii id and glutaraldehyde 6 he fered OsO,, dehydrated in acetone, embedded and polymerized in an Epon-Araldite mixture, and pro- cessed according to standard methods for ultrathin sectioning and eventual photographing with a trans- mission electron microscope Part of the prefixed material was dehydrated with ethanol, embedded into histowax, sectioned with a sliding microtome, stained with resorcin blue, and screened with a light microscope. GENERAL DESCRIPTION OF THE SHOOT Transverse sections of the shoot sample studied show its composition as follows (cf. Figs. 1-3): the innermost part is occupied by the pith (Fig. 1: Pi) containing thin-walled parenchyma cells and large marginal, shizogenous secretory sacs surrounded by six epithelial cells (Fig. 1: O). The length of these sacs, however, is restricted to about three to five times their width. The outermost pith cells, i.e., those bordering the xylem, often have lignified cell walls. The xylem, phloem, sclerenchyma, and cortex make up distinct layers proceeding toward the pe- riphery of the section. The shoot is protected by a periderm four to six cells broad (Fig. 1: Pe). The presence of a periderm, as well as the extent and the arrangement of xylem and phloem, demon- strate that this shoot part is already in its secondary growth period. The xylem (Fig. 1: X) shows a strict radial arrangement of its rather small and evenly greatly disturb the radial array (Figs. 2, 3: X). Small multiseriate rays, in cross sections only one cell broad, are found at regular intervals. In a few parts of the studied shoot sample, the phloem (Figs. 1, 2: P) reflects its origin from radially aligned ' This study was supported by grants from the Deutsche Forschungsgemeinschaft. The author is grateful to B. E. Hammel, Missouri Botanical Garden, for providin stem material of Ticodendron. 5 2 Zellenlehre, Universitat Heidelberg, Im Neuenheimer Feld 230, D-6900 Heidelberg, Federal Republic of Germany. ANN. MISSOURI Вот. GARD. 78: 131-134. 1991. 132 Annals of the Missouri Botanical Garden Ficu Ticodendron incognitum. — 1. Cross section of young shoot with и о cortex (C), phloem (Р), i i (х) and pith (Pi). Cortex and pith contain secretory elements (О); x100.— 2. Detail of cross-sectioned vascular cylinder xylem with vessels (X), ео delimited against the pieds by sclerenchyma (S), and a pericycle made + up of transparent cells (*); x 230.—3. Part of cross- aa poen with numerous thick- walled sieve elements (s) and black tannin cells (T); oblique sieve plate a tween arrows; X500.— 4. ag n section as phloem with many ШШ -aligned sieve elements (SE) with sieve plates between arrows; Х570.— , 7. S-type sieve- emet plastids with many small to medium-sized starch grains (s); x 30,000. —6. TEM pido Volume 78, Number 1 1991 Већпке Ticodendron Sieve-element Characters cambial initials. The cambial layer itself is rather small and difficult to delimit. The sclerenchymatous sheath (Fig. 2: S) consists of mostly one layer of thick-walled fibers, but lon- gitudinal sections indicate that thinner-walled stone cells do occur rather frequently, probably bridging the gaps between different fiber groups. A one- celled pericyclic layer of clear parenchymatic cells adjoins the sclerenchyma toward the periphery of the shoot (Fig. 2: *). The cortex is composed of small parenchyma cells and numerous large secretory cells (Fig. 1: C, о). Compared to the рић secretory sacs, these are much shorter, their extension being almost the same in all directions The content of the pith and cortex secretory elements seems to be mucilaginous. Їп addition, small cells containing tanninlike substances (Fig. 3: T) are frequently found in the cortex and are also present in all other tissues (see black cells in Figs. 1—3). Crystal cells (containing clustered crys- tals) are abundant in the cortex but are also fre- quent in the phloem and in the lignified pith cells, which border the innermost xylem. SIEVE-ELEMENT CHARACTERS The phloem of Ticodendron is composed of sieve elements (Fig. 3: s), companion cells (Fig. 6: CC), and phloem parenchyma cells. Comparatively many of the parenchyma cells, including those in phloem rays, contain tanninlike substances (Fig. 3: T) With a diameter of 6-10 um, sieve elements of Ticodendron are very narrow; their length is about 80-100 шт. Secondary sieve elements are some- what smaller than primary ones and are more reg- ularly aligned (see Fig. 4 with sieve plates aligned between arrows). Sieve plates are commonly re- stricted to the end walls and have their pores ar- ranged in a single sieve area (simple sieve plate). The position of the sieve plate is in general at right angles to the lateral wall, and only in some cases is much inclined (Fig. 3, between arrows). The pore diameter is about 0.5 um (Fig. 6, arrows), mostly occluded by heavy callose collars (due to unfavor- able conditions during tissue preparation of the sample used). As seen with the transmission electron micro- scope, sieve elements have comparably thick walls and are enucleate when mature. Plastids and mi- tochondria are about the only organelles present and reside at the periphery of the cells, whereas filaments of P-protein are dispersed throughout the sieve element (Fig. 6: *) and are also trapped within the sieve pores. The sieve-element plastids contain starch grains only and thus belong to the S-type. Their average diameter is about 1 um, and they contain an av- erage of eight round or ovoid starch grains. How- ever, there is much size variation among Ticod- endron sieve-element plastids (compare Figs. 5 and 7); diameters range from 0.7 to 1.5 um. No persistent or nondispersive P-protein bodies were found in the investigated sample. All mature sieve elements screened contained P-protein, the filaments of which were more or less evenly dis- persed over the cell lumina (Fig. 6: ж). COMPARISON OF THE TICODENDRON SIEVE-ELEMENT CHARACTERS TO THOSE OF PUTATIVELY RELATED TAXA The sieve elements of Ticodendron contain S-type plastids as do the great majority of dicot- yledons, including all of the hamamelidalean fam- Шез (Behnke, 1981a, 1989). The plastids of Псо- dendron are smaller (1.0 шт) than the total average of S-type plastids (1.4 um, based on 1,400 S-type species). The families that contain S-type plastids with an average diameter between 1.2 and 0.8 um include 12 of the 28 families of putative hamamelid alliance, five of the six families making up the Fagales and all of those in the Urticales (Behnke, 1989). Variation in the morphological features of the starch grains of Ticodendron falls outside that of the families of Urticales; their starch grains are larger and fewer (compare Figs. 6 and 7 with figs. 6.5 and 6.6 in Behnke, 1989). Also, with the exception of the Cannabaceae, all of the Urticales sieve elements contain a specific globular nondis- persive P-protein body (see fig. 6.1 in Behnke, 1989), lacking in Псодепагоп. Among the families of Fagales, the Nothofaga- ceae differ by their larger-sized plastids and the Fagaceae by their compound-spherical nondisper- sive protein bodies found in all species studied so far. Betulaceae and/or Corylaceae would have the section through part of phloem with two sieve elements (SE) and a companion cell (CC). The sieve elements contain many S-type plastids (S) and evenly dispersed filaments of P-protein n (*). Sieve pores ir did are almost totally occluded by callose deposits (white wall parts) and arranged into simple sieve plates; х 5,000 134 Annals of the Missouri Botanical Garden closest similarities in their sieve-element charac- ters. The Betulaceae s.l. (e.g., in the sense of Cron- quist, 1981, or Takhtajan, 1987) have S-type plas- tids with an average diameter of 1.2 шт and about five medium-sized starch grains. If Corylaceae are held to be a separate family (e.g., Dahlgren, 1989), their plastid characters (diameter 1.1 um; about eight starch grains) would fit nicely with those of Ticodendron. However, considering the compar- atively wide range of measurements on which av- erages are based, none of the discussed families would seem inappropriate. Based on its sieve-element data, 7icodendron would fit best within Betulaceae/Corylaceae, if only hamamelid families are considered. И is obvious that a positive alignment of Ticodendron cannot be given on only sieve- dein characters; these characters can only be used to favor or exclude proposals for a taxonomic position made on account of many other characters. LITERATURE CITED BEHNKE, H.-D. 1974. ap a een go Un- tersuchungen an Siebróhren-Plastiden und ihre Aus- sage über die systematische Stellung von Lophio- carpus. Bot. Jahrb. Syst. 94: 114-119. 75. Hectorella caespitosa: ultrastructural evidence against its inclusion into Caryophyllaceae. Pl. Syst. Evol. 124: 31-34. . 1981a. Sieve-element characters. Nordic J. Bot. 1: 381-400. 1981b. Swartzia: phloem ultrastructure sup- porting its inclusion into Leguminosae- Papilionoi- deae. Iseleya و‎ 16. 982. Geocarpon minimum: sieve-element plastids as additional locus ты its inclusion in the x c us Taxon 3 5-4 a eve- ins plastids, phloem pro- eins, к de a of flowering plants. IV. Ha- mamelidae. Pp. 105-128 in P. К. Crane € S. Black more (editors), lo Systematics, and Fos History of the melidae, Volume ystem hee о Special ‘Volume 40A. Clarendon Press, oran А. 1981. Ап Integrated System of Flow- ering Plants. Columbia Univ. Press, DAHLGREN, G. 1989. The last Men iet System of classification of the dicotyledons. Pp. 249-260 in K. Tan (editor), The Davis and Hedge Festschrift. Edinburgh Univ. Press, Edinburgh. GÓMEZ- quacum 1. . D. GOMEz P. . Ticoden- dron: a new tree from Central America. Ann. Mis- souri Bot. Gard 76: 1148-1151. —— & — ———. 1989b. А mie hamamelid dam from Central America. Pp. 341-342 in P. R. Cran & Blackmore (editors), ue эше and F ossil History of the Hamamelidae, Volume Systematics Association Special Volume 40B. Clar- xfor , H.-D. BEHNKE, I. be EIFERT & T. J. MABRY. 1974. Halophytum meghinoi: a beta- lain-containing and P-type sieve- nike plastid species. xon 23: 537-53 A formaldehyde- glutaralde- yde fixative of high osmolality for use in electron Ј. . Biol. 27: 137А-138А. microscopy. TAKHTAJAN, 987 ystema Magnoliophytorum. Nauka, {шд И Russian.] REPRODUCTIVE Hiroshi Tobe? MORPHOLOGY, ANATOMY, AND RELATIONSHIPS OF TICODENDRON ABSTRACT Ticodendron is predominantly dioecious. Both male and female inflorescences bear some cymules (or partial inflorescences) on their primary axes. Male partial inflorescences are borne verticillately in clusters of three, each a es in each axil of the secondary bracts. In pistillate flowers, concentric bundles are prevalent in the periphery of the = wall; the ovary is inferior, bicarpellate, and four-locular with one ovule in each locule; the e is axile; the ovules are apical, pendulous, hemitropous, unitegmic, and crassinucellate; the e is 20-30 c and na by about 20 longitudinally parallel vascular bundles; and the mature see ave a two-cell- layered endosperm and seed coats composed of thick-walled integumentary cell layers and o tissues. In reproductive morphology Ticodendron is distinct from any other family but is best placed in Fagales. On the basis of evidence from reproductive morphology and from other sources it is suggested that Ticodendron appears to have diverged early from a common ancestor with Betulaceae or Fagaceae. Ticodendron, comprising only 7. incognita, which occurs from northern Nicaragua through Costa Rica to central Panama, has long been a mystery plant because of its uncertain identity (Go- mez-Laurito & Gómez P., 1989; Hammel & Bur- ger, 1991). The present paper provides a detailed description of floral morphology and anatomy of Ticodendron including a few remarks on its em- bryology, in order to shed light upon the position and affinity of the genus. MATERIALS AND METHODS All examined materials were collected at Mon- teverde, Costa Rica, and fixed with FAA; examined were nearly mature male flowers (Haber 7072), young female buds and fruits pes Laurito s.n. 1986; Haber 6840, 6868, 7286). Voucher specimens are deposited at MO and For anatomical study, single flowers, groups of male and female flowers, and a piece of mature seed were sectioned with a rotary microtome fol- lowing standard paraffin methods. Before being sec- tioned, the materials were softened with a mixture of 10 parts of glycerol, 3 parts of 10% Aerosol OT, and 90 parts of water (Schmid & Turner, 1977) as described in Тође & Raven (1984). Sec- tions cut at 6-10 um thickness were stained with Heidenhain's hematoxylin, safranin, and fastgreen FCF OBSERVATIONS Ticodendron is mostly dioecious. Male and fe- male inflorescences both appear to be spikes. How- ever, on the basis of anatomical observations dis- cussed later, it has been found that the "spike" is composed of some one- to three-flowered cymules borne on the primary axis. To facilitate discussions in comparing Ticodendron with other related taxa, “partial inflores- following descriptions (see Abbe, 1974: 164-165, for terminology). MALE PARTIAL INFLORESCENCES AND FLOWERS Male partial inflorescences are borne verticil- lately on the inflorescence axis in clusters of three (Fig. 1). Each male partial inflorescence is sub- tended by a single primary bract, and is composed of 1-3 staminate flowers basically arranged in a ! І am grateful to Jorge Gómez-Laurito, Barry Hamme l, and William A. Haber for collection of fixed materials, to Michael H. m and Luis = аа for valuable ааа оп Ticodendron and related taxa in earlier а of the work, and to т Suppor ort in various iode ‘of Biology, College of “Liberal Arts and Sciences, Keats University, Kyoto 606, Japan. ANN. MISSOURI Вот. GARD. 78: 135-142. 1991. 136 Annals of the Missouri Botanical Garden dichasium. Individual partial inflorescences (as well as individual staminate flowers in the case of three- flowered partial inflorescences) are not readily dis- tinguished from each other, but their distinctness is revealed by vascular bundles. Those supplying each partial inflorescence or each staminate flower depart in separate groups from the vascular cyl- inder of the inflorescence axis. Staminate flowers have 8-10 or more stamens and lack perianth and pistil, or ip they have an abortive pistil (see arrow, Fig. he stamens are tetrasporangiate, and the an- thers dehisce longitudinally. Anther walls comprise a fibrous endothecium and a persistent epidermis at maturity. Pollen is two-celled when shed FEMALE PARTIAL INFLORESCENCES AND FLOWERS Female partial inflorescences are solitary, and each is always one-flowered, although it is sub- tended by one primary and two opposite secondary bracts. These three bracts may fuse basally (Fig. 6B). At each axil of the secondary bracts, some 10 small appendages or scales are produced, each receiving a distinct vascular bundle (Fig. 2). These scales are horizontally scattered in the axil and are attached at somewhat different levels. Pistillate flowers are composed only of a single inferior ovary with a small, inconspicuous perianth on the top of the ovary (Fig. 3). The flowers rarely have abortive stamens on the inner side of the perianth; the abortive stamens receive a distinct vascular bundle as do ordinary ones. The ovary is bicarpellate with two (rarely three) elongate. sep- arate styles and four-locular, each locule contair one ovule (Fig. 4). Styles are stigmatic ошнои with elongate and papillate epidermal cells. А nar- row adaxial sulcus is present on styles throughout. Judging from the position of the two styles in re- lation to the inflorescence axis (see Fig. 6C, I), the two carpels are positioned tangentially, not radially. The floral vascular anatomy of a female flower is presented in Figure 6, which diagrammatically shows the vascular course in a longitudinal section (Fig. 6A) and in selected serial transverse sections (Fig. 68–1). At and above the level at which the scales are supplied with vascular bundles (Fig. 6B, C), the vascular cylinder, which is composed of many discrete collateral bundles arranged in a ring, irregularly diverges into the peripheral vascular system and the central vascular system (Fig. 6D, The peripheral vascular system is about 15 or 16 aggregates of 4—10 small bundles. Included are a few inverted ones throughout the whole length of flower (Figs. 5, 6D-H). This peripheral vascular system supplies the perianth (and abortive stamens, if present) and the styles (Fig. 6F—I). Each style has three (occasionally four) vascular bundles (Fig. The central vascular system is composed of some 20 small collateral bundles, which ascend as pla- cental bundles to supply the ovules (Fig. 6E, F). Ovule placentation is axile. OVULES Although pistillate flowers were collected simul- taneously with nearly mature staminate ones, the oldest ovules were still in the megaspore dyad stage. Probably more than several weeks pass between pollination and fertilization. The oldest ovules avail- able indicate that the ovule is apical, pendulous, and hemitropous with the micropyle upward, and that it twists somewhat as it grows (Fig. 7). The ovule is crassinucellate. By the megaspore dyad stage, about 4—5 parietal cells lie above the dyad (Fig. 8). Obturators are lacking. The ovule is unitegmic (Fig. 7), with the integ- ument about 20-30 cells thick and with its tips forming a long narrow micropyle. The integument is vascularized by some 20 longitudinally parallel vascular bundles (Fig. 9), which are derived from a raphe bundle via its post-chalazal branching. The integument is longitudinally ribbed, each rib con- taining a single vascular bundle (Figs. 9, 10). Outer epidermal cells of the integument are radially elon- gate and contain rich cytoplasm (Fig. 10 MATURE SEEDS Only one of the four ovules in the ovary develops into a mature seed, thus each mature fruit is one- seeded. The embryo of a mature seed is oily and massive with two straight cotyledons. The mature seed has a two-cell-layered endosperm (Fig. 11). As the seed matures, only the outer tissue of the integument, including a vascular tissue and a surrounding integumentary tissue, persist, while the inner tissue completely degenerates. Thus, the mature seed coat is only about 8-15 cells thick (Fig. 11). Those cells, except for xylem cells of the vascular tissue, are thick-walled and flattened with tanninlike contents. The outer epidermal cells of the integument, which were radially elongate and contained rich cytoplasm in ovular stages, do not show any specialization in the mature seed coat. DISCUSSION The above-described inflorescence and floral structures of Ticodendron fit within the range of Volume 78, Number 1 Tobe 137 1991 Ticodendron Morphology, Anatomy, and Relationships at Жул, ‹ xs POY ту А > -—- “Ss ом & СР ЭЛ КУЋА mr У 8 m^ atm 6 E Y us $c FIGURES 1-5. Anatomy of partial inflorescences and flowers of Ticodendron. — 1. Transverse section (TS) of а primary male inflorescence axis showin ree male, one-flowered partial inflorescences. Arrow indicates abortive bract.—3. Longitudinal section of female partial inflorescence, which is always one-flowered. — 4. TS of female flower at level of pistil. — 5. of concentric bundles prevalent in the periphery of the female flower. Abbreviations: br2, secondary bract; f, single male flower; ov, ovule; pe, perianth; s, scale; sy, style. Scale bars equal 1 mm, 500 um, 1 mm, 1 mm, and 100 um, respectively. Annals of the Missouri Botanical Garden pe =, | " мами ан MINA, o. oe p, oe- ame T | Y б». t. ERN EET oce н „га a УРАН = rus Diagrams illustrating the vascular anatomy of female pue кайын and flower in Псодепагоп. — se sections at levels marked b-i in A. Ficu A. Median longitudinal section throug ra Abbreviations: ax, ~ primary inflorescence axis; brl and br2, esas and secondary bracts; ov, ovule; pe, h the line e presented in perianth; s, scale; sy, sty Volume 78, Number 1 Tobe 139 1991 Ticodendron Morphology, Anatomy, and Relationships | | | Jr - ONT ETAT ма / ae, - : 2 Graf “2 ce ses € $ 6 ESA ALE LG 7 FIGURES 7-11. Anatomy of ovules and seeds in Ticodendron. — 7. Longitudinal section (LS) of ovule.—8. LS of ovule showing a dyad of megaspores in а nucellus. — 9. Transverse section (TS) of ovule showing that the integ t is ribbed and contains a vascular bundle in each rib. — 10. Enlargement of part of Figure 9, showing a structure of thick integument. — 11. of mature seed showing seed coat structure. Abbreviations: em, embryo; en, endosperm; in, integument; mc, megaspore; mi, tip of micropyle; nu, nucellus; pr, parietal cell; vs, vascular bundle. Scale bars equal 100 um, 20 um, 200 um, 200 um, and 100 um, respectively. 140 Annals of the Missouri Botanical Garden TABLE 1. Comparisons of Ticodendron with related families in selected reproductive characters. Fagaceae Betulaceae Juglandaceae Myricaceae Characters! Ticodendron (Fagales) (Fagales) (Juglandales) (Myricales) Cupule Absent Present Absent Absent Absent vary position Inferior Inferior Inferior Inferior Superior Concentric bundles in Present ? Present Present ? pistil wall Number of carpels рег 2 (2)3(-12) 2(3) 2(-4) 2 ovar Number of locules in 4 (2)3(-12) 2(3) 1 1 ovar 4 (4)6(-many) 2-4(-6) 1 1 Number of ovules in ovary Ovule position Axile, pendu- Axile, erect Basal, erect Axile, pendu- Axile, pendulous ous u lous Ovule curvature Hemitropous Anatropous Anatropous Orthotropous Orthotropous Number of integu- 1(2) 1 1 ments If unitegmic, thickness 20-30 cells 4-5 cells 4 cells thick 6-10 cells 3-7 cells thick integument thick thick thick Vascular bundles in Present Present Present Present Present integument Ribbed integument Present Absent ? Absent T Mature seed coat Cells thick- Cells thick- ? Cells thick- ? walled walled walled Thickness of endo- 2 cells thick Absent or 1 2-4 cells thick l cell thick 1 cell thick sperm cell thick ' References are: Netolitzky (1926), Vaughan (1970), and Corner (1976) for all families (seed coat anatomy); Langdon (1939, 1947) and Poole (1952) for Fagaceae; Wolpert (1910) for Betulaceae; Karsten (1902), Langdon (1939), Meeuse & Houthuesen (1964), Verhoog (1968), and Manning (1978) for Juglandaceae; and Kershaw (1900), Kühn (1927), and Yen (1950) for Myricaceae. variation for these features shown in Fagales, Ju- glandales, and Myricales (see Abbe, 1974, for re- view) and suggest closer affinity of Ticodendron us, the reproductive morphology and anato- with these orders than with other Hamamelididae my of Ticodendron indicates a close relationship or members of different subclasses. In Table 1, to Fagales, Juglandales, and Myricales in the Ham- reproductive features of Ticodendron, Fagales, amelididae (or “Hamamelidae” sensu Cronquist, Juglandales, and Myricales are presented. 1981). Among Fagales, Juglandales, and Myri- In general, Ticodendron agrees well with Fa- cales, Ticodendron seems to be more at home gales, Juglandales, and Myricales. They share an within Fagales, which have diverse reproductive inferior ovary (except Myricales with a superior morphology, rather than in Juglandales or Myri- ovary), concentric bundles in floral vascular system cales, both of which are characterized by certain (though uncertain in Betulaceae and Myricales), 1 hies. Juglandal d Myricales bot} Ticodendron agrees with Myricales in having a longitudinally ribbed integument. bicarpellate ovaries (though frequently tricarpellate characteristically have а unilocular bicarpellate in Fagales and Juglandales), unitegmic ovules (com- ovary with a single orthotropous ovule, all features monly bitegmic in Fagales), vascular bundles in the that are unknown in 7icodendron and Fagales. integument (in the outer integument only in Fagales Closer affinities of Ticodendron with Fagales are with bitegmic ovules), a similar mature seed coat further suggested by evidence from wood anatomy, composed of numerous thick-walled cell layers and pollen morphology, and other data, as specified in this volume. Among Betulaceae, Coryleae may be the closest vascular tissues, and a thin endosperm in mature seed. Ticodendron further coincides with Fagales in having two (or more) ovules per carpel and axile allies of Ticodendron, since Coryleae, like Ticoden- placentation, rather than а uniovulate ovary with dron, lack a perianth in the male flowers but possess basal placentation as in Juglandales and Myricales. it in the female flowers. Evidence from wood anat- Volume 78, Number 1 1991 Tobe 141 Ticodendron Morphology, Anatomy, and Relationships omy also suggests Coryleae for the closest affinity of Ticodendron (Carlquist, 1991). Yet, in Ticoden- dron the ovule is hemitropous with a thicker in- tegument (20—30 cells thick), whereas in Coryleae it is anatropous and probably with a thinner integ- ument (e.g., four cells thick in Betula; Wolpert, 1910); Ticodendron has scales at the axils of the secondary bracts, which are unknown in Coryleae (and in all other Betulaceae); Ticodendron lacks the tertiary bracts that are positioned decussate to the secondary bracts and form by their fusion a characteristic involucre or hull in Coryleae (Abbe, 1935, 1938, 1974); in Ticodendron stamens are not deeply divided as in Coryleae (and in all other Betulaceae); and in 7icodendron female flowers are solitary, while in Coryleae two lateral flowers of the dichasium usually develop (Abbe, 1935, 1938, 1974). These differences suggest that Tico- dendron is unlikely to have direct relationships with Coryleae or other Betulaceae. Nothofagus is the best candidate in comparing Fagaceae with Ticodendron. In fact, Nothofagus is superficially similar to Ticodendron in having exclusively unitegmic ovules, occasionally solitary staminate flowers, partial inflorescences composed of 1-3 pistillate flowers, and occasionally bicar- pellate pistillate flowers (for data on Nothofagus see Langdon, 1947; Poole, 1952; Abbe, 1974). Ticodendron, however, is clearly distinguished from Nothofagus (and in general from all other Faga- ceae) in having hemitropous ovules with thicker integument (only four or five cells thick in /VotA- ofagus; Poole, 1952) and in lacking characteristic cupular valves, lacking a perianth in male flowers, and lacking the tertiary bracts in female partial inflorescences. Those dissimilarities seem to indi- cate that Ticodendron clearly stands out from Fa- gaceae as well. Balanopaceae (Balanops only), which have been placed by Cronquist (1981, 1988) in Fagales, may also require comparisons with Псодепагоп. Bal- anops shares with Ticodendron a female partial inflorescence comprising a single pistillate flower subtended by а “сирше,” an ovary with locules, an anatropous, unitegmic ovule, and drupaceous fruits. However, the **cupule" of Balanops is com posed of numerous, spirally arranged, deltoid bracts, lacking an elevated axial structure characteristic of Fagaceae (Hjelmqvist, 1948). Its homology with the cupule of Fagaceae is doubted (Cronquist, 1981). Even if the bracts of the “cupule” of Bal- anops are comparable with the scales of Ticoden- dron, the former encloses the whole of the pistillate flower, while the latter is restricted only at the axils of the secondary bracts. Balanops is clearly dis- tinct from Ticodendron in having a basal, erect ovule and lacking a perianth in the pistillate flower. Hjelmqvist (1948) suggests on the basis of com- parison in the “cupule”” structure that Balanops probably stands nearest Juglandales and Myricales. Carlquist (1980) notes that the wood of Balanops is comparable with that of Hamamelidaceae. Za- vada & Dilcher (1986) indicate on the basis of similarity cluster analysis using pollen characters that Balanops is a sister group of Nothofagus. At the present level of our knowledge it is uncertain even whether Balanops is related with Fagales in general. Balanops is poorly known with respect to many reproductive structures, and an intensive and overall investigation on the genus seems to be de- sired prior to critical comparisons with putative allies and Ticodendron. In considering evolutionary relationships with Betulaceae and Fagaceae, the presence of the scales in the secondary bract axils of female partial in- florescences of Ticodendron may have a key role. Although 1-3-(or 7- or more-)flowered partial male and female inflorescences are prevalent in Betu- laceae and Fagaceae (see Abbe, 1935, 1938, 1947), they are generally considered to have evolved from a more complex cymose inflorescence by reduction of branches of higher orders (see Abbe, 1938, 1974). Indeed, reduced outer sterile branches are laborated as cupular valves in Fagaceae (Fey & Endress, 1983); furthermore, vestigials of presum- ably reduced outer sterile branches appear as “ѕи- pernumerary scales” in male partial inflorescences of Chrysolepis chrysophylla (Fagaceae) (Hjelmqvist, 1948) or as blinding vascular bundles in female partial inflorescences of Betula lenta (Betulaceae) (Abbe, 1935). Although it is uncertain whether the scales of Ticodendron are directly relevant to the scales of the cupule of Nothofagus or not, their anatomical features are very similar to each other in their cross sectional shape and in receiving a distinct vascular supply (see Langdon, 947, for cupule-scale anatomy in Nothofagus). The scales of the female partial inflorescence of Ticodendron may also represent residuals of re- duced outer sterile branches of cymose inflores- cences. Regarding nonreproductive characters, Carl- quist (1991) showed that Ticodendron has primi- tive wood compared with other Fagales; Feuer (1991) observed that pollen of Ticodendron had plesiomorphic (tectate-granular) wall structure shared with Betulaceae, Nothofagus, Casuarinales, and Urticales (see Zavada & Dilcher, 1986); and Hickey & Taylor (1991) suggested, on the basis of leaf architectural analysis, that Ticodendron 142 Annals of the Missouri Botanical Garden took a more basal position than other Fagales (and Brunelliaceae). Ticodendron may be considered an early derivative either from a common ancestor of Fagaceae before the reduced outer sterile branches of the cymose inflorescence were elaborated as the cupular valves, as in modern Fagaceae, or from a common ancestor of Betulaceae before the reduced outer sterile branches were lost completely, as in modern Betulaceae. For more critical discussion of relationships, embryological studies in Ticoden- dron, Nothofagus, and Betulaceae, which are still meager compared with those in other genera or families, are needed. LITERATURE CITED Studies in the phylogeny of the Betulaceae. I. Floral and inflorescence anatomy and dpud Bot. Gaz. (Crawfordsville) 97: 1-67. 1938. Studies in the phylogeny of the Betu- bise. II. Extremes in the range of variation of floral and inflorescence morphology. Bot. Gaz. (Crawfords- ini 99: 431-469. . Flowers and inflorescences of the ntiferae." Bot. Rev. (Lancaster) 40: 159-261. быш "9 . Anatomy and systematics of Bal- anopaceae. Allertonia 2: 191-246. 19 Wood and bark ame of Ticoden- dron: a on relationships. Ann. Missouri Bot. Gard. 78: 96-104. Corner, E. J. H 1976. The Seeds of Dicotyledons. 2 volumes. Cambridge Univ. Press, Cambridge. CRONQUIST, А. An Integrated System of Classi- fication of Flowering Plants. Columbia Univ. Press, New Yor : . The Evolution and Classification of Flowering Plants, 2nd edition. New York Bot. Gard., ork. FEUER, 5. 1991. Pollen morphology and the systematic relationships of ee incognitum. Ann. Mis- souri Bot. Gard. 78: 1. Fey, В. 5. & Р. К. actes 1983. Development and morphological interpretation of the cupule in Faga- ceae. Flora –468. GÓMEZ-LAURITO, J. & L. D. Gómez P. 1989. Ticoden- dron: a new tree from Central America. Ann. Mis souri Bot. Gard. 76: 1148-1151. HAMMEL, В. & W. С. BURGER. 1991. Neither oak nor alder, but nearly: the history of Ticodendraceae. Ann. Missouri Bot. Gard. 78: 89-95. Hickey, Г. J. & D. W. Тахгок. 1991. The leaf ar- chitecture of Ticodendron and the application of foliar characters in discerning its relationships. Ann. Missouri Bot. Gard. 78: 105-130 HjELMQVIST, Н. 1948. Studies on the floral morphology nd phylogeny of the Amentiferae. Bot. Not., Suppl. 171 KARSTEN, С. 1 902. Ueber die Entwicklung der weibli- n Blüthen bei einiger Juglandaceen. Flora 90: | -333. KERSHAW, E. M. 1900. The structure and development of the ovule of Myrica gale. Ann. Bot. (London) 23: 353-362. Конм, С. 1927. Beiträge zur Kenntnis der intrasemina- len Leitbündel bei den Angiospermen. Bot. Jahrb. 61: 325-379. LANGDON, L. M. 1939. о and anatomical studies of the flower and frui Fagaceae and the Juglandaceae. Bot. Gaz. (Crawfordsville) 101: 301- | The comparative morphology of the Fag I. The genus oe Bot. Gaz. (Craw- fordsville) 108: 350-37 MANNING, W. The classification within the Tug andaceae. Ann. Missouri Bot. Gard. 65: 1058- MEEUSE, А. D. J. € J. HoUTHUESEN. 1964. noecium of Engelhardia spicata (Juglandaceae) ie its Рај ове enetic significance. Acta Вог. Neerl. ies t F. 1926. Anatomie a аут Sa- insbauer Handb. d Anat., Volume 10. Ge- br der pU Ber 1952. The пи ВЕРИ of Medus ra Trans. & Proc . Roy. Soc. New Zealand 80: SCHMID, R. & M. D. as 1977. Contrad 70, an d materials for anatom- ical study. Taxon 26: 5 ToBE, Н. & P. Н. RAVEN. ices , An embryological contribution to systematics of the Chrysobalanaceae. x A ibe Chrysobalaneae. Bot. Mag. (Tokyo) 97: 397- e J. G. 1970. The Structure and Utilization of Oil Seeds. Le & Hall, London VERHOOG, Н. А contribution towards the devel. opmental сас аы morphology of Engelhardia spicata Lechen. ex Blume (J nl de Acta Bot. Neerl. 17: 137-150. WOLPERT, G. 1910. Vergleichende Anatomie und Ent wicklungsgeschichte von Alnus alnobetula und Bet- us Flora 100: 37-67. YEN, T. K. 1950. Structure and development of the flower and the fruit of т rubra Sieb. et Zucc Peking Nat. Hist. Bull. Е ТАМАРА, M. S. & D. Г. D. 1986. Comparative pollen morphology and its relationship to phylogeny in the Hamamelidae. Ann. Missouri Bot. Gard. 73 348-381. POLLEN MORPHOLOGY AND THE SYSTEMATIC RELATIONSHIPS OF TICODENDRON INCOGNITUM Sylvia Feuer! ABSTRACT Pollen of Ticodendron incogni to the slight displacement of the apertures toward o rugulate ridges. The exine is organ a columellate/granular зын ши ка sporadic granules, and a venly thicken of tectum and pendent di and granules. pore opening. Pollen otypic Central American tree of uncertain affinity, was examined in detail with the hope of о. гө йез жок x this taxon. Pollen o f the poles, ope rounded-convex, su with slightly aspidote pores. Exine ie дамчы is minutely spinulate w m data, particularly sculpturing and the nature of the interstitial elements, offer strong support for placement of Ticodendron among the **higher" Hamamelidae with close ties to the Myricales and Casuarinales. Ticodendron incognitum Gómez-Laurito & Go- mez P., a dioecious or polygamodioecious Central American tree, has recently been described as a new genus and species with uncertain affinities (Gomez-Laurito & Gómez P., 1989). Based on morphology, anatomy, embryology, and leaf ar- chitecture, 7icodendron has been linked to the "higher" Hamamelidae sensu Cronquist (1981) (Urticales, Casuarinales, Fagales, Myricales, and Juglandales) and to the Dilleniidae. Pollen of 7:- codendron has been examined in detail and com- pared to that of the above-mentioned orders in the hope of clarifying the relationships of this genus. MATERIALS AND METHODS Fresh pollen fixed in FAA from the collection Haber 7072 (MO) was acetolyzed and divided into portions for viewing with the light microscope (LM), scanning electron microscope (SEM), transmission electron microscope (TEM), and in cryosection (cryo-SEM). Pollen for LM was mounted in glycerine jelly and photographed with a Nikon- system at 500 х using Agfapan (ASA 25) black and white film developed in 1 : 25 Rodinal. for SEM was freeze-dried, gold-coated, and viewed with an ISI DS-130 microscope. Pollen for TEM was osmicated in 1% aqueous osmium tetroxide for two hours, washed several times in distilled water, and embedded in 1.5% purified agar. The pollen/agar moiety was minced into cubes smaller than 1 mm, dehydrated through acetone, and embedded in Araldite 6005 resin. Sections were cut with a diamond knife, collected on 75 X 300 uncoated grids, stained with uranyl acetate and lead citrate, and viewed with a JEOL О EX microscope. Pollen for cryosection was prepared according to methods outlined in Feuer (1987), which in- volved mounting a pollen/agar block onto a cry- osectioning stub, surrounding the block with Tis- sue- Tek? and freezing the stub with liquid nitrogen. The sections were cut with an American Optical Cryo-cut freezing microtome at 10 um thicknesses and picked up on cover slips. The c were then placed in double: distilled boiling water over glasses Tissue-Tek® m several times in ‘boiling double-distilled water, cen- trifuged, and prepared for standard viewing in the ut pollen was rinsed Pollen ultrastructural terminology follows that of Faegri & Iversen (1964) in the use of ektexine and endexine, and Walker & Walker (1981) in the use of interstitium. Families included in the comparative pollen dis- cussion follow the system of Cronquist (1981) ex- cept for the maintenance of the Celtidaceae and Ulmaceae as distinct families. Palynological com- ' Department of Biological Sciences, DePaul University, 1036 W. Belden, Chicago, Illinois 60614, U.S.A. ANN. Missouni Вот. GARD. 78: 143-151. 1991. Annals of the 144 Missouri Botanical Garden "едер Зицотиоо ; "ајаејтеле елер ou 0 "js1g pasip јојовлецо јиешшорәз4 ‘exe əy} jo әшоѕ ut уповола лојовлецо +/— *—/+ "(эвээв[п1эЯ ) Lag ‘(эеэзовэаАИ) YAW (9eooeuuense)) суд (овооеривјапј) Of '(гезовоја оцу) OHH '(9es»eqeuue?)) МУ) '(гезоврцје)) TAD аваовш ћу) АТП (Prode o") yon әвәәвәціп) LYN овооедоивјед) Tyg '(оваовАодлеа) дуд ‘(эвээвтфрицоу) LOY (eLə) па (овоовдед) үң :sjoquiás pue suonerAa1qqe 01 Лә Sl -/* VI VI + + + + + + + + + + + + + + + eI + + + + + + + 6 + + + + 8 0 too + + + + 8 9 0 0 9 S S Y 7 = 0 == Y € влајовлецо ролеце [e10 sjueuro[o впошхојмје juapued Á[uo jo pesoduroo əmyədy ıaAe[ 100] зпоцим әлцләдү pausxory1 Аүиәләитп лаХвј 1004 snonunuoo лодвј 1004 juepued иәцо ev*[[eurnjo?) jueururo1d *peuexor A|1e]nZauu *as1e00 av[[ourno7) o1e[nu e 13 eje[[aumn[oo ummns1oju] peugep-[[ao4 штицалејиј s[ouueqooJorur гетрел ‘моллеи Чим штаоој, Á[uo autxo1X^] sedpu эзеп8пл 1104s uo ‘aly зјаше ur 'Алеујов запис Suunid[nos axepnuids аи одору у aod o3ue1 ədeys əejqoqns 1e[odosıqns 01 1ejodos] YAW SFO ЭГ ОНУ NFO THAD WTA ЧОИ ЧП РА РЯ LOK па IFA и01риэро21] "воцишеј уивухә pue иогриәроэтү иээмзэ4 suosueduroo |еотдојошАје4 `I ATAV], Volume 78, Number 1 1991 Feuer 145 Morphology and Relationships of Ticodendron incognitum parisons between Ticodendron and extant taxa are listed in Table 1 in order of increasing palynological similarity to Ticodendron. PoLLEN DESCRIPTION TICODENDRON INCOGNITUM (Figs. 1-11). Symmetry: radial. Polarity: isopolar to subiso- polar; subisopolarity due to slight displacement of apertures toward one polar face. Size: polar X — 22.7 um (21.3 um-24.5 um); equatorial X= 27.3 ит (25.2 um-30.7 um). Shape: amb triangular slightly rounded-convex; suboblate, rarely oblate- spheroidal. Apertures: three, rarely four pores; pores small, X = 3.4 um (2.6 um-4.5 um), slightly aspidote; exine thickening near apertures. Sculp- turing: tectate-perforate, minutely spinulate; spi- nules typically restricted to short rugulate ridges bearing 3—5 spinules, spinules rarely occurring sin- gly; perforations minute, circular, numerous, clus- tered. Exine structure: only ektexine present. /Von- apertural ektexine organized into tectum, interstitium and foot layer. Tectum: evenly thick- ened, thick (0.5 ит-0.7 ит) continuous with solid, acute-tipped minute spinules (0.2 um-0.3 шт), tectum 0.30 um-0.35 um thick in nonspinulate areas, exhibiting narrow, radial channels often con- tinuous through entire width of tectum. /ntersti- пит: well-defined, 0.4 ит-0.8 um thick; colu- mellate/granulate, columellae numerous, clustered, undulate, in section often appearing pendent from the lower surface of the tectum and rarely fusing with foot layer, irregularly thickened, occasionally bifurcating at the base or fusing laterally in the upper half of the interstitium; granules sparse, spherical, coarse, variable in size, occurring spo- radically throughout interstitium. Foot layer: un- evenly thickened, continuous, upper surface un- even, lower surface smooth. Apertural ektexine: in an area approximately 3 um in diam. surround- ing the pore: (1) interstitium thickening slightly composed of larger and longer pendent columellae and numerous granular elements; granules clus- tered along lower surface of the interstitium; (2) foot layer absent, abruptly terminating after a slight thickening and lamellation at the periphery of this region; and (3) pore opening encircled by a ca. 1 шт rim of solid thickened exine resting on a narrow, finely granular zone. In the light micro- scope, exine thinning in equatorial areas adjacent to pores. Ratio of tectum and interstitium to foot layer: 5: 1 in nonspinulate areas. DISCUSSION The following sections compare pollen charac- ters of Псодепагоп to members of the Fagales, Dilleniideae, Urticales, Juglandales, Casuarinales, and Myricales as well as to the fossil pollen form group Normapolles. Palynological comparisons of Ticodendron to extant orders are discussed from least to most similar. FAGALES (BALANOPACEAE, FAGACEAE, EXCLUDING BETULACEAE) From the available description and one published TEM micrograph by Zavada & Dilcher (1986), pollen of the monotypic Balanopaceae differs from that of Ticodendron in aperture number, aperture type, and in several key features of the exine. Pollen of Balanopaceae is 3-5-colpate with a col- umellate/granulate interstitium. The columellae are short, irregularly thickened, and sporadic with the apical portions interconnected by a narrow, infra- tectal granular zone composed of randomly fused granules arranged just below the tectum. Unlike Ticodendron, endexine is prominent and continu- ous around the grain. Pollen of Fagaceae is even more distinct from pollen of Ticodendron than are the Balanopaceae. Excluding Nothofagus (see below), pollen of Fa- gaceae is tricolporate to rarely tricolporoidate and prolate-spheroidal, and spinulate sculpturing does not occur. Sculpturing ranges from coarsely to finely scabrate to striato-rugulate to papillo-ver- rucate to entwined rodlets. The exine is organized into an irregularly thickened, highly perforate tec- tum, a strictly columellate interstitium and a prom- inent, often discontinuous, foot layer. Endexine is continuous around the grain with a gradual, slight thickening near the aperture (Crepet & Daghlian, 1980; Praglowski, 1982). The 3-10-stephanocolpate pollen of Nothofa- gus with its uniquely raised apertural margins and narrow granular interstitium is also very different from pollen of Ticodendron. Though pollen of Nothofagus and Ticodendron are both spinulate, spinules in the former are twice as large, sparsely distributed, solitary, and often extend through the tectum to project into the interstitium. Unlike Ti- codendron, the coarsely granular interstitium in Nothofagus exhibits no columellar elements and the endexine is thick, continuous around the grain, and abruptly and significantly thickened at the aperture (Praglowski, 1982). Volume 78, Number 1 Feuer 147 1991 Morphology and Relationships of Ticodendron incognitum DILLENIALES data are available on the presence of endexine in (DILLENIACEAE S.S., ACTINIDIACEAE) Dilleniaceae and Ticodendron have very differ- ent pollen. Pollen of Dilleniaceae extends into the subprolate shape range and is tricolpate, tetracol- pate, tricolporate to inaperturate with typically psi- late, perforate to reticulate sculpturing. Though the tectum is perforate as is Ticodendron, the perforations are numerous and broad. The inter- stitium in Dilleniaceae, though well-defined, is com- posed of large, numerous, uniform columellae rest- ing on a thick foot layer. Endexine is continuous around the grain and is particularly well-developed at the aperture (Dickison et al., 1982 Pollen of Actinidiaceae is also very different from that of Ticodendron. Pollen is tricolporate with specialized exine bridges associated with the endoaperture. Pollen shape extends into the prolate range, the sculpturing is psilate to scabrate, and the exine is composed of a highly perforate tectum, a thick continuous foot layer, and endexine con- tinuous around the grain and prominent at the aperture. The interstitium in Actinidiaceae is col- umellate/granulate to granulate, and uniquely highly undulate due to an irregularly thickened foot layer (Dickison et al., 1982) URTICALES (BARBEYACEAE, URTICACEAE, MORACEAE, ULMACEAE, CELTIDACEAE, CANNABACEAE) Relatively little is known palynologically about the major families Moraceae (ca. 1,000 species) and Urticaceae (ca. 700 species), but the available pollen data indicate significant differences between ollen of most Urticales and Ticodendron. Urti- calean pollen is typically di-, stephano-, periporate or rarely tricolporate with scabrate sculpturing or minute spinules clustered together on small raised islands of exine. Ultrastructurally, the interstitium in Ulmaceae is spongy, characterized by randomly anastomosing strands of ektexine or columellate/ granulate in the Celtidaceae and Moraceae with thin columellae interspersed among the granules. The foot layer is extremely thin in most taxa and absent in Barbeyaceae (Zavada, 1983). Conflicting Celtidaceae and Ulmaceae. Endexine has not been reported in the Ulmaceae or Celtidaceae by Zavada (1983), though it appears to be present in Zavada's illustration of Parasponia (Celtidaceae) (fig. 11 in Zavada, 1983). In contrast, Takahashi (1989) re- ported endexine in both the Celtidaceae and Ul- maceae. Endexine has been found in Barbeyaceae and Cannabaceae, but in the latter its occurrence is still problematic. Zavada & Dilcher (1986) de- scribed endexine in their familial pollen description, but in Appendix II of the same paper, which details the coding of character states used in their phy- logenetic analysis, endexine is coded as absent in Cannabaceae. Of the Urticales examined to date, the Canna- baceae are palynologically the most similar to Ti- codendron. The grains in this family are triporate with aspidote apertures and spinulate sculpturing. In thin section (TEM), the tectum is irregularly thickened, channeled, and the interstitium is large- ly columellate (Zavada & Dilcher, 1986). Though Cannabaceae and Ticodendron exhibit a similar overall organization of the interstitium, the details of these elements are quite different. Columellae in Cannabaceae are short, thin, typically apically and basally flared and consistently attached to both the tectum and foot layer. JUGLANDALES (RHOIPTELEACEAE, JUGLANDACEAE) Though pollen of the monotypic Rhoipteleaceae exhibits spinulate sculpturing and a relatively thick, narrowly channeled tectum similar to that of Ti- codendron, other pollen characters differ signifi- cantly. Pollen of Rhoipteleaceae is tricolporate, the spinules are solitary, shorter than those in Ticoden- dron, and tend more toward scabrae than spinules. The narrow interstitium, composed of extremely thin, sparsely distributed columellae and sporadic small granules, is supported by a thick, continuous foot layer and endexine continuous around the grain with significant thickening at the aperture (Stone & Broome, 1971). Like Ticodendron, pollen of Juglandaceae is por- URES 1-5. Ticodendron incognitum (cryo-SEM, SEM).— 1. Polar view (SEM); x 2,950.— 2. Equatorial view IG including one of the small aspidote pores (SEM); x2,860.— 3. Detail of exine surface. Acute-tipped spinules are confined to rugulate ridges separated by sparsely microperforate exine (SEM); omposed of numerous, typically pendent, irregularly shaped, 00 occasionally laterally fused скае and sparse granules (сгуо-ЅЕМ); х 25,000. Bar in each micrograph equals l um. 148 Annals of the Missouri Botanical Garden ate and spinulate with an exine characterized by a narrowly channeled tectum and low to moderately protruding aspidote apertures composed only of tectum and granular, spongy interstitial elements. However, there are also significant differences from Ticodendron. Pore number among Juglandaceae is highly variable (range 2-20 pores) and, in several genera, the pores are nonuniformly distributed over the surface. The spinules are solitary and evenly spaced. The interstitium is predominantly finely granulate with the granules randomly anastomosing to form a spongy network in contrast to the isolated granules found among the other families. Colu- mellae are sporadic, extremely thin, and typically fused only with the tectum. In several taxa, the interstitial elements are suspended from the tectum and separated from the foot layer by a broad exine- free zone resulting from the outward curvature of the tectum (Stone & Broome, 1975). CASUARINALES (CASUARINACEAE) Pollen of Casuarinaceae and Ticodendron ex- hibit a significant number of pollen similarities. Pollen in both taxa is triporate with slightly to moderately protruding aspidote pores. Exine sculp- ture is minutely spinulate with the spinules ar- ranged in single file confined to rugately arranged short ridges. The exine is characterized by a thick, evenly thickened tectum with narrow radial chan- nels and perforations and a wide interstitium com- posed of broad, prominent columellae supported by a thick foot layer. The columellae typically extend up to, but do not fuse with, the foot layer. The apertural exine is composed only of pendent columellae and granules with the foot layer ter- minating approximately 2-3 um back from the ore. Pollen differences between Casuarinaceae and Ticodendron include more protruding aspidote pores and a more prominently rugulate exine in Casuar- inaceae. Ultrastructurally, endexine is absent in Ticodendron and the columellae are (1) more ir- regular in shape; (2) often associated with granular elements; (3) frequently pendent from the tectum extending only half the width of the nonapertural interstitium; and (4) occasionally laterally fused in localized regions just below the tectum. MYRICALES (MYRICACEAE) Pollen of Myricaceae (only Myrica examined) is strikingly similar to that of Ticodendron. Myrica pollen is triporate with aspidote apertures, spinulate with thinner exine in equatorial than polar regions, and a wholly ektexinous exine organized into a thick, evenly thickened tectum with narrow, radial channels and perforations, a prominent interstitium often equal in thickness to the tectum, and a thick continuous foot layer. Apertures are similarly com- osed of only imperforate tectum and pendent, slightly larger columellae and granules, the latter tending to cluster along the lower surface of the interstitium (Coetzee & Praglowski, 1984). erences between Myrica and Ticodendron center on pore and sculpturing characters and el- ements of the interstitium. Pores of Myrica are consistently prominently aspidote as opposed to low-lying to moderately protruding in Ticodendron. Though Myrica and Ticodendron exhibit spinulate sculpturing, each spinule in the former is restricted to a small elevation, which usually anastomoses with other spinule elevations to form irregularly shaped islands. Interstitial elements in Myrica are predominantly large, ovoid coarse granules, which extend the entire width of the interstitium but do not fuse with the tectum or foot layer. The overall morphology of the myricaceous interstitial element, despite its granular nature, approaches that of the nonapertural pendent columellae of Ticodendron. FAGALES (ONLY BETULACEAE) In contrast to the Fagaceae, the Betulaceae, particularly Carpinus and Ostrya, exhibit several pollen characters similar to those of Ticodendron. — FIGURES 6- 'odendron incognitum (LM, TEM).— 6. Equatorial section through two apertures (TEM); x 3,400 mh Equatorial view including three slightly subequatorial pores DX 7. Detail of nonapertural The columellae are typically pendent from the tectum, broad, irregularly shaped and rarely connected to the foot layer (TEM); x18,000.— on ural exine composed of thick, channeled tectum, columellate/granulate interstitium and continuous пи thickened foot layer; x8,600.— 9. Detail of exine sectioned directly through abruptly (arrow). Approa finely ES к exine (arrowhead) surrounds the pore opening (TEM). x12,000.— 10. exine is organized into dense, loosely attached, irregularly shaped segments region m back from the pore the foot layer thickens slightly, becomes lamellate then terminates ae the pore opening, the columellae elongate and granules become more num erous. Thin, Section through apertural interstitium significantly reduced in thickness compared to that seen in other nonapertural regions (TEM); x 14,200. Bar in each micrograph equals 1 um. Volume 78, Number 1 1991 Feuer Morphology and Relationships of Ticodendron incognitum 149 150 Annals of the Missouri Botanical Garden Betulaceae pollen is characterized by three low- lying to moderately projecting aspidote pores. Sculpturing is strictly spinulate with the spinules typically restricted to short rugulate ridges, which tend to be more prominent in Alnus and Betula. The tectum in all genera is uniformly thickened and exhibits narrow radial channels and perfora- tions. Pollen of Carpinus and Ostrya exhibits the greatest overall similarity to 7icodendron. In all three taxa the interstitium is columellate to colu- mellate/granulate (though the columellae are ex- in Betulaceae), the organization of the reduced ektexine in equatorial areas is identical, tremely sma endexine is absent and the apertural exine is com- posed simply of granular and, more rarely, colu- mellar elements pendent from the tectum with the foot layer ceasing several micra back from the pore opening (Feuer & Crane, 1984; Feuer, unpub- lished data) Differences between Betulaceae and Ticoden- dron center on characters of the interstitium, foot layer, and organization of the aperture. In Betu- laceae: (1) the interstitium in unreduced areas 15 narrower, often У; the thickness of the tectum dron; (2) the foot layer, except in Alnus and Os- tryopsis, is evenly thickened with smooth upper and lower surfaces; and (3) Betula, Alnus, and Corylus exhibit vestibulate apertures associated with prominent endexine. SYSTEMATIC RELATIONSHIPS OF TICODENDRON In assessing the relationships of Ticodendron, characters of exine sculpturing were used to es- tablish the initial link between 7icodendron and the “higher most of these families have predominantly spinulate pollen. Pollen aperture type and shape (ratio of polar to equatorial axes) were also considered of prime importance since they are usually more ev- olutionarily conservative, particularly at the fa- " Hamamelidae. Excluding Fagaceae, milial level, than other pollen features. Similarities in ultrastructural elements of the exine were ranked next in diagnostic value. These included, in order of importance, the occurrence of an evenly thick- ened tectum with extremely narrow, radial chan- nels and perforations, the presence or absence of endexine, the organization of the apertural exine, the nature of the interstitial elements, the width of the interstitium, and the thickness and continuity of the foot layer. The pollen data offer strong support for place- ment of Ticodendron among the “higher” Hama- melidae (Casuarinales, Fagales, Myricales, Juglan- dales) as suggested by Tobe (1991) and Hickey (1991), and not the Dilleniidae as жө by Wolfe (1989). Wolfe al hat Ticoden- dron must exhibit the compound ice and oth- ntained er specialized features of the fossil Normapolles- type pollen in order to be considered a member of the Hamamelidae rather than the Dilleniidae (pers. comm.). Evidence for a link between Normapolles and Ticodendron is not a requisite for considering Ti- codendron to be a “higher” hamamelid. Taxo- nomically, Normapolles genera are in a confused state because of the considerable morphological variation among the 100 or so currently recognized forms. Evolutionary relationships based on these highly variable forms would be tenuous at best. Though some Normapolles pollen types have been shown to be juglandalean by their association with macrofossils (Friis, 1983), the majority remain as form genera without established evolutionary re- lationships. Some of the more atypical forms may even be of nonhamamelid origin (Batten, 1989). Ticodendron, with its triporate aspidote aper- tures, minutely spinulate exine, tectum with narrow radial channels, prominent columellate/granulate interstitium, continuous foot layer, and apertures composed only of tectum and pendent interstitial elements belongs among the hamamelids with its closest relationships near Casuarinales and Myri- cales. Pollen data suggest little or no link to the Fagales and Juglandales. Though Urticalean pollen appears significantly different from that of Ticoden- dron, the pollen data are too incomplete to allow for adequate comparison and analysis Pollen of Casuarinales and Deo endroi exhibit moderately protruding aspidote pores, suboblate to oblate shapes, spinules confined to short, rugately arranged ridges, and a wide interstitium composed mainly of prominent columellae, which typically are not fused to the foot layer. Aperture structure is identical in Casuarinales and Ticodendron. Casu- arinales can only really be distinguished palyno- logically from Ticodendron at the ultrastructural level by its longer columellae, which typically ex- tend the entire width of the interstitium, the ab- sence of nonapertural interstitial granules, and the presence of endexine. The occurrence of endexine in Casuarinales may not be as significant a differ- ence as it at first appears since the endexine is extremely thin and occurs only in nonapertural £e ~ Pollen similarities between Myricales and Ti- codendron include suboblate shape, the presence Volume 78, Number 1 1991 Feuer 151 Morphology and des права of Тісоаепагоп incognitu, of coarse columella-like granules in a wide inter- stitium, the occurrence of a thick, continuous foot layer, and the absence of endexine. Like the Casua- rinales, Myricales and Ticodendron exhibit an iden- tical aperture structure including the presence of clustered granules in basal regions of the aperture (Fig. 10, arrow). Pollen of Myricales can be dis- tinguished from that of Псодепагоп by its more protruding aspidote pim isolated spinules tending to cluster on small islands of exine, and large, coarse, granular о which approach colu- mellae in overall morphology but are not fused to the tectum or the foot layer. The Juglandales, with their tricolporate (Rhoip- teleaceae) and 2-20-porate (Juglandaceae) pollen, uniformly distributed, solitary, spinulate sculptur- ing and spongy, granular, interstitium composed of a randomly anastomosing network of fine gran- ules, are remote from Ticodendron. Connections to the Fagales are even more re- mote than those of the Juglandales. Pollen of Bal- anopaceae and Fagaceae, including the distinctive Nothofagus, differs in every fundamental aspect from that of Ticodendron. Only Betulaceae could be considered as possibly linked to Ticodendron. A relationship in Betulaceae would have to lie with the simple aperturate, endexine-free pollen of Car- pinus and/or Ostrya. Alnus, with its multiaper- turate, arcuate pollen exhibiting a spongy, granular interstitium and compound apertures, is far re- moved from Ticodendron as is Betula with its multiporate, similarly compound vestibulate aper- tures characterized by well-developed endexine. LITERATURE CITED BATTEN, D. Systematic e between Хаз pénis and the Hamamelidae. Pp. 9- . R. Crane & S. Bla о (editors), Evo- ie бук зета св, and Fossil History of the melidae, Volume 2: The “Higher amamelidae. Bu matics eain Special Volume 40B, Clar- endon Press, Oxfor т. J.A. & J. PRAGLOWSKI. 1984. Pollen Я г the occurrence of Casuarina and о n the Tertiary of South Africa. Grana 23: CREPET, У. L. & C. P. DAGHLIAN. 1980. p inflorescences from the Middle Eocene of Tenn and the diagnostic value of pollen (at the Subfamily level) in the Fagaceae. Amer. J. Bot. 67: 739-757 CRONQUIST, А. 1981 An Integrated System of Classi- fication of Flowering Plants. Columbia Univ. Press, k. New Yor Dickison, W., J. W. NowICKE & J. J. SkvanLA. 1982. Pollen morphology of the Dilleniaceae and Actinidi- aceae. Amer. J. Bot. 69: 1055-1073. FaEGRI, К. & J. IVERSEN. 1964. Textbook of Pollen Analysis. Munksgaard, Copenhagen. FEUER, S. 1987. Combined cryo- and thin sectioning TEM) in the elucidation of mimosoid terad/ а ultrastructure. Rev. Paleobot. Palynol. 52: 367-374 P CRANE. 1984. Palynology of extant and fossil Betulaceae. Sixth International Palynologi- ~ mapolles pollen. Rev. Paleobot. Palynol. 39: 161- 188. GOMEz-LauRITO, J. & L. D. GóMEz P. 1989. Ticoden- dr Mis- го new tree from Central America. Апп. souri Bot. Gard. 76: 1148-1 Hickey, L. D. W. TAYLOR 1991. The leaf ar- chitecture of Ticodendron and the application of um characters in px. its relationships. Ann Missouri Bot. Gard. 78: 105-130. Practovs J. 1982. Fagaceae L.: Fagoideae. World and Spore Flora 11: 1-28. STONE, р. Е. & С. В. BROOME. 1971. Pollen ultra- structure: evidence for relationship of the Juglan- =: and the Rhoipteleaceae. Pollen & Зрогез 13: -14. ———— € ——. 1975. Juglandaceae Е Rich ex Kunth. World Pollen and Spore Flora 4: 1-35. TAKAHASHI, M. 1989. Pollen morphology of TRE Ulmaceae: a reinvestigation. Pp. 253-265 in P. R. Crane & S. Blackmore (editors), Evolution, Systematics, and Fossil History of the Hamamelidae, Volume 2: The “Higher” Hamamelidae. Systematics Association Special Volume 40B, Clarendon Press, Oxford. Тове, H. 1991. Reproductive morphology, anatomy, and relationships of Ticodendron. Ann. Missouri Bot. Gard. 78: 135-142. WALKER, J. W. & A. G. WALKER. 1981. Comparative pollen morphology of the Madagascan genera of эй risticaceae (Mauloutchia, Brochoneura, and Hae matodendron). Grana 20: 1-17. Leaf architectural analysis of the dae. Pp. 75-104 in P. R. Crane & S. Blackmore (editors), Evolution, Systematics, and Fos- sil History of the Hamame elidae, Volume 1: Intro- sociation Special Volume 404A, Clarendon Press, xford. ZAVADA, M. S. 1983. Pollen morphology of Ulmaceae. Grana 22: 23-30. . L. DILCHER. 1986. Comparative pollen morphology and its relationship to phylogeny of pol- len in the Hamamelidae. Ann. Missouri Bot. Gard. 73: 348-381. NEW SPECIES OF PANICUM SUBGENUS DICHANTHELIUM (POACEAE: PANICEAE) FROM SOUTH AMERICA Fernando Zuloaga! and Osvaldo Morrone ABSTRACT Six new species of the pantropical genus Panicum subg. Dichanthelium (Poaceae: Paniceae) are described, illustrated, and compared with the putative most closely pd species. Five of these species, P. heliophilum, Р. adenorhachis, P. cabrerae, P. petropolitanu aense, are found in campos rupestres, which are cu altitude rocky fields of Brazil. Panicum и is че аг to elevated savannas of Venezuela and northwestern Brazil. During the revision of Panicum for America by the senior author, six new species were discovered in South America, all of which belong in subg. Dichanthelium А. Нисћс. & Chase. Subgenus Dichanthelium was traditionally characterized (Hitchcock & Chase, 1910, 1915; Pilger, 1940; Hsu, 1965) as having two vegetative phases: one with a rosette of short, broad leaves formed during autumn and winter that produce primary panicles with chasmogamous spikelets on elongated shoots in spring; and the other with fas- cicles of short branches that produce secondary panicles with и spikelets during spring, summer, and autum Gould (1974) ИВТ Dichanthelium at the generic level, an idea followed by a number of different authors (e.g., Brown & Smith, 1975; Clark & Gould, 1975; Gould & Clark, 1978; Gould, 1980; Hansen & Wunderlin, 1988). Pohl (1980) and Zuloaga (1987) reported vari- ation in the vegetative characters among Central and South American species of Dichanthelium. Lelong (1984), Zuloaga (1987), and Webster (1988) recognized Dichanthelium as a subgenus of Panicum. Due to the absence of rosette and chasmogamous and cleistogamous panicles on the Central and South American species of the sub- genus, it can be recognized by having inflores- cences lax and diffuse, spikelets ellipsoid to obovoid, with the upper anthecium indurate, smooth, with simple papillae on its surface, and apiculate, the apicule covered with stomata and with or without unicellular macrohairs. The leaf anatomy is uni- form in the subgenus and is non-Kranz, with two bundle sheaths around the vascular bundles, the mesophyll irregularly radiate around the vascular bundles, and with extensions of the parenchyma- tous sheath present toward both surfaces of the ade. Within subg. Dichanthelium, the new species are placed in sect. Dichanthelium because of their membranous-ciliate ligules, spikelets with the upper glume and lower lemma 7-9(-14)-nerved, the up- per anthecium apiculate and papillate, with simple papillae regularly distributed, and the upper lemma apically pilose or scabrous. Five of these species are found in Brazil, campos rupestres, rocky fields at moderate ха tudes 1,000-1,800 m (Енеп, 1978), or in rocky mountains of higher altitudes (over 2,000 m) near Rio de Janeiro. The other species grows in elevated savannas in Venezuela and northwestern Brazil. Panicum heliophilum Chase ex Zuloaga & Morrone, sp. nov. TYPE: Brazil. Minas Gerais: Chapeo de Sol, Serra do Cipo, 110 km NE of Belo Horizonte, 900 m, Маг.-1 Apr. 1925, A. Chase 9147 (holotype, US; isotypes, F, GH, MO). Figure 1. Panicum pycnoclados Tutin affine sed spiculis hirsutis sine stipite, et dum rigidis, lineari-lanceolatis, basi sub- cordatis differ Plants perennial, short-rhizomatous. Culms 50- 70(-100) cm tall, many-noded, tangled, leaning on the adjacent vegetation, the basal portion de- cumbent and geniculate, the upper portion erect, freely branching; internodes cylindric, pilose, solid toward the base, otherwise hollow; nodes villous. Sheaths striate, 0.5—4 cm long, strongly distichous, ' Instituto de Botánica Darwinion, Casilla de Correo 22, San Isidro (1642), Argentina. ANN. Missouni Вот. GARD. 78: 152-163. 1991. Volume 78, Number 1 Zuloaga & Morrone 153 1991 Panicum from South America FIGURE 1. Holotype of Panicum heliophilum. —a. Нађи. —b. Ligule.— c. Spikelet, ventral view. — d. Spikelet, lateral view.—e. Spikelet, dorsal view. —f. Lower palea.—g. Upper anthecium, dorsal view. —h. Upper anthecium ventral view. —i. Upper palea with lodicules, stamens, and stigmas. —j. Caryopsis, embryo side. —k. Caryopsis, hilum side. 154 Annals of the Missouri Botanical Garden hispid with short hairs, the margins long-ciliate; auricles small, pilose; collar brownish to purplish, covered with short, whitish hairs; ligule a ciliate membrane, 0.2-0.7 mm long, the cilia 0.1-0.6 mm long; blades linear-lanceolate, 4-5(-12) cm long, 0.3(-1) cm wide, subcordate, densely hirsute on both surfaces, flat or folded, the margins car- tilaginous, scabrous, and short-ciliate. Primary panicles lax, diffuse, 4-16 cm long, 3-13 cm wide; peduncles hispid, up to 8 cm long; branches op- posite or alternate, the lower ones sometimes whorled, diverging from the axis; rachis flexuous, sparsely hirsute, the axes of the branches short- hirsute or glabrous, the axils of the branches brown- ish, hispid; pedicels smooth, glabrous, 2-13 mm long. Secondary panicles similar to the terminal one, smaller. Spikelets solitary, ellipsoid, 2.6-3.1 mm long, 1-1.1 mm wide, hirsute, greenish, with- out a stipe between the lower and upper g lower glume 2-2.7 mm long, % the length of the spikelet, acuminate, 3-nerved, the nerves anasto- mosed toward the distal portion; upper glume and lower lemma subequal, 7-nerved; lower palea el- liptic, 2 mm long, 0.3 mm wide, hyaline; lower flower absent; upper anthecium long-ellipsoid, 2.4— 2.7 mm long, 0.8-0.9 mm wide, stramineous, brownish at maturity, papillose, with simple papillae evenly distributed, densely pilose, with appressed, long hairs toward the apex of lemma and palea; upper lemma 5-7-nerved; rachilla prolonged or not into a short mucro above the upper anthecium; lodicules 2, fleshy, conduplicate, truncate, ca. 0.6 m long; stamens 3, the anthers ca. 1.3 mm long, Suid. stigmas plumose. Caryopsis ellipsoid, 1.4 um oblong; glume; mm long, mm wide, brownish; embryo У, the length of the caryopsis. Paratypes. yn MINAS GERAIS: Diamantina, sum- mit of Serra de San Antonio, 1,400 m, 27-30 Dec dr dod A. Chase 10354 (US); Mun. Jaboticatubis, Lagoa San а Conceição do Mato Dentro, 27 July 1977, Sendulsky & Burman 1920 (SI, SP); km 114 ao largo da rodovia m, Sen- leste da cidade de Diamantina, 1,050-1, 1982, Burman & о 696, 697, 709 (SI, SP); Serra do Cipó, ao longo da rodovia Lagoa Santa—Con- ceicao, 9 Dec. 1971, d. 446 (SP); Serra do Cipo, 25 Feb. 1979, Burman 525 (SP), 24 June 1979, Burman 485 (SP). Panicum heliophilum is most closely related to Р. pycnoclados Tutin, from which it differs by having hirsute, nonstipitate spikelets and stiff, lin- ear-lanceolate blades with subcordate bases. Pan- icum pycnoclados has glabrous to sparsely hispid spikelets and lax, lanceolate to ovate-lanceolate blades that are cordate and amplexicaulous. Panicum heliophilum is also related to P. as- surgens Renvoize; it differs by having leaf blades linear-lanceolate, rigid, usually folded and densely hirsute, and spikelets hirsute 2.6–3.1 mm long. Panicum assurgens has leaf blades lanceolate, flat, herbaceous, sparsely hispid, and spikelets puber- ulous to glabrous, 3.3 mm lon Some of the specimens examined have the ra- chilla prolonged above the upper anthecium as a short mucro, a character also present in other genera of Paniceae [Lasiacis grisebachii (Nash) A. Hitche. and L. ruscifolia (HBK) A. Нисћс. (Davidse, 1978); Brachiaria tatianae Zuloaga & Soderstrom (Zuloaga & Soderstrom, 1985)]. In Panicum heliophilum the lower leaf blades abscise, leaving the dry leaf sheaths on the culms. This is probably an adaptation to the seasonally dry habitat. It also occurs in P. adenorhachis. P. cabrerae, and P. cucaense, species that grow in a similar habitat. Panicum heliophilum grows in campos ru- pestres of the state of Minas Gerais, Brazil, between 900 and 1,400 m elevation. It is common in sandy soils, Were it is found, according to notes on Chase 9147, “їп dry rocks, leaning out, m with Echinolaena inflexa and resembling it Panicum adenorhachis Zuloaga & Morrone, sp. nov. TYPE: Brazil. Bahia: Mun. Rio de Contas, 6-10 km ao NW de Rio de Contas, na estrada para o Pico das Almas, 13?32'S, 41*53'W, 1,000 m, 21 July 1979, 5. А. Mori, R. M. King, T. S. dos Santos & J. L. Hage 12451a ор. СЕРЕС; isotypes, MO, US). Figure 2 Panicum stipiflorum Renvoize affine sed foliis lanceo- latis, glandulosis et spiculis 2.7-3 mm longis sine stipite, gluma inferiore I aci у, vel minus longiore, enervia vel uninervia differ Plants perennial. Culms trailing, leaning over the adjacent vegetation, many-noded, sparingly branched; internodes cylindric, hollow, glabrous; nodes brownish, short-pilose or glabrous. Sheaths striate, caducous, glabrous, covered with small glands, the margins long-ciliate; collar short-pilose, ligule a ciliate membrane, 0.2—0.6 mm long; blades narrowly lanceolate, flat, 4-6 cm long, 0.3-0.5 cm wide, subcordate, the adaxial surface scabrous, the abaxial surface glabrous and covered with small glands, the basal margins long-ciliate with tuber- culate hairs but otherwise scabrous. Primary pan- icles lax, diffuse, few-flowered, 5-7 cm long, 4-5 cm wide; rachis wavy and glandular, the nodes distant; branches widely divergent, solitary at each 155 Zuloaga & Morrone Volume 78, Number 1 1991 Panicum from South America — а. Habit.— b. Ligule.—c. Axis of the panicle with glands. — ventral view.—j. Upper palea with lodicules and stamens. — , lateral view.—f. Spikelet, dorsal view.—g. Lower palea.— |. anthecium, hilum side. Holotype of Panicum adenorhachis. FIGURE 2. d. Pedicel and spikelet, ventral view.— e. Spikelet Upper anthecium, dorsal view. —i. Upper k. Caryopsis, embryo side. —1. Caryopsis, 156 Annals of the Missouri Botanical Garden node with scabrous and glandular axes and long- pilose axils; pedicels claviform, short pilose. Sec- ondary panicles similar to the terminal one. Spike- lets solitary and widely spaced, long-ellipsoid, 2.7— 3 mm long, 0.9 mm wide; lower glume 0.7-0.8 mm long, less than /, the length of the spikelet, acuminate, enervate to l-nerved, hyaline; upper glume and lower lemma 9-nerved, short-hispid, subequal, longer or shorter than the upper anthe- cium; stipe lacking between the lower and upper glume; lower palea lanceolate, 1.8 mm long, 0.4 mm wide, hyaline, ciliolate on the margins; lower flower absent; upper anthecium long-ellipsoid, 2.3 mm long, 0.8 mm wide, stramineous, brownish at maturity, papillose with simple papillae in longi- tudinal rows; upper lemma short-mucronate and with small macrohairs at the apex; rachilla pro- longed beyond the upper anthecium as a short mucro; lodicules 2, conduplicate, truncate, ca. 0.3 mm long; stamens 3, the anthers mm long; stigmas plumose. Caryopsis ellipsoid, brownish, 1.5 mm long, 0.7 mm wide; hilum punctiform; embryo less than У, the length of the caryopsis. Paratype. BRAZIL. BAHIA: Alto do Pico das Almas, 1,850 m, 20 Feb. 1987, Harley et al. 24460 (K). Renvoize (1984) treated the collection Mori et al. 12451a as P. stipiflorum Renvoize. Panicum adenorhachis differs from P. stipiflorum by hav- ing lanceloate blades that are attenuate at the base, scabrous adaxially, and glandular abaxially; non- stipitate spikelets 2.7-3 mm long; enervate or -nerved lower glume that is less than У, the length of the spikelet; and the evidently glandular rachises of the panicles. Panicum stipiflorum has leaf blades ovate-lan- ceolate, cordate and amplexicaul, densely to sparsely hirsute on both surfaces, without glands; spikelets 2-2.6 mm long, with a manifest stipe between the lower and upper glume, lower glume 3-5-nerved, № (occasionally %) the length of the spikelet, and rachis of the inflorescence without glands. Panicum cumbucana Renvoize differs from P. adenorhachis by having nonglandular, short (1- 2 cm), and narrow (0.3-0.6 cm) blades that аге cordate at the base; nonglandular rachises; stipitate spikelets 1.9-2.4 mm long; and lower glumes (1-)3- nerved, У, the length of the spikelet. Panicum adenorhachis has glands on the leaf blades and panicles, a feature also present in other species of subg. Dichanthelium, e.g., P. sabulo- rum Lam., P. stigmosum Trin., P. strigosum Muhl., and P. sphaerocarpon Elliott. Panicum adenorhachis inhabits campos ru- pestres of the state of Bahia, Brazil. Panicum cabrerae Zuloaga & Morrone, sp. nov. TYPE: Brazil. Bahia: Mun. Rio de Contas, Pico das Almas, a 18 km ао SNW de Rio de Contas, 13%33'S, 41°57'W, 1,600-1,850 m, 22 July 1979, S. A. Mori, R. M. King, T. S. dos Santos & J. L. Hage 12475 (holotype, CE- PEC; isotypes, MO, US). Figure 3. Panicum stipiflorum Renvoize affine sed reclinatum, foliis lineari-lanceolatis, basi attenuatis, longe pilosis sine marginibus ciliatis differt Plants perennial. Culms ca. 50 cm tall, leaning over the vegetation, many-noded, sparingly branched; internodes cylindric, hollow, sparsely pi- lose with whitish hairs; nodes densely pilose with long, whitish hairs. Sheaths striate, 1.5-2 cm long, longer than the internodes, long-hispid, the margins pilose with whitish hairs, especially toward the up- per portion; auricles small, pilose; collar pilose; ligule a ciliate membrane, ca. 0.2 mm long; blades linear-lanceolate, 4-6 cm long, 0.2 cm wide, ba- sally attenuate, apically acuminate, densely pilose, with long, whitish hairs on both surfaces, caducous at maturity, the margins involute. Panicles lax, diffuse, 4.5-6 cm long, 3.5-6 cm wide; peduncles hispid, 5-7 cm long; rachis cylindric, densely hispid on the lower portion, otherwise glabrous, the nodes distant; branches alternate, divergent and reflexed, basally hispid, distally glabrous, the axils pilose to glabrous; pedicels claviform, smooth, glabrous, 2- 9 mm long. Spikelets solitary, ellipsoid, 2–2.3 mm long, 0.8–0.9 mm wide, plano-convex, hirsute with whitish hairs on the glumes and lower lemma; lower glume 1.1-1.3 mm long, nearly У, the length of the spikelet, 3-nerved, acute, a small stipe present between the lower and upper glume; upper glume 9-nerved, the nerves anastomosing toward the apex; lower lemma 9-nerved; lower palea lanceolate, 1.5— 1.8 mm long, 0.3 mm wide, glabrous, truncate, hyaline; lower flower absent; upper anthecium el- lipsoid, 1.9 mm long, 0.7 mm wide, papillose, with simple papillae all over its surface, glabrous or with a few microhairs toward the apiculate apex; rachilla prolonged or not beyond the upper anthecium as a short mucro; lodicules 2, truncate, ca. 0.2 mm long; stamens 3, the anthers 1.2 mm long; stigmas plumose. Caryopsis obovoid, 1.1 mm long; hilum oblong; embryo less than У, the length of the cary- opsis. Panicum cabrerae is related to P. stipiflorum, P. adenorhachis, and P. congestum Renvoize. Panicum stipiflorum has ovate-lanceolate, cor- date, glabrous to short-pilose blades with ciliate margins, upper anthecium pilose at apex, and culms ecumbent and prostrate at the base, then becom- Volume 78, Number 1 Zuloaga & Morrone 157 1991 Panicum from South America FicURE3. Holotype of Panicum cabrerae. —a. Habit. — b. Sheath and lower portion of a leaf blade. — c. Шеше. — d. Spikelet, ventral view.— e. Spikelet, lateral view. — f. Spikelet, dorsal мем. — в. Lower palea. —h. Upper anthecium, dorsal view. —i. Upper anthecium, ventral view. —j. Upper palea with lodicules and stamens. —k. Caryopsis, embryo side. —1. Caryopsis, hilum side. 158 Annals of the Missouri Botanical Garden ing erect. Panicum adenorhachis has, in relation to P. cabrerae, bigger, short-hispid spikelets, glands on the leaf blades and panicles, a lower glume less than Y, the length of the spikelet, and spikelets nonstipitate between the lower and upper glume. Panicum congestum differs from P. cabrerae, P. d pe: and P. adenorhachis by having spike- lets 2.8-3.2 mm long, with a lower glume Y, the bs of the spikelet, panicles partially included in the leaf sheaths and few-flowered, with only two or three spikelets per panicle. The holotype of P. cabrerae, Mori et al. 12475 (CEPEC), was treated as P. stipiflorum by Ren- voize (1984). Panicum cabrerae grows in a habitat similar to that of P. adenorhachis. We name this species in honor of Angel L. Cabrera, major professor of the senior author, who is especially grateful for the training in systematics provided by Prof. Cabrera. Panicum davidsei Zuloaga & Morrone, sp. nov. TYPE: Venezuela. Bolivar: Сађапауеп, La Gran Sabana, wet inundated savanna, 1,300 m, 3 Dec. 1973, Davidse, M. Ramia & R. Montes 4796 (holotype: МО). Figure 4. Panicum assurgens Renvoize affine sed spiculis stipi- tatis, gluma superiore 12-14-nervia et lemma inferiore 10-12-nervia; lemma superiore conspicue apiculata cum marginibus superioribus membranaceis differ Plants perennial. Culms 0.50-1.80 m long, de- cumbent, rooting and branching at the lower nodes, the upper portions erect; internodes 8-22 cm long, cylindric, striate, hispid, compressed; nodes hir- sute. Sheaths 5-8 cm long, sparsely hirsute with long tuberculate hairs, one of the margins long- ciliate, the other membranous; collar long-pilose with whitish hairs; ligule membranous-ciliate, ca. 0.2 mm long; blades narrowly lanceolate, 9-13 cm long, 1-1.3 cm wide, flat, subcordate, the apex acuminate, adaxial surfaces long-pilose toward the base, otherwise glabrous, abaxial surfaces sparsely pilose, the margins scabrous and ciliate toward the base, the midnerve conspicuous. Panicles lax, dif- fuse, 12-20 cm long, 7-12 cm wide; peduncle 10-40 cm long; rachis cylindric, striate, pilose near the branches, otherwise glabrous; lower branches whorled, the upper ones subopposite or alternate, the axes of the branches flexuous, glabrous, the axils of the branches pilose; pedicels glabrous, wavy. Spikelets solitary, ellipsoid, 3-3.3 mm long, 1 mm wide, scaberulous, attenuate toward the base and with a small stipe between the lower and upper glume; lower glume ovate-acuminate, 1.5-1.8 mm long, nearly У, the length of the spikelet, 3-nerved; upper glume 2.4-2.7 mm long, not covering the apex of the upper anthecium, 11–14-пегуед; up- per glume and lower lemma subequal with manifest nerves anastomosed toward the apex; lower lemma ca. 2.7 mm long, 10-12-nerved; lower palea lan- ceolate, 2.1-2.4 mm long, 0.3 mm wide, glabrous, hyaline; lower flower absent; upper anthecium el- lipsoid, 2.4-2.7 mm long, 1 mm wide, papillose, apiculate; upper lemma with a green, scabrous crest on the apex and the upper margins mem- branous, prolonged toward the apex as small wings; rachilla prolonged beyond the upper anthecium as a small mucro; lodicules 2, conduplicate, truncate, ca. 0.4 mm long; stamens 3, the anthers ca. 1.2 mm long; stigmas plumose. Caryopsis ellipsoid, 1.8 mm long; 0.9 cm wide; hilum punctiform; embryo И the length of the caryopsis. Paratypes. BRAZIL. RORAIMA: aldeia do Tuchana Rondón s.n., July 1927 (RB 110786). Шош BOLÍVAR: entre Piedra de La Virgen y la parte alta de La Escalera, carretera a La Gran Sabana, 6°0’М, 61°2 1,300 m, 12 Ago. 1989, Zuloaga et al. 4406 (MO, SI, VEN); a 1 km al sur de Puente Sakai 1,200 m, 13 Ago. 1989, Zuloaga et al. 4427 (MO, SI, EN). This species is allied to P. assurgens Renvoize. Panicum assurgens differs from P. davidsei by the nonstipitate spikelets, the 9-nerved upper glume and lower lemma, the lower glume %—% as long as the spikelet, and the apiculate upper lemma with indurate, not membranous upper margins. The large number of nerves on the upper glume and lower lemma of this new species suggest a relationship to P. itatiaiae Swallen, a Brazilian species described from Rio de Janeiro with spikelets 4.6-4.9 mm long, 1.6-1.7 mm wide. Panicum davidsei has spikelets 3-3.3 mm long, 1 mm wide. Panicum davidsei grows in savannas of the Gran Sabana, Venezuela, and Brazil, Roraima, at the edge of the forests, together with plants of P. pycnoclados. We name this species in honor of its collector, Gerrit Davidse, renowned agrostologist of the Mis- souri Botanical Garden. Panicum cucaense Zuloaga & Morrone, sp. nov. TYPE: Brazil. Rio de Janeiro: Mun. Petrópolis, Morro do Cuca, entre Vale dos Videiras e Araras, campo de altitude, 1,600 m, 27 Jan. 83, G. Martinelli & E. Simonis 9011 (ho- lotype, RB; isotypes, MO, SI). Figure 5. Panicum petropolitanum affine sed caespitosum, bre- viter rhizomatosum, vaginis conspicue distichis, foliis in- volutis et paniculis breviter pedunculatis differt. Volume 78, Number 1 Zuloaga & Morrone 159 1991 Panicum from South America FIGURE 4. Holotype x Panicum davidsei. —a. Habit. —b. Spikelet, ventral view. — с. Spikelet, dorsal view. — d. Upper glume.—e. Lower lemma.—f. Lower palea. —g. Upper anthecium, dorsal view.—h. Upper e ventral view. —1. Upper tip “of the lemma with apicule and membranous upper margins. —j. Caryopsis, embryo si . Caryopsis, hilum side. 160 Annals of the Missouri Botanical Garden FIGURE 5. Holotype of a cucaense. —a. Habit. — b. Spikelet, lateral view.—c. Detail of base of spikelet. — Ж Spikelet, iin view. —e. Spikelet, енер мем. —f. Lower раеа.— р. Upper anthecium, dorsal view. —ћ. Upper nthecium, ventral view. —i. Upper palea.—j. Upper palea with lodicules and кен rs. —k. Upper palea.—1. Upper ales with lodicules ad anthers. — m. Caryopsis, embryo side. —n. Caryopsis, hilum side. Volume 78, Number 1 Zuloaga & Morrone Panicum from South America Cespitose perennial, shortly rhizomatous. Culms 0-30 cm tall, many-noded, erect to geniculate- ascending and rooting at the lower nodes, freely branching; internodes 1-4 cm long, cylindric, stri- ate, glabrous; nodes compressed, glabrous, brown- ish. Sheaths striate, ca. 1.2 cm long, longer than the internodes, strongly distichous, long-pilose at the distal portion, otherwise glabrous, one margin pilose with long, whitish hairs, the other membra- nous; ligule a ciliate membrane, the membrane ca. .1 mm long, the cilia 0.4 mm long; blades linear, 1.5-5.5 ст long, 0.1–0.2 cm wide, folded, atten- uate toward the apex, pilose adaxially toward the ligule, otherwise glabrous, the borders scabrous. Primary panicles lax, diffuse, few-flowered, 2-3.5 cm long, 1.5-3.5 cm wide; peduncles to 6 cm long, glabrous; rachis striate, wavy, glabrous, with some glands; branches alternate, divergent and widely spaced, glabrous, the axils of the branches pilose; pedicels glabrous, 1-6 mm long. Secondary pan- icles similar to the primary one. Spikelets solitary, ellipsoid, 2.2-2.7 mm long, 0.8-0.9 mm wide, open at maturity, glabrous; lower glume lanceolate, 1.5-1.9 mm long, %-7, the length of the spikelet, acuminate, 1(-3)-nerved, a small stipe present be- tween the lower and upper glume; upper glume and lower lemma subequal, 7-9-nerved, acute; lower palea lanceolate, 1.7-1.9 mm long, 0.4-0.7 mm wide, hyaline, glabrous; lower flower absent; upper anthecium long-ellipsoid, 2.1-2.3 mm long, 0.7-0.8 mm wide, papillose, with simple papillae evenly distributed, apiculate, the apex scabrous, otherwise glabrous; lodicules 2, cuneiform, trun- cate, ca. 0.2 mm long; stamens 3, the anthers 1.2 mm long; stigmas Pues purplish. Caryopsis el- lipsoid, brownish, 1.4 mm long, 0.7 mm wide; hilum oblong; embryo less than У, the length of the cary- opsis. Paratypes. BRAZIL. RIO DE JANEIRO: Teresopolis, Ser- ra dos Orgàos, Pedra do Sino, 2,100-2,170 m, Feb. 1953, J. Vidal 11-6467, 11-6485 (В, SI). Panicum cucaense is related to P. petropoli- tanum but is distinguished by its cespitose habit; erect, many-noded culms; strongly distichous leaf sheaths; stiff and folded leaf blades; and shortly exserted panicles. Panicum petropolitanum differs from P. cu- caense in its stoloniferous habit; culms geniculate, rooting and branching at the lower nodes, then becoming erect, few-noded; leaf blades linear-lan- ceolate, flat; and inflorescence with a peduncle to 16 cm lon Panicum cucaense inhabits open and dry hab- itats in rocky, granitic soils of high-altitude campos of the mountains in eastern Brazil. Eiten (1978) indicated that these rocky areas over 2,000-2,500 m (planaltos), found on mountains such as Itatiaia or the Serra dos Orgaos and with frosts occurring at night in the cold season, should be distinguished from the campos rupestres, which occur from Ba- hia to the state of Sào Paulo at lower elevations. Panicum petropolitanum Zuloaga & Morrone, sp. nov. TYPE: Brazil. Rio de Janeiro: Mun. Petrópolis, Araras, base da Pedra Maria Com- prida, saxicola, heliofila, crescendo em beira de rio, 23 Mar. 1968, D. Sucre & P. J. Braga 2553 (holotype, RB). Figure 6. Panicum cucaense affine sed stoloniferum, vaginis non conspicue distichis, foliis planis et paniculis longe pedun- culatis differt Plants perennial?, stoloniferous. Culms genicu- late, rooting and branching at the lower nodes, then becoming erect, freely branching, 10-25 cm tall, few-noded; internodes 1—4 cm long, cylindric to compressed, hollow, glabrous; nodes com- pressed, brownish, covered with long hairs or gla- brous. Sheaths striate, 1.5-5.5 cm long, longer than the internodes, one of the margins long-ciliate with whitish hairs, the other membranous or with long hairs near the ligule, glabrous on the rest of the surface; collar brownish, glabrous; ligules mem- branous-ciliate, ca. 0.2 mm long; blades linear- lanceolate, 2-7 cm long, 0.1-0.3 cm wide, flat, glabrous, the margins scaberulous, the apex atten- uate. Panicles lax, diffuse, 2-6 cm long, 1–3.5 cm wide; peduncles to 16 cm long; rachis wavy, cylindric, glabrous, the branches alternate, diver- gent and distant, the axes of the branches glabrous; axils of the branches glabrous; pedicels triquetrous, scaberulous, 1-7 mm long. Spikelets solitary, long- ellipsoid, 2.4-2.6 mm long, 0.9 mm wide, glabrous, without a stipe; lower glume 1.8-2 mm long, 7, the length of the spikelet, lanceolate, 1(-3)-nerved, the midnerve scabrous toward the apex; upper glume (7-)9-nerved, acute, embracing the lower lemma; upper glume and lower lemma subequal; lower lemma 9-nerved, embracing the upper an- thecium; lower palea lanceolate, 1.5 mm long, 0.3- 0.4 mm wide, hyaline, glabrous; lower flower ab- sent; upper anthecium long-ellipsoid, 2.3 mm long, 0.7 mm wide, glabrous, shiny, with simple papillae evenly distributed all over its surface; upper lemma apiculate, the apicule ca. 0.2 mm long, pilose; lodicules 2, truncate, ca. 0.3 mm long, embracing the lower margins of the palea; stamens 3, the anthers 1 mm long; stigmas plumose. Caryopsis ellipsoid, 1.2 mm long, 0.6 mm wide; hilum punc- 162 Annals of the Missouri Botanical Garden FiGURE 6. Holotype: of аек petropolitanum. — а. Нађи. —b. Spikelet, ventral уіем. — с. Spikelet, dorsal view. — d. Lower palea nthecium, dorsal view. —f. Upper anthecium, ventral view. — р. Caryopsis, embryo side. —h. Caryopsis, hilum side. tiform; embryo less than У, the length of the caryop- borders of streams. See the discussion under P. SIS. cucaense. . | : . Distincti 0 ecies a Panicum petropolitanum is a delicate plant that EE € ng the че а bá s and related ab od : : ones are summarized in the following key: grows on rocky soils in open and humid places in A. Upper glume and lower lemma 11-14-nerved; panicles 12-20 cm long. B. Spikelets 4.6-4.9 mm long, 1.6-1.7 mm wide; lower glume 5-nerved P. itatiaiae BB. Spikelets 3-3.3 mm long, 1 mm wide; lower glume 3-nerved P. davidsei Zuloaga & Morrone Volume 78, Number 1 1991 Panicum from South America AA. Upper glume and lower lemma 7-9-nerved; panicles 1-16 cm C. Leaf blades ovate-lanceolate to lanceolate, cordate, amplexicaulous, length : q ratio 3-5: P. cumbucana D. Panicles few-flowered, 1-2.5 cm long; leaf blades 1-2 cm long, 0.3-0.6 с ا‎ DD. Panicles multiflowered, 3-10 cm long; leaf blades 2-12 cm long, 0.5-1. ва eaf blades with the margins long-ciliate, cartilaginous; lower glume to Y (occasionally %) кө length of the spikelet iflorum EE. = blades with the margins ciliate only at the base, otherwise scabrous, not cartilaginous; e Ию Y, the length of the spike let lower glum . py CC. Leaf blades ilia to dar lanceolate (lanceolate in P. assurgens), narrowed to Pandai: not — length : width ratio 12-20: F. Plan cnoclados wit y a glan: t on s PRAE blades and axes of the inflorescences; lower glume less than Y, the length of the FF. Plants without glands oceans pres cucaense), lower glume ' G. Spikelets 2-2.3 m adenorhachis P. few glands on the axes of the los of Е y, the length of the spikelet. m long; lower glume to У, the length of the spikelet _________ - P. cabrerae GG. Spikelets 2.4-3.3 mm long; lower glume %-% the length of the spikelet. pikelets with a conspicuous "p aig en the lower and upper I. Spikelets glabrous, II. лея hirsute, scicles at the top of the culms, de ale branc 2.2-2.1 evenly distributed slong the culm glum E lower glume 1(-3)- сенн) leaf pid ucaense mm lon g; lower glume 5-7-nerved; leaf blades i in ед... Р. сопдезїит НН. Spikelets манал а stipe between the lower and upper glume. J. Plan bs culms 10-25 cm tall; blades 2-7 cm long, 0.1–0.3 cm к pus cles 2) cm long, 0.3 1) с cm wide cum mm lon cm tall; blades 4-12 cm long, 0.3-1 cm m long; spikelets me to glabrous, 2.6-3.3 mm lon m long; leaf blades linear-lanceolate, rigid, eliophilum KK. Spikelets glabrous to puberulous between the nerves, 3.3 mm ne leaf blades lanceolate, herbaceous, 8-12 cm long, 0.7-1 cm wide _ 2- ong; spikelets glabrous, 2.4-2. JJ. Plants "with culms 200 panicles 4-16 K. Spikelets E 4-5(- LITERATURE CITED BRowN, W. V. . SMITH. 1975. The genus ر‎ E (Gramineas). Bull. Torrey Bot. Club 102(1): CLARK, С F. "w. Сошо. 1975. Some epidermal characteristics of paleas of r, Pani- m, and Echinochloa. Amer. J. Bot. 62: 743-748. йр. С. A systematic study of the genus Lasiacis (Gramineae: Paniceae). Ann. Missouri Bot. Gard. 65: 1133-1254. А sketch of the е of Central razil. ue dos trabalhos: [Abstract - paper hn at the II ee eso Latino-Am cano de Botanica, 22-29 Jan. 1978. Brasilia, Brazil. GOULD, Е omenclatural changes in the Poaceae. Brittonia 26: 59-6 e Mexican "s of Dichanthelium кеч Brittonia 32(2): 3 364. 8 . CLARK. ur о асеае) in m United States and Canada. Ann. Mis- souri Bot. Gard. 65: 1088-1132. HANSEN, B. F. UNDERLIN. 1988. Synopsis of Dichanthelium (Poaceae) i in Florida. Ann. Mis- souri Bot. Gar 6 657 1910. “The North Amer- 1. Herb. HITCHCOCK, А. S. & " Teen ican species of Panicum. Contr. U.S. Nat : 96. Tropical North American 1915. species of Panicum. Contr, U.S. Natl. Herb. 17: 39. P. assurgens Hsu, C. C. 1965. The classification of Panicum (Grami- neae) and its allies, with special reference to the characters of lodicule, style-base an © lemma. J. Fac. Sci. Univ. Tokyo Sect. 3, Bot. 9 0. LELONG, M. 1984. New i adl а Ponui sub- genus Panicum and subgenus Dichanthelium (Po- е ~ the southeastern United States. Brittonia 36: -273. PILGER, a Gramineae. In: A. Engler, Die Na- türlichen Planeta, 2nd edition, 14e: 8-25. Рони, К. W. . Gramineae. /n: W. Burger (editor), Flora quw Fieldiana, Bot. n. ser. 4: 1-608 RENVOIZE, S. The Grasses of Bahia. Royal Botanic iR Kew. WEBSTER, В. D. 1988. Genera of the North American Pani (P Panicoideae). Syst. Bot. 13: 576- 609. ZULOAGA, F. O. 1987. Systematics of the New World 306 in T. R. Soderst & M. E. Barkworth (editors), Grass Systematics and Evolution. Smithsonian Institution Press, Washing ton, D.C. . R. SODERSTROM. 1985. — ‘a the outlying species of New World Pani aceae: Paniceae). Smithsonian it Bot. 3. NEW SPECIES OF CHUSQUEA (POACEAE: BAMBUSOIDEAE) FROM COSTA RICA! Yvonne Widmer? and Lynn G. Clark? ABSTRACT ee new species of Chusquea from the upper montane forests of the Cordillera de Talamanca in Costa Rica are described and illustrated. Chusquea tomentosa and C. subtilis belong to Chusquea sect. Longifoliae L. G. Clark and are closely related to C. foliosa L. G. Clark. The third species, C. talamancensis, is a member of Ch usquea sect. Swallenochloa (McClure) L. G. Clark and shows similarities to both C. tonduzii Hackel and C. vulcanalis (Soderstrom & C. Calderón) L. С. Clark. Revised keys to the species of sect. Longifoliae and sect. Swallenochloa in Costa Rica are provided Species of the woody bamboo genus Chusquea Kunth are often important components of montane forests in the New World (Veblen et al., 1977; lark, 1989), but little ecological data relating to Chusquea is available. In Costa Rica, the upper montane forests along the Cordillera de Talamanca and volcanoes of the Cordillera Volcánica Central (northern slopes) are dominated by species of Quer- cus L., with a number of species of Chusquea found as elements in the understory or along the forest margins. A Swiss Forestry Project, in association with the Centro Agronómico Tropical de Investigación y Ensenanza (CATIE) in Turrialba, Costa Rica, is developing a management plan for these forests. As part of this project, the ecology of two species of Chusquea in these forests was studied (Widmer, in prep.). When the two species began to flower in 1987-1989, we were able to confirm that they were both undescribed. Material of a third unde- scribed species in flower was also collected during the ecological study. In this paper, we describe and illustrate the three new species, C. tomentosa, C. subtilis, and C. talamancensis. 'The first two species are members of Chusquea sect. Longifoliae, and the third be- longs to Chusquea sect. Swallenochloa. Revised keys to the species of these two sections in Costa Rica are included. KEY TO THE SPECIES OF CHUSQUEA SECT. LONGIFOLIAE IN CosTA RICA ns) (based on vegetative specimens la. Thin, curly leafless fibrillar branchlets quic with the ee кар subsidiary branches; internodes С. abrous; foliage leaf blades with the bas d blades with the base attenuate to rounded-atten unded to rounded-tru scabra Fibrillar branchlets absent; internodes ОНЫ uen rarely bos just below the nodes; foliage leaf 2a. Foliage leaf blades 0.6-1.3 cm wide; tuber) branches 18-30 per node; culm leaf i 0. 7-3 times as long as blades а ~ = sheaths 1.5-10.5(-14) times as long as bla . Foliage leaf blades 0.3-0.9(-1.1) cm wide; subsidiary branches 24-80(-150) per node; culm leaf 3a. Culm leaf sheaths abaxially scabrous to pilose, or sometimes glabrous; subsidiary branches 24- 65 per node; inner ligules of foliage leaves 0.5-4 mm long; foliage leaf blades 0.4-0.9 c m wide atens w o node; inner ligules of foliage leaves to C. p . Culm leaf sheaths abaxially glabrous to ies at the base; subsidiary branches 50-80(-150) per mm long; foliage leaf blades 0.3-0.7(-1.1) cm wide. 4a. dioec leaf blades abaxially a fimbriae at the apex of foliage leaf sheaths straight, . tomentosa few tuft 4b. Foliage leaf blades abaxially glabrous to sparsely pilose with scattered hairs; fimbriae at the ' Fieldwork was supported by the Swiss Directorate for Development Cooperation and Humanitarian Aid (Widmer) and the National Geographic Society (Clar ). » CATIE/PFAF, Apartado 89, 7170 Turrialba, Costa Rica. Present address: Geobotanisches Institut ETHZ, 8044 Zürich, Switzerla nd. * Department of Botany, Iowa State University, Ames, Iowa 50011, U.S.A. Address correspondence to this author. ANN. Missouni Bor. GARD. 78: 164-171. 1991. Volume 78, Number 1 1991 Widmer & Clark New Chusquea from Costa Rica 165 apex of foliage leaf sheaths often curly, several to many per t 9a. "Way leaf blades with the base attenuate, L: W — (26- Ка 48(- 54); culm leaf sheaths 4.4-6.6 times as long as the blades C. s 5b. Foliage leaf blades with the base rounded-attenuate, L: W — 1.5-5.4 times as long as the blades . foliosa sheaths 1 ubtilis 20-40(-45); culm leaf C. fol KEY TO THE SPECIES OF CHUSQUEA SECT. LONGIFOLIAE IN COSTA RICA (based on flowering specimens la. Primary branches of panicle strongly spreading C. patens lb. ied branches of panicles appressed or ascending. ql lemma П %—% the spikelet length; thin, Е leafless fibrillar branchlets interspersed with the ormal, leafy subsidiary branches; internodes scabro C. Sterile lemma II equalling or ише equalling the suelos length; fibrillar branchlets absent; internodes ко > „> y smooth, rarely scabrous jus per node; culm leaf sheaths 0.7 3 Spikelets 8.4-15 mm w c scabra e nodes. еј 10.4-20.6 mm p foliage leaf blades 0.6-1.3 cm wide; subsidiary branches hys 30 -3 times as long as the blades mm long; foliage leaf blades 0.3-0.7(-1.1) cm wide; subsidiary же 50- C. longifolia 80(-150) per node; culm leaf sheaths 1.5-10.5(-14) times as long as the blades 4a. Spikelets 11.5-15 mm long; sterile lemma I %-Y the spikelet length; foliage leaf blades abaxially tomentose . tomentosa ج c‏ . Spikelets 8.4-11.8 m mm long; sterile lemma I %-Y, the spikelet length; foliage leaf blades abaxially glabrous to y ena pilose with scattered hairs. Spikelets 8.4-11 гоа. L: :W = сл P wn ый attenuate; L: W = Chusquea tomentosa Widmer & L. G. Clark, sp. nov. TYPE: Costa Rica. Cartago: Villa Mills, road to Piedra Alta, 2,880 m, 30 May 1989 (fl), Clark, Widmer & Stein 500 (holotype, CR; isotypes, ISC, MO, NY, US). Figure 1А-Е Culmi 1.3-4 cm фат., 6-9 m alti. Folia culmorum 20.8-29 cm longa; vaginae 11.8-25.4 cm longae, 4- 10.5(-14) plo longior quam laminae, abaxiales glabrae; laminae 1.4-5.4 cm longae, erectae, adaxiales pubes- centes, abaxiales glabrae. Ramificatio ca Va- ginae foliorum fimbriatae, fimbriae paucae, 1-2 mm lon- gae, rectae; laminae foliorum 15-27 cm longa 0.7(-1.1) cm latae, L: W — 25-40, abaxiales tomento- , Squamiformes. Lemmata sterilia 2, subulata; үсү sterile I 7.2-9.7 mm longum; lemma pela 1111.7 mm longum. Lemma fertile 11.5-13.5 mm longum, Area Palea 9.5-12.1 mm longa, sul- cata. Culms 1.3-4 cm basal diam., 6-9 m tall, erect at the base, arching above and scandent. /nter- nodes (11-14-)18-38(-51) cm long, more or less terete, shallowly sulcate above the central bud, usually smooth to scabrid below the node. Сшт leaves 20.8-29 cm long, the juncture of the sheath and blade abaxially a + horizontal and distinct line; sheaths 11.8-25.4 cm long, 4-10.5(-14) times as long as the blades, abaxially glabrous, the mar- gins glabrous, one side occasionally minutely ciliate toward the apex; blades 1.4–5.4 cm long, trian- m long; fertile lemma subulate; foliage leaf blades with the base 5 . foliosa Spikelets 9.8-11.8 mm long; fertile lemma aristate; foliage leaf blades with the base (26-)30-48(-54) C. subtilis gular, erect, usually persistent, adaxially pubes- cent, abaxially glabrous, the apex subulate, the base narrower than the apex of the sheath; girdle 3-5 mm wide, densely pubescent; inner ligule 1— 4 mm long, apically ciliolate. Nodes with one tri- angular central bud subtended by 50-60 subsidiary buds; sheath scar dipping away markedly below the bud/branch complement. Branching infra- vaginal; central bud developing tardily at the mid- dle and upper nodes; leafy subsidiary branches 18- 32 cm long, frequently rebranching basally, 60- 80(-150) in the mature complement. Foliage leaves 4-9 per complement; sheaths glabrous, the overlapping margin ciliate, the apex with a tuft of fimbriae on either side of the summit, the fimbriae few per tuft, 1-2 mm long, straight, eventually deciduous; blades 15-27 cm long, 0.3-0.7(-1.1 cm wide, L : W = 25-40, adaxially glabrous, some- times with some scattered hairs, abaxially tomen- tose and often glaucous, not tessellate, the apex short-setose, the base attenuate to rounded-atten- uate; pseudopetiole 1-1.5 mm long; outer ligule an irregular, glabrous rim to 0.5 mm long; inner ligule to 1 mm long, truncate to rounded, pubes- cent. Panicles 8-16 cm long, narrow, the base often retained within the subtending sheath; rachis triquetrous, glabrous, the edges scabrid; branches loosely appressed, angular, scabrid, the lower ones 3-4 cm long; pedicels 2-6 mm long, angular, scabrid. Spikelets 11.5-15 mm long, scabrid. Glumes 2, scalelike, usually acute, nerves lacking 166 Annals of the Missouri Botanical Garden | Sar Wah үч 273. ОРИ МИА МИ | А IS и UN FIGURE 1. Chusquea tomentosa and C. subtilis. A-E. C. іотепіоѕа. — А. Bud complement.— B. Panicle with subtending foliage leaves. — C. Spikelet. — D. Culm leaf, abaxial view. — E. Apex of foliage leaf sheath, showing inner and outer ligules, fimbriae, and pseudopetiole. (A, D based on Clark & Clark 274; B, E based on Widmer 502; C based on Clark et al. 500.) F-H. C. subtilis. —F. Panicle with subtending foliage leaf. — С. Spikelet. —H. Apex of foliage leaf sheath, showing inner and outer ligules, fimbriae, and pseudopetiole. (F-H based on Widmer 508.) A, B, D, F: ; C, E, G, H: bar equal to 1 mm. Volume 78, Number 1 1991 Widmer & Clark 167 New Chusquea from Costa Rica TABLE 1. A morphological comparison of Chusquea foliosa, C. subtilis, and C. tomentosa. Character C. foliosa C. subtilis C. tomentosa Culm leaf sheath : blade 1.5-5.4 4.4–6. 4-10.5(-14) Foliage leaf sheath apex fimbriae often curly, 1-2 fimbriae often curly, 1 fimbriae straight, 1-2 mm mm long, many /tuft 1.5 mm long, many/ long, few/tuft tuft 20-40(-45) (26-)30-48(-54) 25-40 Foliage leaf length : width Foliage leaf base rounded-attenuate Foliage leaf abaxial pu- glabrous to sparsely pi- bescence mi" Spikelet length 4-11.2 mm Sterile lemma I/spikelet v Y) Fertile lemma apex subulate attenuate attenuate to rounded-atten- uate glabrous to sparsely pi- tomentose lose 9.8-11.8 mm 11.5-15 mm Y %-% aristate subulate or 1-пегуед; glume I ca. и; the spikelet length, 0.5-1.3 mm long; glume II ca. У, the spikelet length, 0.7-1.5 mm long. Sterile lemmas 2, su- bulate, abaxially scabrid; sterile lemma I %-Y, the spikelet length, 7.2-9.7 mm long, 5- ог 7-nerved; sterile lemma II extending the full spike- let length, 11.7-14 mm long, 5-, 7-, or 9-nerved. Fertile lemma 11.5-13.5 mm long, subulate, abaxially scabrid, 7—9-nerved. Palea 9.5-12.1 mm long, sulcate for most of the length, apiculate, scabrid, 6-8-nerved. Stamens 3; anthers 6-8 mm long. Fruit unknown. Additional specimens examined. Costa RICA. CAR- TAGO: S slope of Volcán Turrialba, 1-2 km E of Hacienda Central, Pohl & Davidse 10867B (ISC); Villa Mills, at Quebrada Voltea, Widmer 500 (CR, ISC), 501 (CR, 150); on the path to Cerros Cuerici, Widmer 502, 503, 504, 506 & s.n. (CR, ha La Esperanza del Guarco, Km 61.8 on the Carretera Interamericana, 4 km NE, 19 Jan. 1990 (8), Widmer 505 (CR, ISC). SAN JOSÉ: along the Carretera Interamericana hersmen Km 77 78, Clark & Clark 274 (ISC, MO, US); Cordillera de Talamanca, Km post 107, ca. 20 km N of San Isidro del General along the Carretera Interamericana, Davidse 761 (ISC); Trinidad de Dota, Km 62.5 along the apice Interame- ricana, 6 Feb. 1990 (fl), Widmer 507 (CR, ISC). Chusquea tomentosa is characterized by the abaxially tomentose foliage leaf blades and the sca- brid spikelets 11.5-15 mm long. This species oc- curs in the Cordillera de Talamanca in montane forests at elevations of 2,500 to 3,000 m. It grows on slightly drier sites with southern to southeastern exposures. Vegetative specimens of C. tomentosa were pre- viously assigned to C. foliosa L. G. Clark (Clark, 1989). With the advent of flowering in 19 tomentosa could be clearly distinguished from C. foliosa based on the large. spikelets. Vegetatively, the two species are very similar in overall aspect, but may be differentiated by the abaxially tomen- tose foliage leaf blades and the few, straight fim- briae at the apex of the foliage leaf sheaths of C. tomentosa. In C. foliosa, the foliage leaf blades are abaxially glabrous to sparsely pilose with scat- tered hairs, and the fimbriae are more numerous and usually curly. Chusquea foliosa, C. tomentosa, and C. subtilis form a complex of closely related species within sect. Longifoliae; a comparison of the three species is presented in Table 1. Chusquea subtilis Widmer & L. G. Clark, sp. nov. TYPE: Costa Rica. Cartago: Villa Mills, El Sitio, confluence of Quebradas Siberia and Voltea, 2,550 m, 9 Dec. 1988 (fl), Widmer x (holotype, CR; isotypes, CATIE/SBN, C, ZT). Figure 1F-H. Culmi 1.5-3 cm diam., 2-6 m alti. Folia culmorum inae foliorum (9-)1 2- А L:W = (26-)30- "a 54), е p sellatae. Paniculae (7-)11-1 .2 mm longum, subulatum; lemma sterile II 9.3-11.8 mm longum, subulatum-aristatum fer tile 10-11.7 mm lon ови | aristatum. Palea 7.4-9 mm longa, sulcata tantum ad a Culms to 1.5-3 cm diam., 2-6 m tall, erect at the base, arching above. /nternodes to 35 cm long, terete, shallowly sulcate above the central bud, usually scabrid to smooth. Culm leaves 19.7-24.3 ст long, the juncture of the sheath and blade abaxially a + horizontal and distinct line; sheaths 17-20 cm long, 4.4-6.6 times as long as the blades, abaxially pilose toward the base, glabrous toward the apex; blades 2.6-4.5 cm long, trian- gular, erect, persistent, adaxially antrorsely pu- 168 Annals of the Missouri Botanical Garden bescent between the nerves, abaxially sparsely pi- lose, the apex subulate, the base narrower than the apex of the sheath; girdle 0.5-1 cm wide, densely pilose; inner ligule 1-2 mm long, ciliolate. Nodes with one triangular central bud; sheath scar dipping away markedly below the bud/branch complement. Branching infravaginal; central bud developing tardily at the middle and upper nodes; leafy sub- sidiary branches to 34 cm long, occasionally re- branching basally, 70-80(-100) branches in the mature complement. Foliage leaves 3-5 per com- plement; sheaths glabrous, the overlapping margin ciliate, the apex with a tuft of fimbriae on either side of the summit, the fimbriae numerous per tuft, 1-1.5 mm long, usually curly; blades (9-)12-21.6 cm long, 0.3-0.6 cm wide, L: W = (26-)30-48(- 24), adaxially usually glabrous, sometimes with scattered hairs near the midrib, abaxially glabrous, sometimes with scattered hairs, not tessellate, the apex short setose, the base attenuate; pseudopetiole 1-2 mm long, not distinct; outer ligu glabrous rim О m long; inner ligule to 1 e a minute, mm long, rounded to шл. abaxially basally pubescent. Panicles (7-)11-18 cm long, narrow, fully exserted from the subtending sheath; rachis triquetrous, glabrous, the edges glabrous to scabrid; branches loosely appressed, angular, glabrous, the lower ones 2-4.5 cm long; pedicels 1-7 mm long, angular, glabrous. Spikelets 9.8-11.8 mm long, scabrid. Glumes 2, scalelike, nerves lacking; glume Г ca. % the spikelet length, 0.5-0.8 mm long; glume II ca. У, the spikelet length, 0.6-1 mm long. Sterile lemmas 2, abaxially scabrid-pubes- cent, (5—)7-nerved; sterile lemma I 7;-?, the spike- let length, 7.5-9.2 mm long, subulate; sterile lem- ma П equalling or nearly equalling the spikelet length, 9.3-11.8 mm long, subulate-aristate. Fer- tile lemma 10-11.7 mm long, aristate, abaxially scabrid-pubescent, 7-nerved. Palea 7.4—9 mm long, sulcate only toward the apex, apiculate, scabrid, 4- or 6-nerved. Stamens 3; anthers 4.2-5.2 mm long. Fruit a caryopsis, 4.7-6 mm long, flattened, grooved along the hilum, the hilum dark reddish brown, the style base persistent. Additional specimens examined. COSTA RICA. CAR- TAGO: Cordillera de Talamanca, on the path to Cerros Cuerici, Clark, Widmer & Stein 503 (CR, ISC, MO, NY, US); Villa Mills, Quebrada Voltea, 9 Dec. 1988 (fl), Wid- mer 509 (СЕ, 150); 19 Aug. 1989 (8), Widmer 511 (CR, 150); 15 Nov. 1989 a Widmer 513 (CR, ISC); Cerro iet near Hio Angeles, 18 Jan. 1989 (fl), Widmer 510 (CR, ISC); Cuericí, 9 Nov. 1989 (fl), Widmer жч (СВ, 15С); Villa Mills, im Sitio, confluence of the Que- bradas Voltea and Siberia, 15 Nov. 1989 (fl), Widmer 514 (CR, ISC). $АМ x Villa Mills, S of La Georgina, 15 Dec. 1978 (8), Pohl & Gabel 13725 (ISC, MO). The specific epithet of C. subtilis refers to the fine leaves and delicate aspect of the panicles. This species is characterized by the narrow foliage leaf blades and scabrid-pubescent spikelets 9.8-11.8 mm long with the fertile lemma aristate (Table 1). Chusquea subtilis occurs in the Cordillera de Tala- manca in montane forests along streams or at hu- mid sites at elevations of 2,550 to 3,000 m. KEY TO THE SPECIES OF CHUSQUEA SECT. SWALLENOCHLOA IN CosTA RICA ns) (based on vegetative specimens la. Foliage leaf blades 0.4–0.6 cm wide C. paludicola lb. Foliage leaf blades 0.7-2.4 cm wide. 2a. | ligule of foliage leaves 1-7 cm long, taper Cu m leaf sheaths (1.5-)4.5-7 times as b d ; the blades, the blades adaxially glabrous; foliage С. a mistadensis leaf } blades usually ч green 3b. Culm leaf sheaths blades green -13 times as long as the blades, the blades adaxially pubescent; foliage leaf C. lo ongiligulata ко > Inner ligule of foliage leaves 0.5-4(-1 4a. Foliage leaf blades times as long as t the blades 5) mm long, truncate to taper 3-14.5 cm long; culm leaf blades 0.35- 1. 5 =. lone the sheaths (7-)12-27 subtessellata 4b. Foliage leaf ер a 5-)10-29 cm long; culm leaf blades 1-6.5 cm long, the sheaths a. 5-)2- 1 times as lon the blades ба. Foliage leaf blades with " W = 7-12(-17), the base rounded to rounded-truncate ..................... C . vulcanalis 5b. Foliage leaf blades with L: W — Inner ligules of foliage leaves 1-2.5 mm y when young; culm leaf blades adaxially scabrid 6b. Inner ligules of foliage leaves (2-)3-15 mm ы usually tapering; foliage leaf sheaths not farinose; culm leaf blades adaxially pubescen Са often farinose, езреста (7-)10-20(-24), the base rounded to attenuate. long, truncate to rounded; foliage leaf sheaths C. talamancensis onduzii KEY TO THE SPECIES OF CHUSQUEA SECT. SWALLENOCHLOA IN CosTA RICA ) (based on vegetative and flowering specimens la. Foliage leaf blades 0.4–0.6 cm wide C. paludicola lb. Foliage leaf blades 0.7-2.4 cm wide. Volume 78, Number 1 1991 Widmer & Clark 169 New Chusquea from Costa Rica 2a. Panicles open, pyramidal, with stiffly spreading branches гр. Panicles + congested, often spicate, narrow, with app 3a. Spike elets glabrous, often slightly falcate; inner ligules of fo des 14-33 cm long, 0.7-2.3 ст 5-13 times as long as the blades, the blades adaxially pubescen 4b. Foliage leaf blades 6-20.5 cm long, 0.7-1. . Foliage “ bla C. tonduzii ressed or ascending branches liage leaves 1-7 cm long, tapering. iai green; culm leaf sheaths ongiligulata -16), parra wide, L: W = 4(-1.8) cm wide, L: ‘We -14 yellowish green; culm leaf sheaths (1.5-)4.5-7 times as long as the blades, the xs a w c the blades glabrous . Spikelets with some degree of pubescence, linear; inner ligules of foliage leaves 0.5-4(-7.5 Я тис е, the branches appressed; foliage leaf blades 3-14.5 cm long, 0. js m wide; culm leaf blades 0.35-1.5 cm long, the sheaths (7-)12-27 times as long C. hen bem 5b. Panicles somewhat narrow but not spicate, the branches vs and ascending but not tightly фер». foliage leaf blades (5.5-)10-29 cm long, 1-2.4 .5 с m wide; сит leaf blades 1.5- ong, the sheaths (1.5-)2-11 times as long as the bla det. 6a. Foliage leaf blades with L: W = 7-12(-17), the od rounded to rounded-truncate, the sheaths not farinose; spikelets completely pubesc -)10-20(-24), ne base rounded to attenuate- куйда. 6b. rapt leaf blades with L: W = (8 vulcanalis e sheaths often farinose, bid when young; spikelets scabrid-pubescent except С чине the bases of the bra talamancensis Chusquea talamancensis Widmer & L. C. Clark, sp. nov. TYPE: Costa Rica. Cartago: Cordillera de Talamanca, on the path to Cerros Cuerici, 3,030 m, 30 May 1989 (fl), Clark, Widmer & Stein 502 (holotype, CR; isotypes, ISC, MO, NY, US). Figure 2A-F Culmi 0.7-3.8 cm diam., 3-6(-9) m alti. Folia cul- rya 15-24 cm longa, farinosae initio; vaginae 10.6- cm longae, (1.5-)2-11 plo longior quam laminae, E glabrae; laminae 1.5-6.5 cm longae, erectae, adaxiales scaberulae, abaxiales glabrae. Ramificatio intra- vaginalis. Vaginae foliorum сте farinosae initio; laminae foliorum 13-29 cm longae, 1-1.8 cm latae, L: W = (8-)10-20(-24), glabrae, bd tessellatae. Pan- Q E 3 ge Ф сре ее Lemmata sterilia 2, subulata; lemma sterile I m longum; lemma sterile II 5.3-6.3 mm lon- gum. eda fertile 6-7.1 mm pis nt жоне Рајеа 5.6-6.2 mm longa, sulcata ad apic Culms 0.7-3.8 cm basal diam., 3-6(-9) m tall, erect at the base, arching above. /nternodes (17— 19)27-37(-43) cm long, += terete, smooth, waxy to farinose. Culm leaves 15-24 cm long, the junc- ture of the sheath and blade an irregular, slightly sloping line, abaxially indistinct, farinose when young; sheaths 10.6-18 cm long, (1.5-)2-11 times as long as the blades, abaxially glabrous; blades 1.5-6.5 cm long, erect, persistent, apiculate, adax- ially scabrid, abaxially glabrous; girdle 1-6 mm wide, glabrous; corky ridge present at the juncture of the sheath and girdle; inner ligule 1-2 mm long, stiff, glabrous, irregular. Nodes with one triangular central bud subtended by 5-11 subsidiary buds; sheath scar dipping slightly below the branch com- plement. Branching intravaginal; leafy subsidiary branches 35-69 cm long, frequently rebranching basally; 7-30 branches in the mature complement. Foliage leaves 6-11 per complement; sheaths gla- brous, often farinose о when young; blades 13-29 cm long, 1-1.8(-2.4) cm wide, L: = (8-)10-20(-24), adaxially and abaxially glabrous, abaxially tessellate, the apex acuminate to short the base rounded to attenuate-rounded; pseudopetiole 2-4 mm long; outer ligule a stiff, glabrous rim 0.3-1 mm long; inner ligule 1-2.5 mm long, truncate to rounded, glabrous, charta- ceous. Panicles 11-29 cm spicate, often the base not exserted from the sub- tending sheath; rachis somewhat complanate, gla- brous to scabrid, the edges scabrid; branches loose- ly appressed, ascending, angular, scabrous, the lower ones 3-9 cm long; pedicels 1.5-4 mm long, angular, scabrid. Spikelets 5.6-7.5 mm long. Glumes 2, scalelike, abaxially scabrid to pubescent; glume I са. У, the spikelet length, 0.6-1 mm long, erves absent or l-nerved; glume П са. 4 the spikelet length, 0.9-1.6 mm long, 1-nerved. Ster- ile lemmas 2, subulate, abaxially pubescent on the upper YY, otherwise E sterile lemma I ca. 7, the spikelet length, -5 mm long, 3- or 5-nerved; sterile lemma II ca. 7, the spikelet length, 5.3-6.3 mm long, 5-nerved. Fertile lemma 6-1. mm long, apiculate, abaxially pubescent on the upper /-'4, otherwise glabrous, 7- or 9-nerved. Palea 5.6-6.2 mm long, sulcate toward the apex, the sulcus pubescent, 4-nerved. Stamens not seen. setose, long, narrow but not Fruit a caryopsis. Additional specimens examined. COSTA RICA. CAR- TAGO: along the Carretera Interamericana just below pára- mo, Clark & Clark 273 (ISC, MO, US); Villa Mills, road to Piedra Alta, Clark et al. 501 (CR, ISC, MO, NY, US); along Interamerican Highway near Cerro de la Muerte, Fisher R354 (US). CARTAGO/SAN JOSE: Cerro Jaboncillo, along access road to transmission line tower on summit, 170 Annals of the Missouri Botanical Garden FIGURE 2. Chusquea talamancensis and C. tonduzii. A-F. C. talamancensis. — A. Culm leaf, abaxial view. — B. Branch complement and foliage leaves. — C. Bud/branch complement. — D. Panicle. — E. Spikelet. — Е. Caryopsis. (A-C based on Clark & Clark 273; D based on Pohl 15676; E based on Clark et al. 502; F based on Widmer 5.п.)—С. C. tonduzii, culm leaf, abaxial view (based on Clark et al. 497). A-D, С: bar equal to 1 cm; Е-Е; bar equal to 1 mm. Volume 78, Number 1 1991 Widmer & Clark 171 New Chusquea from Costa Rica TABLE 2. A morphological comparison of Chusquea vulcanalis, C. talamancensis, and C. tonduzii. Character C. vulcanalis C. talamancensis C. tonduzii Culm leaf sheath, abaxial Culm leaf blade, adaxial Foliage leaf blade color Foliage leaf blade width Foliage leaf length : width glabrous glabrous to scabrid yellowish green .1-)1.1-2.4 cm 7-12(-17) 1-4(-7.5) mm Foliage leaf inner ligule length Foliage leaf sheath not farinose Panicle shape Spikelet length Spikelet pubescence narrow 5.7-8.5 mm completely pubescent A (5,-)1 Sterile lemma I/spikelet Sterile lemma II/spikelet glabrous only basal Y2 pubescent to completely glabrous scabrid densely pubescent green (0.7-)1-1.5 cm (7-)10-19.5 (2-)3-15 mm green 1-1.8(-2.4) cm (8-)10-20(-24) 5 mm often farinose when not farinose ung pyramidal (5.4-)6- 7.6 mm glabrous except bracts abaxially apically pubes- bases of bracts cent 2, scabrous/pubescent ex- pt glabrous toward Cerro de la Muerte, Horn 129, 130 (ISC); power lane 1 km W of Cerro Jaboncillo, Cerro de la Muerte, Horn 242 (ISC). LIMON: on the path to Cerro Сћипро, 2 Apr. 1989 (8), Widmer 517 (ISC). PUNTARENAs: Cordillera de Tal- manca, upper slopes of Cerro Echandi, Davidse et al. 23885, 23961 (ISC, MO). saw JOSÉ: La Georgina, road- side W of Carretera Interamericana, of restaurant, forest margin, 28 Jan. 1989 (fl), Pohl 15676 (ISC, MO); La Georgina, Pohl & Clark 13924 (ISC, MO); La Geor- gina, opposite the restaurant, Pohl & Clark 14627 (150). Chusquea talamancensis is characterized by the often farinose foliage leaf sheaths, green foliage leaf blades with the inner ligules 1-2.5 mm long, and pubescent spikelets with sterile lemma II 7, the spikelet length. As indicated by the specific epithet, this species is found in the Cordillera de Talamanca, in upper montane forests at elevations of 2,600 to 3,200 m, usually on northern or northwestern ex- posures. Vegetative specimens of C. talamancensis were previously assigned to C. tonduzii Hackel (Clark, 1989), even though the inner ligules were much shorter than is typical for C. tonduzii. The panicle and spikelets of C. talamancensis are very similar to those of C. vulcanalis, from which it is vege- tatively distinct. The three species are compared in Table he description of the culm leaves of C. tonduzii in Clark (1989) is based on specimens of C. talamancensis. А complete recent vegetative collection of C. tonduzii (Clark et al. 497) showed that its culm leaves are very similar to those of C. paludicola L. С. Clark (Fig. 26), but the blades of C. tonduzii are adaxially pubescent as in longiligulata (Soderstrom & Calderón) L. G. Clark. LITERATURE CITED CLARK, L. G. 19 Systematics of Chusquea section Swallenochloa, section Verticillatae, section Ser- pentes, and section Longifoliae (Poaceae: Bambu- , D. HLEGEL & A. T. 1977. Distribution and Borninence, r spe- Ha duous VEBLEN. TT forest in south-central Chile. J. Ecol. 65: 15-830. PANAMANTHUS, А NEW MONOTYPIC GENUS OF NEOTROPICAL LORANTHACEAE! Job Кий? ABSTRACT The e monotypic genus врење са Kuijt (P. p date t Юлан eviously known as Strut s are provided. It is suggested that Pan genus җи ү in approximately the same ecological zone. Y due a panamensis it) E Barlow & Viens an expanded description and amanthus is related to Gaiadendron G. Don, another monotypic The mistletoe species known presently as Stru- thanthus panamensis (Rizz.) Barlow & Wiens was originally described as a member of the genus Phrygilanthus (Rizzini, 1960). As made clear by Barlow & Wiens (1973), Eichler’s genus Phry- gilanthus cannot be maintained for systematic and nomenclatural reasons. The only species of Phry- gilanthus that Rizzini mentioned as a possible rel- ative to his new species has subsequently been removed to Psittacanthus (Psittacanthus palmeri (S. Watson) Barlow & Wiens; Barlow & Wiens, 1973). Rizzini’s species was transferred to Stru- thanthus by Barlow & Wiens (1973), although these authors did recognize its unusual nature. There are two major morphological features in which Rizzini’s species differs from all other known species of Struthanthus. First, it has bisexual flow- ers, whereas the other species are strictly dioecious. Second, the inflorescence is consistently monadic, while that of the other species is basically or entirely triadic. Also, the prophyllar bracteoles associated with each flower are fused above the flower so as to hide the entire ovary, a feature not known in any other neotropical continental Loranthaceae. The first two features are else ribbean Dendropemon Blume, and strict logic would necessitate placement of Rizzini’s species in one о those genera. Of these, only the last genus needs to be considered, as various other structural char- acters prohibit placement in any of the first three. ndropemon often has pedunculate monads, and the bracteoles are sometimes fused in a cupulate fashion. The anthers of Rizzini’s species are nearly dorsifixed, but Dendropemon has unequivocally basifixed anthers. Also, the filaments of Dendro- pemon are laterally excavated, a feature shared with the continental Phthirusa but not with other genera. The stem roots of Rizzini’s species have no equivalent in Dendropemon, and it would be extraordinary to have a low-elevation, strictly Ca- ribbean genus represented by a single species at high elevations in Chiriqui. No one has suggested affinities to Dendropemon, however, which is clear- ly very closely related to Phthirusa and probably to Oryctanthus as well; the species would clearly be out of place there at least as much as it has been in Struthanthus. The disposition of S. pan- amensis has been a continual problem since its description, and I here propose that the species be treated as a monotypic genus with possible, but rather distant affinities to another monotypic genus found in the same ecological zone, i.e., Gaiaden- dron. Panamanthus Kuijt, gen. nov. TYPE: Panaman- thus panamensis (Rizz.) Kuijt, comb. nov. ee panamensis Rizz., Ann. Missouri Bot. Gard. 0. Sudan p FE ) Barlow & Wiens, Brit- onia 25: 39. 1973 WI МАР continuing financial support from the Natural Sciences and Engineering Research Council of Canada. Karel U. K ramer of Zürich kindly provided the Latin diagnosis. Several collectors have made special efforts in Chiriqui to collect this elusive species, and several curators have extended special courtesies; these efforts have adde d они substance to this presentation uri Botanical Garden; Mailing address: Department of Biology, University of Victoria, Victoria, British Columbia, Canada V8W 2 ANN. Missouni Вот. GARD. 78: 172-176. 1991. Volume 78, Number 1 Kuijt 173 Panamanthus Panamanthus panamensis. —a. Нађи of plant in fruit. —b. Inflorescence in bud. — c. Monad in bud EA). E 1. E van Myr Werff & Herrera 6330, LEA; b, c, McPherson 9380, arcae ien indeterminata, monadica, bracteis nu- with occasional epicortical roots formed from osis caducis suffulta; flores hermaphroditi, a branches; internodes to 6 cm long, somewhat quad- ше, cupulam formantes et ovarium occultantes. . rangular when young, soon becoming terete. Leaves Scandent plants to З m diam., with rather to 14 х 7 cm, decussate; petiole 1.5 cm long, straight, pendant branches 1 m or more in length, blade thin, with evident pinnate venation, base 174 Annals of the Missouri Botanical Garden FIGURE 2. Panamanthus panamensis. —a. Open flower, = two fused > visible below. —b. Longitudinal section of mature bud, with stamen shown separately, dorsal view.—c. Base of flower.—d. Mature fruit.—e. Young inflorescence bud still covered by s f. Base of а with old persistent scale leaves (a-c, McPherson 9380, LEA; d, Wilbur үч Teeri 13109, DUKE; e, f, Luteyn 3786, DUKE). Volume 78, Number 1 1991 Kuijt 175 Panamanthus truncate or very obtuse, apex contracted into con- spicuous, slender tail to 1.5 cm long. Inflorescences 2—4 cm at anthesis, somewhat quadrangular, race- mose, indeterminate, in small axillary clusters, in dividually subtended by several pairs of caducous scale leaves and set in a craterlike corky rim, inflorescence peduncle very short (2-3 mm), fol- lowed by 6-8 pairs of monads; monad peduncle 2-2.5 mm long, fused along its length with a bract extending 4 mm beyond and sharply acute; brac- teoles ca. 2 mm long, fused along the lower % of their length where investing the nearly hidden ova- ry, the free, acute tips 1 mm long; both bracteoles and bracts with conspicuous whitish margins when dry; infrutescence elongating to 6-9 cm at ma- turity. Flowers hexamerous, pale yellow, with pleas- ant odor, the petals very slightly dimorphic; mature bud 8-9 mm long, thickest (2 mm) just above the middle, with prominently acute tip; ovary short (1 mm), calyculus inconspicuous, with slightly undulating rim; anthers dimorphic, basifixed or nearly so, lower ones reaching to top of stigma, upper ones almost entirely beyond; pollen sacs four, long and slender, connective extending beyond as a small spur; filament extremely short (0.5 mm), continuing downward as a long buttress on petal. Pollen isopolar, diplosyndemicolpate, glabrate. Style straight, ca. 5 mm long, terete below but com- nthers above; stigma undifferen- very baccate, nearly 1 cm diam., spherical, yellowish orange, its calyculus inconspicuous, viscin tissue insignificant. Embryo small (2 mm long), slender, lacking a swollen radicular apex at maturity. Apparently a Chiriqui endemic, but to be looked for also in nearby Costa Rica, especially in the region east of San Vito. Materials examined. PANAMA. CHIRIQUÍ: Boquete District, Bajo Chorro, rainforest, 6,000 ft., Davidson 431 (holotype, US; ик as Е, GH), 392 (F not seen); cloud forest, northeastern ridge leading to Cerro Horqueta, 1,800-1,900 m, go 3786 (DUKE, F); Cerro Colora- Cerro Horqueta, in ft., Proctor 31935 (LL); mossy along Rio Chiriqui Viejo, about 2 E of C ridge of Cerro Respinga, Wilbur E Teeri 13109 (DUKE); disturbed cloud forest about 8 km W of Cerro Punta, in the vicinity of Las Nubes, 6,100-6,400 ft., Almeda & Nakai 3536 (CAS, LEA); along old road from Boquete to Cerro Punta on cloud forest trail up to the N slope of Volcan Barú, 1,750-1,900 m, 8°50'N, 82°30'W, Al- meda, de Nevers & McPherson 6166 (CAS, LEA); Distr. Bugaba, Cerro Punta, from STRI house to 29 of moun- tain across the river, 8%52'N, 82%33'W, m, van der Werff & Herrera 6330 (LEA, мој o on path up N go 3 мш of Volcan ды, а along impassable road for- merly linking Boquet Cerro Punta, 8%50'N, 82°30'W, 1,750-1,900 m, La 11343 (LEA, МО); vicinity of Cerro Punta, above Guadalupe, forested slopes above STRI cabin, 2,300-2,450 m, McPherson 9380 (LEA, MO). DISCUSSION The several structural peculiarities of Pana- manthus panamensis outlined above, when taken in aggregate, sharply distinguish it from all other neotropical Loranthaceae. In terms of the inflo- rescence and associated features, the Australian genus Atkinsonia would seem to be the most sim- ilar. However, such an affinity could only be re- mote, as it would be qualified by the strictly ter- restrial habit of that genus, and the various primitive features associated with it in the seedling stages at least some of which we can safely extrapolate from those of the closely related Gaiadendron (Kuijt, 1963, 1965). Any affinity of this sort, therefore, can only be tenuous. Unfortunately, nothing is known about the chro- mosomes of Panamanthus panamensis, and it is precisely here that significant information might be expected. The primitive trio of Atkinsonia, Gaiadendron, and Nuytsia shares a basic number of n — 12, the chromosomes being very small for Loranthaceae (Barlow & Wiens, 1971); in con- trast, the small-flowered neotropical genera of Lo- ranthaceae like Struthanthus have n = 8, the chromosomes being among the largest known in the angiosperms (Wiens, 19 e caducous inflorescence bracts referred to in the above diagnosis have not, as far as I am aware, been mentioned in the literature, and con- trast strongly to all other species of Struthanthus except two or three species that are clearly not related to Panamanthus panamensis, such as S. leptostachyus (Kuntze) G. Don. Similarly incon- spicuous and caducous bracts are normal in Gaia- dendron, where they may occasionally become elongated, foliar, and persistent (pers. obs.), and also in Atkinsonia, where they are always per- sistent (Barlow, 1966; Kuijt, 1981) e pollen of Panamanthus is similar especially to that of Struthanthus oerstedii Standl. (Feuer & Kuijt, 1985) but lacks distinctive, specialized features. There are no significant similarities with the primitive “Gaiadendron trio" genera which, however, shows extreme palynolog- ical contrasts among its component genera, as be- tween Nuytsia and Atkinsonia (Feuer & Kuijt, 1980). Thus it appears again that great palyno- logical divergence is compatible with close affinities in some mistletoe groups. of monotypic 176 Annals of the Missouri Botanical Garden Altitudinal preferences are often consistent with- in individual genera of neotropical mistletoes. Thus, Gaiadendron punctatum tends to be a subpáramo species both in Central America and in its major area, Апдеап South America (Kuijt, 1989). In contrast, Struthanthus has strong preferences for low and middle elevations, nowhere occurring at the elevations noted for Panamanthus panamen- sis (1,200-2,450 m). In summary, the monotypic generic status here roposed more adequately recognizes the unusual nature of the species. In this respect it might be noted that most accepted genera of small-flowered neotropical Loranthaceae are more weakly sepa- rated from each other than Panamanthus is from other genera. For example, Phthirusa and Den- dropemon are separable only on the basis of triadic vs. monadic inflorescences, respectively. The morphological isolation of Panamanthus would seem to warrant status as a monotypic genus, however, no matter what its affinities to primitive or other mistletoe genera are. At the same time, if Panamanthus is related to Struthanthus it may, because of its bisexual flowers and monadic inflo- rescence, be regarded as significantly more prim- itive than that genus. LITERATURE CITED BARLOW, B. A. А revision of the e of Australia and New Zealand. Austral. J. Bot 499. D. WiENs. 1971. The cytogeography of the loranthaceous mistletoes. Taxon 20: 291-312. 1 . The classification of the generic segregates of pue Cop i (7 Notanthera) of the Loranthaceae. Brittonia 25: 26-39. ‚ 5. . Кшт. 1980. = structure of mistletoe ‘pollen. III. Large-flowered neotropical Lo- ranthaceae and their Australian relatives. Amer. J. Bot. 67: 34-50. FEUER & 1985. Fine structure of mistletoe pollen. VI. Small- flowered neotropical Loranthaceae. Ann. Missouri Bot. Gard. 72: 187-212. Кишт, J. 1963. On the ecology and parasitism of the Costa Rican tree mistletoe, Gaiadendron punctatum (Ruiz & Pavón) С. Don. Canad. J. Bot. 41: 927- 65. The anatomy of haustoria and related organs of Gaiadendron (Loranthaceae). Canad. J. Bot. 43: 687-694. 1981. Inflorescence morphology of Loran- thaceae—an evolutionary synthesis. Blumea 27: 1- —. 1989. Additional notes on the parasitism of New World Loranthaceae. Beitr. Biol. Pflanzen 64: 1960. Loranthaceae. In: К. E. Woodson . Schery ce Flora of Panama. Ann. Missouri Bot. Gar 263-290. Wiens, D. 196 а: numbers in North Amer- ican Loranthaceae (Arceuthobium, Phoradendron, Psittacanthus, Struthanthus). Amer. J. Bot. 51: 1-6. RIZZINI, C. T. & NOTES ON SWARTZIA (LEGUMINOSAE: SWARTZIEAE) PRELIMINARY TO THE FLORA OF THE VENEZUELAN GUAYANA! Rupert C. Barneby? ABSTRACT e new species of Swartzia (Leguminosae: Swartzieae) recently collected within or at the periphery of the Fiv Venezuelan Guayana are described and discussed in and S. triptera are presen terms of relationships and m orpho aymardii Barneby, S. oedipus Barneby, S. palustris Barneby, and S. er nted. These du are a prelude to an account of Swartzia by R. S. appear in J. A. Steyermark's Flora of the Venezuelan Guayana, currently in preparation at the Missouri Botanical Garden. At the time that Richard S. Cowan, the premier student in this century of the genus Swartzia, retired from the Smithsonian Institution and from direct participation in systematic work on neotrop- ical Leguminosae, several collectors were actively exploring the flora of Venezuelan Guayana, and others continue to do so. Cowan saw only a small part of the rich collections of Swartzia obtained by the Cerro de la Neblina Expedition (1984— 1985), and little or nothing collected subsequently. Many of the more recent collections of Swartzia from the Guayana Highland have found their way to the New York Botanical Garden, and inciden- tally, to the author of this article for identification. With the aid of Cowan's monograph (1967) and supplementary papers, most important of which is his as yet unpublished treatment of the genus for Steyermark's Flora of the Venezuelan Guayana, I have been able to name, at least to my own satisfaction, the greater part of this material. It will not surprise anyone familiar with the patterns described taxa. The purpose of this article is to place on record those that have been found in Venezuelan Guayana, and to provide names for use in the forthcoming flora of the region. Swartzia (sect. et subsect. Swartzia ser. Touna- teae) alato-sericea Barneby, sp. nov. TYPE: Brazil. Amazonas: Third Serra da Neblina Ex- pedition, between Tatú and Camp Ш, 22 Dec. 1965 (fr), Nilo T. Silva & Umbelino Brazáo in Maguire 607 18 (holotype, INPA; isotypes, K, NY, US). Proxime ut videtur S. steyermarkii Cowan affinis, sed foliolis 6-7(пес 2-3)-jugis subduplo angustioribus (+ 2.3-3.3, nec 4-6 cm latis) margine eximie undulatis diversa. A S. anu lua Schery, quoad foliolorum numerum, ambitum, texturam, colorem necnon marginem undulatam congrua, sed foliolorum rhachi late alata, et a S. sericea Vog. distantius rhachi alata et calyce intus glaberrima abstat. A tree 18 m with trunk 1.5 dm DBH, the young branches and raceme axes densely silky pilosulous with narrowly ascending, basifixed, yellowish hairs, the plurifoliolate leaves except for dorsal side of leaf stalk and pulvinules glabrous, the thinly papy- raceous leaflets lustrously olivaceous on upper face, dull pale brown on lower, the racemes arising singly and geminate from annotinous leaf axils. Stipules erect, linear-lanceolate 4-9.5 x 0.4-0.7 mm, де- ciduous. Leaf stalks 11.5-19 cm, the interfoliolar segments 1-3 cm, each margined with a herba- ceous, finely venulose wing 1.5-2 mm wide, and, at insertion of each pair of leaflets, produced in front of the pulvinules into a deltate, sharply mu- cronate tip; pulvinules 2-2.5 mm; leaflets 13-15 per leaf, the proximal pairs slightly smaller, the rest subequilong, the blade of all leaflets oblong- elliptic from obtuse, at point of insertion enn cordate base, contracted into an obtuse acumen + 1-1.5 mm, undulate ЈЕ: the iar blades 7.9-10.5 x 2.3-3.3 cm; midrib canaliculate ven- ' [llustrated by Bobbi Angell (NY) w York Botanical Garden, | New York 10458-5126, U.S.A. ANN. Missouni Bor. GARD. 78: 177-183. 1991. 178 Annals of th Missouri Bond Garden trally, cariniform dorsally, the numerous and slen- der secondary nerves widely ascending to anasto- mosis close within the margin and giving rise to a reticulum of nervules finely prominulous on both faces. Racemes subsessile, 6-14 cm, densely many- flowered; bracts lanceolate 2-2.5 mm, deciduous; pedicels 4.5-6 mm, ebracteolate; calyx sericeous externally, glabrous within, splitting into 4 re- curved sepals + 4.5 mm; petals, androecium, and young gynoecium unknown, but the gynoecium pubescent (giving rise to a pilosulous pod). Pod declined, the silky pilosulous stipe 2.5-3.5 mm, the body plumply ellipsoid, broadly rounded at each end, + 23-27 mm, slenderly carinate by the ventral suture, the fleshy valves orange when fresh, atrocastaneous and crumpled when dried, thinly pilosulous overall, dehiscent along both su- tures; seeds 2, filling the pod cavity and obliquely truncate where mutually distorted by crowding, the aril + 10 mm long, the rugulose testa ochraceous, highly lustrous. Distribution. In lowland forest on terra firme, known only from the base of Serra da Neblina near 0%40'N, 66°20'W, close to the Venezuelan border in Amazonas, Brazil, and expected in adjacent Ter- ritorio Federal Amazonas, Venezuela. Ripe fruits ecember—January. Until the flower of S. alato-sericea is obtained, it cannot positively be asserted to be apetalous; yet habital and other characters of foliage and fruit almost certainly attach the species to the apetalous subser. Tounateae. In this context the species re- sembles the widespread S. sericea Vog. in habit, form, and number of leaflets, but differs in broadly winged, not terete leaf stalks, and in sepals glabrous on the inner face. Seemingly closer, at least in the winged leaf stalks, is S. steyermarkii Cowan, en- demic to Venezuelan Guayana, but this differs in leaflets about half as many and twice as wide (as enumerated in the diagnosis). The leaflets of 5. alato-sericea closely match the shape, number, and undulate leaflet margin of 5. angustifoliola Schery, a species localized, so far as presently known, on the upper Rio Negro in Venezuela; however, 5. angustifoliola has leaf stalks stipellate at insertion of leaflet pairs and no trace of wing along the stalk between successive pairs. Swartzia aymardii Barneby, sp. nov. TYPE: Ven- ezuela. Bolivar: Dtto. Sifontes, a lo largo de la carretera El Pauji-Icabarú, 80 km al SW de Sta. Elena de Uairén, 26 Sep. 1986 (fl), G. Aymard C. 4888 (holotype, PORT). Swartziae roraimae Sandw. arcte affinis, primo intuitu foliolis dorso praeter costam glaberrimis (nec dense ap- abes pilosis), ulterius antheris majusculis paucioribus (5, nec 11) ac majoribus (2.5-3, nec 2 mm longis) diversa. Tree + 5 m tall, with terete branchlets, the young stems and all axes of inflorescence densely pilose-tomentulose with matted, grayish or brown- ish, basifixed hairs, the leathery, strongly bicolored leaflets glabrous except for midrib dorsally strigu- lose on dorsal face, dark olivaceous and highly lustrous above, paler dull beneath, the few-flowered racemes borne in uppermost leaf axils of hornoti- nous branchlets. Stipules linear-lanceolate 1.6–3 mm, of firm texture, long persistent. Leaf stalks 2—6 cm, shallowly sulcate ventrally, not wing-mar- ginate, the lower of two interpinnal segments 2.2- 3 cm, the distal one shorter (or in lowest leaves absent); leaflets (3-)5, the deeply wrinkled and pilosulous pulvinules 3-4 mm, the blades broadly or narrowly elliptic-acuminate from cuneate base, (4.5-)5-10 x 1.5-3.5 ст, the lanceolate acumen 7-10 mm; midrib canaliculate above, cariniform beneath, giving rise to narrow, crowded secondary nerves little differentiated in strength and incurved- ascending at = 45°, these further generating a close reticulum of veinlets, the whole venation im- mersed on upper face, bluntly prominulous be- neath. Racemes 4-8-flowered, the rachis including short peduncle 6-10 cm; bracts obovate 1.5-2 mm, deciduous; bracteoles 2 at base of pedicel subulate + 2-2.5 mm; pedicels at anthesis 9-14 mm, a little widened and ancipital distally; flower buds immediately prior to anthesis globose, densely brownish silky-pilosulous; sepals 4, glabrous inter- nally, 10-11 mm; vexillum (caducous) white, fla bellate from short claw, 12 mm, thinly hirsute dorsally, undulate-crispate at margin; androecium glabrous, the anthers of 5 larger stamens 2.5-3 mm, that of many smaller ones 1-1.5 mm, the connective of all muticous; gynoecium glabrous, the gynophore 8-9 mm, the lunately elliptic ovary in profile 8 x 3 mm, the uncinately recurved style 2.5 mm, obliquely truncate at apex. Fruit un- known In brush savanna at 750 m, known only from the type locality in the northern foothills of Sierra de Pacaraima in southeastern olivar, Venezuela. er. Distribution. Flowering September-Octo- In foliage and inflorescence 5. aymardii closely resembles 5. roraimae Sandw., as yet known only from the southwestern escarpment of Mt. Roraima (im Thurn 67, К = МУ Neg. 940), but is different in dorsally glabrous, not densely appressed-pilose Volume 78, Number 1 1991 Barneb y 179 Swartzia from the Venezuelan Guayana leaflets. Further differences are fewer (5 vs. 11) large anthers 2.5-3 (vs. 2) mm long, and slightly shorter stipules and bracts, characters that wi deserve close scrutiny when more collections of each species become available. Swartzia roraimae was found by im Thurn at 1,500 m, an elevation unusually high for the genus, and S. aymardii at 750 m, suggesting ecological differentiation. Swartzia (sect. et ser. Possira) oedipus Barneby, sp. nov. TYPE: Brazil. Amazonas: Mun. Cucui, in terra firme forest on Rio Xié, 1 hour by motorboat upstream from confluence with Rio Negro (0°58’N, 67?10'W), 25 Oct. 1987 (young ~ D. С. Daly (with P. Ј. M. Maas, а & R. P. Lima) 5490 (holotype, INPA io isotypes, F, MO, K, NY, US). Figure 1. Foliorum rachi exalata foliolisque parvis (5 cm usque longis) praeter costam нн cum S. foliolosa Cowan, S. brachyrrhachide Harms et S. longistipitata Ducke comparabilis, ultimae тенек proxime affinis, a prima foliolis dimidio minoribus necnon tertia foliolis 4-5(nec 1-3)-jugis, ab omnibus pedicellis calycem versus incrassatis et 2 mm usque crassis diversa. Trees attaining 20 m with trunk 2 dm DBH, the annotinous and older branchlets ashen glabrate, the hornotinous ones together with terete leaf ra- chises, petiolules, and dorsal costa of leaflets gray silky-strigulose, the foliage bicolored, the leaflets glossy dark green above, paler dull beneath, the densely brownish sericeous few-flowered racemes axillary to 1—3 distal leaves of each new branchlet, ud tist pom: all shorter than the ipu es linear-attenuate 2.5— mm, caducous. . Leaf stalks 4–6.5 cm, the petiole proper and the interfoliolar segments 8-13 mm; petiolules 2-2.5 mm, the stipels subobsolete; leaflets 9-11, varying in outline from narrowly ovate- to lance- acuminate from rounded base, including the api- cally emarginate acumen (5-6 mm) 35-50 x 14- 18 mm, the costa shallowly impressed above, ca- riniform beneath, giving rise to many crowded, incurved-ascending secondary veins and a finer reticulum, all prominulous on both faces of blade. Racemes 3-5-flowered, the axis including short peduncle 3-4 cm; bract and bracteoles at base of each pedicel narrowly triangular 1.5-2 mm, de- ciduous; pedicels stout, 8-12 mm, clavately thick- ened upward, at apex + 2 mm diam.; sepals 8- 10 mm, densely sky Sateen yy glabrous within; petal and complete , the anther of shorter Pome + 0.6 mm, glabrous; ovary glabrous; style subulate, gently incurved, + 3 mm, the stigma poriform; stipe of immature fruit 10- 14 mm. Distribution. In lowland, noninundated for- est. Known only from the upper Rio Negro basin astride the Venezuela-Brazil frontier near Piedra de Cucuy in Territorio Federal Amazonas, Vene- zuela, and adjacent state of Amazonas, Brazil. Flow- ering + September-October. Additional material examined. bic xe TERRI- TORIO FEDERAL AMAZONAS: margem o Rio Negro no caminho para Pedra de Cucui (1? 15 N, 66°51'W), 28 Oct. 1987 (immature fr), ЈУ. А. Rodrigues 10795 (Е, GH, К, MO, NY, US). Swartzia oedipus has the essential characters of section and series Possira and falls within the circle of affinity defined by couplets 14-20 in Cow- an's (1967: 26) key to that group. Leaflets glabrous beneath except for strigulose costa make it most nearly comparable to S. foliolosa Cowan (of south- eastern Colombia), to the somewhat polymorphic and widely dispersed S. brachyrrhachis Harms, and to S. longistipitata Ducke (local in central Brazilian Amazonia), which may be its closest known ally. It differs from S. foliolosa, as described by Cowan (1967: 190), in leaflets only half as large and in densely brownish sericeous (vs. glabrous) axes of inflorescence, from S. brachyrrhachis in having 8-11 (vs. 1–5) leaflets per leaf, and from S. longistipitata by leaflet number (8-11 vs. 3- 7 per leaf) and by sepals glabrous (not thinly silky) internally. It differs from all of these in the stout pedicels clavately enlarged upward and about 2 (уз. 1 or <1) mm diam. at apex, the character that suggested the epithet oedipus. The small glossy leaflets, the contrast between gray indumentum of stem and leaf stalk and brownish silky indumentum of inflorescence, and the perfectly glabrous, long- stipitate gynoecium are characteristic features. Swartzia palustris Barneby, sp. nov. TYPE: Ven- ezuela. Territorio Federal Amazonas: hills 2.5 km SW of Base Camp, 400-500 m, SW side of Cerro de la Neblina (00*49'N, 66?10'W), 20 Feb. 1985 (fl, fr), M. Nee 30979 (holo- type, VEN; isotype, NY). Inflorescentia e ramulorum coaevorum axillis orta, ped- icellis ebracteolatis, styloque laterali brevissimo sect. Swartziae subsect. Swartziae о mage autem in- ter Pittierianis et Racemosis ambigua, cum illa bracteo- larum cum hac foliis unifoliolatis congrua, sed foliolis ок inferne aureo- Pio dam in utraque serie anomala, affinitatis verae petenda Slender trees 3-15 m tall with sometimes sar- mentose branches, the annotinous branches gla- 180 Annals of the Missouri Botanical Garden Swartzia oedipus Barneby. Center, branchlet just after anthesis; lower left, stipule and base of leaf stalk; lower right, gynoecium from branchlet, enlarged. (All from Daly 5490.) brate, the young ones, the lower face of leaflets, reticulate above, the short racemes arising singly and the whole inflorescence densely golden seri- or 2—3 together from hornotinous leaf axils. Stip- ceous with contiguous and parallel, forwardly ap- ules (few seen) triangular-subulate, less than 1 mm, ressed hairs, the leaves unifoliolate, the charta- —fugacious. Leaf stalks terete, including the pulvinus ceous leaflet olivaceous, glabrous, and lustrously 5-13 mm, at middle 0.7-1 mm diam.; stipels 0; Volume 78, Number 1 1991 Barneby 181 Swartzia from the Venezuelan Guayana pulvinule of solitary leaflet 2.5-5 mm; leaflet broadly elliptic from rounded or broadly flabellate base, contracted into a slender, apically emarginate acumen = 1-2 cm, the blade including acumen 7.5-15 x 3.5-8 cm; venation pinnate, the midrib shallowly sulcate ventrally, cariniform dorsally, giving rise on each side to 7-9 major secondary nerves incurved-ascending to anastomosis shortly within the plane margin and to many weaker in- tercalary ones, these all generating a reticulum of nervules finely prominulous on both faces. Primary axis of racemes 4-12 cm; bracts ovate, scarcely 1 mm; bracteoles 0; pedicels 5-9 mm, a little compressed but not thickened upward; flower buds just prior to anthesis globose, 4-5 mm diam., densely golden sericeous; sepals 4, 4.5 mm, glabrous internally; vexillum = 6 mm diam., thinly pilose dorsally, caducous; androecium glabrous, the anther of the + 104-114 smaller stamens 0.7-1 mm, that of 3 stout abaxial ones 2.2-2.7 mm, arched backward, apiculate; gynoecium glabrous, the gynophore at anthesis 3-4 mm, the lunate ovary 3-3.5 х 1.6-1.9 mm, incurved and con- tracted into an almost obsolete style, the stigma poriform. Pod obliquely oblong- or obovoid-ellip- soid, cuneately contracted into a stipe 4-8 mm, the body in profile (26-)30-34 x 15-21 mm, the style base becoming infra-apical, the iue valves orange, when dried crustaceous brittle, namon brown; seed solitary, the fimbriate ш + 15-17 x 17-20 mm, the body narrowly reniform + 25 x 9-11 mm, the testa castaneous lustrous. Distribution. In swampy forest with Mauritia and in rainforest on low hills, 130—400 m, known only from the headwaters of rios Baria- Mawarinu- ma at 0%50-53'N, 66?*10'W, near SW foot of Cerro de la Neblina, Territorio Federal Amazonas, Venezuela. Flowering July, January-March. Additional specimens examined. ENEZUELA. TERRITORIO FEDERAL AMAZONAS, all from the vicinity of the type locality: 30 Mar. 1984 (fr), R. Liesner 17035 (МО, NY); 20 Apr. 1984 (fr), W. W. Thomas 3162 (NY); 8 May 1984 (fr), №. W. Thomas 3411 (NY); 27 Nov. 1984 (fr + young fl buds), D. Bell 328 (NY); 27 Jan. 1985 (fl), M. Nee 30564 (NY); 16 Feb. 1985 (fr), Boom & Weitzman 5873 (NY Swartzia palustris neatly fits Cowan's (1967: 13, 18) definition of sect. Swartzia subsect. Swart- zia, but is not easily assimilated into any series of the subsection. The combination of glabrous gy- noecium and absence of bracteoles leads directly to the small group Pittierianae, but these have regularly 2-4-jugate, almost glabrous leaves, and consequently a facies very different from S. pal- ustris, notable for unifoliolate leaves clothed be- neath with a dense indumentum of appressed gold- en hairs. If its lack of bracteoles be disregarded, S. palustris would be compatible with the character of series Racemosae, where it finds a superficial resemblance in S. racemosa Benth. This last diffs greatly, however, in very short leaf stalks comp entirely of pulvinus and terminal pulvinule, in lick of golden indumentum, in conspicuous develop- ment of stipules, floral bracts and bracteoles, and in almost glabrous (not golden silky) flower buds and a distinctly developed style. The true relation- ship of S. palustris is unknown. Swartzia triptera Barneby, sp. nov. TYPE: Ven- ezuela. Territorio Federal Amazonas: Depto. Rio Negro, in lowland tropical evergreen for- est, 140 m, Base Camp on 5 side of Rio Baria (= Mawarinuma), SW side of Cerro de la Neblina, 00?49'30"N, 66?06'20"W, 13 Feb. 1985 (fr), M. Nee 30839 (holotype, VEN; isotype, NY). Figure 2. Bracteolarum defectu styloque obliquo brevissimo sect. tziae referenda, foliorum rachi legumine secus suturas seminiferam late 2- et abaxialem l-alato (alis 3.5-7 mm latis) insignis. Trees 4-12 m with smooth fuscous branchlets, glabrous up to the thinly minutely silky strigulose inflorescence, the ample leaflets subconcolorous, dull olivaceous or on lower face dull brown-oliva- ceous, the racemes of small flowers borne singly or 2—3 together in the ах! of coeval leaves. Stipules lacking. Leaf stalks 1-2.9 dm, terete, estipellate; leaflets in all but an occasional depauperate leaf either 5 or 7, the cylindric black wrinkled pulvin- ules 6-14 x 1.1-1.6 mm; leaflets ovate or elliptic- ovate from broadly cuneate or flabellate base, abruptly contracted distally into a caudiform acu- men cm, the blade of lateral ones 10-19 x (3-)4-6.5 cm, that of the stalked terminal one 13-20 x (4-)5-9.5 cm; midrib impressed above, cariniform beneath, giving rise to 5-6 pairs of slender major secondary nerves incurved-ascend- ing to anastomosis well within the plane, finely corneous margin, the intervenium of both faces reticulate with a mesh of incised veinlets. Racemes ascending, 10-25-flowered, the axis including peduncle (7-)10-24 cm; bracts subulate 1.5-3 mm, deciduous after anthesis; bracteoles 0; pedi- cels at anthesis 4—5 mm, in fruit 5-12 mm; flower buds prior to anthesis obnapiform, bluntly pentag- onal above middle, thinly puberulent; sepals 5, sharply reflexed, lance-ovate 6-7 mm, glabrous internally; vexillum membranous white glabrous, 182 Annals iuo ы Garden ВЕ 2. Swartzia triptera Barneby.— A. Leaf and inflorescence. — B. Calyx and ovary. — C. Pod. — D. Cross FIGUR section of pod. — E. Se ed. — F. Seed testa. (A, B from Davidse 27068; C-F from Nee 30839.) Volume 78, Number 1 1991 Barneb 183 Swartzia from the Venezuelan Guayana caducous with the androecium as the calyx ex- pands, in outline ovate beyond the short claw, + 7 X 4 mm; androecium glabrous, the filaments yellow, the + 70 small anthers 0.6–0.7 mm diam., the 4 large ones 1.5 mm; gynophore 2.5-2.8 mm, puberulent; ovary = 3 mm, glabrous except for a few hairs at base and along sutures; style obliquely terminal + 1 mm, incurved, the stigma punctiform. Pod ascending, stipitate, the stipe 7-11 mm, the body in profile asymmetrically elliptic 5.5-8 x 2.7-3 cm, cuneately attenuate at base, more abruptly contracted at apex into a short triangular beak, laterally compressed but turgid over the 2 seeds, girdled lengthwise by wings 3.5-7 mm wide arising 2 from the adaxial and one from the abaxial suture, the stiffly leathery valves glabrous, verti- cally shallow-grooved over the convex fertile seed chamber, this in transverse section + 18-20 x 9-10 mm; exocarp smooth papery fuscous, the endocarp crustaceous 0.6 mm thick; seeds + 20 X 11 x 10 mm, the fimbriate aril 4-5 mm diam., the fragile papery testa lustrously castaneous. Distribution. In terra firme forest at + 140 m, known only from the SW base of Cerro de la Neblina in Territorio Federal Amazonas, Venezue- la. Flowering July-August; fruits ripe November- February. Additional specimens examined. АП from the im- mediate vicinity of Cerro de la Neblina Base Camp on io Mawarinuma: 6-7 July 1984 (8), С. Davidse & J. S. Miller 27068 (NY); 6 July 1984 (in bud), J. S. Miller 1735 (NY); 27 Nov. 1984 (fr), Boom & Weitzman 5165 (NY); 1 Feb. 1985 (fr), Boom & Weitzman 5476; 7 Feb. 1985 (sterile), Boom & Weitzman 5660. By virtue of habit, terete estipellate leaf stalks, absence of bracteoles, and almost glabrous gynoe- cium with very short incurved style, S. triptera appears related to ser. Benthamianae Cowan of subsect. Swartzia, where it closely resembles, when flowering or sterile, S. benthamiana Miquel and S. laevicarpa Amshoff. The fruit is nearly that of the latter as to size and outline, but unlike that of any described species of the genus is doubly winged along the ventral and singly so along the dorsal suture. The ripe fruit is strangely similar in struc- ture to that of the African Erythrina greenwayi Verdcourt and has no precedent in Swartzia. LITERATURE CITED Cowan, В. S. 1967. Flora Neotropica Monograph No. 1. Swartzia (Leguminosae, Caesalpinioideae, Swart- zieae). Hafner Publishing, New York СУКТОСАКРА КОМТН (ANACARDIACEAE) IN SOUTH AMERICA! John D. Mitchell? and Douglas C. Daly? ABSTRACT Cyrtocarpa has previously been known only from Baja California and western Mexico; here we report two species of the genus from combination, was originally published as a Bursera diaceae and Burser South America. Cyrtocarpa velutinifolia (Cowan) Mitchell & Daly, presented here as a new and later transferred to Tapirira (Anacardiaceae). The Anacar- aceae show frequent morphological convergence in the drier tropics; moreover, in several genera they display strong similarities in staminate flower structure. Cyrtocarpa velutinifolia occurs in savannas and granitic outcrops in Colombia, Venezuela, Brazil, and Guya na. Cyrtocarpa caatingae Mitchell & Daly, a new species represented by four collections from the dry thorn-scrub vegetation of Bahia, Brazil, resembles the Mexican C. procera Kunth and appears to be most closely related to it Cyrtocarpa Kunth is a genus heretofore known from two Mexican species, C. procera Kunth and C. edulis (Brandegee) Standl. Now, two additional species are recognized in South America. One of them was originally described in the Burseraceae; the other was discovered recently in collections from Bahia, Brazil. The Burseraceae and Anacardiaceae are ac- cepted as being closely related by most contem- porary authors (e.g., Cronquist, 1981; Takhtajan, 1980; Thorne, 1976). The two families show con- siderable morphological convergence in dry trop- ical habitats, where species of both families display a densely branched, tortuous habit; succulent branchlets; leaves that are imparipinnate, con- gested toward the branchlet apices, and deciduous; and drupaceous fruits (eventually dehiscent in Bur- seraceae). For example, Bursera microphylla A. Gray and Pachycormis discolor (Benth.) Coville are sympatric in Baja California and are very sim- ilar in appearance (pers. obs.). In arid portions of East Africa, some species of Commiphora Jacq. (Burseraceae) and Lannea A. Rich. (Anacardi- aceae) are difficult to distinguish (J. B. Cillett, pers. comm.). In southern Africa, some species of Сот- miphora and Rhus L. show a strong resemblance The occasional similarities in the structures of the staminate flowers of these two families led Cow- an (1952) to describe an anacardiaceous plant from Guyana as Bursera velutinifolia. This species has since been collected in Brazil, Colombia, and Ven- ezuela (Fig. 1). In the staminate flowers, the only characters that place the species out of the Bur- seraceae are the imbricate (induplicate-valvate in Bursera) petals and the pistillode that is 3-5-lobed from the base (always 3-lobed in Bursera). The deciduous habit of this species is also characteristic of Bursera. After examining pistillate flowers, Marcano-Ber- ti (1986) transferred B. velutinifolia to Tapirira Aublet (Anacardiaceae). This transfer to Anacar- diaceae is supported by the solitary, apotropous ovules, apically suspended within the locules (vs. two epitropous, laterally attached ovules in Bur- seraceae); carpels more than three (2-3 in Виг- sera); and mesocarp and endocarp distinct. More- over, the styles b parated in the developing fruit, whereas in all Burseraceae there is a single, occasionally branched style. We agree with the implication of Marcano-Berti's transfer that the species belongs in the tribe Spondiadeae (Engler, 1882) because of its apically suspended ovules and the ovary with usually more than three carpels. Its affinity with the Spondiadeae was recognized by the indigenous population in Venezuela: it is called "jobillo," and the common name for Spondias is “jobo.” However, examination of the available ma- terial has brought to light several fundamental ! We thank J. A. Kallunki, S. A. Mori, апа W. W. Thomas for critically reading the manuscript; D. W. Stevenson for enlightening comments on fruit morphology; R. P. о for assistance with the Latin diagnosis; В. Angell for the revealing illustrations; and A. Kostermans, P. Berry, and one anonymous reviewer for editorial assistance. We are also grateful to the staffs of the following Pune for the loan of specimens: F, K, MO, P, PORT, VEN. The Fund for Neotropical Plant S aid of the New York Botanical Garden covered some expenses associated with the production of this manuscri ? New York Botanical UN Bronx, New York 10458, U.S.A. ANN. MISSOURI Bor. GARD. 78: 184-189. 1991. Volume 78, Number 1 Mitchell & Daly 185 1991 Cyrtocarpa in South America TROPICAL AMERICA я | F >> = EA a) 30 | A "T а UN E A | (A 24 i о ~ or: = => State University of Utrecht, The Netherlands. characters that preclude its placement in Tapirira and necessitate its transfer to Cyrtocarpa. These characters are summarized in Table 1. Since we have reexamined Cyrtocarpa, we pre- sent below a key to the neotropical genera of Spon- diadeae, a description of Cyrtocarpa, and a ke to the species of the genus; one new combination is made and a new species is described. KEv TO THE NEOTROPICAL GENERA OF SPONDIADEAE la. Leaflets with intramarginal vein; stone consist- ing of bony endocarp enclosed in a fibrous ma- trix Spondias c mt O & Eb o “+ Uu 5. = m о = + jua 5 et H w 3 > а) 99. 5 >. . < o, 5 wn ao ‚ © 3 % о دم‎ Ky gested at branch apices; pistil(lode) glabrous, stone i opercula (lids t ntrocaryon 3b. Buttresses absent; in pistillate flowers, only one locule containing an ovule; stone slightly oblique-oblong, obovoid, Distribution of Cyrtocarpa species. This map is Flora Neotropica base map no. 1, published by the or ovoid; opercula l-5; seed 1; dry forests, savannas, and thorn scrub ... 2b. Plants evergreen; leaves not congested at ranch apices; pistil(lode) pubescent; stone without opercula Tapirira Cyrtocarpa Kunth in HBK, Nov. Gen. et Sp. 7: 20, t. 609. 1824. TYPE: Cyrtocarpa procera. Dasycarya Liebm., Vidensk. Meddel. Dansk Naturhist. Foren. Kjøbenhavn 1853: 98. 1854. TYPE: Dasy- carya mexicana. Small to medium-sized polygamodioecious trees of dry to arid habitats, branchlets somewhat suc- culent (drying wrinkled), resin apparently nonal- lergenic. Leaves deciduous, congested toward the branchlet apices, imparipinnate; lateral leaflets ses- sile or short-petiolulate, margin entire, densely pu- bescent on both surfaces, venation eucamptodro- mous. Inflorescences subterminal, paniculate (pseudospicate in C. velutinifolia), initiated in leaf- less condition. Flowers 5-merous, aestivation im- bricate; petals patent at anthesis; diplostemonous; disk intrastaminal, annular, crenulate, fleshy; pis- 186 Annals of the Missouri Botanical Garden TABLE l. Comparison of Tapirira, Cyrtocarpa velutinifolia, and Cyrtocarpa. Character Tapirira Cyrtocarpa velutinifolia Cyrtocarpa Leaves Phenology eciduous deciduous Leaflets petiolulate subsessile subsessile Venation usually brochidodromous eucamptodromous eucamptodromous Indumentum along veins all of surface all of surface Flowers Pistil(lode) pubescent glabrous glabrous vary 1-locular 1-3(-4)-locular 2-5-locular* Fruit Shape subsymmetric oblique oblique Stone usually cartilaginous Operculum absent bony ny present (1) present (2-5) * Engler (1882) misinterpreted the ovary of Cyrtocarpa as 1-2-locular. tillode consisting of five styllodes; pistil doliform, styles separate, glabrous or papillate (vs. densely pubescent in Tapirira), ovary 1—5-locular. Fruit a drupe, purple or yellow to orange at maturity, obliquely obtuse-oblong, styles often persistent, sur- face often pubescent (vs. glabrous in Spondias), exocarp thin, mesocarp fleshy, endocarp bony, with 1-5 opercula; seeds apparently 1, testa with sad- dle-shaped patch corresponding to the hilum, cot- yledons reniform. Кеу TO THE SPECIES ОЕ CYRTOCARPA la. Secondary axes of inflorescence dii ни C. velutinifolia lb. Secondary axes of inflorescence paniculate. 2a. Lateral leaflets broadly elliptic to broadly (ob)ovate; anther connective not glandular- ко d Lateral leaflets lanceolate to narrowly el- liptic; anther connective glandular- mu- cronulate; flowers 4-6 mm За. Ovary 2-5-locular; upper leaflets fre- quently alternate; filaments 1.8-2 mm long, anthers 1 mm long; pes not exceeding disk; western Mexico ......... “с procera . Ovary 1-locular; upper leaflets always opposite; filaments 1.3-1.5 mm long, anthers 0.5 mm long; pistillode ex ceeding disk; Bahia, Brazil . C. caatingae [55] c Cyrtocarpa velutinifolia (Cowan) Mitchell & aly, comb. nov. Bursera velutinifolia Cow- an, Brittonia 7: 401-402. 1952. Tapirira velutinifolia Marcano-Berti, Pittieria 13: 23. 1986. TYPE: Guyana: Sand Creek, Rupununi River, Sep. 1948, Wilson-Browne 112 (F.D. 5650) (holotype, NY). Figures 1, 2. Loxopterygium gutierrezii Barkley, Lloydia 25: 121- 122. TYPE: Colombia near Masinga, near Santa M H. Smith 911 (holotype, NY; isotypes, А, ВМ, Е, К, MICH, US). The principal characters that have been used to separate genera in the Spondiadeae are the leaf venation; number of carpels; number of styles; degree of fusion of the styles; number of locules; structure of the mesocarp; presence or absence of opercula in the stone; other characters of the en- docarp; and number of seeds developing per fruit (Engler, 1882; Hill, 1933, 1937; Kostermans, 1981; von Teichman, 1987; von Teichman & Robbertse, 1986; von Teichman & van W yk, 1988; Wannan, 1986). The generic limits in the Spon- diadeae are not clear in many cases, but there are consistent qualitative characters that can be used to distinguish Tapirira and Cyrtocarpa. Tapirira is characterized by evergreen leaves; petiolulate lateral leaflets, usually with brochidod- romous venation; densely pubescent pistil and pis- tillode; consistently unilocular ovary; subsymmetric fruit; and Inppercule te Alone: m pom de ie al leaf- carpa h lets with d OI. venaliau: iin pistil and pistillode; 1 —5-locular ovary; oblique fruit; and operculate stone. ursera velutinifolia possesses all the charac- ters of Cyrtocarpa. Removal of the exocarp and mesocarp from fruits of Boom & Eisenberg 6056 (MO, NY) and of Aristiguieta 6090 (MO, NY) revealed a single operculum, located subapically on the anterior face of the stone (see Fig. 2). The genus most closely related to Cyrtocarpa, Oper- culicarya H. Perrier from Madagascar, also has Моште 78, Митбег 1 Mitchell & Daly 187 1991 Cyrtocarpa in South America FIGURE 2. Cyrtocarpa velutinifolia (Cowan) Mitchell & Daly.— A. Part of mature leaf, including detail of abaxial leaflet surface. — B. Staminate inflorescence, showing pseudospicate structure. — C. Detail of staminate inflorescence, showing reduced tertiary axes. — D. Bud. — E. Top view of staminate flower. — Е. Longisection of staminate flower, showing pistillode not exceeding disk.— C. Pistillate flower, post-anthesis, including top view of pistil. — H. ryo; dorsiventral and lateral views of stone, showing operculum. — I. ion of infructescence; transse or ction of drupe. Based on Bunting 4777 (A), Goodland 1046 (B, C), Wilson-Browne 518 (F.D. 5919) (D-F), Wingfield 1011 (G, I), and Aristiguieta 6090 (Н). 188 Annals of the Missouri Botanical Garden unioperculate stones, but it is evergreen and has horseshoe-shaped (vs. reniform) cotyledons and usually five stamens. The operculum of the stone is a structure as- sociated with a specialized germination mechanism (Hill, 1933, 1937). It functions as a door in the protective woody tissue of the stone through which the seed germinates. In the Anacardiaceae, oper- culate stones are restricted to 10 of the 18 genera in the tribe Spondiadeae. Additional specimens examined. VENEZUELA. APURE: Distrito Pedro Camejo, Piedra de San Vicente, 9 airline of Cano Colorado, vic. of Panare Indian village, 30 km SW of National Guard post Maniapure, са. 200 m, 6°45’N, 66°37'W, 5- ae — зе Boom & Eisenberg 6056 (K, MO, NY); Represa Сип, ca. 55 km NE of Ciudad Piar, ley. 250- 300 1 m, с a 7°35" N. 63°07'W, 4–5 Apr. 1981, Liesner & las 11243 (MO, NY); upstream on Rio Asa from Raudal Cotúa, 2 Aug. 1960, 6, Wurdack & Monachino 41240 ov. : vic. of Valencia, 400-800 m, Dec. 1919 Pii 8679 AM d с Distrito Democracia, Pueblo N 20 km rumaco, 520 m, 24 June 1982, R. Wingfield Fon (Ў GUARICO: ise of Tamanaco, Altagracia de Oritu mas, May 1966, Aris- yas 6090 (NY): 1 near jud Real, June 1966, Aris- tiguieta 6165 (NY). PORTUGUESA: Distrito Guanare, Mesa ELLEZ, PORT). ZULIA: Distrito Miranda, | Consejo de Ciru r spring of El Cardón, 150-200 m, 7 Aug orther NY); Iramaipang, Kanuku , Wilson-Browne 518(F.D. 5919) (NY). BRAZIL. pp. Rio Uraricoera, Cachoeira Urubu, 18 Nov. 1979, Pires et al. 16709 (NY). Cyrtocarpa caatingae Mitchell & Daly, sp. nov. TYPE: Brazil. Bahia: Mun. Manoel Vitorino, 1-5 km W of Massel Vitorino on road to Catingal, 20 Nov. 1978, S. A. Mori, T. S. Santos & C. B. Thompson 11248 (holotype, CEPEC; isotypes, ILL, K, NY, RB). Figures кз. Cyrtocarpae procerae ob inflorescentiam paniculatam, foliola lanceolata vel anguste elliptica, et connectivum antheridii glanduloso-mucronulatum simili autem ob ovarium uniloculare (nec 2-5-loculare), sta- mina breviora, et pistillodium discum superans differt. Deciduous, dioecious shrub or small tree ca. 3.5 m tall. Branch apices somewhat fleshy (drying wrin- kled). Leaves (only immature ones known) flushing at time of flowering, congested near the apices, 3- apiculate, camptodromous, leaf surfaces sericeous, with dense, long. Inflores- m long, secondary axes 0.5- 2 cm long, tertiary axes 1.5-6 mm long, pedicels 1.5-2 mm (staminate flowers) and 2-2.5 mm (pis- tillate) long. Flowers greenish, 5-merous and diplo- stemonous, 4 mm diam.; calyx teeth semicircular, fleshy, 1 x 0.7 mm, pubescence as on inflores- cences, margin ciliate; petals oblong, 2-2.4 x 1 ous) and 2 mm (antesepalous) long, anthers broadly elliptic, 0.6 mm long, connective glandular-mu- cronulate, apiculum drying black; disk thin; pistil- lode consisting of 3—5 stylodes, each 0.6 mm long, greatly exceeding disk. Pistillate flowers with fila- ments 1.5 and 2 mm long, anthers in dorsiventral view lanceolate, 0.4—0.5 mm long; pistil doliform with rounded apex, 1-1.5 mm x 0.7-1 mm, gla- brous, styles 5, 0.2-0.25 mm long, separate, spreading, stigmas discoid; locule 1, ovule 1, pen- dent. Developing fruit obliquely oblong, зуби ob- tuse, styles persistent, mature fruit unknown atypes. BRAZIL. s.l., 1816-1821, Saint-Hilaire s.n. d^ BAHIA: Igreja Velha, 1841, Blanchet 3418 (МУ; = 2 sheets); caatinga near Caldeirao, Oct. 1906, Ule 7256 (K). Distribution. Caatinga (dry thorn-scrub) of the interior of Bahia. Because the flora of the caatingas of northeastern Brazil is still not well represented in herbaria (Mori, 1989), it is not possible to зау just how rare Cyrtocarpa caatingae is. In addition, the herbarium labels on the older collections that represent this species provide almost no detai about its habitat, habit, or phenology. Cyrtocarpa caatingae resembles С. ргос because of its paniculate inflorescence, lanceolate to narrowly elliptic leaflets, and glandular-mucror ulate anther connective. The new species d differs by из unilocular (vs. 2—5-locular) ovary, $ shorter stamens, and pistillode greatly exceeding (not ё ceeded by) the disk. While the occurrence of Cyr tocarpa in the semiarid climate of the caatingas is consistent with its habitat preference elsewhere; the ranges of the four species of Cyrtocarp? из markedly discontinuous (see Fig. 1), interrup! over long distances by moister habitats, or by wa in the case of the two Mexican species. era Volume 78, Number 1 1991 Mitchell & Daly Cyrtocarpa in South America 189 FIGURE 3. Cyrtocarpa — Mitchell & реу" ening flower. —C. Stamin wer at anthesis. dk, — E. Stamens, showing glandular mueronulate anther co connective. —F. Pis —H. Stamin gisection of ovary, showin М g single loc odes A-E), and St. Hilaire s.n. r5 LITERATURE CITED roid HES Pes A prod. of G. Wilson- кі 2 J: a. I. Kanuku Moun- она "n ttonia 7. ju "Ti uid soi C. 1981. An Integrated System of Clas- cation “= Flowering Plants. Columbia Univ. Press, dpa E a А. & С. de Сап- йш T Monogr. Phan 171- va AMOR 933. The сан of es of seeds бе: т stony endocarp. Апп. Bot. (London) — Mi The method of germination of seeds : ee a stony endocarp II. Ann. Bot. (n.s.) 1: 05 Ene. A. J. С.Н. Mu Notes on Spondias L. E accen) Quart. J. Taiwan Mus. 34: 105- Ma вон І. 1986. Tapirira eret unie una Moni, vin combinación. Pittieria 13: 23-2 P A. 1989, Eastern, extra- rode dit р. 427-454 in D. С. Campbell & Н. D. Hammon an Brazil. d —A. Staminate- flowering branchlet with immature leaves section of flower, showing pistillode эзелде te tillate flower, post-anthes —G. . Immature fruit. Based on Mori et al. 11248 imis (editors), Floristic Inventory of Tropical Counties. New York Botanical Garden, New York. TAKHTAJAN, А. ks 1980. Outline of the classification of flowering plants е Bot. Rev. (Lan- caster) D ЈЕ 225-3 THORNE, В. К. vehe 7 phylogenetic classification of the Angiospermae. Evol. Biol. 9: 35-10 06. уох TEICHMAN, I. 1987. рока and structure of the pericarp of Lannea discolor Pagers Engl. (An- acardiaceae). Bot. J. Linn. Soc. 9 & P. J. ROBBERTSE. 1986. ‘eae and structure of the drupe in Sclerocarya birrea (Rich- ard) Hochst. subsp. caffra Kokwaro (A Anacardiaceae), vu special reference to the ба A nd the oper- culum. Bot. J. Linn. Soc. 92: 303- — — — & A. E. van Wyk. 1988. ть ontogeny and structure of the pericarp and seed-coat of Harpe- puo caffrum Bernh. ex Krauss (Anacardiaceae). J. Linn. Soc. 98: 159- -17 76. 5 olu it nd Bot. WANNAN, B. r£ the affinities of the "Australian genera. Austral. N 6-8 t. Soc. Newslett. 49: STUDIES IN NEOTROPICAL Alan Graham? PALEOBOTANY. VIII. THE PLIOCENE COMMUNITIES OF PANAMA —INTRODUCTION AND FERNS, GYMNOSPERMS, ANGIOSPERMS (MONOCOTS)! ABSTRACT Thirty-one fossil spores and pollen of ferns, gymnosperms, and monocotyledonous angiosperms have been identified from the Pliocene Gatun Formation of Panama together with one unidentified Pyrrophyta. These are Lycopodium (types 1-4), Selaginella, Alsophila, Cnemidaria, Cyathea (types 1, 2), Ophioglossum, Grammitis, Ceratopteris, with no previously known fossil records) and 27 unknowns presently under study make the the largest Tertiary floras in northern Latin America. The remaining descriptions and interpretation of the flora will be presented in concluding papers In 1962 the Panama Canal Commission drilled a series of wells in Gatun Lake as part of the Trinidad Dam Project to test the foundation for increasing the water storage capacity of the lake. Samples from along these cores were obtained by Elso Barghoorn in 1962 and by the author in 1963. The upper portions of the cores were used by Bartlett & Barghoorn (1973) for study of Quater- nary sea level and climatic changes in central Pan- ama during the past 12,000 years. The lower por- tions penetrated subsurface Gatun Formation sediments, and these samples serve as the basis for the present study. It is fortunate that samples were preserved because subsequently all cores drilled by the Commission, representing millions of years of vegetational and paleoenvironmental history, were discarded because of storage limitations. All know surface exposures of the Gatun Formation are ma- rine sediments (Fig. 1), and the core material is presently the only source of data on the vegetation of Panama during Gatun time. Core SL-49 was drilled at latitude 9%80'N + 4,113, longitude 79°57'W + 1,987, and levels 222.5, 223, and 233.5 (in feet, following original log data) yielded plant microfossils. Core SL-103 was drilled at lat- itude 9°16'N + 5,945, longitude 79%52'W + 2,963, and levels 250, 253, 255.5, and 257 yield- ed plant microfossils. Numbers following latitude and longitude (e.g., + 4,113) are the number of feet between the minute cited (e.g., 80') and the next minute. They are used to locate precisely a site within a small or restricted locale. The Gatun Formation was earlier assigned to the middle Miocene (Cooke et al., 1943; Stewart & Stewart, 1980; Woodring, 1957-1982), but marine invertebrate faunas now indicate a Pliocene age (Vokes, 1983; van den Bold, Stewart, pers. comm.). Other details on the location and geology of the Gatun and other Tertiary formations in Pan- ama are summarized in Graham et al. (1985). The Gatun Formation is the last and youngest in a series of five Tertiary pollen- and spore-bearing strata being studied from the Canal region of Pan- ama. The others are the middle(?) to upper Eocene Gatuncillo Formation (Graham, 1985), and the lower Miocene Culebra, Cucaracha, and La Boca formations (Graham, 1988a, b, 1989). [n addition, lower Miocene palynomorphs are known from the Uscari sequence of Costa Rica (Graham, 1987). Quaternary studies include those of Bartlett & Barghoorn (1973) on sediments from Gatun Lake, and Horn (1985) on sediments from DSDP site ' The орка ке acknowledges field assistance and information on the geology of Panama provided by S. Stewart a wart (Chief Geologists, Panama provided by Alice Tryon Ha rvard Universit 82055926, BSR-8500850, and BSR-881977 R. al Commission, retired), and on spores of the Filicineae Can 208 Research supported by NSF grants GB-5671, DEB-8007312, DEB- ? Department of Biological Sciences, Kent mM University, Kent, Ohio 44242, U.S.A. ANN. Missouri Bor. GARD. 78: 190-200. 1991. Volume 78, Number 1 1991 Graham 191 Pliocene Communities—Monocots 565 off western Costa Rica. In 1989, members of van der Hammen's group (pers. comm.) completed transects through the modern vegetation of Costa Rica preliminary to pollen and spore analysis of high-altitude peat deposits. Collectively, these stud- ies will provide a more complete data base on Cenozoic vegetation and paleoenvironments for Panama and adjacent Costa Rica than for any other region of northern Latin America. MATERIALS AND METHODS Extraction and processing techniques are de- scribed in Graham (1985). Slides are labeled ac- cording to core number, depth, and slide number (e.g., SL-103, 253', 1). Location of the specimens on the slides is by England Slide Finder coordinates (e.g., ESF G-40). All materials are deposited in the palynology collections at Kent State University. SYSTEMATICS Thirty-one palynomorphs are described from the Gatun Formation. One unidentified Pyrrophyta (di- noflagellate) was also recovered. Several are com- mon in other fossil floras, and for these the infor- mation is synoptic. Present ranges and ecological data are based primarily on Tryon & Tryon (1982) for ferns, and on Croat (1978) and D'Arcy (1987) for gymnosperms and monocots. General strati- graphic ranges for the Caribbean region follow Ger- all are marine sediments, sea e, and cores drilled by the Panama Canal ommission between 1950 and the 1960s have been discarded. Diameter of coin, mm. meraad et al. (1968) and Lorente (1986), with records from our previous studies listed separately. Other details on the systematics of the microfossils are discussed in Graham (1989: 51) LYCOPODIACEAE Lycopodium (type 1, Fig. 2). Amb oval-tri- angular to nearly circular, apices rounded; trilete, laesurae straight, narrow, ca. 12 um long, extend- ing % distance to spore margin, inner margin en- tire; distal surface reticulate, muri low, narrow (ca. m wide), smooth, sinuous, lumina irregular, occasionally appearing as narrow, sinuous, slitlike openings as viewed through the spore from the proximal surface, proximal surface laevigate; wall 2 um thick, slightly thinner at apices; 36-39 um. The reticulate spores of Lycopodium are char- acteristic of subg. Lycopodium (Wilce, 1972; Tryon & Tryon, 1982: 807), but the Gatun spec- imens cannot be referred to any individual species. Lycopodium (type 2, Fig. 3). Amb oval-tri- angular, apices rounded; trilete, laesurae straight, narrow, ca. 33 um long, extending to spore margin, inner margin entire; distal surface with narrow, sinuous slitlike openings, proximal surface laevi- gate; wall 3 um thick; 54-58 um Lycopodium (type 3, Figs. 4, 5). Amb tri- angular to slightly concavo-triangular, apices rounded; laesurae straight, narrow, 22-24 um long, 192 Annals of the Missouri Botanical Garden FIGURES 2-13. Fossil spores from the Gatun Formation. — 2. Lycopodium type 1, SL-103, 253' 4, ESF С-40. — 3. Lycopodium type 2, SL-103, 253', 5, ESF 1-22. —4, 5. Lycopodium type 3, SL-103, 253', 7, ESF K-21; SL- 103, 253', 10, ESF P-26.—6. Lycopodium type 4, SL-103, 253', 4, ESF R-26.— 7. Selaginella, SL-103, 253', Volume 78, Number 1 1991 Graham 193 Pliocene Communities—Monocots extending 34 to nearly to spore margin; distal sur- face with low verrucae/rugulae to nearly flat sur- face with spaces in between appearing as circular to narrow, elongated, slitlike openings, proximal surface laevigate; wall 2-3 um thick; 50-60 um. These spores are similar to those of subg. Cer- nuistachys (Wilce, 1972; Tryon & Tryon, 1982: 807), including such species as L. (Huperzia fide Ollgaard, 1987) hippuredeum Christ. Lycopodium (type 4, Fig. 6). Amb oval.tri- angular to nearly circular, apices rounded; trilete, laesurae straight, narrow, ca. 18 um long, extend- ing ?4 to nearly to spore margin, inner margin entire; distal surface foveolate, pits circular, 0.5 um to slightly less in diameter, moderately dense, evenly spaced, proximal surface laevigate; wall 2 ит thick; 45-50 um These spores are similar to those of subg. Selago (Wilce, 1972; Tryon & Tryon, 1982: 807), in cluding L. linifolium L. There are about 125 species of Lycopodium in the Neotropics growing as terrestrial or pendant epiphytes in a variety of habitats, including wet to damp sites in savannas, rainforests, cloud forests, and paramo. The stratigraphic range for the various types of Lycopodium spores has not been summarized. Spores of Lycopodium have been reported pre- viously in our studies from the San Sebastian, Us- cari, Culebra, La Boca, and Paraje Solo formations. SELAGINELLACEAE Selaginella (Fig. 7). Amb oval-triangular to nearly circular, apices rounded; trilete, laesurae straight, narrow, ca. 12-15 um long, extending 34 to completely to spore margin, inner margin entire; distal surface echinate, spines ca. 3-6 um long, base broad, occasionally giving reticulate appear- ance to spore surface, proximal surface laevigate to sparsely and minutely echinate; wall 2 um thick; 30-38 um (excluding spines). There are about 250 species of tropical Amer- ican Selaginella, growing primarily in mesic, damp, shaded forests and oak-pine woods, usually at el- evations between sea level and 1,500 m, but with some species extending to 4,000 m (Tryon & Tryon, 1982: 820). Spores of the Selaginella type are known in the stratigraphic literature as Echitriletes and range from Paleozoic to Recent. They are present in low percentages in the Gatuncillo, San Sebastian, Us- cari, Culebra, Cucaracha, La Boca, and Paraje Solo formations. CYATHEACEAE Alsophila (Figs. 8-11). Amb triangular, api- ces rounded; laesurae straight, narrow, ca. 18 um long (in smaller specimens, to ca. 28 um in larger ones), inner margin entire, bordered by lip 4-6 um wide, outer margin of lip frequently bordered by small (0.5 um or less), circular pits; laevigate; wall 2 um thick; 45-65 um. This description is based on the type illustrated in Figure 8, which is the most common one re- covered from the Gatun sediments. Variations in- clude larger size (Fig. 9), lip faint (Fig. 11) to nearly absent (Fig. 9), pits bordering laesurae faint to conspicuous, laesurae closed (Fig. 9) or open (Fig. 10), and occasional branching of the laesurae api- ces (Fig. 11). The designation of fossil spores as Alsophila vs. Cyathea is almost arbitrary because of the differing classifications of tree ferns (e.g., Holttum 1964, 1965; Tryon, 1970; Tryon & Tryon, 1982: 182) and the consequent labeling of herbarium speci- mens from which spore reference material is ob- tained. Laevigate or nearly laevigate fossil spores are provisionally referred to A/sophila, while mi- cropunctate ones (Fig. 14) are referred to Cyathea (Graham, 1989: 55). Alsophila is a genus of about 13 tropical Amer- ican species. It usually grows in cloud forests, wet montane or elfin forests, primarily as an understory species, at elevations from about 1,000 to 2,000 m. Spores of Alsophila, as defined here, are known from the San Sebastian and Paraje Solo formations (as Cyathea in both). Cnemidaria (Figs. 12, 13). Amb oval-trian- gular, apices rounded; trilete, laesurae straight, narrow, 20-25 um long, extending ca. 34 to nearly to spore margin, inner margin entire, bordered by occasional faint lip 3-5 um wide; foveolate, smaller pits (1-6 um diam.) numerous, irregularly scat- tered over spore surface, larger pits (12-15 um <— 9, ESF S-41.— 8-11. Alsophila, SL-103, 253’, 11, ESF 0-22; 51-103, 253’, 4, ESF N-17; 51-103, 253’, 2, ESF J-42; SL-103, 253', 16, ESF Е-32. — 12, 13. Cnemidaria, SL-103, 253', 15, ESF Y-24; SL-103, 257', C-12. All photographs taken at 400 х; size in microns given in descriptions. 194 Annals of the Missouri Botanical Garden typically 3, symmetrically arranged one each be- tween arms of trilete scar is spore periphery; wall 2-3 um thick; 40-5 Cnemidaria is a кй poem tree fern genus of 25 species (Tryon & Tryon, 1982: 209), typically growing in deep-shaded, wet, montane forests, and along stream banks and waterfalls, at elevations up to about 1,500 m Spores of Cnemidaria have been reported from the San Sebastian (under the name Hemitelia), Uscari, and Paraje Solo (as Hemitelia) formations. Cyathea (type 1, Fig. 14). apices rounded; trilete, laesurae straight, narrow, 30 um long, extending ?4 distance to spore margin, inner margin entire, bordered by faint lip 3-4 um wide; distal surface micropunctate, proximal sur- face laevigate; wall 2 um thick; 58-65 um. hese fossil spores are similar to spores of C. petiola (Hook.) Tryon and C. speciosa Willd. (Tryon & Tryon, 1982: 207, fig. 2 Cyathea (type 2, Figs. 15-17). Amb oval- triangular, margin lobate due to projecting sculp- ture elements, apices rounded; trilete, laesurae straight, narrow, 18-20 um long, extending 34 to nearly to spore margin, inner margin entire; distal surface with conspicuous, moundlike verrucae, 6– 7 um diam., smooth, proximal surface laevigate; wall 3 um thick; 45-62 um. These spores are similar to those of several species of Cyathea, including C. caracasana (КІ.) Domin (Tryon & Tryon, 1982: 207, fig. 17). As defined by Tryon & Tryon (1982: 204) Cyathea is a tropical American genus of 40 spe- cies, growing primarily in montane and cloud for- Amb triangular, ests, more rarely in lowland rainforests, and typi- cally at elevations between 1,500 and 2,000 m Spores of Cyathea are known from the Culebra, Cucaracha, La Boca, and Paraje Solo formations. OPHIOGLOSSACEAE Ophioglossum (Fig. 18). Amb oval-triangular to nearly circular, apices rounded, margin lobate due to projecting sculpture elements; trilete, lae- surae straight, narrow, 32- m long, extending to spore margin, inner margin entire; verrucate, apices of verrucae rounded to broadly pointed, spore surface between verrucae with irregular to elongated, slitlike punctae; wall 3-4 um thick; 70— um. About 10 species of Ophioglossum occur in the American tropics. They grow in a wide variety of habitats, often in open woods, savannas, and marshy areas, and usually at elevations between 500 and 3,000 m (Tryon & Tryon, 1982: 36). Stratigraphic palynologists place Ophioglossum spores in the artificial genus Foveotriletes, and in the Caribbean region these range from the Upper Cretaceous (Maastrichtian) to Recent. They have not been reported previously from our Gulf/Ca- ribbean Tertiary assemblages. POLYPODIACEAE Grammitis (Fig. 19). Amb oval-triangular, apices rounded, margin faintly lobate due to pro- jecting sculpture elements; trilete, laesurae straight, narrow, 17-19 um long, extending to or nearly to spore margin, inner margin entire; verrucate, ver- rucae small (ca. 2 x 1.5-2 um), apex often trun- cate, densely arranged; wall 2 um thick; 42-46 There are about 175 American species of Gram- mitis, growing primarily in wet habitats in cloud and elfin forests at elevations typically between 2,000 and 3,000 m. The fossil spores are similar to those of several living species, including С. kal- breyeri (Baker) Morton, common in Central Amer- ica. Spores of Grammitis have not been reported previously from our Gulf/Caribbean assemblages. PTERIDACEAE Ceratopteris (Fig. 20). Amb oval-triangular to circular, apices rounded; trilete, laesurae straight, narrow, 30-32 um, extending Yê to % distance to spore margin; striate, striae in numerous, broad, flat, smooth bands 4-5 um wide, area between bands 1-2 um wide; wall 2-3 ит thick; 68-72 um. In tropical America Ceratopteris consists of two closely related species of floating ferns [C. pteri- FIGURES 14-25. SL-103, 253', 6, ESF X-4 L-103, 253', 12, ESF M-4 ب fr e Gatun Еогтапоп. — 14. Cyathea 1, SL-103, 255.5', 5, ESF М-32. — 15-17. Cyathea type 2, st 103, 253, 9, ESF H-22, D-41; SL-103, SF N-4 —19. Grammitis, SL-103, 253', 6, ESF Q-38 Q-30.—21. Pteris type 1, SL. 103, 253', 1, ESF U-41.— 22. Pteris type 2, SL- Pteris type 3, SL-49, 222. 2 1, ESF J-18.— 24. Pteris type 4, SL-103, 253', 1, ESF H-4 18. Ophioglossum, 0 antena: SL-48, 162’, 1, 103, "Pu 7, ESF K-24.— 23. — 25. Pteris type 5, Volume 78, Number 1 Graham 195 Pliocene Communities—Monocots 196 Annals of the Missouri Botanical Garden doides (Hook.) Hieron. and C. thalictroides (L.) Brongn.] commonly found in fresh water but oc- casionally in brackish water habitats. They occur in ditches, lagoons, rivers, lakes, or rooted in wet soil or shallow water, usually between sea level and 300 m (Tryon & Tryon, 1982: 315). No consistent differences could be detected between the spores of the two modern species. Fossil specimens are known in the stratigraphic literature as Magnastriatites and in the Caribbean area were previously listed as ranging from the base of the Oligocene to Recent, but the Gatuncillo report (Graham, 1985) extends the genus back to the middle(?) to upper Eocene. Other records are from the Cucaracha and Paraje Solo formations. Pteris (type 1, Fig. 21). rounded; laesurae straight, narrow, 20-24 um long, extending to spore margin, inner margin entire; conspicuous flange present, 8- 10 um wide, smooth, outer flange margin entire; distal surface laevigate, proximal surface with few low, flat, shieldlike ver- rucae; wall 2 um thick; 52-65 um. Amb oval, apices Pteris type-2 spores (Fig. 22) are larger (80– 86 um), and have a few, large verrucae scattered over the distal surface. Type 3 (59-64 ит, Fig. 23) have numerous, low, flat verrucae with sinuous slitlike spaces between the verrucae. Type 4 (47- 66 um, Fig. 24) have low, smooth, elongated, sin- uous verrucae/rugulae. Type 5 (47-51 um, Fig. 25) have numerous, blocky, angular verrucae. There is a considerable range in variation among modern spores, both within and between species, and these fossil morphotypes cannot be referred consistently to any single species. out 55 species of Pteris occur in the Amer- ican tropics in a wide variety of habitats. Most are found *'in wet forests, at the edge of clearings, in thickets, sometimes in cloud forests or in gallery forests, or occasionally on cliffs," generally at lower altitudes between sea level and about 2,000 m (Tryon & Tryon, 1982: 338). The spores are known in the stratigraphic lit- erature as Polypodiaceoisporites (Lorente, 1986: 154) and are known from the Gatuncillo, San Se- bastian, Uscari, Culebra, Cucaracha, La Boca, and Paraje Solo formations. OTHER TRILETE FERN SPORES Several trilete fern spores were recovered that could not be identified. Two of the most common and/or distinctive ones are described below. Туре 1 (Figs. 26, 27). Amb triangular, apices rounded; trilete, laesurae straight, narrow, 20-24 um, extending to or nearly to spore margin, bor- dered by faint lip 6-8 um wide, inner margin entire; spore surface with irregular, slightly raised seg- ments 9-12 um diam., each with numerous, minute baculae/clavae; wall 3 um thick; 38-52 um. ese specimens were originally interpreted as slightly corroded Cyathea spores, but ultimately many were recovered with the same characteristic ornamentation. The spore has not been recorded previously in our Gulf/Caribbean Tertiary assem- blages. Type 2 (Fig. 28). Amb circular; trilete, lae- surae straight, narrow, 16-18 um long, extending 34 to nearly to spore margin, inner margin entire; echinate, echinae small (ca. 0.5 um), numerous and densely distributed; wall 1-2 um thick; 43- 47 um. DRYOPTERIDACEAE Ctenitis (type 1, Fig. 29). Reniform; mono- lete, laesurae straight, narrow, 19-23 um long, inner margin entire; echinate, echinae numerous, delicate, narrow, ca. 2 um long, curved, frequently appearing appressed to spore surface; wall 3 um thick; 38-40 x 24-29 um. In type 2 (30-38 x 28-32 ym, Fig. 30) the spines are more straight and do not appear ap- pressed to the spore surface. In type 3 (40-45 x 34-36 um, Fig. 31) the spines are longer (ca. 3- 4 um) and somewhat fewer in number. Spores of several modern species also exhibit these variations [Ctenitis (= Megalastrum) acrosorum (Hieron.) Smith & Morgan (1987)], and the fossils cannot be referred to any one species. About 75 species of Ctenitis occur in the American tropics, primarily in wet forests, and most frequently at elevations between about 500 and 1,800 m (Tryon & Tryon, 1982: 463). Spores of Ctenitis have not been reported previously in our Gulf/Caribbean assem- ages. OTHER MONOLETE FERN SPORES In addition to Ctenitis, several other monolete fern spores were recovered that could not be iden- tified. Several of the most common and/or dis- tinctive ones are described below. Type 1 (Fig. 32). Reniform; monolete, lae- surae straight, narrow, 16-20 um long, inner mar- gin entire; laevigate; wall 2 um thick; 55-48 x 33-39 um. Monolete laevigate spores lacking the orna- mented perine (Laevigatisporites) are common in fossil deposits, ranging from Paleozoic to Recent, Volume 78, Number 1 Graham 1991 Pliocene Communities—Monocots 197 FIGURES 26-38. Fossil spores and pollen from the Gatun Formation.— 26, 27. Trilete fern spore type 1, SL- 103, 253’, 6, ESF 41; SL-103, 253’, 2, ESF H-48.— 28. Trilete fern spore type 2, SL-103, 253’, 9, ESF 0-28. — 29. Ctenitis type 1, SL-103, 253’, 10, ESF N-52.— 30. Ctenitis type 2, 51-103, 257’, 1, ESF N-15.— 31. Ctenitis type 3, SL-103, 253’, 12, ESF М-32. — 32. Monolete fern spore type 1, SL-103, 253’, 1, ESF 0-30. —33. Monolete fern spore type 2, SL-103, 253’, 1, ESF G-46.—34. Monolete fern spore type 3, SL-103, 253’, 9, ESF Q-26.— 35. Monolete fern spore type 4, SL-103, 253’, 15, ESF T-44.— 36. Monolete fern spore type 5, SL-103, 253’, 1 ESF T-32.— 37. Podocarpus, SL-103, 253', 7, ESF J-30.— 38. Gramineae, 51-103, 255.5', 5, ESF Q-36. 198 Annals of the Missouri Botanical Garden and are produced by numerous Blechnaceae, Poly- podiaceae, and Pteridaceae. Type 2 (69-88 x 45-68 um, Fig. 33) differs from type 1 in being moderately verrucate, while type 3 (85-88 x 55-58 um, Fig. 34) has more conspicuous verrucae that taper to blunt apices. Both are placed in the artificial genus Verrucati- sporites, range from Paleozoic to Recent, and are produced by numerous Blechnaceae, Polypodi- aceae, and Pteridaceae. Туре 4 (48-52 ит in length, Fig. 35) is similar to some Ctenitis spores, but type 4 spores are larger, and the spines (3 um) are more coarse (broader) and the apices are blunt. Type 5 (42-48 x 30-36 ит, Fig. 36) has small (ca. 0.5 um), fine, clavate to baculate/echinate sculpture elements. None of these can be referred to types presently in our reference collection. Types 1-3 are common in all our Gulf/Carib- bean assemblages, while types 4 and 5 have not been recovered previously. PODOCARPACEAE Podocarpus (Fig. 37). Віѕассаїе; monocol- pate, colpus situated between air sacs, margin dif- fuse, 25-30 um long, body amb circular, finely verrucate, some sculpture elements elongated forming rugulae, occasionally anastomosing, outer margin lobate, 40-44 um, body wall 1 um thick; air sacs two, hemispheric, 45-48 x 36-39 um, large in relation to body size (compared to other vesiculate gymnosperm pollen), reticulate, muri low, narrow (ca. 1 um), smooth, lumina irregular in shape and diameter (larger 5-6 um); overall di- mensions (including air sacs) 75-90 x 40-50 um. ree species of Podocarpus occur in Panama (D'Arcy, 1987): P. guatemalensis Standl. (also in Costa Rica and Guatemala), P. magnifolius Buch. & N. E. Gray (Сшапаз, Venezuela), and P. olei- folius D. Don. (Costa Rica, Colombia, Ecuador, Peru, Venezuela). The trees grow primarily in mid- to high-altitude cloud forests at elevations between about 1,000 and 2,000 m and are frequently as- sociated with temperate elements such as Alnus, Пех, Myrica, and Quercus. The fossil specimens are similar to P. oleifolius, but pollen of the other species are not represented in our reference col- lection, or by pollinating material at F, MO, and US Podocarpus has been reported previously from the San Sebastian, Uscari, and Paraje Solo for- mations. It is also known from the Oligo—Miocene Simojovel Group of Chiapas, Mexico (Langenheim et al., 1967). GRAMINEAE (Fig. 38) Spherical, amb circular; monoporate, pore cir- cular, 3-4 um diam., inner margin entire, sur- rounded by annulus 3-4 um wide; psilate to sca- brate; tectate, wall 1-2 um thick; 36-44 um. The Gramineae are a stenopalynous family, and the fossil pollen cannot be referred to individual genera. In the stratigraphic literature they are known as Monoporites, and in the Caribbean re- gion range from the Paleocene but become abun- dant beginning only in the lower Еосепе. Grass pollen occurs in the Culebra, La Boca, and Paraje Solo formations. It has also been reported (as Gra- minidites) from the Eocene of Cuba (Areces-Mal- lea, 1988). PALMAE There is considerable diversity in pollen of mod- ern palms (Ferguson, 1986; Thanikaimoni, 1970), but still many types are common to more than one genus and most fossils cannot be identified below the level of family. Two types were recovered from the Сашп sediments. Type 1 (Figs. 39, 40). Amb cuneiform (wedge- shaped, broadest above equator); monocolpate, col- pus straight, 40-44 wm long, extending nearly entire length of grain, sides parallel for most of colpus length, apices blunt to slightly rounded, inner margin minutely dentate; scabrate; tectate, wall 2 um thick; 47-58 x 33-38 um. The pollen is similar to that of Aiphanes, Mani- caria, Reinhardtia, and the monocolpate forms of predominately trichotomosulcate genera such as Acrocomia. Similar pollen has been reported pre- viously from the Gatuncillo, Culebra, Cucaracha, La Boca, and Paraje Solo formations. Type 2 (Figs. 41, 42). Amb oval; monocol- pate, colpus straight, 33-35 шт long, extending nearly entire length of grains, sides parallel for most of colpus length, apices blunt to slightly round- ed, inner margin entire to minutely dentate; retic- ulate, muri low, smooth, relatively thick (1.5 um) compared to diameter of lumina (2 um), underlying columellae evident giving reticulum faintly beaded appearance, occasional free sexine elements within umina; tectate-perforate, wall 2 um thick; 43-45 x 35-37 um. This pollen type is similar to that of Colpothri- nax, Cryosophila, and Neonicholsonia. Pollen with these characteristics has not been reported previ- ously in our Gulf/Caribbean assemblages. Моште 78, Митбег 1 1991 Graham 199 Pliocene Communities—Monocots FIGURES 39-42. SL-103, 253', 6, S-44.— 41, 42. Palmae type 2, SL-103, 253', 9, ESF Q-23; SL-103, 253', 5, ES In addition to the 31 spore and pollen types described here, 51 kinds of dicotyledonous pollen have been identified from the Gatun Formation, including 13 with no previously known fossil rec- ord. There are also 27 unknowns presently under study. This makes the Gatun assemblage, with 110 types, including one unidentified Pyrrophyta (di- noflagellate), the second largest Tertiary fossil flora known from northern Latin America (the largest is the middle Pliocene Paraje Solo flora from Ve- racruz, Mexico, with approximately 124 types; Graham, 1976). The remaining descriptions, and interpretation of the flora, be presented in subsequent papers in the series. LITERATURE CITED ARECES-MALLEA, А. 1988. Palinomórfos de la costa del o de Norteamérica en el Eocene medio de Cuba. Revista Tecnológica 17: 15-26. BARTLETT, А. S. & Е. S. BARGHOORN. 1973. Phyto- geographic history of the Isthmus of Panama during Graham (editor), Vegetation and Vegetational n of Northern Latin America. Elsevier Publ., Am dam СООКЕ, C. W., J. GARDNER & W. P. WoopRINc. 1943. Correlation of the Cenozoic Er pue of the Atlantic nd the Caribbean region. Bull. Geol. Soc. Amer. 54: 1713-17 197 Flora of Barro Colorado Island. Stanford Univ. Press, Stanford, California D'Arcy, W. С. 19 ora of Рапат n Спасића and сане oe Syst. Bot. Missouri Bot Gard., Vol- сток I. K. 1986. Observations on the variation in pollen morphology of Palmae and its significance. nad. J. Bot. 64: 3079-3090. фм жар. J. H., C. A. Hoppinc & J. MULLER. 1968. Palynology of Tertiary sediments from di areas. Rev. Palaeobot. Palynol. 6: 189-348. GRAHAM, А. 1976. Studies in neotropical pe I. The Miocene communities of Veracruz, Mexico. Ann. Missouri Bot. Gard. 66: 572- 576. Fossil pollen from the Gatun Formation. — 39, 40. Palmae type 1, SL-103, 253', 7, L P-23; Е J-4 1985. Studies in neotropical paleobotany. IV. The Eocene communities of Panama. Ann. Missouri Bot. Gard. 72: 504-534. —— ———. 1987. Mi iocene communities and paleoenvi- ronments of southern Costa Rica. Amer. J. Bot. 74: о Studies in casus ine does V. The lower Mods comm of Pan —the Culebra Formation. Ann. Missouri Bot. Card. 75: 1440-1466 8b. Studies in е ет МТ. "The lower Miocene commun of Pan the Cucaracha Formation. Апп. Missouri Bot. Gard. 75: 1467-1479 . 1989. Studies in neotropical a. € УП. The lower Miocene communities o ma — е La Boca Formation. Ann. Missouri Bot. "Gard. 76: 50- 66. ‚ В. Н. STEWART & J. L. STEWART. 1985. Stud- ies in neotropical e III. The Tertiary com- munities of Panama— geology of the pollen-bearing тоне Апп. Missouri = Gard. 72: 485-503. Ногтт В. Е. tree ferns of the genus lil in Australasia b the Pacific. Blumea 12: 241-274. 965. Tree ferns of the genus Cyathea Sm. in Аде (excluding Malaysia). Kew Bull. 19: 463- 487. Horn, S. P. 1985. Preliminary pollen analysis of Qua- > rnary — from Deep Sea Drilling Project e 565, western Costa Rica. Initial Reports Deep Sea Drilling Projet 84: 533-547. LANGENHEIM, J. H., B. L. HACKNER & A. S. BARTLETT. 1967. Mangrove дај а at the depositional site of Oligo-Miocene amber from Chiapas, Mexico. Bot Mus. Leafl. 21: 289-324. LoRENTE, M. А. 1986. Palynology and palynofacies of the upper Tertiary in Venezuela. Dissertationes Bo- tanicae, Bd. 99, J. Cramer, Berlin. OLLGAARD, B. 1987. A revised classification of the Ly- copodiaceae s. lat. Opera Bot. 92: 153-178. SMITH, А. К. & К. C. Moran. 1987. New combinations in v an e (Dryopteridaceae). Amer. Fern J. TU B 3 H. & 1. L. STEWART (with the мр ти of W. P. Woodring). 1980. Geologic Map o Panama Canal and Vicinity, Republic of Panama, Scale: 1:100,000. U.S. Geol. Surv. Misc. Invest. 200 Annals of the Missouri Botanical Garden Map 1-1232. [Map also included i | Woodring, 1982, Profess. Pap eol. Surv. тшшн, С. 1970. ез d een palynologie | о Inst. Franc. Pondichéry, Trav. Sec i. Tech. XI: 1-286. азон. 'R. 1970. The classification of the Cyatheaceae. Contr. Gray Herb. 200: 3-53. & A. F. Tryon. 1982. Ferns and Allied Plants, with Special Reference to Tropical America. Spring er-Verlag, New York Vokes, E. Н. 1983. Additions to the Typhinae (Gas- T Muricidae) of the n Formation, Pana- a. Tulane Studies Geol. Pd 17: 123-130. Wice, J. H. Lycopod spores, 1. General spore joue and the generic segregates of Lycopodium. : 65-79. и. W. P. 1957-1982. Geology and paleon- tology of Canal Zone and adjoining parts of Panama. Profess. Pap. U.S. Geol. Surv. 306A-F. STUDIES IN NEOTROPICAL Alan Graham* PALEOBOTANY. IX. THE PLIOCENE COMMUNITIES OF PANAMA—ANGIOSPERMS (DICOTS) ABSTRACT Fifty-one fossil d T pollen types and 27 unknowns are described from the Pliocene Gatun Formation of Panama. These are Cymbopetalum, llex, cf. Aguiaria, Bernoullia, Ceiba, Pseudobombax, Bursera, cf. Bucida, Combretum /Terminalia, Mutisieae type, cf. Cionosicys, Ericaceae (types 1, 2), Alchornea, cf. Glycydendrum, cf. Jatropha, cf. Stillingia, Ве Alfaroa / Oreomunnea, Acacia, Crudia, Erythrina, na Loranthaceae (types 1, 2), Malpighiaceae (types 1, 2), Hampea / Hibiscus, Melastomataceae, Cedrela, abomba, Hauya, izophora, Chomelia type, Cosmibuena, Faramea (types 1, Allophylus, Capania, Paullinia, Serjania, cf. Bumelia, Symplocos пр 1, 2), Mortoniodendron, Aegiphila, and Petrea. Spores of the lycopods and ferns, and pollen of the gymnosperms and monocotyledenous angiosperms, have been described previously. Paleoecological and paleoenvironmental ale will be presented in a concluding paper This paper completes the systematic treatment of plant microfossils recovered from the Pliocene Gatun Formation of Panama. Introductory mate- rial and the systematics of other groups (ferns, gymnosperms, and monocots) are given in Graham (1991), new records of previously unreported taxa are provided in Graham (1990), and the geology of the pollen-bearing deposits is summarized in Graham et al. (1985). Fifty-one dicotyledonous pollen types, together with 27 unknowns, are rec- ognized here. Thirty-one fern, gymnosperm, and monocot microfossils, and one Pyrrophyta (dinofla- gellate), were previously reported, bringing the total number of taxa to 110. Numerical representations, and paleocommunity and paleoenvironmental re constructions will be presented in a concluding MATERIALS AND METHODS Extraction and processing techniques are de- scribed in Graham (1985). Slides are labeled ac- cording to core number, depth, and slide number (e.g., 51-103, 253’ [feet], 2). Location of the specimens on the slides is by England Slide Finder coordinates (e.g., ESF L-44). Details of interwall morphology (e.g., homogeneous; columellae just visible) are described as seen in median optical section with light microscopy at the magnification indicated (e.g., 400 x). All materials are deposited in the palynology collections at Kent State Uni- versity. SYSTEMATICS ANNONACEAE Cymbopetalum (Figs. 1, 2). Oblate, amb cir- cular; nonaperturate(?); psilate; tectate-perforate, tectal perforations large, circular to irregular in outline, constituting ca. 40–50% of grain surface, columellae conspicuous, ca. 15 um long; 180 um. Cymbopetalum consists of 11 to 13 species of trees and sh distributed from Mexico to tropical South America, with three species in Panama: C. brasiliense (Vell.) Benth., C. costaricense (J. D. Smith) R. E. Fries, and C. languipetalum Schery. Pollen of the modern species has been studied by Walker (1971a, b, c, 1972). Cymbopetalum is known only from the Gatun assemblage. | c pup و‎ acknowledges field assistance and information on the geology of Panama provided by R. art (Chief Geologists, Panama C S. Ste tew Botanical Garden). x William Ek (Exxon Com n Маша. Research was genial by NSF ¢ Би св. 5671, РЕВ- 8007312, DEB-82055926, BSR-8500850, and BSR. 8819 анин ын of Biological Sciences, Kent State University, Kent, Ohio 44242, U.S.A. ANN. MISSOURI Bor. Garp. 78: 201-223. 1991. 202 Annals of the Missouri Botanical Garden ma SS ET S, ез ПАЛИ ^ я, уче, 53 { > K я S AS < Yo rZ рт > © A > С 2 ћ b 49 AZ 29 tf 8 FIGURES 1-8. Fossil pollen from the Gatun Formation.—1, 2. Суть ', 16, ESF N-26; ; .— 1, 2. Cymbopetalum, SL-103, 253', 16, і 51-103, 253', 9, ESF J-29.— 3. Ilex, 51-103, 253’, 5, ESF Е-18.—4, 5. Combretum/ Terminalia, SL-103, 253, 7, ESF D-18.— 6. cf. Aguiaria, SL-103, 253', 6, ESF O-40.— 7. Bernoullia, SL-103, 253', 9, ESF сл е Pseudobombax, SL-103, 253', 14, ESF Р-44. All photographs taken at 400 x; size in microns given in descriptions. AQUIFOLIACEAE 3 I- straight, 18-20 um long, apices acute, inner ma Пех (Fig. 3). Oblate-spheroidal, amb oval to gin frequently appearing undulate due to overlying circular, tricolporoidate (colpi occasionally ob- sculpture elements, pore poorly defined, frequently scured by sculpture elements), colpi meridionally obscure, situated at midpoint of colpus; intectate, elongated, equatorially arranged, equidistant, clavate, wall 3-4 um thick; 32-34 * 29-3 | | Volume 78, Number 1 1991 | Graham | 203 Pliocene Communities—Dicots llex is widespread in northern Latin America, occurring in premontane to montane, wet to drier forest habitats. Pollen of Пех is found in low percentages in all of our Tertiary microfossil assemblages (Gatuncillo, San Sebastian, Uscari, Culebra, Cucaracha, La Boca, and Paraje Solo formations). Other occur- rences are summarized in Graham (1985, 1987). BOMBACACEAE cf. Aguiaria (Fig. 6). Oblate to рего ме, amb circular; tricolpate, colpi equatorially arranged, meridionally elongated, equidistant, situated in in- terapical area, short (6-8 um apex to equator), costae colpi 3—4 um wide; finely reticulate (di- ameter of lumen 1 um or less); tectate-perforate, wall 2 um thick, individual columellae evident in median section (400 х magnification); 34-36 um. Pollen of the Bombacaceae is frequent in low percentages in Gulf/Caribbean Tertiary deposits. Assignment of some types to modern genera is difficult because of overlap with the Sterculiaceae and Tiliaceae, and the occurrence of several fossil lorms that do not correspond to the pollen of any modern genus in our reference collection or in the literature (Graham, 1989: 59). The specimens are of the Aguiaria type (Tsukada, 1964, pl. VII, figs. 10-14; Nilsson & Robyns, 1986, figs. 7f, g), a monotypic genus of Brazil. Similar specimens are known from the La Boca Formation (Graham, 1989), and a smaller, poorly preserved one from the Uscari Formation (Graham, 1987) Bernoullia (Fig. 7). Oblate to peroblate, amb triangular, apices rounded; tricolpate, colpi equa- torially arranged, meridionally elongated, equidis- tant, situated in interapical area, short (6-8 арех to equator), costae colpi 4-5 um wide; finely reticulate (diameter of lumen 1 um or less), tectate- perforate, wall 3 um thick, columellae evident in median section (400 x magnification); 43-47 um. Ber noullia comprises three species of large, tropical trees distributed from Mexico to South America, The pollen specimens are similar to the pollen of B. flammea Oliver of Mexico and Central America, А specimen from the San Sebastian For- mation (Graham & Jarzen, 1969, fig. 15) described as Bernoullia is different and probably is not Ber- поша. echa (Fig. 9). Oblate to oblate-spheroidal, amb fes; E tricolpate, colpi equatorially arranged, ig "ond elongated, equidistant, straight, short 5 um apex to equator), costae colpi 4-5 um wide; reticulate, muri narrow (1.5 um), straight, smooth (appearing beaded from underlying colu- mellae), lumina polygonal, varying in size, larger commonly in polar and mesocolpial areas and 6– 8 um diam., free sexine el ts in lumina; tectate- perforate, wall 2 um thick, individual columellae evident in median section (400 x magnification); 68-74 um. The specimens are similar to pollen of Ceiba pentandra (L.) Gaertner, a large pantropical tree occurring in a wide variety of habitats. In Panama it commonly grows in the tropical moist and tropical dry forests, but also in other habitats. Microfossils are known from the La Boca Formation (Graham, 1989, fig. 34, as cf. Ceiba). Pseudobombax (Fig. 8). Oblate to peroblate, amb triangular, apices rounded; tricolpate, colpi equatorially arranged, meridionally elongated, equidistant, situated in interapical area, straight, short (8-10 um), costae colpi 5-6 um wide; re- iculate, muri smooth, narrow (l um or less), straight, lumina polygonal, diameter ca. 3 um in polar area, diminishing to 1 um or less at apices; tectate-perforate, wall 2 um thick, individual col- umellae evident in median section (400 x magni- fication); 54-58 um. In addition to difficulties noted earlier, the re- ferral of Bombacaceae microfossils to certain mod- ern genera is further complicated by nomenclatural changes among the extant taxa. In particular, Bombax, as treated in earlier taxonomic literature, included New World species. Herbarium material, pollen reference collections based on this material, and subsequent reports of microfossils from the Americas as Bombax, are frequent in the litera- ture. Robyns (1964; Nilsson & Robyns, 1986: 6), however, placed species of Bombax in Pseudo- bombax, Bombacopsis, and other genera, with Bombax considered an Old World, primarily Af- rican, genus. The Gatun specimens are similar to pollen of such species as Bombax ellipticum HBK in our reference collection (Barkley 1 7M103, Mexico, TEX), now treated as Pseudobombax el- lipticum (HBK) Dugand. In Central America Pseudobombax is repre- sented by P. septenatum (Jacq.) Dugand, distrib- uted from Nicaragua to Brazil and Peru. In Panama it occurs typically in the tropical moist forest, but is also known from tropical wet, tropical dry, and premontane moist forests (Croat, 1978: 591). Sim- ilar pollen occurs in the San Sebastian (Graham & Jarzen, 1969, fig. 22, as Bombax), and the La Boca (Graham, 1989, fig. 35) formations. Annals of the Missouri Botanical Garden 204 ` M С т | x “Ау; 457 VAM + IARA Ix Ал б NSS р: > 6 ^. ` я, post Um dum .. т rie ~ J ¿FAN 77, СА he. > 5 Ме ra 5% ALIS rm © y ak SACA А>, h > Of AREAS DON 20 12 rcd S: € BA RG #6 Е Dy ae N ROS I ee RS Su * vhi Bn ^ r, 1 1 MA YAT ! БЕКИ, > Жу С. | УА 2 ~ {А М Tt INS > AR Би + 15 ON A > uL "de rc -— LES ~. ' 19 ~ 18 Ра t < FIGURES 9-20. Fossil pollen from the Gatun Formation. —9. Ceiba, 51-103, 253’, 8, ESF D-41.— 10, 11. Bucida, SL-103, 253’, 6, ESF G-25; SL-103, 253', 1. ESF М-39. — 12. Bursera, SL-103, 253', 6, ESF V-49.— 13. Alchornea, SL-103, 253', 1, ESF T-47.— 14. Mutisieae type, SL-103, 253', 14, ESF J-35.— 15. cf. Cionosicys, SL-103, 253', 4, ESF J-21.— 16. Quercus, 51-103, 253', 9, ESF L-47.— 17. Alfaroa/ Oreomunnea, SL-103, 253', .— 18, 19. Ericaceae, 51-103, 253', 5, ESF N-26; SL-103, 253', 10, ESF K-45.— 20. cf. Jatropha. 9, ESF R-39 SL-103, 253', 1, ESF Х-34. Volume 78, Number 1 1991 Graham 205 Pliocene Communities—Dicots BURSERACEAE Bursera (Fig. 12). triangular to nearly circular; triporate, pores equa- torially arranged, equidistant, circular to slightly elongated equatorially, 3-4 um diam., faint costae pori; striato-reticulate; tectate-perforate, wall 2 um thick, individual columellae evident in median sec- tion (400 х magnification); 28-30 um. Modern pollen of the Mexican species of Bur- sera has been studied by Palacios Chávez (1984). The fossil specimens are similar to B. simaruba (L.) Sarg. et al., a tree or shrub of widespread distribution from southern Florida, through Mexico and the Antilles, to Colombia, Venezuela, and the Guianas. It is typical of the tropical moist forest, but also occurs in tropical wet, tropical dry, pre- montane moist, and premontane dry forests. Bur- sera pollen is known from the San Sebastian (Gra- ham & Jarzen, 1969, fig. 23) and Paraje Solo (Graham, 1976, figs. 63, 64) formations. Oblate-spheroidal, amb oval- COMBRETACEAE cf. Bucida (Figs. 10, 11). Prolate-spheroidal, amb circular, lobate due to thinning of wall at apertures; tricolporate with three pseudocolpi, colpi equatorially arranged, meridionally elongated, equidistant, straight, ca. 15 um long (apex to equa- tor), colpus margin entire to slightly dentate, pseu- docolpi similar, slightly shorter (ca. 12 um apex to equator), pores circular, 3 um diam., situated at midpoint of colpus; scabrate; tectate, wall 2 um thick, individual columellae just visible in median section (400 х magnification); 30 um. Bucida is a genus of four species distributed from southern Florida to Central America and the Antilles. One species, B. buceras L., is listed for Panama (D'Arcy, 1987), a small tree or shrub growing in wet forests and strand thickets at ele- vations from sea level to about 1,000 m. It has been reported as a megafossil from the Eocene Wilcox Formation of Kentucky (Berry, 1941) and from the Tertiary of the Dominican Republic (Ber- ry, 1921); however, these megafossil records have not been reevaluated using modern methods of foliar architecture and cuticular anatomy. As a microfossil it is known only from the Gatun For- mation. Similar scabrate, tricolporate grains with three pseudocolpi occur in several families of the Myrtales (Combretaceae, Lythraceae, Melasto- mataceae); hence, the provisional cf. identification. Combretum/ Terminalia (Figs. 4, 5). Prolate- spheroidal, amb circular, lobate in polar view; het- erocolpate, colpi three, equatorially arranged, me- ridionally elongated, equidistant, straight, 8-10 um long (apex to equator), colpus membrane faintly and minutely granular, margin entire to slightly diffuse, pores faint, situated at midpoint of colpus, 3—4 um diam., pseudocolpi three, alternating with colpi, slightly shorter (6-8 ит); psilate; tectate, wall 3 um thick, wall homogeneous (400 x mag- nification); 18-24 um. ollen of the Combretaceae was surveyed to determine the possibility of distinguishing between the common tropical trees Combretum and Ter- minalia (Graham, 1980). Pollen of the two genera could not be distinguished consistently, especially d in the dispersed state, and the microfossils are referred to Com- bretum/ Terminalia. Combretum is widespread in the New and Old World tropics, with about seven species in Panama. It commonly grows as a coastal or shore plant in tropical wet, tropical moist, pre- montane wet, and premontane moist forests. Ter- minalia is also widespread and is represented in Panama by about nine species, typically growing in tropical wet, tropical moist, premontane wet, and premontane moist forests. Microfossils are known from the Gatuncillo, Culebra, and Paraje Solo formations. in the limited numbers COMPOSITAE Mutisieae type (Fig. 14). Prolate to prolate- spheroidal; tricol(por?)ate (pores obscure), colpi equatorially arranged, meridionally elongated, equidistant, ca. 45-48 um long; echinate, echinae ca. 2 um long, broad at base, tectate, wall 8-9 um thick, individual columellae and two wall layers evident in median section (400X magnification), inner wall 4—5 ит thick, outer wall (including са. 2-um spines) 3-4 um thick; 68 x р Among the pollen types surveyed in our refer- ence collection and in the literature, this specimen most closely resembles Mutisia. However, this is a South American genus of high altitudes and/or arid habitats, and these conditions are not known to have existed in southern Central America during the Tertiary. Because the survey of modern genera of the Mutisieae is not complete, the specimen is provisionally referred only to the Mutisieae type of pollen. CUCURBITACEAE cf. Cionosicys (Fig. 15). cular; triporate, pores equatorially arranged, equi- distant, circular, large (12-15 um), inner margin entire; surrounded by faint, granular costae pori 3 Spherical, amb cir- 206 Annals of the Missouri Botanical Garden um wide; echinate, echinae 3-4 um long, broad- ening at base, surface between echinae faintly sca- brate; tectate, wall thin (3 ит) in relation to size of grain, frequently folded, individual columellae just visible in median section (400 X magnification); 88-94 ит. The Cucurbitaceae are eurypalynous (Ayala- Nieto et al., 1988; Campos, 1962; Dieterle, 1976; Marticorena, 1963; Melhem, 1966), and although several genera have large, echinate pollen grains, many of these are periporate. The Gatun specimens are similar to the triporate pollen of Cionosicys, but differ in that the spines are not as robust, hence the cf. identification. Cionosicys is a Caribbean genus of three species with one [C. macranthus (Pittier) C. Jeffery ] in Panama. The plant is a large vine growing in moist forests and open thickets, often along streams from Mexico to Panama (Wun- derlin, 1978). The Cucurbitaceae are poorly represented in the Gulf/Caribbean microfossil record. Probably be- cause of their large, relatively fragile pollen. pro- duced in small amounts, and their entomophilous pollination. The oldest pollen record for the family is from the Oligocene of Cameroon (Muller, 1981: 39). The family has not been reported previously from the Gulf/Caribbean Tertiary, although some large ornate forms reported as unknowns (e.g., Fig. 85) may belong to the Cucurbitaceae. Bartlett & Barghoorn (1973) reported a different form as Sechium edule-type pollen from Quaternary de- posits in Panama. ERICACEAE Type 1 (Fig. 18). Tetrahedral tetrad; individ- ual dp o bo dein in tetrad), amb rcular; colpi equatorially arranged, tact between grains); scabrate; tectate, wall 2 um thick, homogeneous (400 х magnification); indi- vidual grains 30-34 um, tetrad ca. Type 2 (Fig. 19). These grains are similar to type 1, but are smaller. The individual grains are 25-28 um diam. and the tetrads are 35-40 um. The occurrence of Ericaceae pollen in Gulf/ Caribbean Tertiary deposits has been discussed by Graham (1987: 1511). Briefly, the specimens can- not be identified to genus, but the larger grains fall within the range of variation of Cavendishia, a large genus of about 150 species of epiphytic and terrestrial shrubs distributed from southern Mexico through Central America into South America as far as Bolivia. About 30 species are recognized for Panama, where they are most common at moderate to higher altitudes. Ericaceae pollen has been re- ported previously from the Uscari Formation. EUPHORBIACEAE (Fig. 13). Oblate, amb circular; tri- colpate, colpi equatorially arranged, meridionally elongated, equidistant, straight, 10-12 um long, extending within 5 um of pole, inner margin entire, 412 | distinct operculum 3 шт wide; faintly scabrate; tectate, wall 2 um thick, homogeneous to colu- mellae occasionally just visible in median section (400 x magnification); 26-30 um Alchornea is widespread in the Neotropics, rep- resented by about eight species in Panama. It grows in tropical moist, premontane wet, and premontane rainforests at altitudes from about 300 to 2,000 Alchornea is frequently present in low per- centages in Gulf/Caribbean Tertiary microfossil floras (San Sebastian, Uscari, Culebra, Cucaracha, La Boca, and Paraje Solo formations). cf. Glycydendrum (Figs. 21, 22). Oblate to oblate-spheroidal, amb circular; tricolpate, colpi equatorially arranged, meridionally elongated, equidistant, straight, 8-10 um long, inner margin entire to minutely 2 um lobate; clavate, clavae fine, ca. long; intectate, wall 2-3 um thick; 42-45 m. Glycydendrum is one of the few genera pres- ently confined to South America (Amazonian Brazil) that appears to be represented in southern Central American Tertiary deposits. Similar pollen occurs FIGURES 21- T SL-103, 253', 2, ESF K-22.— 23. ESF K-27. us Crudia, = 103, 253', 11, Melastomataceae, 51-103, 253', zophora, 1, SL-103, 253', 8, ESF R-51.— 32. Loranthaceae type Fossil aa from the Gatun Formation.—21, cf. Stillingia, 51-103, 253', 14, ESF С-4 SF J-45.— 26. Erythrina, SL- 103, 253, 4, ESF N. 20.— genia/ Myrcia, SL-103, 253', 1, ESF K-37.— 29. E 3, 253', 7, ESF E-14.— 30. Utricularia, SL-103, 253', 12, ESF M-31.— 31. Loranthaceae type e 2, SL-103, 253', 7, ESF D-33.— 33. Malpighiaceae type 2, ESF G-28.— 28. Eu — 22, ооа 51-103, 253', 8, p . Acacia, SL-103, 257', Es 2, SL-103, 253', 4, ESF V-24.— 34. Malpighiaceae type 1, SL-103, 253', 9, ESF J-21.— 35. Cedrela, SL-103, 253', 1, ESF F-52 207 Volume 78, Number 1 Graham Pliocene Communities—Dicots Pi 4 5» „> D ЈАТ SHUN "t A < | Я r5 } + 4 + + à 208 Annals of the Missouri Botanical Garden in Tetrorchidium, but the columellae in that genus are finer and nearly echinate. Glycydendrum-type pollen is present in the Uscari Formation, and Bartlett & Barghoorn (1973) reported Glycyden- drum from Quaternary deposits in Gatun Lake, Panama. cf. Jatropha (Fig. 20). Spherical, amb cir- cular; nonaperturate; intectate, clavate, clavae rel- atively large and coarse, 3 X 4 um, frequently triangular in outline; wall 4 um thick; 56-60 um. The specimens have the wedge shape or trian- the stellate clusters or **crotonoid" pattern of these genera. The specimens cannot be identified below the level of family but are similar to pollen of Jatropha. The genus is widespread in Latin Amer- ica, and four species are recorded for Panama. The plants range from the tropical moist forest to drier vegetation types. This type of Euphorbiaceae pollen has not been reported previously from Gulf/ Caribbean Tertiary deposits. cf. Stillingia (Fig. 23). Prolate, amb oval; tricolporate, colpi equatorially arranged, meridio- nally elongated, equidistant, straight, 59-62 um ong, inner margin entire, pore situated at midpoint of colpus, equatorially elongated, 2 x 4—5 um; finely reticulate, lumina ca. 1 um diam.; tectate- perforate, wall 3 um thick, individual columellae evident in median section (400 x magnification); 68-72 x 49-51 um. These large, prolate, triangular, reticulate grains are similar to Stillingia, a primarily New World tropical genus of 25-30 perennial herbs, shrubs, or small trees (Webster & Burch, 1967). The genus is represented in Panama by S. zelayensis (HBK) Muell.-Arg., which produces pollen similar to the Gatun specimens. Stillingia has been re- ported previously from the Paraje Solo Formation (as cf. Stillingia; Graham, 1976, fig. 114). FAGACEAE Quercus (Fig. 16). Prolate-spheroidal, amb circular; tricolpate, colpi equatorially arranged, meridionally elongated, equidistant, straight, 15- um long, inner margin entire; scabrate; tectate, wall 2 um thick, homogeneous; (400 X magnifi- cation); 38 um. This is the earliest record of Quercus in Tertiary deposits of southern Central America, and it is rare in these Gatun sediments. It was more abundant in the Pliocene Paraje Solo Formation of Veracruz, Mexico, and first appeared in South America about 900,000 years ago (van der Hammen in Living- stone & van der Hammen, 1978). It was not re- corded from Gatun Lake Quaternary deposits in Panama (Bartlett & Barghoorn, 1973), but was present in late Quaternary sediments from west coastal Costa Rica (Horn, 1985). Thus, the north- to-south migration of Quercus during the Tertiary and its relatively recent introduction into southern Central America and particularly South America (late Quaternary) seems clear, based on present data (Crepet & Nixon, 1989). JUGLANDACEAE Alfaroa/Oreomunnea (Fig. 17). Oblate, amb circular; triporate, pores equatorially arranged, equidistant, circular, 3—4 um diam., inner margin entire; psilate to faintly scabrate; tectate, wall 2 um thick, homogeneous (400 x magnification); 27- 30 um. Englehardia sensu lato was earlier considered a component of warm temperate vegetation in both the New and Old Worlds. Following the treatment of Stone & Broome (1975), however, Englehardia is now treated as an Old World genus, with related taxa in the New World assigned to Alfaroa and Oreomunnea. Alfaroa is a genus of about seven species growing in the highlands of Mexico, Central America, and northern Colombia. Oreomunnea comprises two species, **one ranging from the pre- montane cloud forests of southern Mexico to cen- tral Costa Rica (O. mexicana with two subspecies) and the other (O. pterocarpa) endemic to the Rio Reventazón Valley of central Costa Rica" (Stone & Broome, 1975: 2). Alfaroa/ Oreomunnea-type pollen is known m the stratigraphic literature as Momipites and is present in the Gatuncillo, San Sebastian, Cucara- cha, La Boca, and Paraje Solo formations. It 1S also reported from the Oligo-Miocene Simojovel Group of Chiapas, Mexico (Langenheim et al. 1967). The fossil record of the Juglandaceae, 1n cluding Alfaroa and Oreomunnea, has been sum- marized by Manchester (1987). LEGUMINOSAE Mimosoideae Acacia (Fig. 24). Sixteen-cell polyad; individ- ual grains square to polygonal (compressed in рогу“ te; tec- inut 120 x 88 um (broken). Volume 78, Number 1 1991 Graham Pliocene Communities—Dicots 209 Twelve species of Acacia are found in Panama (fide Woodson & Schery, 1950), and five are listed for Barro Colorado Island (Croat, 1978). They are common in the tropical moist forest, but range into more open, drier habitats. Acacia pollen occurs in the San Sebastian, Culebra, and Paraje Solo for- mations. Caesalpinioideae Crudia (Fig. 25). Prolate; tricolporoidate, col- pi equatorially arranged, meridionally elongated, equidistant, straight, 30-35 um long, extending nearly entire length of grain, pore area faint, cir- cular, situated at midpoint of colpus; distinctly and coarsely striate, striae generally oriented parallel to long axis of grain, surface of striae psilate, mar- gin entire, occasionally appearing beaded from un- derlying pores in foot layer /endexine; tectate but with occasional separation between sculpture ele- ments, wall 2 um thick, homogeneous (400 x mag- nification); 44–48 x 28-32 um. Pollen of the Caesalpinioideae, including Cru- dia, has been studied by Graham & Barker (1981). Crudia is a South American, mainly Amazonian, genus of about 10 mostly riverine species. It is recorded in the stratigraphic literature as Stria- tocolpites cataumbus and in northern Latin Amer- ica is known from the Gatuncillo, Cucaracha, and La Boca formations. Herendeen & Dilcher (1988) reported Crudia-type pods from the Eocene of western Tennessee. Papilionoideae Erythrina (Fig. 26). Oblate, amb oval-trian- gular; triporate, pores equatorially arranged, equi- distant, circular (10-12 um, compressed open), Inner margin entire, surrounded by narrow margo са. 4-5 um wide formed by diminution of sculpture bordering pores; reticulate, muri thin (1 um), low, oth, sinuous, lumina polygonal, ca. 6-7 um diam. in polar area, becoming finer toward equator especially toward pores; tectate, wall 2-3 um homogeneous (400 x magnification); 34-37 с Pollen of the Erythrineae has been studied by raham & Tomb (1974, 1977). This type of Er- qns pollen is produced by several Central "pn species, including E. caribeae Krukoff E arneby, E. castillejiflora Krukoff & Barneby, + ш Krukoff, E. cobanensis Krukoff K arneby, E. fusca Loureiro, E. guatemalensis ili off, E. williamsii Krukoff & Barneby, and "tie The plants occur through a wide range of tats, including tropical wet, tropical moist, and premontane rainforests, from low to moderate el- evations. Erythrina fusca sometimes forms pure stands in freshwater marshes (Holdridge, 1970 in Croat, 1978: 473). LENTIBULARIACEAE Utricularia (Fig. 30). 'Oblate-spheroidal, amb circular; stephanocolpate, colpi equatorially ar- ranged, meridionally elongated, equidistant, ca. 15- 18 um long (apex to equator), 12 in number, inner margin entire; psilate; tectate, wall 2-3 um thick, homogeneous (400 x magnification); 42—45 um. Sixteen species of Utricularia are listed for Pan- ama (D'Arcy, 1987), with two (U. foliosa L.; U. obtusa Sw.) occurring in the waters surrounding Barro Colorado Island. It is a freshwater, perennial, stoloniferous herb of widespread distribution from temperate to tropical regions in both the New and Old Worlds. Pollen has been reported from the Miocene of Senegal (Médus, 1975), and from the Pliocene Paraje Solo Formation of Mexico (Gra- ham, 1976). LORANTHACEAE Type 1 (Fig. 31). Oblate, amb concavo-tri- angular, apices blunt; tricolpate, colpi equatorially arranged, meridionally elongated, equidistant, straight, 25-27 um long (pole to apex), syncolpate, inner margin entire; mesocolpium baculate to slightly echinate, becoming psilate at apices; in- tectate in mesocolpium, tectate at apices, wall 1.5 pm thick, tectate portion homogeneous (400 х magnification); 42-46 um. syncolpate, inner margin entire; finely striate with minute echinae between striae; tectate, wall 1.5 um thick, homogeneous; 34-36 um. Both specimens are of the Loranthaceae type but cannot be referred to any one modern genus. Type 1 is generally similar to Aetanthus, and type 2 to Psittacanthus. Modern neotropical Lorantha- ceae pollen has been studied by Feuer & Kuijt (1979, 1980, 1985), and fossil pollen has been reported previously from the San Sebastian and Paraje Solo formations. MALPIGHIACEAE Type 1 (Fig. 34). Oblate-spheroidal, amb cir- cular; periporate with colpal arcs connecting pores, pores circular, 4-6 um diam., inner margin entire, 210 Annals of the Missouri Botanical Garden inner margin of colpal arc minutely dentate, oc- casional free sexine elements on colpal membrane; scabrate; tectate, wall 3-4 um thick, individual columellae evident in median section (400 х mag- nification); 36-38 um. Type 2 (Fig. 33). Oblate-spheroidal, amb cir- cular; periporate with colpal arcs connecting pores, pores circular, 3 um diam., inner margin entire, inner margin of colpal arc minutely dentate, oc- casional free sexine elements on colpal membrane; scabrate, with occasional larger, low verrucae; tec- tate, wall 4-5 um thick, homogeneous (400 х mag- nification); 44—46 The Malpighiaceae are frequent in Gulf/Carib- n low percentages. Although the family is moderately eurypalynous bean Tertiary deposits, but i » most fossil pollen cannot be referred consistently to any one modern genus. Malpighiaceae pollen has been reported previously from the Gatuncillo, Uscari, Culebra, La Boca, and Paraje Solo for- mations. MALVACEAE Hampea/ Hibiscus (Fig. 36). Spherical, amb circular; periporate, pores circular, 3-4 um diam., inner margin entire; echinate, echinae 5-7 um long, occasionally curved, broad at base, surface between echinae scabrate; tectate, wall 3 um thick, individual columellae evident in median section (400 х magnification); 75 um (excluding spines). There is sufficient overlap in the pollen of Ham- pea and Hibiscus that individual dispersed micro- fossils cannot be referred consistently to one genus. Hampea is represented in Panama by six species (D'Arcy, 1987), including the common H. appen- diculata (J. D. Smith) Standley, which also occurs in Honduras and Costa Rica. It is a shrub to mid- size tree of tropical wet, tropical moist, premontane rain, and premontane moist forests. Hibiscus is represented by 12 species in Panama, including H. sororius L.f., a species common in marshy areas and coastal strand vegetation. Similar specimens have been recovered from the Culebra and Paraje Solo formations. Melastomataceae (Fig. 27). Oblate-spheroi- dal, amb circular; heterocolpate, tricolporate with three pseudocolpi, colpi equatorially arranged, meridionally elongated, equidistant, straight, 6-7 um long (apex to equator), inner margin entire, ores obscure, situated at midpoint of colpus, pseu- docolpi slightly shorter to as long as colpi; psilate, tectate, wall 2 um thick, homogeneous (400 x mag- nification); 18-20 um. Dispersed fossil pollen of the Melastomataceae cannot be referred to any single modern genus, and in some cases they overlap morphologically with the pollen in other heterocolpate myrtalean families. The specimens serve only to record Me- lastomataceae-type pollen in Gulf/Caribbean Ter- tiary deposits. Similar pollen has been recovered from the Uscari and possibly the Culebra (unknown type 5) formations. MELIACEAE Cedrela (Fig. 35). Prolate-spheroidal; tetra- colporate, colpi equatorially arranged, meridionally elongated, equidistant, straight, 32-34 um long, costae colpi 3-4 um point of colpus, slightly elongated equatorially, 4 wide, pores situated at mid- X 6 um; psilate; tectate, wall 3 um thick, homo- geneous (400 х magnification); 50-52 x 42-44 m. The pollen of some species in Flacourtiaceae, Meliaceae, and Sapotaceae is similar and often difficult to separate in the dispersed fossil state. The relatively large (50 um), tetracolporate spec- imens are most similar to Cedrela in our reference collection. Four species are listed for Panama (D'Arcy, 1987), and two are widespread in north- ern Latin America. Cedrela odorata L. occurs from northern Mexico to South America and the Antilles. In Panama it grows in tropical moist, tropical dry, and premontane wet forests. Cedrela angustifolia Sessé & Мосто occurs in the tropical moist and tropical dry forests and is most abundant along the Pacific slope. Both grow at low to mod- erate elevations. Cedrela has been reported pre- viously from the Paraje Solo Formation. FIGURES 36-50. ب Fossil pollen from the Gatun Formation.—36. Hampea/ Hibiscus, 51-103, 253', 5, ESF N-37.— 37. Hauya, SL-103, 253', 14, ESF H-35.— 38. Cabomba, SL-103, 253', 1, ESF J-51.— 39, 42. Guarea, SL-103, 253', 13, ESF L-39; SL-103, 253', 6, ESF 5-29. —40. Chomelia type, SL-103, 253', 7, ESF F-23. —41. Faramea, SL-103, 253', 7, ESF D-3 4.—43. Cosmibuena, SL-103, 253', SL-103, 253', 1, ESF H-37; SL-103, 253', 9, ESF H-20; SL-103, 253', 2, ESF F-4 1, ESF эн — 44, 45, 48. Faramea, 2.—46, 47, 49. Casimiroa, 51.103, 253', 1, ESF Р-40; 51-103, 253', 1, ESF H-43.— 50. Paullinia, SL. 103, 253', 7, ESF U-45 Volume 78, Number 1 Graham Pliocene Communities—Dicots 211 AT э Poti n: Ф 212 Annals of the Missouri Botanical Garden Guarea (Figs. 39, 42). Oblate-spheroidal, amb circular, tetracolporate, colpi equatorially ar- ranged, meridionally elongated, equidistant, straight, relatively short (12-14 um), faint costae colpi 4- um wide, inner colpus margin entire to minutely dentate, pores conspicuous relative to colpi, situ- ated at midpoint of colpus, circular to slightly elon- gated equatorially, 6 um diam., inner margin en- tire; psilate to faintly scabrate; tectate, wall 3 ит thick, homogeneous to individual columellae just visible in median section (400 х magnification); 36 x 3 Guarea is a large genus of about 150 species listed for the American tropics, with about 10 re- corded for Panama that encompass a large number of synonymies (D'Arcy, 1987). Two common spe- cies in Panama are G. glabra Vahl and G. mul. tiflora Adr. Juss., which range from Mexico to South America. In Panama these species occur in tropical wet and tropical moist forests at low to moderate elevations. Guarea has been reported previously from the San Sebastian and Paraje Solo formations. MYRTACEAE Eugenia/ Myrcia (Fig. 28). Oblate, amb tri- angular; tricolporate to tricolporoidate, syncolpate, colpi equatorially arranged, meridionally elongated, equidistant, straight, 12-15 um long, inner margin minutely dentate, pore obscure, situated at mid- point of colpus, ca. 1.5 um diam.; scabrate; tectate, wall 2 um thick, homogeneous (400 X magnifica- tion); 24-28 um The pollen of these genera cannot be distin- guished consistently (Graham, 1980), and it is not possible to identify dispersed fossil pollen below the rank of family. The genera are widespread in Latin America and range through a variety of habitats. Similar pollen has been recovered from the Са- tuncillo, San Sebastian, Uscari, Culebra, Cucara- cha, and Paraje Solo formations. NYMPHAEACEAE (CABOMBACEAE) Cabomba (Fig. 38). Reniform; monosulcate, sulcus 45 um long, bordered by faint, narrow (ca thick, homogeneous (400 х magnification); 58 х um. Cabomba comprises six species of aquatic herbs in warm temperate to tropical regions of the New World. In Panama it is represented by G. piau- hyensis Gardner, which has pollen similar to the Gatun specimen. Megafossils are known from the Paleocene Ft. Union Formation of North Dakota (see listings in LaMotte, 1952: 89). Microfossils are known presently only from the Gatun For- mation. ONAGRACEAE Hauya (Fig. 37). Oblate, amb oval-triangular (aspidate); triporate, pores equatorially arranged, equidistant, large (10-12 um diam.), slightly pro- truding, surrounded by conspicuous costae pori, elements of costae pori granular to elongated (stri- ate), oriented parallel to equator; psilate to faintly scabrate; tectate, wall 3-4 um thick, homogeneous (400 x magnification); 102 um. Hauya comprises two species in Mexico and Central America: H. heydeana J. D. Smith (Chia- pas, Mexico, and Guatemala) and H. elegans DC. (Hidalgo and Guerrero, Mexico to Costa Rica). It grows **with oaks and other lush middle elevation vegetation. It does not range upward into the pine belt or downward into the seasonally drier decid- uous vegetation" (Graham & Jarzen, 1969, from Raven, pers. comm.). Hauya has been reported previously from the San Sebastian Formation. RHIZOPHORACEAE Rhizophora (Fig. 29). Prolate to prolate- spheroidal; tricolporate, colpi equatorially ar- ranged, meridionally elongated, equidistant, straight, 12-15 um 2-3 um wide, pores situated at midpoint of colpus, elongated equatorially (costae transversalis), 1 x 4 um; finely reticulate; tectate-perforate, wall 3 um thick, homogeneous to individual columellae (400 x magnification); long, inner margin entire, costae colpi just visible in median secti 18-20 x 22-24 um. Rhizophora is not abundant in the Gatun sam- ples. In the absence of other mangrove gener (e.g., Pelliciera), together with the diversity of upland and freshwater types, this suggests depo- sition in a swamp, shallow lake, or marsh only occasionally influenced by marine or brackish wa- ters. Its geologic record in the Gulf/Caribbean region was summarized by Graham (1989). It is known from all our Tertiary formations in northern Latin America (Gatuncillo, San Sebastian, Uscari, Culebra, Cucaracha, Culebra, La Boca, and Paraje Solo). RUBIACEAE Chomelia type (Fig. 40). Spherical, amb cir- cular; nonaperturate (apertures obscure?); coarse- ly reticulate, muri 2 um wide, straight, lumina Volume 78, Number 1 1991 Graham 213 Pliocene Communities—Dicots polygonal, 2-8 um diam.; tectate-perforate, wall 4-5 ит thick, individual columellae evident in median section (400 x magnification); 28-34 um. This pollen type is common to several genera of the Rubiaceae (e.g., Chomelia, Guettarda, Ter- ebraria) but is most similar to Chomelia in our reference collection. Nine species are listed for Panama (Dwyer, 1980; eleven in D'Arcy, 1987). The Panama species are trees, shrubs, or scandent shrubs growing in a variety of habitats, including swamp forests, tropical moist, and secondary for- ests. Similar pollen has been recovered from the Culebra, La Boca, and Paraje Solo formations [as the smaller Terebraria (22 um); Graham, 1976, figs. 187-188]. Cosmibuena (Fig. 43). Oblate, amb circular; tricolporate, colpi equatorially arranged, meridio- nally elongated, equidistant, oriented parallel to the equator (as observed in polar view), equidistant, 12 um long, inner margin entire, pore situated at midpoint of colpus, circular, 3-4 um diam., bor- dered by beaded annulus of fine, discontinuous ektexine elements; psilate; tectate, wall 2 um thick, homogeneous (400 X magnification); 28 um e genus comprises about 12 species in Central and tropical South America, with two occurring in Panama: C. ovalis Standley, and C. skinneri (Oer- sted) Hemsley. The latter is most widespread, dis- tributed from Mexico to Colombia and possibly Brazil. It is a component of tropical moist, pre- montane wet, and premontane rainforests. Cos- mibuena is known only from the Gatun Formation. Faramea type 1 (Fig. 41). Oblate, amb oval- triangular; triporate, pores equatorially arranged, equidistant, slightly protruding, 3—4 um diam., an- nulus 3-4 um wide, faint costae pori(?); finely reticulate, muri ca. 1 um wide, lumina 1 um or less in diameter; tectate-perforate, wall 4-5 um thick; individual columellae evident in median sec- tion (400 x magnification); 34-39 um. Faramea type 2 (Figs. 44, 45, 48). Oblate, amb oval-triangular; triporate, occasionally dipor- ate or tetraporate, pores equatorially arranged, equidistant, circular to slightly oval, ca. 4 um diam., inner margin entire to slightly dentate, surrounded by conspicuous annulus 5-6 um wide; scabrate; tectate, wall um thick, wall Е. to individual a just visible in median section (400 х magnification); 27-36 um The pollen of Faramea is diverse, including di- porate (e.g., Graham, 1985, figs. 64, 65) and tetraporate forms (Fig. 48), and is somewhat poly- morphic with both thick-walled and thin-walled forms occurring on the same modern reference slide. Type 1 is similar to F. talamancarum Stan- dley in our reference collection, and type 2 is similar to F. occidentalis (L.) A. R is a large genus of about 125 species widespread in Latin America (Dwyer, 1980). Nineteen species are listed for Panama (D'Arcy, 1987), and these row in a variety of habitats, including tropical wet, tropical moist, premontane wet, and premon- tane moist forests. Various pollen forms of Fara- mea have been recovered from the Gatuncillo, San Sebastian, and Paraje Solo formations. Posoqueria (Figs. 51-53). Spherical to ob- late-spheroidal, amb circular; tricolporate, colpi equatorially arranged, meridionally elongated, ich. Faramea equidistant, short (12 um apex to equator), broad at midpoint (9-12 um), colpi faint and frequently obscure, pore circular, conspicuous (7-9 um diam.), situated at midpoint of colpus, bordered by con- spicuous annulus 4-5 um wide; reticulate, retic- ulum irregular, lumen circular, 6-8 um diam. in equatorial region, becoming finer and disrupted at poles, columellae supporting reticulum long (3-4 um) clavae with smaller clavae-baculae occupying lumina of reticulum, muri smooth, slightly sinuous; tectate-perforate, wall 3-4 um thick, individual columellae evident in median section (400 X mag- nification); 68-70 um. Posoqueria includes about 15 species of trees and shrubs in Mexico, South America, and the Antilles. The most widespread species is P. latifolia (Rudge) Rose & Standley, growing primarily in the tropical moist forest, but also in tropical wet, pre- montane wet, premontane moist, and tropical dry forests. Posoqueria is known only from the Gatun assemblage. RUTACEAE Casimiroa (Figs. 46, 47, 49). Prolate-spheroi- dal, amb circular; tricolporate, equatorially ar- ranged, colpi meridionally elongated, equidistant, 30 um long, colpus margin entire to faintly dentate, costae colpi ca. 4-5 ит wide, pore situated at midpoint of = spain i uin 2x3 um, bordered by ae pori m wide; finely striate-reticulate, онин. їп me pattern, striae/ muri occasionally appearing minutely beaded due to underlying columellae; tectate-perforate, wall 2 um thick, columellae evident in median section; (400 x magnification); 45-40 x 35-40 um Six species of Casimiroa occur in Mexico and Central America, but none are listed for Panama (D'Arcy, 1987). Microfossils are known presently only from the Gatun Formation. 214 Annals of the Missouri Botanical Garden Volume 78, Number 1 Graham 215 1991 Pliocene Communities—Dicots SAPINDACEAE triangular; tricolporate, colpi equatorially ar- атару ће ЖОЕ, ОГО ГГ: ranged, meridionally elongated, equidistant, straight, triangular; triporate, pores equatorially arranged, ови ashes cin РРА i dnd equidistant, slightly elongated equatorially, 4 x 2 E 2 Es А COP d " fo. um, faint costae pori; finely reticulate; tectate- и оше nies i. кие; gc 0 wall 2 um dick individual кена ке қ - “rg ши ы шне d just visible in median section (400 x magnification); VPI о nemen няк и а ка 21-24 um ең AE Twenty species of Serjania are recorded for Allophylus is a large genus of about 190 spe- P D'A 1987), hrough cies, with 13 listed for Panama (D'Arcy, 1987). anama гсу, jı айй these occur throug a wide range of habitats at low to moderate ele- The plants grow at low to moderate elevations in о, А . . : vations. Similar pollen has been reported from the tropical wet, tropical moist, tropical dry, premon- : x Gatuncillo and Paraje Solo formations. tane wet, and premontane moist forests. Similar pollen has been reported from the Culebra and Paraje Solo formations. SAPOTACEAE Cupania (Figs. 54, 55). Oblate to peroblate, cf. Bumata (Fig. 38). Prolate; tetracolporate, triangular; tricolporoidate, colpi equatorially ar- colpi ly arranged, meridionally elongated, ranged, meridionally elongated, equidistant, straight, equidistant, straight, 18-21 um long, inner margin 10-12 um long, syncolpate, apices of colpi oc- entire to minutely dentate, costae colpi ca. 3 um casionally branched to include triangular polar area; Wide, pores situated at midpoint of Тыш elon- scabrate; tectate, wall 2-3 um thick, homogeneous gated equatorially (colpi equatorialis), 2-3 = (400 x magnification); 26 um. ит, inner pore margin entire; psilate; tectate, wall The pollen of Cupania is difficult to distinguish З 4m thick, homogeneous (400 x magnification); from Matayba. Those grains with a moderately 32-35 x 23-26 um. thick wall, often including a triangular polar area, Pollen of tetra-aperturate Flacourtiaceae, Me- are most similar to C. latifolia HBK in our ref- liaceae, and Sapotaceae is difficult to distinguish erence collection. Eleven species are listed for Pan- Consistently, especially in the dispersed fossil state. ama (D'Arcy, 1987), and these grow primarily in The Gatun specimens are most similar to Bumelia tropical wet, tropical moist, and premontane moist in our reference collection. Bumelia is a New World forests. Similar grains are known from the Culebra genus of about 23 species of trees and shrubs, and Paraje Solo formations. ranging from the United States to Mexico, South America, and the Antilles (Blackwell, 1968). A Paullinia (Fig. 50). Oblate to peroblate, amb single species, B. persimilis Hemsley, is listed for triangular; triporate, pores equatorially arranged, Panama (D'Arcy, 1987) and is distributed from equidistant, 4-6 um diam., costae pori ca. 4 um Veracruz and Oaxaca, Mexico, to Venezuela. This wide; finely reticulate; tectate-perforate, wall 2-3 um thick, individual columellae just visible in me- dian section (400 х magnification); 34-38 um. Paullinia is a large genus of about 194 neo- tropical species, with 49 species listed for Panama SYMPLOCACEAE (D'Arcy, 1987). The plants typically grow in moist forests at low to moderate elevations. Similar pollen pollen type was not reported previously in our Gulf/ Caribbean assemblages. Symplocos type 1 (Fig. 59). Oblate, amb oval- triangular to nearly circular; triporate (or tricol- porate but with colpi short and obscure), pores equatorially arranged, equidistant, circular, 3 um Serjania (Fig. 56). Oblate to peroblate, amb diam., annulus 3 um wide; psilate to faintly sca- has been reported from the Gatuncillo and Paraje Solo formations. < FIGURES 51-64. Fossil pollen from the Gatun Formation. — 51-53. Posqueria, 51-103, 253', 4, ESF Р-19; SL- 3, 253', 8, ESF U-27.— 54, 55. Cupania, 51-103, 253', 8, ESF P-44; 51-103, 253', 13, ESF J-42.— 56. Sdn 51-103, 253’, 15, ESF 5-49. — 57. Allophylus, SL-103, 253', 2, ESF H-50.— 58. cf. Bumelia, SL-103, 253', 12, ESF T-47.— 59. Symplocos type 1, SL-103, 178', 2, ESF E-35.— 60. Symplocos type 2, SL-103, 253', 5, ESF V-33.— 61. Mortoniodendron, SL-103, 253', 13, ESP W.23 —62, HA Асар, 51-103, 253', 12, ESF J-26; SL-103, 253’, 9, ESF T-48.— 64. Petrea, SL-103, 255.5', 5, ESF 5-3 216 Annals of the Missouri Botanical Garden brate; tectate, wall 3 um thick, homogeneous (400 x magnification); 28 um. Symplocos type 2 (Fig. 60). This specimen is similar to type 1 but is larger (38 um). Type 1 is comparable to S. chiriquensis Pittier, and type 2 to S. pychantha Hemsley, in our reference col- lection. mplocos about 350 species of trees and shrubs of pantropical distribution growing in a wide variety of habitats. Megafossils have been reported from the late Eocene Goshen Formation of Oregon (Chaney & Sanborn, 1933), and mi- crofossils from the Maastrichtian of California (Chumura, 1973). It is known only from the Gatun Formation in the Tertiary of northern Latin Amer- ica. TILIACEAE Mortoniodendron (Fig. 61). Oblate, amb cir- cular; tricolpate, colpi equatorially arranged, me- ridionally elongated, equidistant, short (6-7 um apex to equator), inner margin minutely dentate, costae colpi 4-5 um wide; finely reticulate; tectate- perforate, wall 2 um thick, individual columellae evident in median section (400 X magnification); 34-38 um. The pollen of Mortoniodendron and its fossil record has been summarized by Graham (1979). The genus comprises about five species of small shrubs to tall trees growing from southern Mexico through Central America in tall moist forests at low to moderate elevations. Similar pollen has been recovered from the Gatuncillo and Paraje Solo formations. VERBENACEAE Aegiphila (Figs. 62, 63). Oblate, amb cir- cular; tricolpate, colpi equatorially arranged, me- ridionally elongated, equidistant, short (9-1 apex to equator), inner colpus margin entire to slightly diffuse; echinate, echinae short (ca. 1 um), moderately dense and uniformly distributed; tec- tate, wall ca. 1.5 um thick, individual columellae just visible in median section (400 х magnification); 5 um. Aegiphila comprises about 160 species of trees, shrubs, and lianas distributed from Mexico to South America (Brazil, Colombia, Venezuela, the Guianas, Peru). Twenty-one species are presently listed for Panama (D'Arcy, 1987), primarily in tropical wet, tropical moist, premontane wet, and premontane moist forests. The Gatun specimens are similar to A. elata Sw., a widespread species found along the margins of woods and streams. Fossil pollen of Aegiphila is known only from the Gatun Forma- tion. Petrea (Fig. 64). Oblate, amb oval-triangular, tricolpate, colpi equatorially arranged, meridionally elongated, equidistant, short (6-8 um apex to equa- tor), margin entire to minutely dentate; scabrate; tectate, wall 3 um thick, individual columellae just visible in median section (400 х magnification); 52-56 um. There are about 30 species of Petrea distributed from northern Mexico to southern Brazil and the Antilles. Four species are known from Panama (D'Arcy, 1987), including the widespread P. as- pera Turcz., which produces pollen similar to the specimens. Megafossils identified as Petrea have been reported from the late Eocene (now dated as earliest Oligocene) La Porte flora of California (Pot- bury, 1935). However, the holotype has an in- flated, possibly cross-striated petiolule, which in- dicates that it is a legume (Doyle et al., 1988). Microfossils are known presently only from the Gatun assemblage. UNKNOWNS In addition to the specimens identified from the Gatun Formation, a number of unknowns were recovered. Twenty-seven of the more common and/ or distinctive ones are described below. Unknown 1 (Figs. 65, 66). Oblate, amb cir- cular; tricolpate, colpi equatorially arranged, me- ridionally elongated, equidistant, straight, 16-18 um long (apex to equator), inner margin finely dentate; finely reticulate, width of muri 1 um or less, diameter of lumina ca. 1.5 um; tectate-per- forate, wall 2 um thick, individual columellae ev- ident in median section (400 x magnification); 35- um. These grains are similar to Avicennia, but all are oriented in polar view (Avicennia is prolate to prolate-spheroidal and is usually seen in equatorial view), and the wall is slightly thinner in the spec- imen. Unknown 2 (Figs. 67, 68). Oblate, amb cir- cular; tricolpate, colpi equatorially arranged, me ridionally elongated, equidistant, straight, short (3- 4 um apex to equator), inner margin finely dentate; finely reticulate, width of muri and diameter of lumina ca. 1 um; tectate-perforate, wall 2 um thick, individual columellae evident in median section (400 X magnification); 35-39 um. Volume 78, Number 1 Graham 217 1991 Pliocene Communities—Dicots Oe e Ж, ¢ * B ~ e FIGURES 65-84. Fossil pollen from the Gatun Formation. — 65, 66. Unknown 1, SL-103, 253', 1, ESF F-47.— 61, Unknown 2, SL-103, 253’, 1, ESF H-52.—69, 70. Unknown 3, SL-103, 253', 7, ESF W-40.—71. Unknown 4, SL-103, 253’, E $S-40.— 72. Unknown 5, SL-103, 253’, 1, ESF P-37.— 73. Unknown 6, SL-103, 257', 1, ESF P-34.— 74-77. Unknown 7, SL-103, 253’, 14, ESF Q-53; SL-103, 253’, 12, ESF M-40.— 78. Jules SL-103, 253', 14 ESF Q-41.— 79, 80. Unknown 9, SL-103, 253', 10, ESF J-42; SL-103, 253', 13, ESF N-30.—81. Unknown 10, SL-103, 253’, 7, ESF T-27.—82. Unknown 11, SL-103, шщ 5, ESF F-19.—83. Unknown 12, 51-103, 253', 5, ESF O-14.—84. Unknown 13, SL-103, 253', 7, ESF U-2 218 Annals of the Missouri Botanical Garden Unknown 3 (Figs. 69, 70). Oblate, amb cir- cular; tricolpate, colpi equatorially arranged, me- ridionally arranged, equidistant, short (3 um apex to equator), inner margin finely dentate; finely re- ticulate, width of muri and diameter of lumina ca. 1 um; tectate-perforate, wall 3 um thick, individual columellae evident in median section (400 x mag- nification); 31-34 um. he wall of unknown 2 is slightly thinner and the columellae finer than in unknown 3. The short colpi and fine reticulum of these unknowns are suggestive of the Bombacaceae, but no exact match could be found among modern analogs in our ref- erence collection. Unknown 4 (Fig. 71). tricolpate, colpi equatorially arranged, meridionally elongated, equidistant, short (3-4 um apex to equa- tor); scabrate; tectate, wall 2 um thick, columellae just visible in median section (400 х magnification); 23-26 um. Unknown 5 (Fig. 72). tricolpate, colpi equatorially arranged, meridionally elongated, equidistant, short (6–7 um apex to equa- tor), costae colpi 3 um wide; scabrate; tectate, wall 2 um thick, individual columellae just visible in median section (400 х magnification); 29-34 um. Oblate, amb circular; Oblate, amb circular; Unknown 6 (Fig. 73). Oblate, amb circular; tricolpate, colpi equatorially arranged, meridionally elongated, equidistant, short (6-7 ит apex to equa- tor), costae colpi 4-5 um wide; finely reticulate, width of muri and diameter of lumina ca. 1 um; tectate-perforate, wall 2 um thick, individual col- umellae just visible in median section (400 X mag- nification); 32-35 um. Unknown 7 (Figs. 74-77). Oblate, amb cir- cular; tricolpate, colpi equatorially arranged, me- ridionally elongated, equidistant, 12-15 um long (apex to equator), broad (10-14 шт at equator, possibly compressed open), inner colpus margin diffuse; striate, striae coarse (3 um wide), slightly sinuous, parallel and oriented at right angle to equatorial plane; tectate, wall 2 um thick, homo- geneous; 22- Unknown 8 (Fig. 78). Oblate-spheroidal, amb circular; tricolpate, colpi equatorially arranged, meridionally elongated, equidistant, So пеше! by sculpture elements; verrucate, verr like, translucent, variable in size, larger ca. 6 um diam., base of verrucae and intervening surface area with small columellae giving granular ap- pearance to portions of grain surface; tectate, wall 2 um thick, individual columellae evident in median section (400 X magnification); 28-3 These grains are somewhat similar to Elizabetha (Leguminosae-Caesalpinioideae), but are smaller and have fime columellae giving a granular ap- pearance to portions of the exine surface. They are also present in the La Boca (Graham, 1989, figs. 55, 56) and Paraje Solo (Graham, 1976, figs. 244, 245) formations. Unknown 9 (Figs. 79, 80). Oblate-spheroidal to spherical, amb circular; tetraporate, pores equa- torially arranged, equidistant, circular, 3 um diam., annulus 3 um wide, outer margin of regular and slightly granular; psilate; tectate, wall 2 um thick, homogeneous (400 X magnification); 22-25 um These grains are very similar to Cosmibuena (Fig. 43) and may represent tetraporate forms of that genus. However, tetraporate grains have not been observed in our modern reference material of Cosmibuena, hence the designation of these specimens as unknown. annulus ir- Unknown 10 (Fig. 81). stephano(5)colporate, colpi equatorially arranged, meridionally elongated, equidistant, short (ca. 5 Oblate, amb circular; um), narrow (1 um), inner margin minutely den- tate, pores situated at midpoint of colpus, equa- torially elongated, ca. 2 х 4 um, costae pori; psilate; tectate, wall 2-3 um thick, homogeneous (400 х magnification); 31-33 um Unknown 11 (Fig. 82). Oblate, amb circular; stephano(5)colporate, colpi equatorially arranged, meridionally elongated, equidistant, 15-18 um long, tapering to prolonged acute apex, inner margin entire, pores situated at midpoint of colpus, 3 x 9 um; scabrate to finely verrucate, scabrae/ ver- rucae low, larger ca. 5 um diam.; tectate, wall 4— 5 um diam., homogeneous (400 x magnification); 45 um. Unknown 12 (Fig. 83). Oblate, amb trian- gular, apices rounded; trilete or trichotomosulcate, individual arms straight, са. 30 um long, extending to or nearly to margin, inner margin entire; sca- brate/punctate, punctae (or spaces between sca- brae) narrow, elongated, sinuous, becoming finely reticulate bordering apertures; tectate to tectate- perforate, wall 3 um thick, individual columellae/ punctae evident in median optical section (400 x magnification); 60 um. This specimen may represent a trilete fern spore, but the wall structure suggests a trichotomosulcate palm pollen. Volume 78, Number 1 1 Graham 219 Pliocene Communities—Dicots Unknown 13 (Fig. 84). Spherical, amb cir- cular; periporate, pores circular, iam., evenly distributed, inner margin entire, costae pori 4—5 um wide; scabrate; tectate, wall 4 um thick, individual columellae evident in median section (400 x magnification); ca. 75 um. Unknown 14 (Fig. 85). cular; nonaperturate; echinate, echinae short (2- Spherical, amb cir- um), somewhat variable in shape, mostly broad at base, moderately densely arranged, distance be- tween echinae 1-3 шт; tectate, wall 2 um thick, homogeneous (400 x magnification); 75-80 um. The size and ornamentation of this grain suggest Cucurbitaceae, but an exact match could not be found. Some of the several folds simulate a trilete mark, but the specimen is provisionally interpreted as nonaperturate. Unknown 15 (Fig. 86). Oblate, amb circular; stephano(6)colpate, colpi equatorially arranged, meridionally elongated, equidistant, straight, 20- 22 um long, inner margin minutely dentate; retic- ulate, muri narrow (1 шт or less), sinuous, lumina polygonal, ca. 3 um diam., slightly smaller at poles; tectate-perforate, wall 3 шт thick, individual col- umellae (appearing as slight clavae) evident in me- dian section (400 X magnification), ca. 1 um wide, spaced ca. 2 um; 70-75 um. The specimen is similar to some Labiatae (e.g., Hyptis), but an exact match could not be found. Unknown 16 (Fig. 92). Spherical, amb cir- cular; nonaperturate; echinate, echinae 3-5 um long, uniform width (2 um) for most of length, then tapering abruptly near apex, occasionally curved; tectate, wall 3 um thick, homogeneous to individual columellae just visible in median section (400 x magnification); 58 um (excluding spines). An irregular fissure extending nearly the entire length of the grain may represent a sulcus, but the margins are irregular, suggesting a split in the exine, and the specimen is interpreted provisionally as nonaperturate. Unknown 17 (Fig. 91). Spherical, amb cir- cular; nonaperturate; echinate, echinae 7-9 um long, 3 um wide at base, occasionally curved, more or less widely spaced (distance between echinae 6– 10 um), surface between spines scabrate; tectate, wall 3 um thick, homogeneous to individual colu- mellae just visible in median section (400 x mag- nification); 52 um (excluding spines). Unknown 18 (Figs. 87, 88). Oblate, amb cir- cular; nonaperturate; reticulate, reticulum more or less obscure, shallow, muri narrow, sinuous, lumina polygonal; tectate-perforate, wall 3 um thick, in- AME. 1 11 ЧА | + т 5H + ANO ¥ magnification); 60 um. Unknown 19 (Figs. 97, 98). Oblate, amb cir- cular; tricolpate (?, apertures obscure), colpi equa- torially arranged, meridionally elongated, equidis- tant; reticulate, reticulum shallow (supporting columellae short, ca. 1.5 um), muri narrow (1 um or less), sinuous, lumina polygonal, larger lumina 3-5 um diam.; tectate-perforate, wall 2 um thick, individual columellae evident in median section (400 X magnification); 65 um. Unknown 20 (Figs. 89, 90). Oblate, amb square; apertures obscure, four(-six) colpi or colpi with pores, apertures equatorially arranged, merid- ionally elongated, equidistant, with two additional apertures possibly at each pole; finely reticulate, width of muri and diameter of lumina 1 um or less; tectate-perforate, wall 3 um, individual columellae evident in median section (400 х magnification); um. This specimen may be an aberrant form. There appear to be four colpi (with pores?) about the equator, and one at each pole. Grains with a similar number and arrangement of apertures were not encountered in the modern reference collection or in the literature examined. Unknown 21 (Fig. 95). spheroidal; tricolporate, colpi equatorially аг- ranged, meridionally elongated, equidistant, straight, 22-24 um long, extending nearly entire length of Prolate to prolate- grain, inner margin entire; reticulate, reticulum heavy (columellae ca. 2 um thick), width of muri and diameter of lumina ca. 1 um, lumina circular to polygonal; tectate-perforate, wall 3 um thick, individual columellae evident in median section (400 x magnification); 28 x Unknown 22 (Fig. 94). Prolate to prolate- spheroidal; tricolporate, colpi equatorially ar- ranged, meridionally elongated, equidistant, colpi straight, narrow, 18-20 um long, inner margin entire, costae colpi 3-4 um wide, broadening at pores, pores situated at midpoint of colpus, slitlike (1 x 4—5 um); psilate to faintly scabrate; tectate, wall 3 um thick, individual columellae just visible in median section (400 X magnification); 27 x 23 um. Unknown 23 (Fig. 96). Prolate to prolate- spheroidal, tricolporate, colpi equatorially ar- ranged, meridionally elongated, equidistant, colpi Annals of the 220 Missouri Botanical Garden n 25 : 44 e ہے‎ у ми Me + та жє a A 1X Te E ^ Y у ^ AL re ту, ) Jut RM у of ^ g V uL ight, V E Lenin d Ук " „Ж. CM “era Ау Piet m. њи а ~...“ P. کک e Xm‏ яя 1‏ > a x „ш‏ бс‏ © £95 | ہے ~ а yA‏ E >‏ MS‏ , ۴ , > 24, SL-103, 253’, 13, ESF H-46.— 94. Unknown 22, 85. Unknown 14, SL-103, 253’, 15, ESF K-55.— 253’, 1, ESF M-39.—96. Unknown 23, SL-103, 253’, , Fossil pollen from the Gatun Formation.— SL-103, 253', 2, ESF T-41.— 93. Unknown 1, ЕЗЕ F-36.— 95. Unknown 21, SL-103 Volume 78, Number 1 1991 Graham Pliocene Communities—Dicots 221 FIGURES 97-103. straight, narrow, 22-24 um, extending nearly еп- ше length of grain, inner margin entire, costae colpi #74 шт wide, pore situated at midpoint of colpus, slitlike (1 x 4—5 um); psilate; tectate, wall 3 um thick, homogeneous to individual columellae Just visible in median section (400 x magnification); 26 X 23 um | "dedi 22 and 23 may represent variations e same pollen type. The former has costae Olpi that broaden at the pores, while the latter is more psilate and the wall more homogeneous. on 24 (Fig. 93). Prolate; tricolporate, Pl equatorially arranged, meridionally elongated, equidistant, straight, narrow, 24 um long, extend- чи nearly entire length of grain, costae colpi 4— ic ce (obscure), pore situated at midpoint of E e ongated equatorially, 3 x 5 um; finely E 1 ate, width of muri and diameter of lumina ш "d E epos wall 3 um thick, in- Magnification); 45 xX 30 ит. ae grains are similar to those of certain An- ea ee and Euphorbiaceae, but an exact match E. oi found. They are frequent in low per- b. + s in Gulf /Caribbean Tertiary deposits, and c en reported from the Gatuncillo, La Boca, araje Solo formations. iss Fossil pollen from the Gatun Formation.—97, 98 a aa Unknown 27, SL-103, 253’, 2, ESF N-46.—100, 101. Unknown 25, 51-103, 253’, 1, ESF 0-37; SL- > , 5, ESF K-45.—102, 103. Unknown 26. SL-103, 253', 12, ESF P-22; SL-103, 253', 13, ESF 0-51. . Unknown 19, SL-103, 253', 6, ESF Unknown 25 (Figs. 100, 101). Prolate; tri- colporate, colpi equatorially arranged, meridionally elongated, equidistant, straight, narrow, 20-22 um ong, inner colpus margin entire, costae colpi 3-4 um wide, thicker at роге (4-5 um), pores situated at midpoint of colpus, slitlike (1-2 x 4-6 um); psilate to scabrate; tectate, wall 3 um thick, ho- mogeneous to individual columellae just visible in median section (400 х magnification); 37-44 х 24-28 um. These specimens may include some tetracol- porate forms and are similar to several Sapotaceae, including Caesaria, a pantropical genus of about 250 species. Eight were recognized for Panama by Robyns (1968), and D'Arcy (1987) listed 17 species. Most grow in moist habitats, but some range into drier premontane or tropical dry forests. Pollen of Caesaria includes both tricolporate and tetracolporate forms. It has been reported from the Gatuncillo, San Sebastian, Culebra, and Paraje Solo formations. Unknown 26 (Figs. 102, 103). Prolate to pro- late-spheroidal; tricolporate, colpi equatorially ar- ranged, meridionally elongated, equidistant, straight, narrow, 21-24 ит, inner margin entire, costae colpi 3 um wide, thicker at pores (4 um), pores situated at midpoint of colpus, slitlike (1-2 x 3- 222 Annals of the Missouri Botanical Garden 5 um); psilate to scabrate; tectate, wall 2 um thick, individual columellae just visible in median section (400 x magnification); 36-41 x 33-38 um. Unknowns 25 and 26 are similar and may rep- resent variations of the same pollen type. The former has slightly thicker walls and costae colpi. Unknown 27 (Fig. 99). colpi equatorially arranged, meridionally elongated, equidistant, straight, narrow, 28-30 um long, inner colpus margin minutely dentate; scabrate; tectate, Prolate; tricolpate, wall 2 um thick, homogeneous to individual colu- mellae just visible in median section (400 X mag- nification); 4 e specimen may represent a Quercus grain that is slightly corroded, hence the granular ap- pearance of the colpus margin. LITERATURE CITED AYALA-NIETO, M., В. LIRA SAADE & J. L. ALVARADO. 1988. Morfología polínica de las Cucurbitaceae de la pennas de Yucatán, Mexico. Pollen & Spores 30: 8. BanTLETT, А. S. & E. S. BARGHOORN. 1973. Phyto- geographic history of the Isthmus of Panama during the past 12,000 years (a history of vegetation, cli- mate, and sea-leve ge) Pp. 203-299 in A Graham (editor), Vegetation and Vegetational ани P orthern Latin America. Elsevier Publ., ster Bini “E W. 1921. Tertiary fossil plants from os Dominican Republic. Proc. U.S. Natl. Mus. 59: 1 1941. Additions to the Wilcox flora from i pu Texas. Profess. Pap. U.S. Geol. Surv. 193E: BLACKWELL, W. H 1968. Sapotaceae. In: R. E. Wood- n & R. W. Schery (ешге), Flora of Panama. Missouri Bot. Gard. Ga 3 M. DE. 1962. Pollen grains E the* “cerrado.” IV. Bombaceae, Connaraceae, Cucurbitaceae, Dil- leniaceae, Erythroxylaceae, Gesneriaceae. Revista Brasil. Biol. 22: 307-315. CHANEY, К. W. ‚ 1. SANBORN. 1933. The Goshen flora of west central Oregon. Publ. Carnegie Inst. Wash. 439: 1-103. CHUMURA, C. А. 1973. Upper Cretaceous ا‎ aN an- Maastrichtian) angiosperm pollen from the western San Joaquin eek nes U.S.A. ie. phica B 141: 89-17 . C. NIKON 1989. Earliest mega- Flora of Barro Cds Idand. Stanford Univ. Pies: Stanford, California. D'Ancv, W. G. 1987. Flora of Panama Checklist and Index. Monogr. Syst. Bot. Missouri Bot. Gard., Vol- umes n DIETERLE, J. V. . Cucurbitaceae. /n: Flora of Guatemala. ШШ 24: 306-395. SCHORN, В. Н. TIFFNEY А С. В. The La Porte Flora — Ear- liest Oligocene of North- central California. Field Guide 1988 Meeting Paleobot. Sect. Bot. Soc. Amer., Da- vis, California Dwyer, J. D. 1980. Rubiaceae. In: R. E. Woodson & R. W. Schery er Flora of Panama. Ann. Mis- -522. FEUER, S. M. & J. Kv шт 979. Pollen morphology and evolution in Рашани ћи (Loranthaceae). Bot. Not. 132: 295-309. 1980. Fine structure of mistletoe pollen. III. Large-flowered neotropical Loranthaceae and their relatives. Amer. J. Bot. 67: 34-50. & 1985. Fine structure of mistletoe pollen. VI. Small-flowered neotropical Loranthaceae. Ann. Missouri Bot. Gard. 72: 187-212. 76. Studies in و‎ p I. The Miocene communities of Ver , Mexico Ann. Missouri Bot. Gard. 63: 787 842 197 Mortoniodendron Старе) апа И Trichipteris (Cyatheaceae) in Ce- s of the о region. Апп. 80. Morfologia del oen Ed Eugenia / Myr- cia (Myrtaceae) y Combretum/ Terminalia (Com- bretaceae) en relaçién a su alcance estratigráfico en el Terciario del Caribe. Biotica 5: 5-14. Studies in neotropical paleobotany. IV. The Eocene communities of Panama. Ann. Missouri Bot. Gard. 72: 504-534 87. Tropical American Tertiary floras and paleoenvironments Mexico, Costa Rica, and Pana- . 74: 1519-1531. . 1989. ‘Studies i in neotropical ا‎ The lower Miocene communities of Panama — the Boca Formation. Ann. Missouri Bot. Card. 76: 50- 66. 19 New angiosperm records from the Ca- ribean ен Атег. Ј. Вог. 77: 897-910. 99]. Studies in — мы pe ind VIII. The Не communities of Panama — introduction and ferns, gymn иав angiosperms (monocots). Ann. Missouri Bot. Gard. 200. & G. F. BARKER. 1981. Palynology and tribal classification in the | Pp. 801-834 in R. M. Polhill & P. H. Raven (editors), Advances in Legume Systematics, Royal Botanic Gardens, Kew. & 69. Studies in neotropical arid I. The Oligocene communities of Puerto . Ann. Missouri Bot. Gard. 56: 308-357. A. S. Томв. 1974. Palynology of Erythrina (Leguminosae: a preliminary survey of the subgenera. Lloydia 37: 8l. & 1 Ош of Erythrina (Leguminosae: Papilion oideae) the subgenera, sec- tions, and generic relationships. Lloydia 40: 413- 43 5 ‚ В. Н. STEWART & J. L. STEWART. 1985. Stud- ies in neotropical айн ке III. The Tertiary com- munities of Panama — geology of the pollen bearing iments. Апп. Missouri Bot. Gard. 72: 485-503. HERD: P. S. € D. L. ОиснЕк. 1988. A fossil egume pod with affinities to dA Detarieae- Amherst- ieae (Caesalpinioideae) from the Eocene of south- ages North America. e Bot. 75, No. 6, pt. 110. e HoLompcE, L. 1970. Manual dendrológico pus a 1000 especies EXE DEAE en la Republica de Pana Government of Panama, Panama City. Volume 78, Number 1 1991 Graham Pliocene Communities—Dicots 223 Horn, S. P. 1985. Preliminary pollen analysis of Qua- ternary sediments from Deep Sea Drilling Project site 565, western Costa Rica. Initial Reports Deep Sea Drilling Project 84: 533-547. ГАМОТТЕ, К. S. 1952. Catalogue of the Cenozoic plants of сне America through 1950. Mem. Ceol. Soc. . 51. Am Pasa J. H., B. L. HACKNER & A. S. BARTLETT. 1967. Mangrove pollen at the ei site of Oligo-Miocene amber Chiapas, Mexico. Bot. Mus. Leafl. 21: 289-3 LIVINGSTONE, D. A. & T VAN DER HAM 1978. Palaeogeography and palacolimatolog ie н Forest Ecosystems, ural н search XIV, UNESCO/UNEP/FAO, Par MANCHESTER, S. . The fossil niet of the т с. асеае. Monogr. Syst. Bot. Missouri Bot. Gard., "зе С. 963. Material para una monografia la А del polen de Cucurbitaceae. Grana b melodia de sédiments Tertiaires Pa enégal méridional. Pollen & Spores 17: 545- MELHEM, Т. 5. 1966. Pollen grains of plants of the Cerrado. 12. Cucurbitaceae, M Brasi 1981. Fossil pollen records E extant an- giosperms. Bot. Rev. ере 47: 1-1 NILSSON, S. & А. ROBYNS. 6. didam Kunth. Pp. 1-59 in S. Nilson (editor), World Pollen and Spore Flora 14. Almqvist & Wiksell, Stockholm PaLacios CHAVEZ, К. 1984. “La morfología de los granos e polen de las especies Mexicanas del genero Bur- 2. 935. The La Porte flora of Plumas اة‎ California. Publ. Carnegie Inst. Wash. 465: 9-81. 1964. Bombacaceae. In: В. E. Woodson W. Schery cine Flora of Panama. Ann. Missouri Bot. Gard. 51 -68. 68. ا‎ KoA In: R. E. Woodson & R. W. Schery (editors), Flora of Panama. Ann. Mis- souri Bot. Gard. 55: 93-144. STONE, D. E. & C. R. BROOME. 1975. Juglandaceae А. Rich. ex Kunth. Pp. 1-35 in S. Nilsson (editor), World Pollen and Spore Flora 4. Almqvist & Wiksell, Stockholm. pen E 1964. Pollen morphology and identifi- I. Modern and fossil tropical pollen with emphasis on Bombacaceae. Pollen & Spores 6: 393- Вовум$, A. & R. WALKER. J. W. 1971a. Elucidation of exine structure and sculpturing in the Annonaceae through combined use of light and scanning electron microscope. Pollen & Spores 13: 187-198. 971b. Pollen morphology, phytogeography, m" phylogeny of the Annonaceae. Contr. Gray Herb. 202: 1-130. 971c. Unique type of angiosperm pollen from the family Annonaceae. Science 172: 565- Я ` 1972. Contributions to the pollen morphology and кок E the Annonaceae. II. J. Linn. Soc., Bot. 65: 173- WEBSTER, G. L. & D BURCH. In: R. E. Woodson 1967. Euphorbiaceae. & R. W. Schery (editors), Flora of Panama. Ann. arn е бя 54: 211-350. жошо, R. E. & R. W. ScH . Leguminosae. . Woodson & К. schen (editors), Flora of . Ann. Missouri Bot. Gard. 37: 184-314. шаш. К. Р. Сконе. In: R. E. Woodson & В. W. Schery (editors), Flora of Panama. Ann. Missouri Bot. Gard. 65: 285-366. NEW ТАХА OF SOLANUM (SOLANACEAE) FROM BRAZIL, COLOMBIA, CENTRAL AMERICA, AND VENEZUELA! L. d'A. Freire de Carvalho? ABSTRACT Following morphological and taxonomic studies described, not all of them in sect. Lepidotum: Solanum hatschbachii (Brazil), S. lep aes var. lepidioc ита (Colombia), S. lepi ut a revision of S olanum sect. Lepido 10 new taxa are nse (Brazil), S. gauss glans 5 davidsei (Venezuela um var. trianae (Colombia), S ). oliveirae (Brazil), S. pereirae (Brazi 1), S. sooretamum (Brazil), and S. о (Guatemala A taxonomic revision of Solanum sect. Lepi- dotum has revealed the new species and varieties described in this paper. The morphological studies of various features of trichomes, the indumentum of the abaxial leaf surfaces, and the structure of the inflorescences as well as the realignment of the group into subsections and a new section are being published elsewhere (Carvalho & Machado, in press; Carvalho & Shepherd, in press). Nine of the ten new taxa fall into two groups. Section Lepidotum sensu strictu is identified by a suite of characters: (1) indumentum appressed- lepidote with a golden or argenteous appearance, formed of peltate and peltate-stellate trichomes; (2) inflorescence short and simple, cymose or cymose- dichotomous with a scorpioid rachis; (3) corolla campanulate-stellate; (4) flowers with pedicels ar- ticulated on the inflorescence branches; (5) anthers subsessile, oblong with large apical anterior pores and dehiscing tardily by longitudinal slits; (6) leaves geminate, discolorous with golden or argenteous indumentum on the abaxial surface. These features are shown by the new taxa Solanum carautae, S. davidsei, S. hatschbachii, S. lepidotum var. lep- idiochlamys, S. lepidotum var. trianae, and S steyermarkii. A second group of new species shares several features: (1) indumentum lepidote-tomentose or lepidote-floccose formed of trichomes with long and ornamented stalks having lateral expansions; (2) macroscopic chaffy laminate or fringed trichomes present on branches, inflorescence, and sometimes outside the wall of the calyx; (3) inflorescence ro- bust, short-stalked, cymose-dichotomous and with erect and pendulous rachis; (4) leaves covered by indumentum of yellowish or whitish coloration on the abaxial surface; (5) fruit globose, totally en- closed in the accrescent calyx. This group includes the new species Solanum caldense, S. oliveirae, and 5. pereirae. A final species, Solanum sooretamum, is more closely allied to S. vellozianum Dunal. 1. Solanum caldense Carvalho, sp. nov. TYPE: Brazil. Minas Gerais: Municipality of Caldas, 1845, J. F. Widgren s.n. (holotype, R ex herb. Imp. Brasil. Regnellium Mus. Bot. Stockholm). Nd 1, 8, 9, 14А, B. rbuscula ] llis albidi fl cosis pilis stellato- peddle se pd paleaceo- fimbriatis, brevi li 1 ato-lanceolata, apice 110 mm 20 mm longa, sicciata castanea, duin glabra aut glabrescentia, subtus dense lepidoto-tomentosa tri- chomis stellatis vel porrecto чеша petiolus ad 25 m longitudine. Inflorescentia multiflora, pendula, robusta, crassa, 90-150 mm longa, ramos perd 60-110 mm ! Part of a doctoral thesis by L. d'A. Freire de Carvalho (1988). Gratitude is extended to Carlos T. Rizzini, botanist of the Botanical Garden of Rio de Janeiro, and to e. Clemente J. Steffen of the University of Vale do Rio dos Sinos, Sao Leopoldo, Estado do Rio Grande do Sul, for e Latin diagnoses, and to the curators of national (CEPEC, HB, RB, UB) and foreign (BH, F, G ‚ К, MO, NY, P, S, W) herbaria for lending important material, in ncluding nomenclatural types. А special acknowledgment i is made to W. G, D’Arcy for critical review of the English manuscript and for other help and suggestion: ? Botanical Carden of Rio de Janeiro, Rua Pacheco Leão, 915 Gávea, Brasil. Fellow of the “Сопзефо Nacional ). de Desenvolvimento Científico e Tecnológico" (CNPq ANN. MISSOURI Bor. GARD. 78: 224-244. 1991. Volume 78, Number 1 1991 Carvalho New Taxa of Solanum 225 FIGUR Sola anum caldense (from Widgren s.n., the petiole densel (abaxial) face. (2) Ventral (adaxial) face. longis, racemosa, pedunculo 20-35 mm longo instructa. Flores globosi pubescentesque. Calyx urceolatus, externe tomentosus pilis dendriticis. Fi ructus haud vidi. Small trees with fistulose branches; indumentum whitish or yellowish, densely floccose on youn branches, the trichomes stellate-pedicellate, den- dritic or paleaceous(chaffy)-fringed, short and soft. 1845, R). Corolla ue stellate, densely tomentose with long-pedicellate stellate trichomes. — — А. Bud oblong, densely ее outside. — В. iole and leaf base, entose with stellate-stalked trichomes. — D. Anthers linear, the filament dL (1) Dorsal Leaves solitary, oblong to oblong-lanceolate, 17— 22 cm long and 9-11 cm wide, glabrate above with stellate-pedicellate or rarely porrect-stellate trichomes on the main veins, densely lepidote-to- mentose beneath with stellate-pedicellate orna- mented trichomes, central cells present, 47—60 um diam., 8-9(-12) radial cells 20-34 um long, free 226 Annals of the Missouri Botanical Garden portion 17-32 um; petiole са. 2.5 cm long, densely tomentose with stellate-pedicellate trichomes. In- florescence short pedunculate, ca. 2-3.5 cm long, extra-axillary, robust, congested, 9-15 cm long, cymose-dichotomous of about 80 flowers in series rather than glomerules, rachis straight, 6-11 cm long, densely floccose with stellate-pedicellate, den- dritic and fimbriate-paleaceous trichomes; peduncle straight, 2-3.5 cm long. Flowers with buds oblong, sessile; calyx urceolate, ca. 1.6 cm long, densely tomentose outside, the lobes equal, acute, ca. 4 m long; corolla campanulate-stellate, 2 cm long with short tube and evident vascularization, densely tomentose with long-pedicellate stellate and den- dritic trichomes inside; anthers linear, ca. 5 mm long. Fruits unknown Paratype: BRAZIL. MINAS GERAIS: Municipality of Caldas, i Feb. 1876 (8), H. Masén 4295 (R). Distribution and habitat. During 1841-1876 the region of Caldas in the state of Minas Gerais hosted a vegetation classified as broad-leaved trop- ical forest. Solanum caldense occurred there at that time together with Solanum cernuum; a more recent collection has not been seen. The flowers appear in February. Affinity. Solanum caldense is similar to So- lanum pachinatum Dunal with respect to the stel- late-pedicellate, dendritic and chaffy-fringed soft trichomes, which are somewhat united like the tuft- orming, densely floccose indumentum on the branches, inflorescences, and on lower parts of the petiole It is distinguished principally by the erect, long rachis with flowers in series and by the ur- ceolate calyx The species is named after the locality of Caldas, where the species was first found. 2. Solanum carautae Carvalho, sp. nov. TYPE: Brazil. Rio de Janeiro: Municipality of Parati, Parati-Mirim, 8 Dec. 1976 (8, fr), D. Araujo, M. D. Vianna, К. F. Oliveira & P. J. Р. Carauta 1421 (holotype, GUA). Figures 2, 3D-G, 14C. Frutex vel arbor indumento lepidoto-adpresso, tricho- matibus peltatis rariusve peltato-stellatis. Folia g eminata, fere conformia, membranacea, late пен а 0-14 mm longa et 42-50 mm lata, indumento argenteo vel obscure cinereo pilisque peltatis in facie superiore, енед peltatis et a уры enge r cym vel dicho a usque ad 6.5 mm a, pedunculo 2- 3. 5 sip sets rachi brevi оош pauciora (floribus ad 10). Fructus kh 0.9 mm diametro, calyce ac- crescente involucrato, indumento adpresse lepidoto or- natus. Shrubs or trees 3-6 m tall, young branches slightly striate and angulate, the indumentum sparsely lepidote, silvery and somewhat dark with peltate trichomes. Leaves sparse, geminate, the larger leaves membranaceous, broadly lanceolate, 7-14 cm long and 4.2-5 ст wide, apically long- acuminate, almost falcate, basally attenuate, slightly asymmetric, the margin slightly inrolled, secondary veins 10-18, prominent beneath, camptodromous, discolorous, the pubescence argenteous ог greenis beneath, appressed-lepidote with peltate trichomes and peltate-stellate trichomes with 8-20 radial cells, sparsely pubescent above; the smaller leaves similar but only 7 cm long; petiole ca. 1 cm long. Inflo- rescences opposite the leaves or extra-axillary, erect, a dichotomous cyme of ca. 15 flowers, ca. 6.5 ст long, the rachis scorpioid, 1.5-2 cm long, ap- pressed-lepidote, the peduncle 2-3.5 cm long. Flowers with pedicels ca. 1 cm long; calyx cam- panulate, 2.4 mm long, the lobes acute, 1-2 mm long, appressed-lepidote outside; corolla white, ro- tate-stellate, 1.2 cm long and 2.5 cm diam., the lobes lanceolate, 8 mm long; anthers oblong, sub- sessile, 4-5 mm long, 1 mm wide, the filaments 0.50 mm free. Berry globose, ca. 9 mm diam.; fruiting pedicels erect with dense, peltate and pel- tate-stellate indumentum; fruiting calyx accrescent to cover most of berry, ca. 1.2 cm long, the lobes lanceolate, sparsely pubescent with peltate trich- omes. BRAZIL. RIO DE JANEIRO: Municipality of Par pr beach, 20 Angra dos Reis, Ilha Grande, road to Palmas Apr. 1980 (fl), M. К. V. Barbosa 31 (GUA). Distribution and habitat. The species is 4 heliophyte from Atlantic forest and sandbank (res- tinga) of the Municipalities of Angra do Reis and Parati in the state of Rio de Janeiro. In the “restinga” the pe was observed in flower and fruit in December, no significant variation was noticed in ај т this shrubby plant formation. olanum carautae is related to S. swartzianum and especially to its subsp. argyrophyllum (Dunal) arv., a taxon also occurring in the southern Bra- zilian forests of the states of Rio de Janeiro. Espirito Santo, and frequently in Minas Gerais and Bahia. It differs in its dark or silvery indumentum» its broadly lanceolate, long-acuminate leaves, whic may grow to 14 cm long and have a membrana- ceous texture, and in its long-stalked inflorescence with scorpioid rachis This epithet бина botanist Pedro J. P. Carauta, researcher of the Fundacao Estadual de Meio Am- Volume 78, Number 1 Carvalho 227 1991 New Taxa of Solanum | | Solanum carautae (from Araujo et al. 1421 GUA).—A, B. Cymose a r amg f а inflorescences, the rachis with a scorpioid cicatrice of pedicel articulations. — C. Attenuate lea ү View. (2) Dorsal view. 228 Annals of the Missouri Botanical Garden FicURE 3. Solanum carautae and Solanum hatschbachii. A-C. Solanum hatschbachii (from Hatschbach 26837 BH).— A. Open flower showing the e calyx lobes. — B. Open flower s C. Fruit, the calyx covers most of the berry. anum carautae (from Araujo et al. 1421 GUA).— D. Open flower showing the campanulate calyx and rotate- еке corolla with lanceolate lobes.— E. Gynoecium.— F. Stamens with oblong, S sil anthers. (1) Ventral face. (2) Dorsal face. —G. Anther apex. Volume 78, Number 1 1991 Carvalho 229 New Taxa of Solanum biente, Departamento de Сопзегуасао Ambiental, Rio de Janeiro (FEEMA), and one of the collectors of the type specimen. 3. Solanum davidsei Carvalho, sp. nov. TYPE: Venezuela. Bolivar: La Gran Sabana, Km 145 along hwy., 2 km S of La Ciudadella, 3 Dec. 1973 (fl, fr), G. Davidse 47 16 (holotype, MO; isotype, VEN). Figures 4, 14D. Frutex 1.5 cm altus, indumento aurantiaco lepidoto- adpresso pilis peltatis rariusque peltato-stellatis formato, rami flexuosi, internodiis usque ad 1-5 m longe cuspidata, basi acuta asymmet et 2-3 mm lata; petiolus ca. 1-2 mm longus. In centia pauciflora, in cymis simplicibus disposita, erecta, rachi valde reducta, prope 3-5; corolla alba campanulato- stellata, laciniis longe lanceolatis. Fructus calyce ampliato inclusus. Shrubs to 1.5 m tall, the appressed-lepidote in- dumentum of peltate trichomes showing an intense golden color; young branches flexible and plane. Leaves unequal-geminate, sparse on branches, the internodes 1.5 cm long, the large blades broadly lanceolate, 3.5-8 cm long and 1.8-3.2 cm wide, the apex long-cuspidate or long-acuminate, the base acute, asymmetric, the margin inrolled, discolo- rous, glabrate or sparsely appressed-lepidote appressed-lepidote with peltate and peltate-stellate —44 um diam. with 33-37 radial cells 16-22 um long, secondary veins 13-15, camp- todromous; smaller leaves ca. 3 cm long and 1.2 cm wide; petioles ca. 1.2 cm long, canaliculate, appressed-lepidote. Inflorescence extra-axillary, 5 cm long, simple cymes of 3—5 flowers , above trichomes 3 terete. Flowers with the pedicel teil basally ar- ticulated, ca. 1 cm long; calyx campanulate ca. 4 mm long, the lobes acute, ca. 2 mm long, indu- mentum appressed-lepidote on both surfaces; co- rolla white, campanulate-stellate, ca. 1.5 cm long, ca. 2 cm diam., the lobes long-lanceolate, ca. 8 mm long, indumentum appressed-lepidote outside; anthers ca. 4 mm long, equal, oblong, subsessile; style with the stigmatic region spatulate. Berry globose; calyx accrescent and covering most of the berry; seeds flattened-reniform, ca. 2 mm long, 2 Paratype. | VENEZUELA. BOLÍVAR: la Gran Sabana, 49 km W of intersection of main road to Sta. Elena and road to ke па hillside with savanna on upper slope, forest in gallery, 1,360 m, Davidse 4752 (MO). Distribution and habitat. This new species is indigenous to the Gran Sabana region of southern Venezuela at an elevation of ca. 1,300 m. It was found growing with dominant Gramineae of the pma Echinolaena, Paspalum, and Trachypo- ^ Affinity. вору созі is ена, related p о its subsp. E (Dunal) Carv. from mie and Ven- ezuela. The two species share an intense golden appressed-lepidote indumentum and few-flowered inflorescences. Solanum davidsei differs from So- lanum swartzianum subsp. chrysophyllum in its broadly lanceolate, long-cuspidate leaves, which occur sparsely along the flexuous branches, and in the long-pedunculate reduced rachis of the inflo- rescence. o Solanum 4. Solanum hatschbachii Carvalho, sp. nov. TYPE: Brazil. Paraná: Municipality Bocaiuva do Sul, Sumidouro, 6 July 1971 (8), G. e 26837 (holotype, BH). Figures 15A. utex indumento lepidoto-adpresso pilis peltatis et pel- tato-stellatis, 17-45 cellulis radiatis interdum apiculatis constitutis. Folia geminata, inaequalia, chartacea, anguste lanceolata, ad 100 mm longa et 2.7 mm lata usque, supra indumento argenteo vel cinereo vestita subtusque glabra. Inflorescentia cymosa aut subdichotoma sive pseudo-cor- ymbosa, usque ad 2.5 mm longa, breviter pedunculata pedunculo 1.5 mm longo rhachi scorpioidea pauciflora (floribus ad 10). Fructus incognitus. Shrubs to 2 m tall, young branches terete and striate, indumentum argenteous or greenish, lepi- dote-appressed with peltate trichomes. Leaves gem- inate, unequal, sparse on the branches; larger leaf blades chartaceous, lanceolate, 7-10 cm long x 2-2.7 cm wide, apically acuminate, basally acute and somewhat asymmetric, the margin entire to revolute, 10-18 secondary veins prominent on the dorsal surface, camptodromous, discolorous, in- dumentum appressed-lepidote with peltate and pel- tate-stellate trichomes on the dorsal surface, gla- brous above; smaller leaves to 3.5 cm long, apically obtuse. Inflorescences opposite the leaves or extra- axillary, erect, ca. 2.5 cm long, simple cymes or subdichotomous to somewhat pseudocorymbiform with ca. 10 flowers, rachis кчы nd elongate, ap- pressed-lepidote; peduncle short, ca. 1.5 cm long Flowers with the pedicel cd to 3 mm long; buds globose; calyx campanulate, ca. 6 mm long, the lobes acute, to 4 mm long, corolla white, rotate- stellate, ca. 1.5 cm long and 2 cm diam., the lobes lanceolate, to 9 mm long, indumentum appressed- lepidote, the outside of the calyx and the corolla with peltate trichomes and their venation formed of 5 median and 10 weak laterals, hardly ramified, acropetally anastomosing; anthers equal, oblong, 230 Annals of the Missouri Botanical Garden puc oW Y ES [ v = + 7, S EY + ч AN 9 Зе va > 16.67. ree e e D C: 5mm 58 се Sh Уа. D a FIGURE Solanum davidsei (from Davidse 4716 MO).—A. Open flower showing the campanulate calyx, the campanulate-stellate corolla, and the long style. —B. Corolla showing the long-lanceolate lobes and appressed-lepidote Volume 78, Number 1 1991 Carvalho 231 New Taxa of Solanum ca. 5 mm long, connivent, subsessile. Fruit un- known. Paratype. BRAZIL. PARANÁ: Municipality Bocaiuva do Sul, uan 10 May 1914 (fl), P. Dusén s.n. (S). ~ Distribution and habitat. This species occurs in the state of Paraná in the municipalities of Ja- guariaiva and Sumidouro. The botanist P. Dusén collected it in primary forest, and Hatschbach col- lected it in secondary formation. Flowering, still with leaf branches, is in May an y. Affinity. Solanum hatschbachii approaches Solanum swartzianum Roem. & Schult., espe- cially its var. Swartzianum, but is distinguished by its lanceolate leaves up to 2.5 cm wide and by its leafy habit with leaves regularly spaced. This epithet honors the collector, botanist Gert Hatschbach, from “Museu Botanico Municipal” of the Curitiba state of Paraná, a great connoisseur of the plant life of this region. л . Solanum lepidotum var. lepidiochlamys Bitter ex Carvalho, var. nov. TYPE: Colombia: Cordillera Occidental, 1,900 m, 9 Sep. 1899 (8), E. Langlassé 58 (holotype, С; isotype, P, fragment, F). Figures 10, 15B. i apo A E subsp. о Вит. mss. in d. herb, Genéve (G), nome Arbuscula indumento adpresse lepidoto aurantiaco ves- tita кыса dene peltatis atque peltato-stellatis. Folia = o ne formaque inaequalia, lanceolata, а. 6 a et 3 cm lata, apice recta vel falcata, basi E tusa ar dida leviter asymmetrica, margine in- tegra et revoluta. Inflorescentia multiflora, erecta, laxa rachi scorpioidea. Fructus haud vidi. Small trees to 3 m tall, indumentum appressed- lepidote, golden or rarely argenteous with peltate and rare peltate-stellate trichomes. Leaves gemi- nate, unequal in size and form, the larger leaves chartaceous, lanceolate, 5.3-10 cm long and 1.5- 3 cm wide, apically acuminate, straight or falcate, basally obtuse, decurrent, somewhat asymmetric, the margin entire to revolute, the indumentum appressed-lepidote above, glabrescent beneath, pel- tate and peltate-stellate, the trichomes, 30-47 um diam. with 27-33 radial cells, 18-25 um long, the free portion 2-7 um. Inflorescence long-stalked, 5-6.5 cm long, erect, cymose-dichotomous of about 100 flowers, straight or scorpioid at the end of the rachis, ca. 2.5 cm long; peduncle ca. 30 cm long. Flowers as in the typical subspecies. Fruit un- known. Paratypes. COLOMBIA. ANTIOQUIA: Angelopsis La Ca- ep 1,800 m, 22 Jan. 1928 (fl), Toro 884 (NY), Mayor 01 (Z). cauca: El Tambo, n B , 22 July 1938 (fl), эЛ ум 1495 (S); Chisquio, os Derrumbos, m, 3 Apr. 1949 (fl), Asplund 10716 (МУ, 5); Hague, 3 Sep. 1844 (8), Goudot s.n. (BM, F, MO, NY, S). VALLE: near River Cali, Pichinche, 1, 700 m, 1946, Duque- Jara. millo 3935 (NY). Distribution and habitat. This plant grows exclusively in Colombia at 1,700-1,900 m altitude in the same place as Solanum lepidotum Dunal var. lepidotum. Flowering is in September. Affinity. Solanum lepidotum var. lepido- chlamys differs from the typical taxon, Solanum lepidotum Dunal var. lepidotum, by the lanceolate form of the leaves and the densely leafy branches. The name of this taxon alludes to the plant's lepidote indumentum. 6. Solanum lepidotum var. trianae Carvalho, var. nov. TYPE: Colombia. Province Quindio: 1,800 m altitude (Nouvelle-Genade), between 1851 and 1857, J. Triana s.n., Voyage de J. Triana, 1851-1857 (fl) (holotype, P; iso- types, BR, C. NY, W). Figure 15C. Arbusculo indumento argenteo adpresse Bir pilis peltatis ornato. Folia solitaria et geminata inaequa a mbranacea lanceolata ad oblongo geni eg usque ad 150 mm longa et 4.5 mm lata apice cuspidata vulgo falcata, basi obtusa margine integra trichomatibus peltatis ca. 47 um diametro, cellulis radialibus fere ommino coali- tis; petiolus ad 10 mm longus. Fructus desunt. Trees, indumentum appressed-lepidote argen- teous of peltate trichomes. Leaves sparse, solitary or geminate on young branches, unequal in size and form, the larger leaves membranaceous, from lanceolate to oblong-lanceolate, to 15 cm long and 4.5 cm wide, apically cuspidate to reflexed (fal- cate), basally obtuse, margin entire, trichomes pel- tate, ca. 47 um diam., with ca radial cells almost totally connected; petioles ca. 1 cm long. Flowers as in the typical subspecies. Fruit un- nown indumentum outside. (1) Apex acute and base acuminate. (2) Apex long-cuspidate and base acute and asymmetrical. — simple cymes with a short rachis and terete peduncle — C. Accrescent calyx covering most of the young berry. — D. Examples of leaf bases and apices. — Е. Inflorescence, 232 Annals of the Missouri Botanical Garden F FIGURE 5. Solanum oliveirae (from R. Е. de Oliveira et al. 455 GUA).— A. Flower bud. — B. Flowering calyx.— C. Corolla, campanulate-stellate. —D. Leaf base and petiole apex.—E. Gynoecium showing ovary pubescence and demarcated stigmatic region. — Е. Anthers, linear with elongate filaments. Моште 78, Митбег 1 ' Carvalho 233 1991 New Taxa of Solanum Ух ~ —! D CNN YE А Ergo Laa 4 "Y « [v ЊУ Solanum pereirae (from E. Pereira 1270 RB). — A. Leaf bases. (1) Dorsal view. (2) Ventral view. — B. Corolla, campanulate-stellate. —C. Calyx, campanulate with paleaceous-laminate or fringed trichomes. 234 Annals of the Missouri Botanical Garden EM -> " LETS ~ NAS SA ом sí ЖЛ > 2% DIN Y * e pt С $7 bree ci y Newnes dE dA У хе = eX, ~ MM 6234 по کم‎ بر‎ ~ + x ag an SS ya o 5, (4 Mov T. v б ` a M ~ ERE My 1 POR ES Ni ИХ YU -. ` FIGURE 7. Solanum sooretamum (from R. P. Belem 1534 UB).—A. Calyx, densely floccose with stellate (f m B. Open corolla, campanulate-stellate with stamens and style.—C. Fruit, a globose berry partly covered by the enlarged calyx. — D. Anthers, linear with short filaments. — E. Seeds. — F. Scorpioid inflorescence branch with cicatrice of pedicel articulations. Моште 78, Митбег 1 1991 Carvalho New Taxa of Solanum 235 FIGURE 8. Solanum caldense (from Widgren s.n. R).— A. General : view "ој stellate- pedicellate trichomes (50 um scale). — B. S tellat gl stalk with dd expansions. Distribution and habitat. The J. Triana col- lection was distributed to various herbaria (BR, C, NY, P, W) without indication of collection locality and sometimes with mention of only the em and altitude. This plant grows at 500-1,500 m Quindio in the department of Huila, and it is known only by the type collection. Time of flowering is Solanum lepidotum var. trianae dif- fers from the typical taxon in its argenteous in- dumentum, which is formed of large diameter pel- tate trichomes with radial cells almost totally connected and also in having a lanceolate to oblong- lanceolate leaf blade up to 15 cm long and 4.5 cm The epithet honors botanist J. Triana, collector of the type specimen. " Solanum oliveirae Carvalho, sp. nov. TYPE: Brazil. Rio de Janeiro: Varre-Sai, 4 Sep. 1984 (8), R. F. de Oliveira et al. 455 (holotype, GUA). Figures 5, 11, 16A Frutex ramis rigidis dense floccoso-paleaceis castaneis, ne paleaceo-fimbria ac inaribus vel stellatis. Folia solitaria E angustata, apice obtuso plus minusque emarginata, 160-260 mm longa et 90-110 mm lata, supra glabres- centia, scabra, pilis stellatis, infra pallidiora, cinerascentia, dense lepidoto-tomentosa, pilis stellatis raris ornata se = a FIGURE 9. Solanum caldense (from Widgren s.n., R).—C-F. Paleaceous-fimbriate trichomes; fringed-chaf- fy, stellate with a central cell, and ornamented pedicel (50 um scale). cundum nervos; petiolus ca. 30 mm longus. гиев strigosus. Bacca globosa, calyce ampliato inclusa. Shrub 2-6 m tall with rigid or fistulose branches, the indumentum whitish or yellowish, densely floc- cose, chaffy-fringed of dendritic and rarely chaffy- laminate and stellate trichomes with multicellular pedicels. Leaves solitary, oblong, 16-26 cm long and 9-11 cm wide, apically obtuse to obtuse-emar- ginate, basally obtuse, glabrescent above with stel- late pedicellate trichomes bearing an apiculate cen- tral cell 66-72 um diam., (5-)8 radial cells, 30— 39 um long and the free portion 25-36 um; densely lepidote-tomentose beneath with stellate trichomes 35-44 um diam. having 14-16(-17) radial hori- zontal cells, 16-21 um long, the free portion 10- 16 um long, short-stalked, ornamented with a small apiculate central cell, rarely with soft, hyaline stel- late trichomes that have ca. 20 radial cells and multiseriate pedicels on the main veins; secondary 236 Annals of the Missouri Botanical Garden та Hp ж. FicunE 10. ys (from Langlassé 58 G trichomes (50 um scale). — (20 um scale). Solanum lepidotum var. lepidiochlam- — А. General view of stellate B. Structure of the trichome veins 15-18; petiole 1.2-5 cm long, densely floc- cose-chaffy, laminate and fringed-chaffy (laminate), and dendritic trichomes. Inflorescence extra-axil- lary, pedunculate, robust, congested, 14-17 cm long, cymose-dichotomous of about 30 flowers in series rather than glomerules, the straight rachis densely floccose, paleaceous, with short-laminate and fimbriate trichomes that do not involve the flowers; peduncle straight, ca. 3 cm long, chaffy- laminate. Flower buds oblong, 5-6 mm long; ped- icels ca. 3 mm long; calyx urceolate or broadly tubular, longer than 1 cm, strigose, chaffy-fringed outside, short, rigid, the lobes equal, acute, ca. 2 mm long; corolla campanulate-stellate, ca. 2.5 cm long, the lobes equal, ca. 1.5 cm long, the tube short, ca. 3 mm long, pubescent outside with stel- late trichomes having ca. 8 i trichomes with ca. 8 radial cells that are rarely stellate multiseriate; anthers linear, ca. 6 mm long, .4 cm long, the stigmatic region demarcated by papillae. Berry globose, ca. radial cells, and also subsessile; style ca. 1 cm diam., enclosed in the expanded calyx. Paratypes. BRAZIL. Without other a Regnell Ш ж, fr) (S). ESPÍRITO SANTO: Mimoso, 1916, Souza FiGURE 11. ra et al. 455 GUA). — A. Structure of dendritic trichome with uniseriate pedicel a um scale). — B-D. Paleaceous- laminar trichomes, the кы cells tufted and the base ornamented (50 um scale Solanum oliveirae (from R. F. de Olivei- Brito 149 (В) (В). va Е JANEIRO: Cacimbas near River Itabapaoana, a 9, Sampaio 906 (R); fields, Bon- sucesso farm, Mar. 8, Sampaio 2864 (R); i dis. Serra da Estrela, M 14123 (fl) (NY). SAO PAULO Reserva do Parque Estadual das Fontes do Ipiranga, 3 Feb. 1983 (fl, fr), Macedo 13 (SP). Distribution and habitat. The yellow flowers, which are infrequent in this group, were indicated only for the holotype, collected in the state of Rio tropical humid forest in the states of Espirito Santo and Rio de Janeiro. Affinity. Solanum oliveirae is distinguished from the similar Solanum cernuum by length of the chaffy-laminate indumentum of the inflores- cence, which is shorter than the flowers so that the flowers are evident. The strigose aspect of the chaffy-laminate covering on the outside of the ca- lyx, the short and floccose pubescence types on stems and branches, as well as the shape and con- Volume 78, Number 1 1991 arvalho 237 New Taxa of Solanum FIGURE 12. Solanum pereirae (from E. Pereira 1270 RB).—A. General view of peltate trichomes (50 um scale). — B. S um scale). — C. Structure of peltate- stellate trichome (20 , um scale), sistency of the leaves, are the main distinctions of this species. The epithet honors botanist Ronaldo F. Oliveira, Fundação Estadual de Engenharia do Meio Am- biente, Departamento de Conservação Ambiental, Rio de Janeiro (FEEMA), a member of the group that collected the type specimen. eo . Solanum pereirae Carvalho, sp. nov. TYPE: Brazil. Rio de Janeiro: Municipality of Santa Maria Madalena, Tamandare, 18 Mar. 1955 (fl), E. Pereira 1270 (holotype, RB). Figures 6, 12, Arbuscula ramis rigidis aut. fistulosis, novellis suba ап- glabra, in inferiore squamosa castanea, pilis peltato- а porrecto-stellatis supra ner- vum medium insertis; iolus floccoso-to omentests, tri- chomatibus stellatis ае dentriticis raro fim- briatis. Flores cymosi, congesti, extra axillares cymis 1- FIGURE 13. Solanum steyermarkii (from J. A. Stey- ermark 33825 Е).— А. General views of peltate and peltate-stellate trichomes (50 ит scale). — B. Structure of peltate trichome (20 um scale). — C. Structure of peltate- stellate trichome (20 um scale). 3 ramosis, я e pig а laminacea; pedunculus m pubescens. Calyx campanulatus, ex- tus и. pilis stellatis 2d IL. vestitus. Fructus ignotus. Shrub 3-3.5 tall, young branches plane, indu- mentum brownish, appressed-lepidote with peltate- stellate trichomes having a multicellular pedicel. Leaves solitary, lanceolate to broadly lanceolate, 14–40 cm long and 6.5-12(-20) cm wide, apically acuminate, basally obtuse-rotund to asymmetric; glabrescent above, densely appressed-lepidote be- neath with peltate-stellate trichomes, 35-39 um diam. with 15-16 radial cells, 15-17 um long, the peltate-multiseriate; secondary veins 14—20; pet- iole 1.5-4 cm long, appressed-lepidote, the stellate trichomes with ornamented long stalks. Inflores- cence robust, extra-axillary, pedunculate, 8-20 cm long, cymose-dichotomous, flowers in series, rarely congested, the rachis straight, 3—4-branched, ca. 6 cm long, densely paleaceous, with long, laminate 238 Annals of the Missouri Botanical Garden FIGURE = Solanum caldense, S. carautae, and S. davidsei. a, b. Solanum caldense (from Widgren s.n. R). . Portion of cymose-dichotomous | s straight rachis. —c. Solanum carautae. Habit with abit. - flower —d. S. davidsei. Habit sta flowers and f Volume 78, Number 1 Carvalho 239 1991 New Taxa of Solanum | ME ње red "yes, Ia a оле 15. Solanum hatschbachii, S. lepidotum var. И Е, S. lepidotum var. trianae, and 5 oliveirae. —a. Solanum hatschbachii (from Hatschbac h 26837 BH). Habit with flowers. — b. d um lepidotum var. Shee are ane (from Langlassé 58 G). Habit with flowers.—c. Solanum lepidotum var. trianae (from Triana s.n. Р). Habit with flowers.—d. Solanum oliveirae ння Oliveira et al. 455 GUA). Habit with feuis -dichotomous flowering. 240 Annals of the Missouri Botanical Garden trichomes enveloping the flowers; peduncle ca. 3 cm long, chaffy-laminate. Flowers with buds ob- long, 2-4 mm long; pedicel ca. 3 mm long; calyx campanulate, ca. 2 cm long, pubescent with stellate trichomes with ornamented stalks and paleaceous- laciniate or fringed trichomes outside, the lobes ca. 7 mm long; corolla campanulate-stellate, ca. 2 cm long, c cm diam., with trichomes with orna- mented stalks, the lobes ca. 1 cm long; anthers linear, ca. 3 mm long, subsessile, the filaments ca. ] mm long; ovary globose, tomentose, ca. 9 cm long. Fruit unknown. Paratypes. BRAZIL. ESPÍRITO SANTO: near Domingos Martins, S. Miguel, 8 Feb. 1973 (fl), Hatschbach 31386 (BH, С, В). PERNAMBUCO: May 1978 (fl), Burle-Max s.n. (RB). вю DE JANEIRO: Municipality of Santa Maria Madale- na, Lisboa s.n. (fl) (RB); farm Mater Bem, 18 Feb. 1981 В), С. Martinelli et al. 7610 (К, MO, ВВ); Martinelli et al. 7610 (fl) (K, MO, RB); 24 Nov. 1977 (В), Mautone 9 (RB); road to Pedra da Agulha, 7 Mar. 1978 (fl), M. C. Viana et al. s.n. (GUA, HB); Serra de Estrela, 14 May 1864, 22 June 1873, Glaziou 1077, 6658 (8) (both P). Distribution and habitat. This plant grows in the forested parts of the city of Santa Maria Mada- lena, flowering in February and March. nity. Solanum pereirae is similar to So- lanum cernuum but differs in its long chaffy-lam- inate trichomes, which are present only on the branches of the inflorescence. It is also suggestive of Solanum vellozianum because of the brownish appressed-lepidote tomentum of the leaf under- sides. The epithet honors Edmundo Pereira, research- er at the Botanical Garden of Rio de Janeiro, col- lector of the type specimen, a great student of the plant life of Brazil, and a teacher of systematic botany. 9. Solanum sooretamum Carvalho, sp. nov. TYPE: Brazil. Espirito Santo: Forest Reserve of Sooretama, 9 Aug. 1965 (fl), R. P. Belem 1534 (holotype, UB; isotype, CEPEC). Fig- ures 7, 17A, B. Arbuscula ramis rigidis aut fistulosis juventude suban- gulatis, castaneis, tomentosis vel floccosis trichomatibus stellatis pedicellis longis Werken expansis, dentriticis fimbriatisve. Folia lanceolata, basi leviter cuneata, 100- 0 mm longa et 50-160 mm lata, supra glabra vel glabrescentia subtusque dense lepidoto-tomentosa colore castaneo imbu 13-40 mm longus. Inflores- centia cimoso-corymbosa, multiflora, longius pedunculata, erecta, floccosa pilis stellatis pedicellatis lateraliter excre- centibis ornatis; rachis apicem versus scorpioidea. Flores subsessiles. Calyx campanulatus, extus floccoso-stellatus, mm longus. Corolla ad 9 mm longa et ca. 20 mm diametro. Bacca globosa puberula, laciniis calycis inae- qualibus paullo supra medium circumvallata. Small tree to 6 m tall, young branches ca. 6 cm diam., subangulate, striate and fistular, indu- mentum brownish, appressed-lepidote and lepidote- tomentose to densely floccose, peltate-stellate, trichomes large, dendritic and chaffy-fringed, with an apiculate central cell and a long or short pedicel with lateral projections. Leaves solitary, not gem- inate, lanceolate, 10-43 cm long and 5-16 cm wide, apically acute or sometimes attenuate at both ends, basally cuneate, the margin entire, secondary veins 16-40, prominent beneath, camptodromous, discolorous; brownish glabrate; beneath densely lepidote-tomentose with peltate-stellate trichomes, 41-46 um diam. having 16-17 radial cells 15-20 um long, the free portion 9-14 um long; petiole lepidote-floccose, 1.3-4 cm long. In- florescence erect, extra-axillary, opposite the leaves or pseudoterminal, corymbiform cymes ca. 10-35 cm long, with scorpioid branches up to 3rd order, rachis elongate, ca. 6 cm long, scorpioid at the ends, each with more than 50 flowers; tomentose to densely floccose, with stellate trichomes having lateral expansions on the long stalks and with large fringed-chaffy trichomes; peduncle erect, terete to flattened-angulate, to 9-15 cm long. Flowers with the buds oblong, 4 mm long; pedicel obsolete or short, basally articulated; calyx campanulate, to 9 mm long, densely floccose with stellate trichomes outside; pedicel articulate subsessile; corolla white, campanulate-stellate, to 1.5 long and 2 cm diam., the lobes lanceolate, to 9 mm long, floccose; an- thers linear to linear-oblong, ca. 4 mm long, equal, the filaments ca. 1 mm long. Berries globose, ca. 1 cm diam., above, the fruit walls with sparse peltate- stellate trichomes, partly covered by the enlarged calyx; seeds ca. 23, 4-5 mm long, the testa little- ornamented, striate-reticulate. Sclerocytes (stone cells) globose or somewhat deltoid geminate, ca. 2 mm long, slightly rugose on the surface, adherent inside the wall fruit apex. FIGURE 16. Solanu GUA). E with straight rachis um oliveirae —а. Sola c. Solanum pereirae (from Е о 1270 RB). a Habit. — anum oliveirae (from R. F. de Oliveira et al. 455 — с. Portion of cymose-dichotomous flowering Volume 78, Number 1 Carvalho 241 1991 New Taxa of Solanum Solaminra SV M. Udo Guy PARA PO m ли Det. Lócie ФА. Freire de Carvalho : 1154 Jerdim Botanico do Rio de Janeiro 242 Annals the MUT Botanical Garden mapano Um! vens ADE ot Wasit & al Я PLANTS OF GUATEMALA t ч бах nima аба АД за. “ b | голо АИ eer О фт: ни FIGURE 17. Solanum sooretamum and S. ое а, b 05). —а. Нађи with flowers. — num sooretamum (from R. P. Belem 1534 ortion of c -dichotomous flowering with scorpioid rachis. — c steyermarkii (from J. А. Steyermark 35389 F). ког branch. -c. Solanum Volume 78, Number 1 Carvalho 243 1991 New Taxa of Solanum Paratypes. BRAZIL. BAHIA: Municipality of Porto 5е- nearly so, oblong-lanceolate to ovate-lanceolate, guro, km 19 on road to Vera Cruz and Vale Verde, 4 Apr. 1979 (fl, fr), Mattos Silva et al. 349 (CEPEC, RB); orto Seguro, Estagáo Ecológica Pau-Brasil, 4 Feb. 1972 (8), Eupunino 206 (CEPEC, NY); Santa Cruz Cabrália, 22 Маг. 1978 (fl), Mori et al. 9823 (СЕРЕС); Conceição da Barra, 13 Nov. 1968, Almeida 232 (CEPEC, RB ESPÍRITO SANTO: Linhares, Reserva Florestal da Compa- nhia Vale do Rio Doce, 17 Jan. 1975 (fl), ү еї ji 406 (RB); Lagoa ар 100-200 m, 3 1 (8, fr), Sucre et al. 10136 (RB). bk PAULO: Sern ” Bocaina, Parque sl da Bocaina near River Jac Pintado, 10 Feb. 1959 (fl, fr), Pabst 4754 (HB); Ribeirão Rico, 28 July 1921, Edwal 2568 (SP). Distribution and habitat. Solanum sooreta- mum occurs in the state of Espirito Santo in the forest reserves of Sooretama and of Companhia Vale do Rio Doce, and also in the state of Bahia in the municipality of Porto Seguro in a forest known as ““mata de Tabuleiro,” which is still pri- mary although disturbed enough to produce vig- orous plants of this taxon. It also occurs in the state of São Paulo in the Parque Nacional de Bo- caina. Affinity. Solanum sooretamum is distin- guished from Solanum vellozianum by the length of the inflorescence and by the lepidote-floccose indumentum with stellate long-pedicellate tri- chomes, which occurs on the abaxial surfaces of leaves and on branches. The geminate, globose sclerocytes are evident up into the apex of the fruit wall. The epithet testifies to the plant's occurrence in the forest reserve of Sooretama. 10. Solanum steyermarkii Carvalho, sp. nov. TYPE: Guatemala. Dept. Quezaltenango: after volcano Santa Maria, between property Piri- neos and Los Positos, 1,300-1,500 m, 8 Jan. 1940 (fl), J. А. Steyermark 33825 (holotype, F). Figures 13, 17C scula ramis teretibus rigidis cum pedunculis, cy- t falcata oblongo -lanceolate vel ovato-lanceolata apex lon- guiscule cuspidata, pe asymmetrica, = integra vel leviter sinuata 50-180 mm longa -75 mm lata, supra glabriuscula inira көше Silis peltato- stallate cel- lulis radialibus ad tertiam em cellat е оро вае multifor orae - tomae erectae seins Bacca globosa о. calyce non ampliato, ornata; seminis testa reticulata. Trees 2-35 m tall and 5-10 m diam., branches terete, leafy, indumentum argenteous lepidote with peltate-stellate and rare peltate trichomes. Leaves solitary оп young branches, membranaceous ог falcate, 5-18 cm long and 2-7 cm wide, apically reflexed, cuspidate to long-cuspidate, basally ob- tuse, slightly decurrent, slightly asymmetric, mar- gin entire to sinuate; glabrescent above with sparse- ly peltate-stellate trichomes, densely lepidote beneath with peltate-stellate or rarely peltate tri- chomes 24-30 um diam. having 15-19 radial cells, 9-12 um long, the free portion 0.1-0.2 um, at least Уз connected; 6-10 secondary veins prom- inent on the dorsal surface, camptodromous; petiole 0.4-1.3 cm long. Inflorescence subopposite the leaves, erect, 6-10 cm long, cymose-subdichoto- mous of about 50-100 flowers, rachis scorpioid, long-pedunculate, ca. 6 cm long. Flower buds ob- long, 1-3 mm long; pedicels 3-4 mm long; calyx campanulate, 2 mm long, the lobes acute, to 1 mm long; corolla rotate-stellate, 4 mm long, the lobes acute, ca. 2 mm long; anthers ca. 1 mm long; ovary apically tomentose, the style curved. Berry globose, glabrescent, ca. 9 mm diam.; calyx per- sistent but not accrescent; seeds reniform with re- ticulate testa. Paratypes. | GUATEMALA. DEPT. SUCHITEPÉQUEZ: after Volcán Zunil, near property Las Nubes, 500-800 m, 2 Feb. 1940 (fr), J. А. Steyermark 35389 (F). ALTA VERA- PAZ: 210-250 m, between Hacienda Yaxcabnal along Río Icvolay and Río Apia, 13 Mar. 1942 (fr), та 45027 (Е). HUEHUETENANGO: between Ixcan and Rio Cerro de los Cuchamatanes, 23 July 1942 Steyermark 49220 (fr), о near San Franscisco Miramar, Apr. 1905 (fl), Pittier 72 (F) b ео т Pirineos and Patzulin, 1,200-1,400 m, 9 Feb. 1 dia Standley 86809 (Е); between Santa Мапа де Jesus and Calahuahe, 1,300-1,400 m, 5 Jan. 1940 (8), oo м ipei [a pis MARCOS: Barranco Eminencia n Rafael, 2,500-2,700 m, 6 Feb. 1961 E poet 86234 (Е). HONDURAS. Without other locality, 7 May 1934 (A, fr), Schipp 8-677 (Е, S). MEXICO. CHIAPAS: June 1913, (8) Purpus 6958, 7010 (both F). NAMA. BOCAS DEL TORO: Buena Vista, 1,500 m, 1 Mar. 1928 (fl, fr), Cooper 620 (5); canal area, Barro Colorado Island, Croat 11320 (MO), Foster 5 (MO); Wetmore 200 (MO) Ко Mendoza, Nee 14 Dwyer 13955 (MO). DARIÉN: Cerro Tacarcuna, 1,500 m, 2 Feb. 1975, Can 14107 (COL). PANAMA: Cerro Azul, 2,000 m, i July 1962, Dwyer 2064 (MO); Goofy ale 1,800 m Aug. 1967 (fr), Dwyer 8046 (COL, MO); El Uan. = July 1972 (8), D'Arcy et al. 6045 (MO); 26 Mar. 1975, Dressler 4941 (MO). VENEZUELA. Colonia Tovar, 1854-1855, Fendler 2609 (GOET, к COLOMBIA. CAUCA: Cordillera Central between Tacuejó a Е Palo, 1,450-1,700 m, 21 Dec. 194 Marta, 1, m, 3 casas 8225 (F). CHOCO: Alto de Limón, Rio Sucio, 4 July 1976 (fr), Forero et al. 1827 (C -1 ‚18 Cordillera Occidental between Marina and La Margarita, 244 Annals of the Missouri Botanical Garden T 4 Nov. 1944 (fl, fr), Cuatrecasas (Е); Alto de Dinde between Cartago and Alaclá, 1,200-1,260 m, 16 Dee: 1946 (fl, fr), Cuatrecasas 22886 (F). ECUADOR. Los RÍOS: Sto. Domingo, Dodson et al. s.n. (fr) (MO). PERU. MADRE DE DIOS: near Мапа, 10-11 Aug. 1974 (fl, fr), Foster et al. 3210 (MO) Distribution and habitat. This plant ranges from Guatemala to Peru and Venezuela at eleva- tions from 500 to 2,700 m. Specimens have flowers and fruits in January and February. Affinity. This species is similar to Solanum lepidotum but is distinguished by the indumentum always argenteous colored and constituted of pel- tate-stellate trichomes with radial cells, which are at least one-third connected, by the falcate, mem- branaceous leaves, and by the many-flowered, con- gested inflorescence. The epithet is in homage to the late J. A. Stey- ermark, botanist at the Missouri Botanical Garden, a great connoisseur of the plant life of this region. LITERATURE CITED CARVALHO, L. D'A. FREIRE DE. 1988. Revisáo Тахопо- ADO. Morphology of indumentum and tichames js pisi of Solanum sections Cer- nuum and Lepidotum (Solanaceae). Contributions of the Third International Solanaceae Congress, Bogotá, Colombia (in press). & С. J. SHEPHERD. A new section for the genus Solanum section Cernuum (Solanaceae). Contribu- tions of the Third International Solanaceae Congress, Bogotá, Colombia (in press). ESPECIES NUEVAS Y Oswaldo Téllez Valdés? y Bernice 6. COLECCIONES NOTABLES DE Schubert DIOSCOREA (DIOSCOREACEAE) EN MESOAMERICA! RESUMEN Se describen dos nuevas especies de Dioscorea, D. carpomaculata y D. sumiderensis, ambas de la región cubierta combinación basada en la información provista por colecciones recientes, y se presentan discusiones de especies notables. ABSTRACT wo new bg on of Је D. carpomaculata and D. sumiderensis, are described from the region encompassed by the Flora Mesoam a. Geographical and ecological distributions and notes on flowering and fruiting are presented. Relationshi ips sectional placement are discussed for each species. Based on the information provided by recent collections, one new combination is made, and Pacem and specimen citations of noteworthy species are provided. A raiz del proyecto de la Flora Mesoamericana walled canyon with tropical deciduous forest y el tratamiento de los grupos que la componen, with Hauya, Euphorbia, Diospyros, d colecciones botanicas sistemáticas e intensivas han Trichilia, and Heliocarpus, 800 m, 20 Aug. sido realizadas, tanto en la región que cubre ésta 1976, D. E. Breedlove а D flora, como en el resto de Mexico. Estas colecciones ENCB; isotipo, CAS). Figura 1. proveen las bases de registros de taxa nuevos para Her И : Ap UU ae Pu pu glabrae. Caules sinistrorsum volu- Mexico, asi como para la ciencia, incluyendo las biles. ан (3-16.5-15.8 om ода, (2-)4-15.5 cm иш, especies descritas en el presente trabajo. ovata vel ell oe ig (7-)9-11 nervata. Inflorescentiae Para cada una de las exsiccata se indica el sexo — staminatae paniculatae 10-30 cm longae; flores 2-6 in del especimen consultado, (8) estaminado y (9) pis- cymis; stamina 6, dist tincta, one tepalorum affixa, antheris tilado introrsis; pistillodium con . Inflorescentiae pistillatae racemosae 15-30 cm saan Capsulae 2.4-3.6 cm lon- ‚ 15 mm | 7-9 mm lat t Dioscorea carpomaculata О. Tellez 4 B. G. ш onga, шышы аы aes Schubert, sp. nov. ПРО: México. Chiapas: Herbácea trepadora generalmente pubescente a Mpio. Chiapa de Corzo, El Chorreadero, steep- glabra. Rizomas 0.3-1.0 m de largo, hipógeos, ' Derivado del proyecto de la Flora Mesoamericana, llevado a cabo entre el Instituto de Biología U.N.A.M. México, The Missouri Botanical Garden, U.S.A., y The Museum of Natural History, Great Britain. Agradecemos a CONACyT (PCECCNA-721259, 723355, 723357), a la D.G.A.P.A. al Jardin Botánico de Missouri en especial a, Enrique а y Gerrit Davidse, al Alfonso Delgado, Instituto de Biología U.N.A.M., al Негђапо Gray у Arnold Arboretum, Universidad е , TEFH, US, W, ХАГ, las facilidades para la sulla de al: A Elvia Esparza, M. Rosa Martine: y Regina О. Hughes por las excelentes ilustraciones a las especies. ? Departamento de Botánica, Instituto de Biologia U.N.A.M., Apartado postal 70-367, 04510 Mexico, D.F. * Arnold Arboretum of Harvard University, 22 Divinity Avenue, Cambridge, Massachusetts 02138, U.S.A ANN. MISSOURI Вот. GARD. 78: 245-253. 1991. 246 Annals of the Missouri Botanical Garden napiformes, teretes, cónicos en el ápice, suculentos, corteza membranosa, pardo-amarillenta a color piel, creciendo verticalmente. Tallos levovolubles, lineados, pubescentes a gla- internamente blancos, brescentes en la madurez. Hojas (3-)5.5-15.8 cm )4-15.5 em de ancho, alternas, ovadas a suborbiculares, base cordada, ápice cortiacumi- de largo, (2- nado, haz esparcidamente estriguloso a glabres- cente, envés generalmente densamente сшегео- estrigoso a glabrescente, lámina con gran cantidad de rafidios visibles a trasluz; (7—)9- 1 1 -nervias, los nervios más exteriores generalmente bifurcados; peciolos 6—10 cm de largo, ligeramente sulcados, puberulentos a glabrescentes. Inflorescencias es- taminadas 1-4, racimos de cimas, de 10-30 cm de largo por axila; brácteas exteriores 1 mm de largo, ovado-lanceoladas, brácteas interiores 0.5— 1 mm de largo, ovadas, ambas agudas, pilosas; flores 2-6 por cima, pedicelos 0.5-1 mm; perianto 2.5-3 mm, blanco o amarillento o crema, pubes- cente, los tépalos 2-3 mm de largo, oblongos a oblongo-lanceolados, patentes, ocasionalmente con puntuaciones glandulares translücidas; estambres 6, ca. ] mm, insertados en la base de los tépalos, anteras introrsas, tecas coherentes; pistilodio ca. 1 mm, cónico, ocasionalmente triacanalado. Inflo- rescencias pistiladas 1-4 racimos, de 15-30 cm por axila; brácteas similares a las estaminadas; flores solitarias, subsesiles; perianto 1-3 mm, blan- quecino a pardo-amarillento, pubescente, los té- palos 1-3 mm, oblongos, patentes; estaminodios 6, ca. 0.7 mm, filiformes, insertados en la base de los tépalos; columna estilar ca. 1 mm, estilos bifidos, teretes, delgados. Cápsulas 2.4-3.6 cm de largo, 1.1-1.8 em de ancho, oblongas a elipticas, base y ápice agudos, discoloras por una macula longitu- dinal central pardo-oscura, con forma lacerada, las valvas densamente cinéreo-tomentosas cuando јо- venes a glabrescentes en la madurez, sublenosas; semillas 10-15 mm de largo, 7-9 mm de ancho, oblongas a elipticas, 2 por lóculo, aladas periferi- camente, pardo-rojizas, lisas. Material adicional examinado. MÉXICO. CHIHUAHUA: Guasaremos, Río Mayo, Gentry 2328 8 (F). SINALOA: liacán, Martínez et al. 4152 8 Santa Мапа del Oro, 6 km al E de Santa Téllez & Dávila 8991 82 mo GUERRERO: o Mpio. E Acahuizotla, Agua de Obispo, Zamudio s.n. 8, Dieg 2927 8 (FCUNAM, MEXU); Mpio. Guayameo, 7 km al 5 r. Cd. Alta- de Organo, Tlatlaya, Matuda 31109 à (MEXU); Los Bejucos, Tejupilco, Matuda 31383 8 (MEXU); Cerro de Ixtapan de la Sal, Машаа 31710 (MEXU); Canada de Nanchititla, Tejupilco, Matuda 37456 8 (MEXU). OAXACA: Dist. Sola de Vega, El Vado, 1 km al W de San Sebastian, (MEXU). CHIAPAS: Mpio. Amatenango, along the creek Chenek’ ha’ near Amatenango, Shilom Топ 2681 9 (CAS, ENCB); Mpio. Arriaga, 6 km N of Arriaga, Breedlove 27292 ? (MEXU); Mpio. Huixtla, 6-8 km NE of Huixtla road to Motozintla, Breedlove 25963 8 (MEXU); Malpaso near Siltepec, Matuda 4534 8 (F, GH, MEXU, NY); Mpio. Ixtapa, Zinacantan paraje of Muctajoc, Laughlin 15899 (А); жы Mexican highway 190 in the Zinacantan j uctajoc, о 11874 9 (A, CAS); Мрю. N of Ocozocoautla along road alpaso, incen 20989 9 (CAS, ENCB, MEXU, W of Ocozocoautla, Breed- love 28929 (CAS, "MEXU) mi alpaso, San Cristóbal Las Casas, km 1 1 20, road Chiapa de Corzo- San Cristóbal Las Casas, Schubert & Gómez-Pompa 1747 (MEXU); Mpio. Tonalá, Tonalá, Guzmán 57 à (ENCB); W side Cerro Vernal, 25-30 km SE of Топаја, Breed- love 25635 8 (MEXU); Mpio. Mpio Ти: Margaritas, Co- mitan, Matuda 15784 8 (Е); o. Tuxtla, Colonia Gri- jalva, cerca de Tuxtla Gutiérrez, "cé ómez-Pompa 139 8 (CAS, MEXU); El Sumidero, Cafetal cerca de La Coyota, Miranda 7912 (MEXU); A the road from El Bosque to Simojovel, Shilom Ton 3090 9 (CAS, ENCB); Mpio. Unión Juárez, Finca Juárez, Escuintla, ae ne 1776 à (F, MEXU); Mpio. Venustiano Carranza, 3 mi. SW of Pinola Las Rosas, Breedlove 11360 9 (A, ENCB): Es- peranza, Escuintla, уз 16543 8 (Е, MEXU); Monte Cristo, Matuda 15939 в (Е, MEXU); Mpio. Villa de Corzo, above El ое Оо, а 515599, 51567 8 (CAS); El Chorreadero, 5 mi. E of Chiapa de Corzo Ton 2975 9 (CAS, ENCB); El Chorreadero Park, 6 mi. SE of Chiapa de Corzo, Thorne 40190b (CAS); P Villa Flores, Rancho Margarita, (MEXU); Rancho Lindavista, 24 km E randa a U). GUATEMALA. Heyde & Lux 6391 10 8 8 (А, GH). ESCUINTLA: не (MEXU, ХАІ), 20 (XAL). HUEHUETENANGO: Paso del Boquerón, along Ri His Nue below La Libert Steyermark 51163 9 (F). ad, EL SALVADOR. SONSONATE: Calderón 1691 à (GH). Dioscorea carpomaculata representa un ca- racteristico elemento de la sección Macrogyno- dium Uline del subgénero Dioscorea, cercana- mente relacionado a D. nelsonii Uline. El material o Dioscorea carpomaculata Téllez & Schubert. — A. Planta estaminada. — B. Planta pistilada con una IGUR infrutescencia. — C. estaminodios y la Semilla alada der date (А olumna estilar. — E. Flor pist г estaminada mostrando los 6 estambres y el e —D. Flor pistilada mostrando los tépalos, tilada, ovario y bráct cula. —G. y С, Breedlove 25635; B, D-G, nd 39690). s.—F. Cápsula mostrando la má 247 Téllez 8 Schubert Volume 78, Number 1 1991 Dioscorea en Mesoamérica Е i-e KE t a2 К ( | 7 В U 4 | Б A NE EA oz A 295 ~ J А AO = ES К \ à A یر‎ UE a i c h^ / 8 ea TEER ج‎ = 5 У AP A ка - ” v A La њи 8 8 3 L [u^ Ns ч A i REN "s RS E PY er EF m : CO — Ы A т Со Ls а eO PERS ms Г E ХА > AP Кс; А, A х >, n A Е у 5 DE 4} eT Áf Ц а^, ill RARI х IK 4 / XU ОУ dex. "н Д @ e АМ 23 74 A „> С. m á ^ Е D 5 Y S. e М © " x у de y 248 Annals of the Missouri Botanical Garden arriba citado ha sido frecuentemente confundido con D. cymosula Hemsl., especie erróneamente ubicada en la misma sección Macrogynodium; esta especie posee semillas aladas posteriormente, por lo que pertenece al subgénero Helmia (Kunth) Benth. En este trabajo se propone su reubicación en la sección Chondrocarpa. Ambas especies com- parten un interesante nümero de с tanto de estructuras vegetativas como reproduc como ha sido mencionado por Burkill (1960), 2. & Schubert (1987), y Medina et al. (en prep.) para algunas otras estructuras de numerosas especies de Dioscorea. Dioscorea pue pra var. cinerea (Uline) Téllez & B. С. Schubert, comb. nov. Dios- corea ол Hemsl. var. cinerea Uline, Notizbl. Bot. Gart. Mus. Berlin-Dahlem 7: 202. 1917. TIPO: México. Oaxaca: Knechtel 223 (B). Esta variedad se caracteriza por su pubescencia tomentosa, principalmente sobre hojas y tallos jó- venes, y por las flores ligeramente таз pilosas que en la var. carpomaculata. Habita principalmente en bosques mesófilos, por encima de los 1,500 m, y sólo se distribuye en Oaxaca y Chiapas. Material examinado. MÉxICO. ree Mpio. Chiapa de Corzo, El Chorreadero State Park, 5.6 mi. SE of Chiapa e Corzo, Thorne & Lathrop 40190a (CAS) Pw de Huixtla, Honduras near Siltepec, Matuda 4382 8 (GH, above Soyalo along road to Bochil, Breedlove 10230 8 (CAS). Dioscorea sumiderensis B. C. Schubert & O. Téllez, sp. nov. TIPO: México. Chiapas: Mpio. of La Trinitaria, 15 mi. S of Сотнап along highway; mixed woodland and open grass slopes with limy rocks, ca. 2,000 ft., 20 Sep. 1952, H.S. Gentry 12192 89 (holotipo, GH; isotipo, MEXU). Figura 2 Herbae scandentes glabrae. Caules sinistrorsum volu biles. Caules, petioli et rhachides alati, dentato-serrulati. 8 cm longa, (2.1-)3.3-6.8 cm lata, e paniculatae 5-15 cm lon re n cymis; stamina 3, distincta, basi tepalorum affixa, antheris extrorsis; staminodia 3; pistillodium co- nicum. Inflorescentiae pistillatae racemosae 8-17 cm lon- ‚ 3 longis, filiformibus, 3 brevibus, trian- 8-12.5 mm longae, (4-)6- latae, verrucosae; semina 4-7 mm longa, 3-5 mm lata, circumcirca alata. Herbáceas trepadoras glabras. Rizomas de 5- 15 cm de largo, de 1-1.5 cm de ancho, hipógeos, subteretes, con crecimiento longitudinal, corteza ardo-oscura, internamente blancos o cremas. Ta- los levovolubles. Tallos, peciolos y raquis de las te alados, dentado-se- rrulados. Hojas de (2. 5-)4.5-8.8 cm de largo, de (2.1-)3 .3–6.8 ст de ancho, alternas, ovado-cor- dadas a suborbiculares, membranáceas, seno basal triangular a campanulado, base cordada, ápice agu- do a estrechamente acuminado, 7–9-пегу1а5, cons- picuamente dentado-serruladas, las más exteriores generalmente bifurcadas; peciolos de (0.8-)2.5-8 cm de largo. Inflorescencias estaminadas | a 3 racimos de glomérulos o fasciculos o 1-2 paniculas compuestas, de 5-15 cm de largo por axila; brac- teas exteriores de 1-2 mm de largo, de 0.5 mm de ancho, ovado-lanceoladas a lineal-lanceoladas, acuminadas, brácteas interiores de ca. 1 mm de largo, de 0.4 mm de ancho, ovadas, acuminadas; flores 2-4(-5) por cima, pedicelos 1-2 mm; pe- rianto de 1-2 mm, verdoso o amarillento o crema, los tepalos de 1-2 mm de largo, de 0.5-0.7 mm de ancho, ovales, ovados o cortamente oblongos u oblongo-lanceolados; estambres 3, de ca. 0.8 mm de largo, insertados en la base de los tépalos, las anteras 2. las tecas coherentes, estamino- ios 3, mm de largo, ligeramente ge- Sd. o curvados casi en ángulo recto, con una pequena prolongación surgiendo en la curvatura cerca de la parte media del estaminodio, espatu- láceos y verrugosos en el tercio apical, insertados en la base de los tépalos; pistilodio de 0.1- de alto, cónico o trilobado. Inflorescencias pistiladas un racimo, de 8-17 cm de largo por axila; brácteas exteriores de 1-2 mm de largo, de 0.5 ancho, ovado-lanceoladas, acuminadas, interiores де ca. 1 mm de largo, de 0.4 mm de ancho, ovadas, acuminadas; flores solitarias, sub- sesiles; perianto de 1-2 mm, verdoso o amarillento o crema, los tépalos de 1-2 mm de largo, de 0.5- 0.7 mm de ancho, ovales, ovados a cortamente FIGURA 2. alados, x1.— B. e estaminada abierta mostrando 3 H. Fe j joven, Х15.—1. 2.— C. Flores estaminadas y bot estambres y З estaminodios, x 15. — E. — К. Flor pistilada abierta mostrando los 6 estaminodios y la columna estilar, x15.— C. orte longitudinal de la cápsula, mostrando los 2 lóculos y las semillas en cada uno, x 6. (A-D, Н. S. Gentry 12191; E-I, Н. S. Gentry 12192). — Dioscorea sumiderensis Schubert & Téllez. — A. Planta jóven con rizoma, mostrando tallos y pecíolos Planta estaminada con inflorescencia, X x15.—D. ones florales, Planta ждей con infrutescencia j joven, Infrutescencia, x Volume 78, Number 1 1991 Téllez & Schubert Dioscorea en Mesoamérica 249 250 Annals of the Missouri Botanical Garden oblongos u oblongo-lanceolados; estaminodios 6, 3 largos de 0.4-0.5 mm de largo, filiformes, 3 cortos de ca. 0.2 mm de largo, cónicos, triangulares o piramidales, insertados en la base de los tepalos; los estilos 3, unidos en una columna estilar de ca. ] mm de alto, bifidos, delgados, teretes, los estigmas 6. Cápsulas 8-12.5 mm de largo, de (4-)6-10 mm de ancho, oblongas a obovadas, ligeramente más amplias arriba del tercio basal, verrugosas; semillas 4—7 mm de largo, de 3-5 mm de ancho, oblongas a suborbiculares, 2 por lóculo, aladas pe- rifericamente, pardo-rojizas, lisas. aterial adicional examinado. MÉXICO. CHIAPAS: Mpio. Berriozabal, 5 km E of Berriozabal, Breedlove 20405 9 s Mpio. Mazapa, Río Motozintla between Amatena e la Frontera and Mazapa, Breedlove & Strother 46149 2 (CAS); Mpio. La Trinitaria, 11 mi. S of La Trinitaria, Breedlo ove а: Сина 13243 82 (CAS); way, H.S.Gentry 12191 CR 162 9, CR 163 9, CR 165 $ (XAL); Sumidero, Cox анта 1856 9 (XAL), P 1858 8 (XAL); Mpio. Venustiano Carranza, Soyatitan, from Las Rosas to Pugiltik Shilom Ton oe 2, 3142 8 (CAS, ЕМСВ); Mpio. Villa de Corzo, road to Nueva Concordia, Breedlove 37596 8 (CAS); Mpio. а Flores. Cuchulote- Magda- lena- Villaflores, Robledal, Cid Rojas CR 169 $ (ХАТ). Especie cercanamente relacionada a Dioscorea matagalpensis en la sección Lychnostemon, am- bas poseen estaminodios más cortos que los estam- bres, curvados casi en ángulo recto con una ligera prolongación surgiendo perpendicularmente en la parte media del estaminodio, asi como periantos pequenos, agrupados en cimas muy cortas. Sin embargo, se diferencian por la presencia del ala dentado-serrulada en tallos, peciolos y raquis, y los frutos verrugosos en D. sumiderensis, contra las mismas estructuras todas lisas en D. matagalpen- sis, con excepción del raquis de la inflorescencia ligeramente dentado-serrulado. En el proyecto Flora Mesoamericana, se han realizado colecciones notables de material de al- gunas especies problemáticas. Numerosas especies de Dioscorea han sido descritas de uno de los sexos, situacion que ha causado muchos problemas, el más comun de ellos siendo la descripción de la misma especie en dos ocasiones, una basada en el estaminado y otra en el pistilado. А continuación se brindan discusiones acerca de algunas de estas especies y el material de referencia consultado. Dioscorea chiapasensis Matuda, Bol. Soc. Bot. Mexico 15: 25. 1953. TIPO: México. Chiapas: En matorral claro, Unión Juárez, falda sur del Volcán Tacaná, alt. 1,400 m, 17 marzo 1939, Matuda 2781 8 (holotipo, MEXU; isotipos, F, GH, NA, US) Dioscorea chiapasensis es claramente un ele- mento de la sección Apodostemon, ya que presenta los 6 estambres casi sésiles, el pistilodio cónico a triangular, las semillas aladas periféricamente y el polen estriado. Sin embargo, su fruto es el más pequeño de la sección, entre 8- 5-7 mm, y difiere considerablemente de los caracteristicos de esta sección anotada, ya que es oblongo-eliptico, con la base y el ápice agudos, a diferencia del resto de las especies que presentan frutos con el àpice redondeado a truncado, y con forma oblanceolada a subcuadrada o eliptico hasta redondeado. Esta especie habita principalmente en bosques de Pinus y bosques mesófilos de montana entre los 1,000-1,700 m, siendo endémica del sur de Mexi- co y Guatemala. Material adicional examinado. MÉXICO. CHIAPAS: Mpio. Unión Juárez, Volcán Tacaná a 500 m al E de Talquian, Martínez & Reyes 20333 82, 20334 $ (MEXU). GUATEMALA. QUEZALTENANGO: Finca Pirineos, below Santa ids de Jesús, Standley 68278 8, 68395 8 (F). RE- ALHULEU: Sierra Siete Orejas, 20 km N of San Sebastián, $55 12546 8, 12566 82 (GH, NA). SAN MARCOS: Río Мора, below Rodeo, Standley 68786 8 (F). SOLOLÁ: Río ravo in vicinity of Finca Moca, 5 slope Volcán Atitlán, Steyermark 47960 9 (F, NA). Dioscorea cyanisticta J. D. Smith, Bot. Gaz. (Crawfordsville) 20: 10. 1895. TIPO: Costa Rica. Cartago: Rio Turrialba, Donnell-Smith 4969 8 (holotipo, US; isotipos, F, СН). Dioscorea cyanisticta, de la seccion Brachys- tigma, está cercanamente relacionada a D. den- siflora de la seccion Oxypetalum, (no obstante pertenecen a diferentes secciones). Se caracteriza por sus estaminodios espatulaceos, mas largos que los estambres y muy similares a los de D. densi- flora. Por otra parte, sus frutos son más grandes que en D. densiflora, suculentos, y producen gran cantidad de liquido mucilaginoso mientras madu- ran, y generalmente son glaucos. Ocasionalmente producen tres semillas por lóculo en vez de las dos normalmente encontradas en la mayoría de las especies de Dioscorea. Material adicional examinado. MEXICO. CHIAPAS: Mpio. de Bochil, along river E of Bochil, Shilom Ton 4005 8 (CAS, ENCB); Mpio. de Jitotol, 5 mi. 5 of Jitotol, Breedlove 11936 & (CAS, ENCB, СН); km 18 Col. Cuauhtémoc, Shilom Ton 8154 (MEXU); Mpio. Huixtla, Escuintla, Matuda 1018 8 (СН); Pinabeto, Motozintla, Matuda 15482 8 (F); Mpio. La Trinitaria, km 18 Col Cuauhtémoc, Shilom Ton 8154 8, 8211 8, 8869 8 (MEXU); 10 km ENE of Dos Lagos above Santa Elena, Volume 78, Number 1 1991 Téllez & Schubert Dioscorea en Mesoamérica 251 Breedlove 58413 8 (ENCB, MEXU); Mpio. Unión ери Volcán Tacaná, Talquian, Téllez & Martínez 8939 8 8945 9 (MEXU), 42546 8 (MEX U); Volcán Tacaná, $06 m al E de Talquian, Martínez & Reyes 20331 (MEXU); Volcán Tacaná, Chiquihuite, Matuda 2817 8 (F, МА); Ejido Eureka, Syntex JP1272 à (MEXU). GUATEMALA. ALTA VERAPAZ: Región del Chicoj, NE of Carchá, Standley 70032 8 (F). QUEZALTENANGO: Finca Azucena, above Co lomba, Standley 68090 8 (F); above Mujuhá, entre San Martin Chile Verde y Colomba, Standley 85481 8 (F). $ (MO), Stevens & Grijalva 16411 (МО); Cerro Отава, Stevens & Grijalva 16130 (MO). MATAGALPA: Macizos Perias Blancas, Stevens et al. 21103 8 (GH, MO). C RICA. GUANACASTE: Santa Elena to Tilaran road, Khan et al. 1131 8 (ВМ). ALAJUELA: Reserva Forestal San Lorenzo, Carvajal 164 6 (MEXU, MO). Dioscorea koepperi Standley, Field Mus. Nat. Hist. 9: 269-270. 1940. TIPO: Honduras. At- lántida near Rio Danto; lower slope of Mt. Cangrejal, near La Ceiba, 16 July 1938, ‘uncker, Koepper & Wagner 8482 8 (ho- lotipo, F; isotipos, ЄН. P, US Dioscorea koepperi presenta una serie de ca- racteres que le ubican claramente en la sección Apodostemon, tales como la presencia de seis es- tambres casi sésiles insertados en el toro, pistilodio conico, tripartido, seis estaminodios cónicos у pe- quenos, estilos amplios carnosos, festoneados o um- канн 7 unidos en una corta columna, asi e encuentra aladas cercanamente relacionada а 1). pumicicola ya que comparten importantes caracteres como son, los rizomas esféricos con fuertes adelgazamientos in- termedios entre los cuerpos principales (como pro- bable estrategia de reproducción vegetativa), la forma semejante de las hojas, y la forma y color de los frutos. Representa a su vez una interesante vicarianza, ya que D. koepperi as endémica de Honduras, en una sola localidad en la vertiente Atlántica, en el Monte Cangrejal, entre 0-50 msnm, habitando la selva alta perennifolia. Dioscorea pu- micicola se encuentra solamente en el estado de Morelos, México, en una región muy reducida, habitando la selva baja caducifolia, alrededor de 800 msnm a ко А HONDURAS. ATLÁNTI- : 5 п Јаз гергезаз де! (Sistema de Aguas y ашыны sobre el Rio Danto, por el c o que sale atrás del poblado, selva alta “а ае 10 junio 1985, Téllez & Martínez 8848 (topotipo, MEXU). > >” Dioscorea lepida С. V. Morton, Carnegie Inst. Publ. 461: 248. 1936. про: Costa Rica. Provincia de San José, Laguna de la Chonta, NE of Santa Maria de Dota, alt. 2,000-2,100 m, 18 Dec. 1925, Standley 42133 (holotipo, US; isotipo, NY). D. racemosa (Klotzsch) Uline var. hoffmannii Uli- ne in Engler’s Bot. Jahrb. 22: 431. 1897. TIPO: Costa Rica. Candelaria, Hoffmann 614 8? (holotipo, B). Dioscorea lepida fue ubicada erróneamente en la sección Trigonobasis por Morton (1936) quien anota “easily distinguished from other species of the section Trigonobasis by the small, acute, readi- ly dehiscent capsules." Sin embargo, estas carac- teristicas son compartidas por numerosas especies n el subgénero Helmia. Esta especie se excluye facilmente por poseer estambres en una columna, extrorsos y los estilos en una columna, a diferencia de los estambres libres, introrsos y los estilos tam- bien libres en la sección Trigonobasis. Por las caracteristicas anotadas, esta especie debe per- manecer conjuntamente con D. standleyi en la sección Arestophyton (ver discusion en D. stan- dleyi). Dioscorea lepida tiene un comportamiento re- productivo particular. En primer término, cuando as semillas son dispersadas, éstas ocupan muy fre- cuentemente las horquetas superiores de los árboles principales y más altos del dosel, donde germinan y desarrollan hasta formar pequenos, pero fuertes rizomas epifitos, al igual que D. standleyi. Por otra parte, aparentemente dependiendo de las condi- ciones del medio, su comportamiento “sexual” es particular. Esta especie ha sido colectada en es- pécimen pistilado con todo y su rizoma, y este mismo rizoma ha producido a estami- nados o monoicos alternativ raves de varios años entre 1983-1988; Téllez ©. 9357 monoica (ME Dioscorea inda tambien presenta un intere- sante patron de distribución discontinua a través de Mesoamérica, en áreas que se relacionan con las denominadas “zonas refugio" en partes de Cen- troamérica y el sur de México. Se ha definido asi, ya que la distribucion de esta especie encaja per- fectamente en varios casos en los patrones de zonas con caracteristicas de refugios, dados por Haffer (1982), Prance (1982), Toledo (1982), y Wendt (1989), quienes senalan que son zonas de gran precipitacion y humedad y alta diversidad biolo- gica, predominando tipos de vegetación como la selva alta perennifolia o el bosque mesófilo de mon- 252 Annals of the Missouri Botanical Garden tana. En algunos casos, D. lepida se encontró en las mismas zonas que estos autores determinaron como probables refugios y en otros casos habitando zonas con caracteristicas similares a las de estos refugios. Su distribución geográfica es del sur de México, de Chiapas hasta Panamá en Centroaméri- ca. Habita en la selva tropical hümeda y los bosques mesófilos y de neblina, entre los 350-2,010 m Florece y fructifica de julio a marzo. Material adicional examinado. MEXICO. CHIAPAS. km N of Berriozabal near Pozo Mpio. la Trinitaria, 15 km ENE of Dos Lagos above Santa Elena, Breedlove 56660 82 (CAS); 10 km ENE of Dos Lagos above Santa Elena, Breedlove 57560 82 (CAS). GUATEMALA. ALTA VERAPAZ: Quebradas Secas, Johnson 869 82 (US). CHIQUIMULA: Cerro Tixixi (Tishishi), 3-5 mi. N of Jocotán, Steyermark 31622 8 (GH). IZABAL: Cerro San Gil, ebria 41955 2 (F). HONDURAS. COMAYAGUA: Cerro Azul de Meambar, Hazlett 1853 8 COSTA E. ALAJUELA: Reserva Forestal San of шо along Rio Grande de Orosi, Lent 1813 82 (Е); 8 km S of Tapanti, Lent 1229 8 (F); Finca El Chaparral, А 5 km 5 of bridge оп Agua Caliente at Lourdes, Liesner & Judziewicz 14621 82 (МО). GUANACASTE-ALAJUELA: Miravalles above Búagua, Gómez et al. 19116 8 (МЕХО, MO). PUNTARENAS: Upper Río Burú, Gómez et al. 21518 8, 21756 (MEXU, MO). SAN JOSÉ: trailside above Rio Honduras, Lent 2261 89 (Е); vicinity of El General, Skutch 3021 9 (GH); La Palma de San Ramón, Brenes 5124 89 (F); 3 mi. NE of El Copey, Stork 1629 8 (F). PANAMA. CHIRIQUÍ: Cerro Colorado, -28 mi. from San Felix, Croat 33354 89 (GH). состЕ: 1.5 mi. from El Cope, n 44553 9 (MEXU, MO); El Valle de Antón, La Mes тая Valle, Nevers et al. 6337 (MO). LOS SANTOS Y VERAGUAS: 10 km SW of El Cortezo (Prov. Los Santos), ко 5380 & (МО). PANAMÁ: top of Cerro Campana, D'Arcy 15086 8 (MEXU, МО); Cerro Campana along trail to summit, Miller & Miller 981 8 ( EXU, MO), E pos 9784 (TEFH) Dioscorea standleyi C. V. Morton, Carnegie Inst. Publ. 461: 252-253. 1936. про: Cos- ta Rica. Provincia San José: in wet forest at La Hondura, alt. 1,200-1,500 m, 9 Mar. 1929, P. C. Standley & J. Valerio 51877 9 (holotipo, US). Esta especie se ubica erróneamente en la nueva sección Arestophyton, en el subgénero Dioscorea, no obstante haber sido descrito con base en material pistilado por Morton (1936). Sin embargo, muchas colecciones recientes demuestran que esta especie posee semillas aladas solamente en su parte pos- terior, caracter que la ubica en el subgenero Hel- mia. Esta especie habita primordialmente los bosques hümedos de neblina, mesófilo y pluvial premontano, entre los 300-2,040 m. Florece y fructifica prác- ticamente durante todo el ano. Material adicional examinado. CosTA RICA. ALAJUE- LA: Reserva Forestal е о саа 356 $ (MEXU, МО); Virgen del Зосо Chacón y Herrera 1187 2 (MEXU, МО); Solomon 2473 8 (GH): Vara Blanca de p-2168 8 (GH); San km N of bridge over Quebrada Volio, Stevens 13891 ô ; ca. 13.8 km N of bridge over Quebrada Volio, Stevens 13747 8 (МО); 5.7 km М of bridge over Quebrada Volio, Stevens 14099 à (МО); 15 km NNW of San Ó ] ] 9 (MO). CARTAGO: ~ HEREDIA: 4 mi. ara Bla 3 mi. S of Cariblanco, Croat 35809 à (GH); Braulio Carrillo National Park, Garwood et al. 366 8 (BM, MEXU); road between San Rafael and Rio las Vuel- tas, Stevens 13928 8 (MO). PUNTARENAS: just N of Sta. Elena on Fila Coton, S of Agua Caliente, Davidse et al. D'Arcy et al. 15887 (MEXU, MO), Churchill 5749 8 ; Fortuna Dam area, between Quebrada Los Chorros rada Hondo, Churchill y Churchill 6084 8 (MO); Cerro rads Folson et al. 4727 8 (MEXU, MO); Bocas del Toro, McPherson 7223 8 (MEXU, МО); Fortuna dam site, above Gualaco, Croat 49966 8, 50057 y Herrera 6470 9 (MEXU, МО); vicinity of Bajo Chorro, Woodson y Schery 652 8, 664 8 (GH) BIBLIOGRAFÍA ires I. H. 1960. The organography and the evo- lution of Dioscoreaceae. The family of the yams. J. Linn. Soc. Bot. 56: 319-412. . 1982. General aspects of the refuge theory. 1979. Columbia Univ. P KNuTH, R. 1924. Dioscoreaceae. [n: A. Engler, Das Pflanzenreich IV 43: 1-387. (Heft 87.) Las Dioscóreas de Mexico. Anales б 0 а de la palinología en la sistemática E Dioscorea ieee (En prep.) Volume 78, Number 1 1991 Téllez & Schubert 253 Dioscorea en Mesoamérica 1936. Notes on Dioscorea, with special MORTON, C. V. ula. Publ. reference to the species of Yucatan Penins Carnegie Inst. Wash. 461: 239-253. 82. А тешен ie the Раоа ای‎ Мео tropics. Апп. Missouri [m Card. 69: E 624. TÉLLEZ, V., O. € B. С. SCHUBERT. 1987. nueva especie del género Dioscorea (Dioscoreaceae) de Estado de EETA México. Ann. Missouri Bot. Gard. 74: 539-54 Тогеро, M. V. M. bus Pleistocene changes of veg- etation in tropical México. Pp. 93-11 l inG. Prance Tropical Biology, Macuto Beach, Caracas, Venezue la, February 8-13, 1979. Columbia Univ. Pise New York. WENDT, T. 1989. Las selvas de Uxpanapa, Veracruz- Oaxaca, México: evidencia de refugios floristicos ce- éxico 20-I- Anales Inst. Biol. Univ. Nac. (Nam. ünico) 29-54. NOZOICOS. 58(1987), Ser. Bot. 9. NOTES NEW COMBINATIONS AND A NEW VARIETY IN THE CYPERACEAE OF MESOAMERICA Studies of sedges for Flora Mesoamericana have resulted in the following innovations. Cyperus manimae Kunth var. apiculatus (Liebm.) C. Adams, comb. et stat. nov. C. apiculatus Liebm., Kongel. Danske Vidensk. Selsk. Naturvidensk. Math. Afh., ser. 5, 2: 220. 1850. TYPE: Mexico: S. Antonio Hua- tusco, Liebmann 14352 (holotype, C). Diplacrum mitracarpoides (Standley & L. O. Williams) C. Adams, comb. nov. Scleria mi- tracarpoides Standley & L. O. Williams, Cei- ba 3: 36. 1952. TYPE: Honduras. El Paraiso: abundant in a wet season bog in a savanna near El Limonal, 650 m, 31 Dec. 1950, Louis O. Williams 17323 (holotype, EAP; isotypes, BM, F, MO). Pycreus flavicomus (Michaux) C. Adams, comb. nov. Cyperus flavicomus Michaux, Fl. Вог.- Amer. 1: 27. 1803. ТУРЕ: United States: “Carolina.” Pycreus fugax (Liebm.) C. Adams, comb. nov. Cyperus fugax Liebm., Kongel. Danske Vi- densk. Selsk. Naturvidensk. Math. Afh., ser. 5, 2: 196. 1850. TYPE: Mexico: Chinanthla, Liebmann s.n. (isotype, K). ANN. Missour! Bor. GARD. 78: 254. 1991. Scirpus analecti Beetle, Brittonia 5: 148, 1944, var. evadens C. Adams, var. nov. TYPE: Mex- ico. Chiapas: steep slope with Quercus and Pinus, 3 mi. south of Aguacatenango along road to Pinola Las Rosas, Municipio of Venus- tiano Carranza, 5,600 ft., 25 June 1965, Breedlove 10570 (holotype, F-1679132). Ab inflorescentia spiculis omnino remotis non fascicu- latis, et pedicelli usque ad 2(-5) cm longi, a S. analecto distinguendus. Distinguished from Scirpus analecti by inflo- rescence with all spikelets well separated, not clus- tered, and pedicels to 2(-5) cm long. Typical S. analecti has the spikelets borne on pedicels less than 1 cm long. Torulinium macrocephalum (Liebm.) C. B. Clarke, Kew Bull., Add. Ser. 8: 20. 1908, var. eggersii (Boeckeler) C. Adams, comb. et stat. nov. Cyperus eggersii Boeckeler, Beitr. Сурег. 1: 53. 1888. ТУРЕ: Dominican Re- public: Santo Domingo, near Batey, Rio Yasi- са, 23 June 1887, Eggers 2627 (holotype, B). — C. Dennis Adams, Department of Botany, The Natural History Museum, Cromwell Road, Lon- don, SW? 5BD, United Kingdom. А NEW BRAZILIAN SPECIES OF DALECHAMPIA SUBSECT. TRIPHYLLAE (EUPHORBIACEAE) Dalechampia allemii is described as a new spe- cies in sect. Dalechampia subsect. Triphyllae. This new species from Bahia, Brazil, is isolated within subsect. Triphyllae because of its unguic- ulate involucral bracts, suppressed adstaminate pis- tillate bractlets, and sparsely lobed pistillate sepals. А key is provided to the Brazilian species of subsect. Triphyllae. During a field trip to Brazil in 1978 with Antonio Allem of EMBRAPA/CENARGEN, Brasilia, I en- countered an interesting species of Dalechampia in the scrubby vegetation in the interior of the state of Bahia. This plant clearly has not been described, and indeed does not appear to have any very close relatives. Dalechampia allemii Webster, sp. nov., sect. Dalechampia subsect. Triphyllae. TYPE: Bra- zil. Bahia: Mun. Andarai, scrub at edge of forest, 50 km NW of Andarai, 15 Nov. 1984, A. С. Allem, G. L. Webster & W. L. Werneck 2980 (holotype, CEN; isotype, DAV). Fig- ure 1 abitu D. clausseniana et D. olfersiana similis, sed differt bracteis involucratis unguiculatis, involucello sub- trifido, sepalis lobis paucis. Twining vine; flowering stems terete, puberulent near tips with retrorsely recurved hairs 0.1-0.2 mm long and sparse, spreading hairs 0.3-0.5 mm long. Leaves: stipules reflexed, deciduous, lanceo- late, strigose (hairs mostly 0.3-0.6 mm long), 3- X 0.7-1 mm; petiole puberulent as the stems, 2.5-5 cm long; stipels at base of blade lanceolate, 0.8-1.7 mm long; lamina chartaceous, 3-foliolate, sparsely hirtellous оп veins and margins; middle leaflet elliptic-lanceolate, apiculate or shortly acu- minate at the tip, pinnately veined (major laterals 5 or б on a side), 4.5-6.5 х 1.7-4 cm; lateral leaflets asymmetrically eint; apiculate, with 2 veins from base 1.5-2.2 cm; veinlets some- what laa beneath; margins entire. Inflo- rescences solitary on axillary shoots 0.3-0.7 cm long, with a single reduced or obsolete leaf (blade < 0.5 cm long) at the single node; peduncle be- coming 1.8-2.5 cm long; involucral bracts pale yellowish green at anthesis, 1-1.3 x 0.7-1 cm, ANN. Missouni Вот. GARD. 78: 255-258. 1991. distinctly unguiculate (claw about V4 length of bract), deeply 3-lobed; lobes entire, with stimulose an simple hairs proximally; middle lobe elliptic-lan- ceolate, 0.6-1 x 0.15-2 cm; lateral lobes some- what falcate, 0.5-0.7 x 0.2-0.3 cm; bracteal stipules lanceolate, slightly unequal, 1.8-2.2 x 0.9-1.2 mm. Pistillate cymule: abstaminate bract ovate, eglandular, 1.2-2.2 x 2 mm; adstaminate bractlets suppressed. Pistillate flowers subsessile or with pedicels up to 1.5 mm at anthesis; sepals generally 6, usually with 1 or 2 lateral lobes below the middle, sometimes with minute intercalary se- paloid projections; lobes minutely apiculate but not glandular-capitate; ovary minutely hispidulous; sty- lar column cylindrical, glabrous or sparsely hispidu- lous below, ca. 5 X 0.5 mm, with stigmatic bands extending 25 the length; stylar apex excavated, eccentric, 0.4-0.6 x 0.6-0.8 mm. Staminate cy- mule with densely hispidulous peduncle up to ca. 2 mm long; involucel 2-lipped, rather deeply 4-lobed (but middle adpistillate lobe reduced or suppressed); lateral adpistillate lobes ovate, 2.5-5 x 3.5-7 mm, margins entire or sometimes remotely and minutely glandular; adpistillate central lobe elliptic, small (1-1.5 mm long) or obsolete; abpistillate lobe 1.5— 2.5 x 3 mm, enclosing the resiniferous gland; gland ca. 2 x 3.5 mm, of 1 or 2 rows of entire laminar ceriferous bractlets; resinous secretion pale, whitish. Staminate flowers subsessile; mature buds angled, 2.3-2.8 mm across, glabrous or nearly so; calyx splitting into З or 4 recurving segments 1.8- 1.3-1.8 mm; staminal column very short (0.5 mm or less), glabrous; stamens 50-80; fila- ments glabrous, 0.2-0.3 mm long; anthers ca. 0.4 mm long. Middle fruiting pedicel 8-10 mm long, lateral fruiting pedicels 2.5-4 mm long; fruiting sepals 3-4 х —0.8 mm, flattened, mostly with a pair of medial lateral lobes ca. 0.2-0.5 mm long; columella ca. 4 mm long; capsule valves ca. 5.5 mm long, inconspicuously strigose-hispidulous; seeds not seen. е specimens examined. BRAZIL. BAHIA: Mun. Andarai, Km 39, Andarai, BR 242, 4 Mar. 1978, Allem & i dm 1801, 1802 (CEN, DAV). pun ru da allemii appears outstanding among all known species of sect. Dalechampia subsect. Triphyllae (Webster & Armbruster, 1990) 256 Annals of the Missouri Botanical Garden FIGURE tillate cymule. min ate subinflorescence. — l. Dalechampia allemii. — E. Abstaminate pistillate I. Res —A. Flowering branch.—B. Stipule.—C. Inflorescence at anthesis.—D. Pis- e bract.—F. Pistillate flower (calyx removed). let siniferous bractlet. — J. Staminate flowe — С. Fruiting sepal.—H. Sta- Volume 78, Number 1 1991 Notes 257 in its small unguiculate involucral bracts, sup- pressed adstaminate pistillate bractlets, staminal involucel with only 3 well-developed segments, and pistillate sepals with usually a single pair of lateral lobes. The small size of the involucral bracts is reminiscent of D. micromeria Baillon of southern Brazil, but in that species the bracts are differently shaped, the stamens fewer (10-25), and styles shorter (2-4 mm). In the treatment by Pax & Hoffmann (1919), D. allemii would key down near D. olfersiana Muell. Arg.; however, that species is more heavily pubescent and has nonunguiculate bracts and clavate styles. Somewhat more similar in appearance to D. allemii is D. clausseniana Baillon, which also has nonglandular pistillate se- pals; however, that species differs in its larger non- unguiculate involucral bracts, well-developed ad- staminate pistillate bractlets, staminate involucel with 2 entire lips, and pistillate sepals with 4-7 pairs of lateral lo It is a great pleasure to dedicate this distinct new species of Dalechampia to Antonio Allem, of EMBRAPA/CENARGEN in Brasilia. Dr. Allem has made substantial contributions to the system- atics of Brazilian Euphorbiaceae and is the leading authority on the taxonomy of the genus Manihot. he dedication is also especially appropriate be- cause Dr. Allem was the first collector of the species in 1978 and has published the first critical reas- sessment of the taxa of subsect. Triphyllae since the treatment of Pax & Hoffmann (Allem & Wae- chter, 1977) Relationships among the species of Dalecham- pia subsect. Triphyllae as defined by Webster & Armbruster (1990) require further study; such spe- cies as D. clausseniana, D. cissifolia, and D. heteromorpha are highly variable, and species lim- its are uncertain. The following key to the Brazilian species of subsect. Triphyllae should aid in locating D. allemii as far as gross morphological characters are concerned. Morphological terminology follows the treatment of Webster & Armbruster (1990). Dalechampia subsect. Triphyllae (Pax & Hoffm.) Webster & Armbruster, Bot. J. Linn. Soc. (in press). Dalechampia sect. а Pax & Hoffm., Das Pflanzenreich IV. XII(Heft 68): 13. 1919. TYPE: eti triphylla Lam. Twining vines, indumentum often sparse; leaves 3-foliolate (rarely 5-foliolate), sometimes also with simple leaves; floriferous shoots axillary; bracts greenish, unlobed or 3-lobed; staminate involucel with 7-11 flowers; stamens 10-100; pistillate se- d 6-12, entire to deeply pinnatifid; seeds smooth. TENTATIVE KEY TO THE BRAZILIAN SPECIES OF SUBSECT. TRIPHYLLAE la. Leaves 3-foliolate; seeds 2.8-3.5 mm long. 2a. Pistillate sepals oblong or obovate, with broad rachis (3-3.5 mm) in fruit, rachis much ۰ than lateral teeth (mostly 0.5-1 mm); foliage and involucral bracts nearly glabrous .. iphylla Lam. 2b. Pistillate sepals I lanceolate, with narrow rachis (not broader than 1.5 mm) iı in 1 fruit, bd teeth mostly broader than rachis. 3a. Pistillate песен not glandular-capitate on margins. Pistillate sepals without glandular-capitate lobes. Involucral bracts unguiculate; adstaminate pistillate bractlets suppressed; pistillate sepals mostly with only 1 pair of lateral lobes; staminate involucel 3-lobed; stamens 50-80 _ allemii Webster 9b. Involucral bracts rounded or cordate at base; adstaminate pistillate bractlets developed; pistillate sepals mostly with 4 or more pairs of lateral lobes; staminate involucel 2-lipped, scarcely lo 6 a. Stylar column slender, not clavate; ovary strigose; stipules 3-4 mm lon D. c ng usseniana Baillon column clavate; ovary villose; stipules 6-8 mm long .. obes. 6b. Stylar 4b. gr psi with glandular-capitate . Fruiting sepals 12-15 mm long; fruiting peduncle mostly not longer than 1 cla . D. olfersiana Muell. Arg. D. cissifolia Poeppig 7b. Fruiting viue shorter; peduncle deii] ae ral bracts usually lo ong v r than 1 cm; stamens mostly more than 25 per flower. es mon a ес ар inv a bracts 3-lobed; stamens 30-60. on omorphic, 10a. Involucral bracts 3-lobed; fruiting sepals 3.5-5 mm meridionalis Muell. Arg. 10b. Involucral bracts unlobed or trifid at tip; fruiting sepals 7-8 mm lon D. 5 - bangii Pax & К. Hoffm. 9b. Leaves dimorphic, both tripartite and unlobed; bracts unlobed or 3-lobed; sta D. mens less 8b. Involucral bracts long, glabrous; ae all tripartite than 30; fruiting sepals mostly = 1 cm long, 3-lobed; stamens 10-25; styles 2 = 5-6 mm lon burchellii pis Arg. D. micromeria Baillon 258 Annals of the Missouri Botanical Garden 3b. Pistillate чарды glandular-capitate on margins; stamens 30 or more; leaves d beneath, -lobed D. all tripartite; involucral bracts 3-lobe epala Muell. Arg. lb. Leaves 5- Ps. Mis longer than 4 mm Fieldwork in Bahia was made possible through a grant from the National Science Foundation and the collaboration of Dr. Allem and personnel of EMBRAPA/CENARGEN, Brasilia. The illustra- tions were drawn by Leslie Randall. LITERATURE CITED ALLEM, А. С. & J. L. WAECHTER. 1977. Notas siste- máticas y nuevos sinónimos en Euphorbiaceae de America del Sur — II. Rev. Brasil. Biol. 37: 91-101 D. теке Muell. Arg. Pax, Е. & К. HOFFMANN. 1919. Euphorbiaceae-Dale- о Das Pflanzenreich. IV. 147. XIKHeft 68): 1 WEBSTER, с L. & W. S. ARMBRUSTER. 1990. А syn- opsis of the acie species of Dalechampia. Bot. ess). J. Linn. Soc. (in p — Grady L. Webster, tment of Botany, Uni- versity of и Don California 95616, U.S.A FOUR NEW SPECIES OF ELAPHOGLOSSUM (ELAPHOGLOSSACEAE) FROM VENEZUELA In a series of papers (Mickel, 1985, 1987, 1990) I published 37 new taxa of Elaphoglossum from Venezuela. Specimens sent recently for identifi- cation included four more novelties, making a total of 107 species and three subspecific taxa known from that country. Elaphoglossum polyblepharum Mickel, sp. nov. TYPE: Venezuela. Territorio Federal Ama- zonas: Dept. Atures, W side of valley of Rio Coro-Coro, 8 km NNW of settlement of Yu- taje, 05?41'N, 66?08'30"W, 1 Маг. 1987, Liesner & Holst 21517 (holotype, NY; iso- type, MO). Figure 1A. Ab Е. erinaceo frondium squamis brevioribus rubro- brunneis necnon laminae facie abaxiali squamosa diver- Rhiz. act, 4–6 mm diam.; rhizome scales linear-lanceolate, orange, ca. 5 mm long, entire; phyllopodia lacking; fronds approximate, 30-40 cm long, 3.5-4.6 cm broad; stipe 4—% the frond length, densely clothed with patent, reddish brown, subulate scales 4–6 mm long; blade narrowly el- к papyraceous, apex caudate, base broadly cuneate; veins evident, free, ca. 1 mm apart, at 75? angle; hydathodes lacking; blade scales reddish brown, subulate, liberally distributed on abaxial surface, 1-2 mm long, more concentrated on mar- gin, 1.5-2 mm long, and costa, 2-3 mm long, lacking on the adaxial surface; fertile fronds un- known. Epipetric in wet forests, 900-1,000 m. ENEZUELA. S: Dept. Atures, same locality as type, Liesner & Holst. 21 405 (MO, NY); Dept. Río Negro, Cerro Arati- tiyope, 02?10'N, 65*34'W, Steyermark et al. 130312 (NY). M: d examined. VE Elaphoglossum polyblepharum differs from E. erinaceum (Fée) Moore in having frond scales red- dish brown rather than black-brown, these scales more dense on stipe, conspicuously present on lam- inar surface, and shorter on margins (1.5-2 mm long vs. 3-4 mm) and costa (2-3 mm long vs. 4– 6 mm), and the rhizome scales lighter orange and shorter (ca. 5 mm long vs. 15-20 mm El 1 1 | БРЕ 1 ckel, sp nov. Mi TYPE: Venezuela. Portuguesa: Dept. Suére, La Divisoria de la Concepción, 09°18'N, 70°06'W, 23-26 Oct. 1985, Ortega et al. 2748 (holotype, UC; isotype, NY). Figure 1B. Lamina utrinque glabra nisi stipite costaque squamis subulatis et margine squamulis lanceolatis provisis praes- tans. Rhizome short-creeping, 3-5 mm diam.; rhi- zome scales linear-lanceolate, orange-tan, 1.5-2 mm long, with sparse hair-teeth; phyllopodia lack- ing; fronds approximate, 30-40 cm long, 2.9-3.4 cm broad; stipe Y4-% the frond length, with scales subulate, patent, 2-2.5 mm long, brown; blade narrowly elliptic, papyraceous, apex acuminate, base narrowly rounded; veins evident, free, ca. 1.5 mm apart, at 60-70? angle; hydathodes lacking; blade scales on adaxial margin lanceolate, brown, 1-2 mm long, appressed, on abaxial costa subulate, brown, spreading, 2-2.5 mm long, abaxial lamina subglabrous, with simple to stellate trichomidia 0.2 mm long; fertile fronds much smaller than the sterile (12-14 the length), stipe Y2 the fertile frond length, blade narrowly elliptic, 8-11 mm broad; fertile blade lacking scales on costa, margin, or among sporangia. Terrestrial in cloud forests, 1,500-1,800 m. Elaphoglossum heteroglossum resembles E. er- inaceum in the subulate scales of stipe and costa t in the small lanceolate marginal scales vs. aye subulate ones, and the much smaller fertile ronds Elaphoglossum xiphoides Mickel, sp. nov. TYPE: Venezuela. Bolivar: a la cuenca media del Rio Paragua, 06%03'N, 63°47'W, 13 June 1987, Stergios 10310 (holotype, NY). Figure 1C. E. rhynchophyllo laminae squamis ovatis gilvis et squamis intersporangialibus nullis abstans. Rhizome creeping, 1-1.5 mm diam.; rhizome scales linear-lanceolate, orange-tan, ca. 3 mm long, entire; phyllopodia present but indistinct and cov- ered by scales; fronds 6-10 mm apart, 5-7 cm long, 1-1.3 cm broad; stipe Y, the sterile frond ANN. MISSOURI Вот. GARD. 78: 259-261. 1991. Annals of the 260 Missouri Botanical Garden а ~ M. pm AAA O Mr и ALI ر‎ а FE SS не lr اپ‎ apiy ھی‎ E v3 ` yt i ЛАА, 24 WA y ef p a УАЙ ft ДУХА 4 РЕА р yr / / Aor 2 x r МаК ти 5 .. РА arep t » à df V FIGURE 1. —A. Elaphoglossum polyblepharum. —B. E. heteroglossum. —C. E. xiphoides. — D. E. zosteriformis. Volume 78, Number 1 1991 Notes 261 length, with scales ovate-lanceolate, orange-tan, spreading, ca. 3 mm long; blade elliptic to lanceo- late, subcoriaceous, apex obtuse, base cuneate; veins evident, 0.8-1 mm apart, at ca. 60? angle; hy- dathodes lacking; blade scales scattered, orange- tan, 1-2 mm long on abaxial costa, otherwise blade glabrous; fertile frond longer than sterile (ca. twice as long), stipe /-% the fertile frond length, blade linear, 5 mm broad; intersporangial scales lacking. Terrestrial in wet forests, 17 Elaphoglossum xiphoides is similar to E. rhyn- chophyllum Christ in blade shape of fertile and sterile fronds, but the blade scales in the former are tan, ovate, sparse on midvein, and intersporan- gial scales are lacking whereas in the latter the blade scales are black, linear, dissected, and in- tersporangial scales are abundant. The phyllopo- dium is indistinct but present in E. xiphoides whereas it is lacking in E. rhynchophyllum and other members of the E. squamipes complex. It also resembles E. minutum (Pohl) Moore in the tan scales and the presence of a phyllopodium, but the fertile blade is acuminate rather than obtuse, the rhizome is longer-creeping, and it occurs at a much lower elevation (175 m vs. over 3,000 m). Elaphoglossum zosteriformis Mickel, sp. nov. TYPE: Venezuela. Bolivar: Rio Caneracuni, entre Caneracuni y la boca, 04?30'N, 63?42'W, 13-26 Арг. 1988, Stergios 11930 (holotype, NY; isotype, UC). Figure 1D. E. laminarioide laminae stipitisque squamis lineari- lanceolatis distans. Rhizome compact, 5 mm diam.; rhizome scales linear-lanceolate, black-brown, lustrous, entire; phyllopodia present; fronds approximate, ca. 95 cm long, 4.8 cm broad; stipe М the frond length, with scales loosely appressed, linear-lanceolate, 1- 3 mm long, cilio-denticulate, tan, imbricate to somewhat deciduous; blade linear-elliptic, charta- ceous, apex lacking, base narrowly cuneate; veins evident, 1 mm apart, at 80? angle; blade scales adaxially lacking except for a few 1-2 mm long, cilio-denticulate scales on costa, abaxially with scat- tered (touching but not overlapping) dissected scales ca. 1 mm long, linear-lanceolate, the teeth longer than the scale body width, orange-tan, costa scales linear, 1-2 mm long, cilio-denticulate with teeth shorter than the scale body width; fertile fronds unknown. Epiphytic in wet forests of berenas," elevation unknown "tierra firme y ri- In the long, narrow blade Е. zosteriformis re- sembles E. laminarioides (Bory) Moore, but the former has linear-lanceolate scales on the blade and stipe whereas the latter has round, peltate scales. LITERATURE CITED MICKEL, J. Т. . The proliferous species of Ela- phoglossum (Elaphoglossaceae) and their relatives. Brittonia 37: 261-278. New species of Elaphoglossum (Ela- phoglossaceae) from northern South America. Brit- топа 39: 313-3 o new species of Elaphoglossum from Amazonas, Venezuela. Ann. Missouri Bot. Gard. 95. 77: 594- —John T. Mickel, New York Botanical Garden, Bronx, New York 10458, U.S.A A NEW SPECIES OF ARISTOLOCHIA (ARISTOLOCHIACEAE) FROM VENEZUELA Very little is published on the Aristolochia spe- cies of Venezuela. No floristic treatment exists and none of the available revisions treat all the Vene- zuelan species (Ahumada, 1967; Hoehne, 1942; Pfeifer, 1966). Bazzolo & Pfeifer (1978) revised the herbaceous, efimbriate species, but all other species must be identified using Hoehne (1942). This makes the identification of Venezuelan aris- tolochias unusually difficult. The following species was first collected in 1962, but only now can it be confidently described as new. Aristolochia melgueiroi Barringer & Guan- chez, sp. nov. TYPE: Venezuela. Territorio Fe- deral Amazonas: Dept. Atures, along road be- tween damsite of Rio Cataniapo and Gavilan, 35-40 km SE of Puerto Ayacucho, 200 m, 11 May 1980, Steyermark, Davidse & Guánchez 122297 (holotype, VEN; isotypes, F, MO). Figure 1. Species Aristolochiae killipianae О. Schmidt affinis, sed foliis ovatis, pseudostipulis nullis, floribus rectis, utri- culis ovoideis, tubis usque ad 12 mm longis differt. Item A. triangulari Cham. affinis, sed foliis ovatis cordatis, pseudostipulis nullis, floribus rectis, tubis ad 12 mm longis, seminibus 6-7 mm longis descrepat Liana; young stems slightly tomentose, glabres- cent, ridged, the nodes swollen. Pseudostipules ab- sent. Leaves alternate; petioles 10-30 mm long; laminas ovate, 7— -8 cm, glabrous above, slightly pilose and glabrescent beneath, the base cordate, the sinus 6-17 mm deep, the apex acu- minate. Flowers solitary in axils of young growth or solitary on axillary branches to 10 cm long, with a basal bract 3-7 mm long; pedicel and ovary 7-10 mm long, pilose; calyx green marked with brown-purple lines, the utricle ovoid, 5-7 x 4 mm, the tube straight, 9-12 mm long, the limb broadly funnelform, not reflexed, broadly ligulate above, spreading, 12-16 x 9-14 mm, apiculate, yellowish green marked with brown lines and brown- purple to madder-brown spots toward the mouth of the tube, with a broad orange-yellow spot near the apex; gynostemium coroniform, 6-lobed. Cap- sule ovoid, to 5 cm long, 3-4 cm wide; seeds flat, subcordate, broadly winged, 6-7 x 9-10 mm Additional specimens examined. VENEZUELA. TERRITORIO FEDERAL AMAZONAS: Dept. Río Negro, forest and roadside 5-6 E of San Carlos de Rio Negro, 120 m, 12 Apr. 1979, Liesner 6561 (МО, VEN); Ups Río Casiquiare, 20 km NE of San Carlos de Rio Neg 120 m, Liesner 7373 (MO, VEN); rr study n 4 е МЕ of San Carlos de Rio Negro, 120 m °52'N, 67%02'W, 24 Jan. 1980; Liesner 8663 (МО, VEN. Dept Atures, Alrededor " margen izquierda del Río Autana, 100 m, 04*48'N, 61*28'W, 13 Nov. 1984, Guánchez et al. 3574 (MO, TFAV, VEN). BOLÍVAR: entre la бөгү ыт Los Brasilieros y el camino hacia Playa Blanca y el Río Uai- parú, SE of Icabarú, 4 ec. 1978, Steyermark et al. 117728 (F, MO, VEN). This species is dedicated to Emigdio Melgueiro, assistant at TFAV, in acknowledgment of his valu- able contributions to the botanical exploration of the Territorio Federal Amazonas in southern Ven- ezuela. Aristolochia melgueiroi has been found in the Venezuelan Guayana between 100 and 450 m, mainly in open and disturbed sites such as roadsides and secondary vegetation (“rastrojos”). Aristolochia melgueiroi is part of the А. trian- gularis Cham. complex (Schmidt, 1935), which also contains А. pseudotriangularis О. Schmidt, a Bolivian species, and 4. killipiana О. Schmidt, from upland Peru. All the species in this complex are woody lianas with axillary flowers. The flowers are usually less than 3 cm long and have a broadly ovate limb marked with purple spots. Aristolochia melgueiroi can be distinguished from all other spe- cies in this complex by its lack of pseudostipules. It can be distinguished from А. triangularis and A. pseudotriangularis by its ovate leaves with deep, basal sinuses and from А. killipiana by its smaller utricle, straight flowers, and ovate leaves. Aristolochia melgueiroi is vegetatively similar to A. acutifolia Duchartre, a native of the Amazon basin. However, 4. acutifolia has flowers with a narrow, ligulate limb that lacks the conspicuous orange-yellow spot at the apex. Aristolochia mel- gueiroi might also be confused with the recently described А. guianensis Poncy from French Gui- ana (Poncy, 1988). The mature leaves of А. gui- anensis have longer petioles and a more triangulate lamina than 4. melgueiroi, but the juvenile leaves are similar. The flowers of the two species are very ANN. Missouni Вот. GARD. 78: 262-263. 1991. Volume 78, Number 1 1991 Notes -— BN К / f$ А2 № | Y РЬ E | FIGURE 1. different, though. The flowers of А. guianensis have a large, unilabiate limb borne perpendicular to the tube and 5-8 cm long peduncles. We thank Paul Berry and Bruce Holst (MO), who helped coordinate our efforts, and the curators of F, MO, NY, TFAV, and VEN, who allowed us to study their collections. LITERATURE CITED AHUMADA, 2. . Revision de las Aristolochiaceas Argentinas. Opera Lilloana 16: 1-145. BazzoLo, T. & Н. У. PFEIFER. 1978. Efimbriate ћег- b istolochias in Brazil and l South aceous t America. Caldasia 12: 19-33 HOEHNE, F.C. 1942. Aristolochiaceas. In: Е. C. Hoehne (editor), Flora Brasilica, Volume 15, Part 2. F. Lan- zara, Sao Paulo \ А T Ду Ly A 7 = » OY Ay т! e. К Aristolochia melgueiroi. — A. Flowering branch.—B. Fruit. PFEIFER, Н. W. 1966. Revision of Aristolochia in North entral America. Апп. Missouri Bot. Gard. 53: 8. Studies on the Flora of the Guianas о ant une espece d'Aublet. Bull. Mus. Hist. Nat. (Paris), sér. 4. 10(sect. B) Adansonia (4): 337-344. SCHMIDT, О. С. 1935. Beitráge zur Kenntnis der Aristo- lochiaceae. VI. Neue andine Aristolochiacien aus der Verwandschaft der Aristolochia triangularis Cham. Repert. Spec. Nov. Regni Veg. 38: 110-112. — Kerry Barringer, Brooklyn Botanic Garden, 1000 Washington Avenue, Brooklyn, New York 11225-0199, U.S.A.; and Francisco Guánchez, In- stituto Botánico, INPARQUES, Apartado 2156, Caracas 1010-A, Venezuela. ON THE TYPIFICATION OF BIGNONIA CRUCIGERA L. (BIGNONIACEAE) In the first edition of Species Plantarum, Lin- naeus (1753: 624) proposed Bignonia crucigera, taking as his polynomial Bignonia foliis conjugatis cirrhosis; foliolis cordatis; foliis imis ternatis, a reorganized form of the phrase name he had first proposed in Viridarium Cliffortianum (1737a: 60) and which was used again in his Hortus Clifforti- anus (1738: 317) and van Royen’s Florae Ley- densis (1740: 289). He also cited two synonyms, one from Gronovius (1739: 73), the other from Morison (1699: 612, erroneously cited as 672). astly, Linnaeus added a brief description: **Caulis volubilis, scaber, transversim dissectus crucem rep- raesentat,” and the statement “Habitat in Virginia and australiori America." There are no specimens in any of the Linnaean herbaria, nor in the Clifford herbarium (BM), as- sociated with this name. There are, however, three elements associated with it. Two are linked to the synonymy, a specimen cited by Gronovius, Clayton 100 (BM), and a Morison (1699) figure, s. 15. t. 3. f. 16. The third element is a plate from Charles Plu- mier, not published until three years after Species Plantarum (Plumier, 1756), and hence not cited directly in the protologue. Plumier had made more than 1,200 drawings on trips to the West Indies in 1689-1690 and 1696-1697 (see Stafleu & Cowan, 1983). The Dutch naturalist Hermann Boerhaave had acquired copies of more than 500 of them, and it was these copies that Linnaeus examined while in Leiden in 1738 (Polhill & Stearn, 1976). Linnaeus recorded his observations in a copy of the first edition of Genera Plantarum (Linnaeus, 1737b), now in the library of the Lin- nean Society of London. While it was known that Linnaeus had examined Plumier plates prior to 1753, Polhill & Stearn (1976) were the first to establish the connection between the Plumier cop- ies in Holland, the annotated Genera Plantarum, and the descriptions in Species Plantarum. Linnaeus wrote the notes on the Bignonia plate opposite p. 179: Bignonia scandens bifolia & trifolia, ligno crucea figurato. Plum. mss. Obs: caulis volubilis scaber, dissectis cruceum depictam exhibet He took the observation, with minor rephrasing, for his descriptive sentence in 1753. Although the plate was not cited until the second edition of Species Plantarum (Linnaeus, 1763), where it was cited from the published source (Plumier, 1756), the commentary provides the necessary link con- firming that the plate (now at Groningen) is original material, and that it influenced Linnaeus's concept of the species as published in 1 The name Bignonia crucigera has been asso- ciated with three taxa in as many genera, each of which is represented in the original material. The information, summarized here, is discussed in greater detail by Gentry (1975). The Gronovian specimen, Clayton 100 (BM), is a collection of what is now known as Bignonia capreolata L., a connection recognized by Small (1903), who treat- ed B. capreolata as a synonym of B. crucigera. The Morison illustration is of a fruit and single seed of Bignonia echinata Jacq. (= Pithecoctenium crucigerum (L.) A. Gentry). The Plumier plate illustrates Tanaecium crucigerum Seem. circumscription, Seemann (1856), citing Linnaeus (1763), excluded all of Linnaeus’s synonymy for B. crucigera except Plumier (1756). Gentry (1975) attempted to resolve the problem by typifying the name and thereby fixing its usage. Several of the options were unappealing. Selecting Clayton 100 would have resulted in the displace- n his ment of the North American Bignonia capreolata and, writing prior to Polhill & Stearn (1976), it was understandable that Gentry should not have regarded the Plumier plate as original material. It therefore seemed impossible to typify the name in the sense of Tanaecium crucigerum, even if such a choice were deemed desirable. Gentry argued that the most reasonable option was to return to Linnaeus’s original concept of the species, reflected in the protologue in Hortus Clif- fortianus. Linnaeus cited two synonyms there, that of Morison (1699), cited again in 1753, and a name from Miller’s (1737) Gardeners Dictionary, Bignonia scandens tetraphylla, fructu maximo echinato. This appears to relate to the same sort of plant illustrated by Morison, and there is a Miller specimen at BM, possibly contemporary, belonging to this species. Gentry argued that the Miller spec- imen was an eligible syntype because it was tied to a citation in Linnaeus’s original, that is, 1738, ANN. Missouri Bor. GARD. 78: 264-265. 1991. Volume 78, Number 1 1991 description. The designation of the specimen rather than the illustration was based on Recommendation 7B-5, in the “Guide to the Determination of Types" then in effect (Stafleu et al., 1972), which stated that specimens were to be selected over illustra- tions. Unfortunately, whatever its connection to Hor- tus Cliffortianus, the Miller specimen cannot, un- der the present Code (Greuter et al., 1988), be considered original material, as Linnaeus never saw it, and it is hence ineligible for selection as the lectotype. Moreover, the expanded synonymy in Species Plantarum indicates that Linnaeus's con- cept of Bignonia crucigera changed between 1738 and 1753, and there is no direct evidence that he still considered the Miller name a synonym by Although the lectotypification on the Miller spec- imen cannot stand, one of the eligible elements, the Morison plate, illustrates the same sort of plant. Therefore, Bignonia crucigera L. is here lecto- typified on PseudoApocynum folliculis maximis obtusis. . . , s. 15. t. 3. f. 16 (Morison, 1699). This typification preserves current usage of Pithecoc- tenium crucigerum (L.) A. Gentry. This study was supported in part by National Science Foundation Grant BSR-881 2816. It is Sci- entific Article no. A-5076, Contribution No. 8136 of the Maryland Agricultural Experiment Station. LITERATURE CITED GENTRY, А. Н. . Bignonia e a case of mistaken identity. Taxon 24: 23. GREUTER, W. ET AL. (editors). TR оа сна of Botanical Nomenclature. Regnum Veg. 118: 328. GRONOVIUS, ү К. 1739. Flora Virginica, Рап I. C. Haak, Leiden. Notes 265 LiNNAEUS, C. 1737a. Viridarium Cliffortianum. Am- sterdam. 1737b. Genera Plantarum. Leyden 1738. Hortus Cliffortianus. Amster daii 1753. Species Plantarum. L. Salvius, Sode —. 17 Species Plantarum, 2nd edition, Vol- e IL с Stockholm Miis. T. 1737. The Cerdeners Dictionary, 3rd edi- tion. Published by the author, London. MORISON, 1699. Plantae Historiae ... pars tertia. xford PLUMIER, C 56. Plantarum Americanum, J. Burman (editor), Fascicle III. С. Potvliet & T. Haak, Am sterdam. Рогнил, В. M. & W. T. SrEARN. 1976. Linnaeus's notes on Plumier drawings with special reference to Mimosa мени: or Taxon 25: 323-325. 1740. Flora Leydensis Prodromus. SN "B. 1856. Revision der Crescentiaceen-Gat- un 128. 903. Flora of the Southeastern United States. Published by the author, New York. STAFLEU, F. ET AL. (editors). 1972. Eu of Botanical Nomenclature. Regnum Veg. 82: 426. . COWAN. 1983. Taxonomic Literature, 2nd edition, Volume IV. Regnum Veg. 110. —Fred R. Barrie, Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166, U.S.A.; James L. Reveal, Department of Botany, Univer- sity of Maryland, College Park, Maryland 20742, U.S.A.; Charles E. Jarvis, The Natural History Museum, Cromwell Road, London SW? 5BD, United Kingdom; and Alwyn H. Gentry, Missouri Botanical Garden, P.O. Box 299, St. Louis, Mis- souri 63166, U.S.A. Mailing address for F.R.B.: The Natural History Museum, Cromwell Road, London SW7 5BD, United Kingdom. STUDIES IN MIDDLE AMERICAN BROMELIACEAE I Smith & Downs (1974) in their monograph of Bromeliaceae subfamily Pitcairnioideae included 56 species of Pitcairnia from Mexico and Central America. Since that time four previously unre- ported species were collected from the area, in- cluding one originally thought to be native to Brazil (Zamudio, 1988), and five new species have been described (Luther, 1987a, b; Rauh, 1986; Smith & Read, 1975). In addition, we herein propose two new taxa, P. beachiae and P. virginalis, bring- ing the number of Pitcairnia species for Mexico and Central America to 67. It is likely that this number will continue to fluctuate as our knowledge of the flora expands, due not only to new species, but also to reinterpretations of species limits (cf. ). McVaugh, 1989 Pitcairnia beachiae Utley & Burt-Utley, sp. nov. TYPE: Costa Rica. Puntarenas: Valle de General, 300 m, collected 12 May 1974, flow- ered in cultivation April 1984, Beach 74-25 (holotype, US; isotypes, CR, F, MO). Fig- ure 1 Folia petiolata persistentia spinescentia. Scapus decur- vus abrupte bracteae persistentes internodia superantes. Inflorescentia 4.5-10 cm longa densiflori. Petala lutea. varium circa % inferum. Epilithic or terrestrial herbs, to about 1.5 m tall in flower. Leaves dimorphic, persistent; outer leaves sheathing, bladeless; inner leaves 1-1.7 m long, subpetiolate to petiolate; sheaths 1–1.5 cm long, 4—4.5 cm wide, serrulate, light brown to stramin- eous with brown shaggy pubescence, contracting abrubtly into the petiole; petiole 20-60 cm long, 0.4-1.2 ст wide, scattered spinose-serrate, be- coming strongly involute on drying; blade with a pronounced midrib, plicate, narrowly elliptic to lance-elliptic, 69-100 cm long, 3-6 cm wide, api- cally attenuate, glabrous adaxially, densely pale pubescent abaxially, marginally scattered spinu- lose-serrulate becoming subdensely serrulate api- cally, chartaceous. Scape 22-32 cm long, 5-7 mm diam. (1-1.5 cm diam. in living material), erect for 5-7 cm and then abruptly decurved, brown floccose pubescent; bracts 3.5-36 cm long, exceeding the internodes but divergent to sub- spreading and the scape exposed, pale pubescent to locally glabrous, lower bracts subfoliaceous and scattered spinose-serrate, middle and upper bracts much reduced and entire. Inflorescence simple, 4.5-10 cm long, 5-6 cm diam., horizontal or slightly ascending, densely spicate to subglobose, polystichously many-flowered. Floral bracts ex- ceeding the sepals, erect, ovate to elliptic, 2.3- 3.5 cm long, (0.7-)1 entire, pink drying brown, finely evanescently pu- bescent. Flowers erect to divergent, sessile; sepals coriaceous, triangular, 2.6-2.9 cm long, 1-1.2(- 1.5) cm wide, short attenuate, brown floccose abax- ially, glabrous adaxially, much thickened basally, thickened to distinctly asymmetrically keeled me- dially; corolla narrowly tubular, curved (abaxially ?) in distal 4 at anthesis; petals basally appendaged, dissimilar, straight to falcate, narrowly obovate, g, 0.7-1.4 cm wide, yellow, basally attenuate, apically acute and mu- cronulate; stamens included, filaments 3-4.6 cm long, flattened, anthers basifixed, narrowly hastate, 3-1.6 cm long; ovary about % inferior, 0.9-1.3 cm long, 0.75-1 cm diam.; style 4.5-5.5 ст long. This species is named in honor of Eloise Beach of Orlando, Florida. In addition to collecting this species for the first time, Eloise's unstinting studies of the Bromeliaceae have added considerably to our horticultural and botanical knowledge of the family. P. beachiae is most similar to P. quesne- lioides L. B. Smith from Colombia and the Peru- vian P. fimbriato-bracteata Rauh in its densely flowered, simple inflorescences, persistent petiolate .2-1.5 ст wide, acuminate, ovate or elliptic, 6–6.4 cm lon armed leaves, and sessile to subsessile flowers with appendaged petals and largely inferior ovaries. It is easily distinguished from these taxa by its per- sistent floral bracts, in contrast to the deciduous or early decomposing floral bracts of the latter two species. It further differs from P. quesnelioides in having persistent upper scape bracts and a shorter inflorescence (4.5-10 cm vs 15 cm). In addition, P. beachiae differs from P. fimbriato-bracteata in its scape bracts, which much exceed the inter- nodes and yellow flowers; in contrast the latter species has red flowers and scape bracts that are conspicuously shorter than the internodes. Despite our numerous attempts we have been unable to ANN. Missouni Вот. GARD. 78: 266-269. 1991. Volume 78, Number 1 1991 Notes 267 FIGURE 1. Petal.—F. Flower post-anthesis. From Beach 74— collect fertile material at the type location, even though vegetative individuals are regularly en- countered. Pitcairnia virginalis Utley & Burt-Utley, sp. nov. TYPE: Mexico. Oaxaca: steep roadbanks along MEX 175, 29.7–30.1 mi. N of N side of Pochutla or 5.1-5.5 mi. М of turnoff to Pluma Hidalgo, 4,900-5,000 ft., 17 July 1989, Utley & Utley 8400 (holotype, MEXU; isotypes, MICH, MO, US). Figure 2. Folia heteromorpha. Laminae magiores deciduae. Sca- pus 32-65 cm longus bracteae infernae foliaceae. Inflo- rescentia 7-10 cm longa ante florenti circa 30 cm ubi fructificans. Bracteae florales 4.5-6.5 cm longae late ovatae vel anguste ellipiticae acutae chloroleucae pubes- centes pallentes. Petala alba appendiculata. Ovarium circa у; inferum Herbs, terrestrial or epilithic on humus-covered Pitcairnia beachiae Utley & Burt-Utley. — A. Нађи. —B. Leaf. — C. Inflorescences. — D. Sepal.— Е. 25. rocks; 60--100 cm tall in flower. Leaves trimor- phic: the outer two types with broadly triangular, tightly overlapping sheaths, these entire proximally becoming serrate near junction with blade, casta- neous, densely white to pale ferrugineous lanate abaxially, glabrous to glabrescent adaxially; some with blades persistent, spiniform, pectinate-serrate, 3-5 cm long, others with apparently linear(?), de- ciduous blades which disarticulate 2-4 cm above the sheath, persistent portion of blade 1-2.5 cm wide, densely serrate (characters of deciduous por- tion unknown): inner leaves and lower scape bracts entire or weakly serrate basally: sheaths ovate, 4— 5.5 ст long, 2.2-3 cm wide, pale, scattered pu- bescent abaxially; blades narrowed above the sheath, often with a tendency to disarticulate at the sheath upon drying, elongate-triangular or elliptic to nar- rowly ovate, 30-36 cm long, 2.5-3 cm wide, scat- tered pale floccose abaxially. Scape curved-erect, 268 Annals of the Missouri Botanical Garden FIGURE 2. S 32-65 cm long, scattered to subdensely pale floc- cose; lower bracts foliaceous, the upper reduced and subfoliaceous, 21-31 cm long, glabrous adax- ially, scattered pale floccose abaxially, sheaths in- flated, clasping, blades divergent to drooping-re- curved above middle. Inflorescence simple, racemose, 7-10 cm long pre-anthesis, elongating to about 30 cm post-anthesis. Floral bracts broadly ovate to narrowly elliptic, 4–6.5 cm long, 1.3-3 cm wide, becoming reduced distally, acute, pale green-white drying brown post-anthesis, adaxially glabrous, abaxially white floccose proximally, densely pale ferrugineous distally. Flowers ascend- ing to divergent or subspreading at anthesis, ped- icels 3-5 mm; sepals broadly triangular to subligu- late, 2.9-3.3 cm long, about 1 cm wide, acute, subdensely brown pubescent; corolla broadly tu- bular to elongate-campanulate; petals basally ap- pendaged, obovate, 5.9–6.3 ст long, 2-2.3 cm wide, white, broadly acute; stamens included, fil- Pitcairnia virginalis Utley & Burt-Utley.— А. Habit. —B. Inflorescence. —C. Flower at anthesis. — . Sepal. — E. Petal. From Utley & Utley 8400. aments 4.1—4.3 cm long, anthers basifixed, has- tate, 1.4-1.5 cm long; ovary trigonous in cross section, about 1.2 cm long, 0.6-0.7 cm diam., about one-half or less inferior; style about 5-5.2 Pitcairnia virginalis is most similar to P. abun- dans L Smith from Nayarit, Mexico in its heteromorphic leaves, with the larger leaf blades deciduous, elongate scape with foliaceous to subfoli- aceous lower bracts and white appendaged petals. It differs in having broadly ovate to narrowly el- liptic, apically acute floral bracts, 4-6.5 cm long, in contrast to the floral bracts of P. abundans, which are triangular to triangular-lanceolate, api- cally caudate-attenuate an Moreover, sepals of P virginalis are generally longer (2.9-3.3 vs 2.6-3 cm) and apically acute rather than long subulate like those of P. abun- dans. Pitcairnia virginalis is unusual among Mex- ong. ican and Central American Pitcairnia species in Volume 78, Number 1 1991 Notes 269 having white, elongate-campanulate corollas, and large, pale green to white, pubescent floral bracts. In addition, the inflorescence elongates consider- ably post-anthesis; this change in morphology is so striking that plants at anthesis and in fruit could, on casual inspection, be taken for distinct species. Additional collections. MEXICO. OAXACA: El Porve- 5,100 ft., 1 Aug. 1990, Utley & Utley 8588 (CAS). This work was supported by a grant from the Standley Smith Horticultural Trust to the Univer- sity of New Orleans and in part by NSF grant DEB-8107868 to the Missouri Botanical Garden. We thank the curators of the following herbaria for the loan of specimens or use of facilities: MEXU, MICH, MO. LITERATURE CITED LurHER, H. E. 1987a. Pitcairnia hammelii, a new Pitcairnia from Panama. J. Bromeliad Soc. 37: 128- 129. 1987b. Two new Pitcairnias from northwest- n Panama. J. Bromeliad Soc. 37: 212-214. MON AUCH, В. 1989. In: У. В. Anderson (editor), Flora ovo-Galiciana 15: 4-79 Bromelienstudien I. Neue und wenig bekannte Arten aus Peru und anderen Ländern (18. экы т Subtrop. Pflanzenwelt 58: 5-63. SMITH, І. B. К. J. Downs. Pitcairnioideae тое Fl. Neotrop. 14: 1-658. . READ. 1975. Notes on Вготећасеае, XXXVII. Phytologia 30: 289-303. ZAMUDIO, S. 1988. Descubrimiento de Pitcairnia un- dulata Scheidweiler аа en Tabasco Mexico. Acta Bot. Mexicana 2: 5-9 —John F. Utley and Kathleen Burt-Utley, De- partment of Biological Sciences, University of New Orleans, New Orleans, Louisiana 70148, S.A STUDIES IN MIDDLE AMERICAN BROMELIACEAE II Mezobromelia is distinguished from other mem- bers of the subfamily Tillandsioideae by epipetal- ous stamens and biappendaged petals that are ag- glutinate or connate for a significant portion of their length. Smith (Smith & Downs, 1977) rec- ognized two species in the genus, and Rauh (1976, 1977) described two more. Recent collections from Costa Rica of flowering material congruent with Vriesea splitgerberi (Utley & Utley 7141) show that the petals of this species are biappendaged and connate. These characteristics, in conjunction with the polystichously flowered inflorescence branches (a condition rarely encountered in Vrie- sea but common in Mezobromelia), strongly sug- gest that this species is better treated as a member of Mezobromelia. Gouda (1987) presented evi- dence that V. splitgerberi is conspecific with Til- landsia pleiosticha and proposed the combination Vriesea pleiosticha. Because we consider the spe- cies to be better placed in Mezobromelia, we pro- pose the following new combination. Mezobromelia pleiosticha (Griseb.) Utley & Luther, comb. nov. Tillandsia pleiosticha Griseb., Nachr. Konigl. Ges. Wiss. Georg Got- tingen-Augusts-Univ. 1864: 19. 1865. Vrie- ANN. Missouni Вот. GARD. 78: 270. 1991. sea. pleiosticha (Griseb.) Gouda, Fl. Guianas, Ser. A 189: 82. 1987. TYPE: Venezuela: Fen- dler 1514 (holotype, GOET). Guzmania splitgerberi Mez, Monogr. Phan. 9: 930. 1896. Thec о splitgerberi (Мег) Pittendr., Evo- 48. Vriesea ыл (Мег) Гу- man B. Smith E Pittendr., J. Wash. Acad. Sci. 43: 403. 1954. TYPE: Surinam: da 966 (ho- lotype, L). LITERATURE CITED Goupa, E. J. 1987. Bromeliaceae. пау Tilland- sioideae. Fl. Guianas, Ser. A 189: 1-11 Raun, W. 1976. Bromelienstudien I. Neue m wenig bekannte Árten aus Peru u BE Trop. Subtr 977. Bromelienstudien I. Neue und wenig ВЕ es Arten aus Peru und anderen Ländern (7. e Top. Subtrop. Pflanzenwelt 21: 5-47. SMITH R. J. Downs. Tillandsioideae о Fl. Neotrop. 14(2): 663-1492. — John Е. Utley, Department оў Biological Sci- ences, University of New Orleans, New Orleans, Louisiana 70148, U.S.A. and Harry Luther, The Marie Selby Botanical Gardens, 800 South Palm Avenue, Sarasota, Florida 33577, U.S.A. ERRATUM ERRATUM FOR XYRIS (XYRIDACEAE) FROM THE PLANALTO OF BRAZIL There has been oversight in regard to disposition and location of holotype material of Xyris species treated in R. Kral & M. С. L. Wanderley's (1988), “Ten novelties in Xyris (Xyridaceae) from the Planalto of Brazil" (Ann. Missouri Bot. Gard. 75: 352-372). The holotypes of five [X. sceptrifera Kral & Wanderley (Irwin et al. 1341 О); X. seu- bertii Nills. var. espinhacae Kral & Wanderley (Irwin et al. 23667); X. obcordata Kral & Wan- derley (Irwin et al. 27763); X. itambensis Kral & Wanderley (W. R. Anderson et al. 35879); X. villosicarinata Kral & Wanderley (Irwin et al. 19570)], incorrectly assumed to be deposited al- ready at UB, had not yet been sent there. The holotype of X. pranceana Kral & Wanderley (Prance & Silva 58195) is in place at MG, not at UB as was stated in the publication. Thanks to a thoughtful check by Jacquelyn Kallunki at NY, these errors were caught, the holotypes are being sent to UB, and the correct location for the ho- lotype of X. pranceana is now given. The oversight is entirely mine! —R. Kral, Box 1705, Sta. B, Vanderbilt Uni- versity, Nashville, Tennessee 37235, U.S.A. Volume 77, Number 4, рр. 607-865 of the ANNALS OF THE MISSOURI BOTANICAL GARDEN was published on November 16, 1990. ANN. Missouri Вот. GARD. 78: 271. 1991. Novon, А New Journal for Botanical Nomenclature Missouri Botanical Garden In 1991, the Missouri Botanical Garden will begin publishing a j 14 d to th blisl t of latural entities and new taxonomic synonymy. The name of the journal, Novon, is a made-up word meaning any nomenclatural entity at its first publication. Novon will include descriptions and names of new taxa, new combinations of existing names, and new names. Although the format of articles will be concise and stereotyped, ample space will be provided to discuss the taxonomic and/or nomenclatural justification for proposed nova. Taxonomic coverage will be restricted to vascular plants and bryophytes. The journal will be automatically distributed quarterly to all who receive the Annals of the Missouri Botanical Garden. Initally, Novon will not be available as a separate subscription or exchange. Beginning in 1991, the editorial policy of the Annals of the Missouri Botanical Garden will be changed to exclude articles devoted to nova. Novon will provide a means to rapidly publish nova. To speed typesetting, authors should submit manuscripts as two printed copies and on diskette. Authors are also requested to submit xerographic copies of basionyms for new combinations and replaced names for new names. Detailed instructions for authors are available from the editorial office of Novon, Department 11, Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166. Each manuscript will be reviewed for nomenclatural correctness under the current edition of the International Code of Botanical Nomenclature, for the preéxistence of any proposed new nomenclature, and for scientific merit. S ui E. and a New Variety in the Cyperaceae of Mesoamerica C. Dennis | Adams 254 А New Brazilian Species of Dalechampia subsect. Triphyllae (Euphorbiaceae) Grady L. Webster 1512298 Four New Species of Elaphoglossum (Elaphoglossaceae) from Venezuela John T. | Mickel 259 | А New Species of Aristolochia (Aristolochiaceae) from Venezuela _ Kerry Barringer & Francisco Guanchez 262 On the Typification of Bignonia crucigera L. (Bignoniaceae) Fred R. Barrie, James І. Reveal, Charles E. Jarvis & Alwyn Н. Gentry —-———- 264 | Studies in Middle American Bromeliaceae 1 John F. Utley & Kathleen Burt-Utley .. 266 . | Studies in Middle American Bromeliaceae II John Е. Utley & Harry Luther ........... X AAA, Erratum for Xyris (Xyridaceae) from the Planalto of Brazil К. Kral - 27 | Monograph in Systematic ME from the Missouri Botanical Garden Number 37 | otiosus of the Flora of Missouri : George Yatskievych and Joanna Turner | xii, 345 pp., 5/2 х 8% inches, softcover | $9.00, plus postage — The Flora of Missouri Project is a collaborative effort between. the Missouri Department of | > and the Missouri Botanical Garden to produce a fully revised edition of Julian 2er s Ns known Flora of Missouri, first published in 1963. The Catalogue represents a first step toward. goal and contains an annotated checklist of the 2, 622 species and named hybrids, as well as the | Pumerous infraspecific эз, that аге presently rec to grow ou outside of sisti in the state. The Catalogue features • Y id nomenclature | e Common names • Cross-referencing to Steyermark's Flora by page. number | * Additions to the flora indicatd with asterisks | i | | Фф Introduced taxa indicated \ with h diamonds rd E Pp money order in U.S. funds, IEEE TEL ET EEE v | pM book; non ee ie ta hk US aT | Orders should be prepaid; a $1.00 fee will be added to o orders requiring i (o shipments - E Es 4 ч : ы ш. "de until payment is received. Май пе Я pus or noises е = и СИ н; | иен Beren Меша Вані бе e St. Мање мо 63166-0299, USA. D Payment enclosed. и: (D Seinen 41.0 wil be илле Torn en; зе . . Notes on Swartzia | CONTENTS А a Centr of Botany: 1889-1989, the 36th Annual Systematics Symposium of the Missouri | Botani " ical Garden One Hundred Years of the Missouri Botanical Garden Emanuel D. Rudolph — 1 — One Hundred Years of Plant Taxonomy, 1889-1989 Barbara С. Briggs —— 19 A History of Botanical Nomenclature Dan Н. Nicolson ... | AE Who Conquered the New World? or Four Centuries of Exploration + in an 1 Indehiscent | ki Neither Oak nor Alder, But Nearly: the History of Ticodendraceae Barry Hammel Ў г psule Joseph Ewan ........ WU LL (un Botanical Explorations in the American West—1889-1989: An Essay on the Last £ quur ES ntury of a Floristic Frontier James L. Reveal . + ais ; mU The Niger and the Nile: Botanical Exploration Around Two African Rivers F. Ni- 2 ; ДА, ; gel Hepper M Man Uu, PEU ramon ea А New Fins of POTE Plants Jorge € Gómez-Laurito & Luis D. s Cine A анални НИ MEUS > & W.G. Burger be ER Bes ord jn = boe and Bark Anatomy x Ticodendron: Comments on Кере "Sherwin A _ Carlquist . E: fee Te Leaf Architects of Reo а the Oe of Ређе Characters in d pd _ Discerning Its Relationships ` Leo J. Hickey & David Winship Taylor - — ~ Sieve-element Characters of Ticodendron | H.-Dietmar Behnke .... j Pep oduct =e = К gy. Anatomy, and Relati » е = Ticodendron _ Hiroshi ———— iT CBS umm A aid. > Е EES bs и Ws : of Pa anicum subgenus Dichanshelinm тасви : Panicese) dion a Soh America о Zuloaga & Osvaldo Morrone .......... A _ New Species of Ch | :B busoide ) f Смал бок Ue Yonne Widmer ; 5 & Lynn G Clark PG: . Бараа, _ Panamantkus, а New Monotypic Genus of НЕ БАБЕ зе : sae ; Swartzieae) Preliminary to the Flora o the Y enezuela RAA Guayana A. Rupert C. Barneby .. Aea et guo Kunth ас in 1 South A America p р. Michell & Douglas اوی تھے و ماه ии‏ ГТ. ncr f AZ IA SKI Ват Annals of the Missouri ۰ = 78 | Volume 78, Митбег 2 _ Annals of the Spring 1994 Missouri Botanical Garden The Annals, published quarterly, contains papers, primarily in systematic botany, con- tributed from the Missouri Botanical Garden, St. Louis. Papers originating outside the Garden will also be accepted. Authors should write the Editor for information concerning arrangements for publishing in the ANNALS. Instructions to Authors are printed in the back of the last issue of each volume. Editorial Committee Marshall R. Crosby · Gerrit Davidse Editor, | 7 - Missouri Botanical Garden наны Botanical Garden | 225 Wacom M John D. d cts E I | Ату da | VET А _ Missouri Botanical Garden #7 | 2 Managing Editor, | veri 23 — Saint Louis University | Missouri Botanical Garden — 2 === pore As | | moie | ~ EN Peter Goldblatt Diana Gunter Bii ct Missouri Зале Garden _ Editorial Assistant, NC CELA d K Missouri Botanical Бараа E a Pcr Dale E. Jouer: БУРЕ Md Lado AEE | Missouri Botanical Garden e "Сеп Dolan EN Macr и алыуы ныз | Uu ` Magdalen Lampe i rcc Mec dunk van nC We Аан ганса зик A а Missouri Botanical Garden. UU ДЖЕР S = Бог. ii за information contact а Department | The. Ал OF THE MISSOURI BOTANICAL бошо | Eleven 299, St. Louis, MO 63166-0299. (ISSN 0026-6493) is published quarterly by ° | Subscription. pee is 35 p volume U. S, $80 ^ Missouri Botanical Garden, 2345 Tower Grove Av: tries. Four enue, St. Louis, MO 63110. Second class postag? | -issues per volume. чаршы 1991, recipients of paid at St. Louis, MO and additonal mailing обот | У ANNALS it секао receive Novon, the |. POSTMASTER: Send address chang $ P ANNAS ae ae wly lishe juarterly. тне MISSOURI BOTANICAL. GARDEN, D: cn Jie кеи Вох 299, St. Louis, мо и | | 28 | Volume 78 Number 2 1991 Annals of the Missouri Botanical Garden NZ BIOLOGICAL EXTINCTION IN WESTERN ECUADOR С. Н. Dodson! and А. H. Gentry? ABSTRACT This paper documents the extent of deforestation in western Ecuador and what is known about its effect on e evidence — extinction of plant species from four florula sites is provided. The naturally fragmented nature of much wes of the w conservation. n Ecuadorian forest is discussed, both in terms of evolutionary effects and implications for long-term Тће forests of western Ecuador have been cited as one of the most severely threatened areas on earth in terms of biological extinction as a result of deforestation and other activities of humans (Meyers, 1986, 1987, 1988; Gentry, 1989; Sim- berloff, 1986). Although few precise data are readi- ly available, clearly deforestation has left only small forest remnants. Extrapolations based on the the- ory of island biogeography (MacArthur & Wilson, 1967; Simberloff & Able, 1976) suggest that bi- ological extinction in this region may have been massive. This report is intended to provide concrete data concerning the status of forests, reserves, and ex- tinction in western Ecuador and to suggest possible means to salvage what remains. GEOGRAPHY, SOILS, AND CLIMATES OF WESTERN ECUADOR The boundaries of western Ecuador have been established by geographers (Anonymous, 1980) as the Pacific Ocean to the west, the Colombian border to the north, the Peruvian border to the south, and the 900-m contour line on the Andean mountains to the east. With this definition, western Ecuador has a land area of approximately 80,000 km”, almost a third of the total 273,000 km? of the present-day Republic of Ecuador. As thus defined, western Ecuador is about the size of North Carolina, while the Republic of Ecuador is similar in size to California. The majority of western Ecuador con- sists of a series of peneplains extending westward ' Missouri Botanical Garden, Р.О. Вох 299, St. Louis, f 63166, U.S.A., and Director ad Honorem, Herbario del Museo Ecuatoriano de Ciencias Naturales, Quito, Ecua ? Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166, U.S.A. ANN. Missouni Bor. Garp. 78: 273-295. 1991. 274 Annals of th Missouri са Garden from the base of the abruptly rising Western Cor- dillera of the Andes. There is also a low range of coastal hills seldom exceeding 800 m in elevation. The cold Humboldt Current moves northward along the coast and collides with the warm Panama Current moving southward. Both then move west- ward toward the Galápagos Islands. The earth's seasonal oscillation provides annual northward and southward shifts in the currents, which affect the climate of western Ecuador. In general, the region nearest the coast is more affected by the Humboldt Current and is drier, while the Andean foothills are moister. The southern half of the coastal region, with less than 1,000 mm annual rainfall, has a rainless period exceeding eight months (early May through early January), although a persistent cloud cover during the dry season ameliorates the tem- pues and dryness. Near the ocean, night fogs bstantial atmospheric humidity, especial- Ў on the crests of the low coastal mountains. Be- ginning about one degree south of the equator, the coastal region becomes increasingly humid north- to annual rainfall of 7,000 mm, with at least 100 mm of rain each month near the Colombian border. The soils of the western peneplains are alluvial and volcanic in origin. Agricultural use of the lands by humans since aboriginal times has sustained large populations, and much of the region is now under intensive agriculture. The soils of western Ecuador are much richer than those of most other tropical lowland areas. Associated high productivity and turnover rates of the native forest may be in part responsible for the dynamic speciation that seems to have occurred. The rich soils make the western Ecuadorian conservational situation very different from that in most tropical regions. Instead of the common alternative of forest conservation, i.e., a few years of crops followed by abandonment, in western Ecuador the alternative to forest con- servation is often productive and sustainable ag- riculture. This situation is somewhat analogous to that of mid-America during the middle of the last century where cultivation of rich soils nearly elim- inated native prairie flora. VEGETATION COVERAGE According to one of the more recent ecological maps of Ecuador (Canadas & Estrada, 1978), based on the Holdridge (1967) system, 12 life zones occur in western Ecuador (Fig. 1). We suspect that one more life zone, a subdivision of pluvial forest, which Canadas & Estrada did not include, may occur in the extremely wet region of northern Esmeraldas Province. Our experience, after years of collecting plant specimens and traveling throughout western Ecuador, suggests that all of the ecological maps of Ecuador are quite generalized, and we feel that many of the life zones are much more fragmented than is indicated even on the Canadas & Estrada map. This natural fragmentation may have helped spawn the many endemics found in small habitat patches. The prolonged dry season in the south and west generally results in a progression of desert, desert scrub, tropical thorn scrub, and a dry to very dry tropical forest as one moves inland away from the cold waters along the coast. The coastal hills are clothed with thorn scrub at the base, premontane dry forest on the slopes, and moist or wet forest on the higher crests. The original forest cover of the southern portion of the central plains was trop- ical dry forest. This gives way to tropical moist forest, without a clearly discernable transition zone, near 2? south latitude (near Babahoyo), and tropical moist gradually gives way to tropical wet forest north of 1? south latitude (north of Quevedo). Near Santo Domingo the moist forest is replaced by premontane pluvial forest. An extensive system of savannas, swamps, and gallery forests once existed along the large rivers that drain the Guayas basin south of 1? south latitude. Only the Santa Elena Peninsula, at the western extreme of Ecuador im- mediately adjacent to the cold Humboldt Current, had true desert conditions similar to those of coastal Peru and was completely lacking forest cover. A narrow strip of perhumid cloud forest on the lowermost Апдеап slopes is included in western Ecuador. This forest is an extension of the Colombi- an Choco pluvial forests, which аге generally char- acterized by extremely high annual rainfall. These wet forests extend southward in increasingly nar- row elevational bands along the flanks of the Andes. True lowland pluvial forest, with perhaps 8,000 mm of rain, reaches south only to near Lita, a few kilometers south of the Colombian border, at about I? north latitude. nearly to the Peruvian border, where it is reduced Premontane wet forest reaches to a strip only a few hundred meters wide near the 900-m contour. (Other montane vegetation types exist above that contour but are not within the scope of this discussion.) In general, the moist vegetation types form broad extensions near the Colombian border but are reduced to narrow bands between 300 and 900 m elevation near the Pe- ruvian border. Volume 78, Number 2 Dodson & Gentry 275 1991 Biological Extinction in Western Ecuador о“ 5 ЕЈ m: 7 / 4 / ~ -— [ • | ) \ 7 p. ” $ 1. Tropical desert scrub. ^ 2. Tropical desert. 2 3. Tropical thorn scrub. ( 4. Premontane thorn scrub. ) „5. Very dry tropical forest. 6. Premontane dry forest. 7. Tropical dry forest. 8. Premontane humid forest. 9. Tropical moist forest. 10. Premontane wet forest. 11. Tropical wet forest. 12. Premontane pluvial forest. FIGURE 1. Мар of life zones of western Ecuador based on Cañadas & Estrada (1978). mated 17,000 species (Morin, pers. comm.) known for all of North America. Ecuador as a whole is very rich biotically. The Of the 12 life zones reported in western Ecuador, current estimate of 20,000 vascular plant species 10 have been studied with varying degrees of in- in Ecuador (Harling, 1986) far exceeds the esti- tensity. Five have been extensively sampled by FLORA AND FAUNA OF WESTERN ECUADOR 276 Annals of the Missouri Botanical Garden using the florula approach (see Dodson & Gentry, 1978). These include the Flora of Rio Palenque (1978) in tropical wet forest, the Flora of Jauneche (Dodson et al., 1985) in tropical moist forest, the Flora of Capeira (in press) in tropical dry forest, the Flora of Centinela (in prep.) in premontane pluvial forest, and the Flora of Tenefuerte (in prep.) in premontane wet forest. The Santa Elena Pen- insula, with three very dry life zones (tropical des- ert, tropical desert scrub, and tropical thorn scrub) was studied by Svenson (1946) and Valverde et al. (1979). Valverde is presently preparing a flora of the Cerros de Colonche. This region of the dry coastal mountains has tropical thorn scrub, pre- montane thorn scrub, tropical dry forest, very dry tropical forest, premontane dry forest, and tropical moist forest. Accurate estimates of the number of species of vascular plants in western Ecuador are not possible at this time, especially since two of the life zones richest in species diversity have not been sampled extensively, i.e., forest. Yet extrapolations of species numbers can premontane humid and pluvial be based on data from the areas sampled for florula preparation and from the relatively thoroughly sampled western Ecuadorian orchid flora. The five sites studied by Dodson and Gentry have a com- bined total of about 2,500 species. The number of orchid species found in the five florula sites is 253, closely following the 10% suggested for tropical floras worldwide (Kress, 1986). We have 2,607 orchid species cataloged for all of Ecuador, thereby also close to 1096 of the estimate of 20,000 vas- cular plant species recently projected by Harling (1986). The 632 orchid species cataloged for low- land western Ecuador, multiplied by ten, give a rough estimate of 6,300 vascular plant species in the region. Overall figures for fauna are not available, but data on birds show more than 1,600 species known for the country (Ridgley & Greenfield, pers. comm.), 650 of these in western Ecuador. The highest bird endemism in the world is probably in the Ecuador / Colombia western slope border area (Terborg & Winter, 1982). ENDEMISM IN WESTERN ECUADOR Endemism can occur at many different levels, from a single ridge top to a continent. Gentry (1986a) used the term “locally endemic" for spe- cies that have total distributions of up to 75,000 km?. Most species endemic to western Ecuador are much more localized, occurring in only one or two ecologically appropriate forest types, with original distributional areas of less than 20,000 km? By comparing distributional patterns of the spe- cies included in the florulas mentioned above, Gen- try (1982a) estimated that about 20% of the flora is endemic. For example, of the 870 naturally occurring flora of Rio Palenque species with known distributions, 2096 (172) are endemic to western Ecuador and another 6% to western Ecuador plus adjacent Colombia. Similar analyses of the distri- butional patterns of the plant species of the dry forest in the flora of Capeira show that 19% are endemic to the dry area of western Ecuador (here taken to include the phytogeographically similar adjacent northwesternmost corner of Peru). Thus, endemism of the dry forest species is essentially as high as for the wet forest ones. Dry forest plants differ in having only five species (176 vs. 676 of the Rio Palenque flora) more widely endemic to coastal Ecuador plus adjacent Colombia. In western Ecuador moist-forest species tend to be less prone to endemism than those of either dry or wet forest. Nevertheless, analysis of the 553 species of the 1985), for which we have complete identifications and moist-forest Jauneche florula (Dodson et al., distributional data, indicates that a significant 15% are endemic to western Ecuador, with an additional 295 endemic to western Ecuador plus southwest Colombia. Identifications are as yet too incomplete for our cloud forest florula sites to attempt similar analyses, but they may have even higher rates of endemism. We conclude from the above data that if lowland western Ecuador has 6,300 naturally occurring species, around 20% of them, or 1,260 species, are probably endemic to the region. Interestingly, even though moist and dry forest florulas have a higher percentage of widespread species (respec- tively 47% and 50%) than does wet forest Rio Palenque (31%), the endemism rate remains sim- ilar, due to reductions in such intermediate distri- butional categories as Central America-to-western Ecuador (8% at Capeira and 12% at Jauneche vs. 18% at Rio Palenque). A taxonomic breakdown of endemism figures is also of interest. At Rio Palenque 28% of the mono- cots are endemic vs. 19% of the dicots, but at both Capeira and Jauneche endemism of monocots and dicots is similar. At all three sites, fern endemism is much less (6-9%), and orchid endemism unusu- ally high (24% at Jauneche, 50% at Capeira, and 28% at Rio Palenque). Much of the excess of endemic monocots at Rio Palenque is due to high representation of speciation-prone orchids. In view of their unusually rapid speciation (Gentry & Dod- Volume 78, Number 2 1991 Dodson & Gentry 277 Biological Extinction in Western Ecuador son, 1987b; Burns-Balogh & Bernhardt, 1988), an estimate of endemism based on orchids might be expected to be on the high side, even though orchids are also famous for long-distance dispersal of their minute seeds. Orchids have been relatively intensively surveyed in western Ecuador and 68 (or 2776) of the 250 orchid species known from the sites are regional endemics. Again, this would seem to agree with our estimate of around 20% endemism for the flora as a whole. The preceding discussion addressed endemism at a regional level. However, within western Ec- uador there are loci of much narrower endemism, especially on isolated ridge tops. Some of these patches of high local endemism may consist of habitat islands no more than 0.5-10 km?. That such extreme local endemism is typical of certain tropical forests has only recently been realized (Gentry, 1982c, 1986a). The best documented case is the Centinela ridge top (see below and Gentry, 1986a). This is the first Andean foothill ridge, about 8 km east of Rio Palenque and isolated from the main Andean range farther east by a broad, flat valley about 15 km wide. The ridge is only about 20 km long and 1 km wide, an area of some 20 km”. At its northern end it merges with the flat but higher altitude terrain formed by the less eroded outwash from the Andes of the Santo Domingo area. It is only 800 m high at its highest point and thus only 500 m higher than the valley that separates it from the Andean foothills proper. Nevertheless, it is high enough to engender a local cloud forest effect, and has a flora dramatically different from adjacent regions. Thirty-eight new species of vascular plants have already been de- scribed from this ridge top, and many additional collections appear to represent new species. In- cluding morphotypes that we have been unable to match with named material in the herbarium, there were 90 plant species endemic to the Centinela ridge, about 10% of the entire Centinela flora. The main ridge is now deforested, and an undetermined number of these species are now apparently extinct. Others still tenuously survive as herbs under the banana, coffee, and cacao plantations that replaced the forest, as epiphytes in cultivated trees, or on a disjunct southern spur of the ridge where we recently discovered a tiny remnant patch of forest. Many of the species at Centinela are strikingly beautiful and quite unlike any known congeners. One noteworthy feature shared by many species of unrelated families is that the upper surfaces of their leaves are black and usually intricately bul- ate. How did so many distinctive species come to exist at Centinela? Gasteranthus of the Gesneri- aceae provides an instructive example. This small genus of about 25 species ranges from Guatemala to Peru, mostly in cloud forests, generally with a few endemic species in each country in which it occurs. In Panama, for example, there are five species—two endemic and three shared with Costa Rica (Skog, 1978). Most Gasteranthus species have rather similar-looking orange flowers with a pouch- like corolla. On Centinela there are (or were) no fewer than six species, a quarter of the world total! Four of these are restricted to the high-endemism area near the southern tip of the ridge, a fifth species is also known from a single collection from north of Santo Domingo, and the sixth occurs also at Rio Palenque. Most of these six species are dramatically different from each other and from their congeners. Indeed, Gasteranthus shows more morphological diversity in the few square kilome- ters of Centinela than in the rest of its range. The endemic species of Centinela represent a high level of in situ evolution. Our floristic knowledge is inadequate to evaluate how many isolated ridge tops shelter large numbers of endemic species as does Centinela. It is possible that many of them do (or did before being defor- ested). For example, preliminary observations at Lita, a high-rainfall region in Esmeraldas near the Colombian border, suggest a very distinctive flora with many species new to Ecuador and presumably to science; however, several of the putative Cen- tinela endemics have also been discovered there. Many new and apparently endemic species are also being discovered at La Planada just across the Colombian border on a 1,700-m ridge that extends from the main Andean Cordillera (Benavides & Gentry, in prep.); at Tenefuerte, another local flo- rula site east of Quevedo at 1,000 m, and at Cerro Azul, a 400-m ridge subject to cloud effect just north of Guayaquil. How many such species exist and how many of them will turn out to be extreme local endemics can only be determined with far more extensive floristic data than currently available. However, we do have previously unreported data comparable to those of Centinela for a different habitat island much farther from the Andes, at Cerro Montecristi. Cerro Montecristi, 12 km east of Manta (see Fig. 2), is one of the characteristic, outlying, iso- lated hills on the extreme western side of Ecuador, and provides a good example of how microhabitat fragmentation and speciation promote extreme lo- cal endemism in western Ecuador. Cerro Monte- cristi was explored and the flora was studied during 1957 and 1960 (Dodson & Frymire, unpublished). 278 Annals of the Missouri Botanical Garden FicunE 2. Altitudinal zonation on Cerro Montecristi. The elevation at the base of Cerro Montecristi is about 150 m and at the summit 620 m. The hill is roughly triangular at the base and about 5 km from point to point on the eastern side with a prolonged western slope about 10 km long. The eastern face rises quite abruptly above the town of Montecristi. At night during the dry season, a cloud cover blankets the upper 100 m of Cerro Mon- tecristi. The result is a nightly deposition of con- siderable moisture along the crest similar to that reported by van der Werff (1978, 1983) in the higher mountains of the Galápagos Islands. Con- sequently, the crest has a cover of wet-forest veg- etation like those of wet forests further inland at much lower elevations. Walking from the sparse, dry scrub at the base to the summit results in crossing a sequence of habitats, each 50 to 100 m across; the lower bands are several hundred meters wide while the upper bands, on steeper slopes, are narrower. The bands probably correspond with lo- cal associations sensu Holdridge (1967). We could recognize nine distinct vegetation types. e were particularly interested in the distri- bution of orchid species in the vegetation types (Table 1). Certain species were distributed through several zones, while others were limited strictly to specific zones and were encountered at the same elevation on any of the three access trails. With Моште 78, Митбег 2 1991 Dodson & Gentry 279 Biological Extinction in Western Ecuador the exception of zone 1, each zone has at least one orchid species restricted to it; most zones also have several species shared with other zones. Two spe- cies of orchid are known only from Cerro Mon- tecristi, Lepanthes manabina Dodson and Stelis manabina Dodson. Both were quite common in zone 9 in a total area of probably less than 4 ha. Search of similar habitat near Jipijapa, approxi- mately 30 km due east of Montecristi and the closest occurring vegetation of the same type, did not reveal any plants of either species. Recent conversations with Alex Hirtz concern- ing the distribution of Bromeliaceae on Cerro Mon- tecristi revealed strikingly similar distribution pat- terns. Hirtz is a mining engineer with a deep interest in orchids and bromeliads. He is the most knowl- edgeable authority on the identification and distri- bution of Ecuadorian Bromeliaceae. He has climbed Cerro Montecristi several times and has indepen- dently confirmed for both orchids and bromeliads the zonational distribution that we encountered in studying the orchids of the site. We have included in Table 1 the bromeliad species cited by Ing. Hirtz. Though Cerro Montecristi may be extreme in its diversity of habitat in a very small area, like Centinela it is probably representative of the sit- uation in much of western Ecuador. Elsewhere we have proposed (Gentry & Dodson, 1987b) that small, fragmented habitats in restricted life zones, especially in cloud forests, can promote rapid spe- ciation, perhaps in as little as 15 years. Since the only two relatively isolated cloud forests for which , Centinela and Cerro Mon- tecristi, both show extremely localized endemism, we strongly suspect that extremely localized en- demism is a common occurrence in western Ec- uador. Thus, from a conservation perspective, to avoid massive extinction it is not only important to conserve patches of the major forest types, but also to conserve individual isolated or semi-isolated habitat islands. we have rea EFFECTS OF HUMAN ACTIVITIES ON THE FOREST COVER HISTORICAL PERSPECTIVE During the centuries preceding the Spanish con- quest in the early 1500s, western Ecuador sup- ported a large population of native peoples. The quantity of artifacts and ceremonial mounds scat- tered across the plains is so extensive that they could only have been produced by a large popu- lation over a prolonged period of time. Towns that apparently had 10,000 or more inhabitants at the time of the arrival of the conquistadores have been discovered recently in Manabi, and it has been suggested that the rural population of western Ec- uador was greater at that time than it is today (P. Norton, pers. comm.). European diseases presum- ably drastically reduced the native population. uring the colonial period, most farming and forest conversion was restricted to the Guayas Riv- er system, the Sta. Elena Peninsula (then appar- ently much more humid than now according to our reading of historical records; M. Aspiazu, pers. comm.), and the seasonally dry coastal region of Manabi. The lucid descriptions of Hassaurek (1967), United States Ambassador to Ecuador from 1861 to 1866, of the region of Guayaquil and Babahoyo indicate that the population was still sparse at that time. For example, Guayaquil had an estimated population of 20,000 in 1861 vs. 1.7 million today. According to Hassaurek, most of the region from Babahoyo to the high country, even flanking the main trail to the sierra, was still unbroken forest. All available evidence suggests that substantial de- forestation did not begin until after World War II, when road construction was accelerated. Figures 3 and 4 illustrate the transport system and asso- ciated deforestation of western Ecuador. Figure 3 (Map 1) shows the river system navigable by launch or canoe, the principal means of inland transport prior to roads. Figure 3 (Map 2) shows the road and railroad system existing in 1938. Because ex- tensive deforestation seldom occurs more than 5 km on either side of a road or river bank, a com- bined map of roads, railroads, and river systems provides a tentative indication of where relatively undisturbed forest existed in western Ecuador prior to 1945 (Fig. 4, Map 1). At that time, we estimate that at least 7576 of western Ecuador was still forested. In 1957, when the senior author first arrived in Ecuador, substantial forest conversion to banana plantations already had occurred, par- ticularly along the post-World War II roads (Fig. 5, Map 3); however, we estimate that approxi- mately 63% of the surface of western Ecuador still retained primary forest cover (Fig. 4, Map 2). From 1957 until 1963 the senior author was in residence in Ecuador for more than three years and spent a major portion of that time studying population dis- tributions and pollination systems of the orchids in the forests and western Ecuador wherever roads or paths made access possible. Therefore, subse- quent to 1957 we have direct evidence, albeit in part anecdotal, of the extent of deforestation in coastal Ecuador Between 1957 and 1988 several factors played important roles in the deforestation of western Ec- uador. The population of Ecuador increased from 280 Annals of the Missouri Botanical Garden TABLE 1. Zones on Cerro Montecristi and their orchid and bromeliad distributions. Not endemic to zone Endemic to zone Zone 9 600-620 m. Dominated by wet-forest elements with a dense load of epiphytes and abundant moss. Orchids: 1. Oncidium obryzatum . Dracula mopsus 2. Cyclopogon epiphytica 7 Lepanthes manabin 3. Masdevallia nidifica 3. Masdevallia cahlbruchneri 4. Sobralia fenzliana 5. Stanhopea dide . Stelis manabin 7. Pleurothallis и Bromeliads: 1. Guzmannia teuscheri Zone 8 575-600 m. Dominated by very humid forest elements with abundant, mesic bromeliads and very little moss. Orchids: 1. Gongora grossa 1. Kefersteinia expansa 2. Ornithocephalus bryostachys 2. Oncidium exasperatum 3. Oncidium obryzatum 4. Cyclopogon epiphytica 5. Masdevallia nidifica Bromeliads: 1. Guzmannia sanguinea var. sanguinea l. Guzmannia wittmackil 2. Tillandsia venusta 2. Tillandsia hammeliana Zone 7 480-575 m. Dominated by humid-forest elements with mesic bromeliads and no moss. Orchids: 1. Gongora grossa 1. Oncidium cardiochilum 2. Ornithocephalus bryostachys Bromeliads: 1. Guzmannia sanguinea var. sanguinea 2. Tillandsia venusta Zone 6 410-480 m. Dominated by dry forest mixed with humid-forest elements. Bromeliads mostly mesic. Orchids: 1. Cattleya maxima 1. Pleurothallis + 2. Trichocentrum tigrinum 2. Restrepia dodso 3. Gongora grossa Bromelia Ss . Tillandsia venusta Zone 5 350-410 m. Dominated by dry-forest elements with mesic and dry bromeliads. Orchids: 1. Catasetum expansum . Campylocentrum pachyrrhizum 2. Cattleya maxima a Psygmorchis pusilla 3. Trichocentrum tigrinum 4. Notylia replicata ess liads: . Tillandsia venusta 1. Aechmea mexicana ^ Bromelia pinguin 2. Aechmea angustifolia Zone 4 280-350 m. Dominated by dry-forest elements with tillandsias as epiphytes. Orchids: 1. Catasetum expansum 1. Macroclinium manabinum 2. Cattleya maxima 2. Brassia jipijapensts 3. Trichocentrum tigrinum 3. Dipteranthus estradae 4. Notylia replicata Volume 78, Number 2 1991 Dodson & Gent ry 281 Biological Extinction in Western Ecuador TABLE l. Continued. Not endemic to zone Endemic to zone Bromeliads: 1. Bromelia pinguin 2. Vriesia barclayana var. minor 1. Tillandsia floribunda 2. Tillandsia multiflora var. decipiens Zone 3 220-280 m. Dominated by very dry-forest elements and thorn forest. Orchids: 1. Catasetum m 2. Cattleya m a 3. a tigrinum Bromeliads: 1. Vriesia oe var. minor Vriesia es 3. Tillandsia no ВИ var. multiflora 1. Oncidium hyphaematicum Zone 2 160-220 m. Dominated by thorn scrub with scattered ceibas and arborescent cacti. Orchids: 1. Catasetum expansum Bromeliads: . Friesia e var. major 2. Vriesia espino 3. Tillandsia a var. multiflora 1. Oncidium onustum (Restricted to Cereus spp.) 1. Tillandsia latifolia Zone 1 160-220 m. Dominated by bare soil or scattered scrub. No orchids encountered. Occasional fences of Bromelia pinguin less than 4 million to 10.2 million. The natural increase rate of the population of Ecuador as of 1988 was 2.8% with a “doubling time" of 25 years (1988 World Population Data Sheet Popu- lation Reference Bureau, Washington, D.C.). Nat- ural increase is birth rate minus death rate, im- plying the annual rate of population growth without regard for migration. “Doubling time" is the num- ber of years it would take to double the population, assuming a constant rate of natural increase. This doubling time is not intended to forecast the actual doubling of a population and is simply the length of time it would take for the population to double in the absence of migration in or out, if the present birth rates and death rates continued over the next 25 years. The 2.8% increase rate is a slight de- crease from the previous growth rate but is still among the highest in Latin America. In Latin America, only Guatemala, Honduras, Nicaragua, and Paraguay, all having much smaller absolute populations, have faster growth rates than Ecua- dor, whereas countries like Brazil (2%), Colombia (2.1%), and Peru (2.5%) have substantially lower growth rates. Land reform programs initiated in the early 1960s effectively promoted colonization of government-owned forested lands. Petroleum export provided more than 60% of the national income as Ecuador became a member of the Or- ganization of Petroleum Exporting Countries (OPEC), selling oil at more than 35 dollars a barrel, and thereby insuring abundant new funds for de- velopment. Major sums were invested in road con- struction to provide communication and transport to cities and markets. Bananas, oil palm, soybeans, rice, corn, and shrimp farming, as well as the traditional cacao and coffee, became sizable export crops, providing substantial agro-industrial income. Constructed and maintained by the Ministry of Public Works, an extensive network of primary and secondary roads was developed throughout most of western Ecuador (Fig. 4, Map 4). In 1957 about 13,000 km of primary and secondary roads existed in the country (Anonymous, 1987). In 1982 (the most recent available information), there were 19,341 km of primary road and 16,276 km of secondary road for a total of 35,617 km in the Republic of Ecuador, almost triple the 1957 total. Estimates by Fausto Lafebre (pers. comm.), Di- rector of Technical Programming of the M.O.P., indicated that as of 1987 there are more than 50,000 km of primary and secondary roads in the country, and well over half are in western Ecuador. n addition, many penetration roads (or neigh- borhood roads, as they are often called) were con- Annals of the Missouri Botanical Garden 282 ¢ 1 e f ^ ? s^ ot Í $ B Quito 7 / ( > \ / ( ^ no! ( [* | . | / 21 т» „ \ | $ Guayaquil | \ Mc y \ Ne f 209. +“ E > 2 $ е rcs 7 Peta “ a d | ) La f 1 P y и е. \ c ) M MAP 2. Road system 1938. MAP 1. Navigable rivers by launch. MAP 4. Road system 1988. МАР 3. Road system 1958. Transportation network of western Ecuador. FIGURE 3. Dodson 4 Gentry Biological Extinction in Western Ecuador Volume 78, Number 2 9 "m 2x | За • р 8 0 oy @ Quito у ( / ~ ЧЕ, Ы | е ) \ Y \ \ \ • ^ / & ЈЕ = € ^u л > ni M МУ ) Vv MAP 2. Forest cover 1958. MAP 1. Forest cover (aboriginal). МАР 3. Forest cover 1988. Forests of western Ecuador from 1938 to 1988. FIGURE 4. 284 Annals of th Missouri рш Garden Pd ; 1988 4 5 8 / 1 [4 / (2 / 1958 РА 2 ГА ۴ MOIST FOREST / DRY FOREST FIGURE 5. Comparison of dry and moist forest conversion between 1958 and 1988. structed either by private enterprise or by 10-12 penetration roads were constructed during the land largely uncoordinated governmental agencies, which reform efforts between 1965 an 5 when a were programmed to help in development of ag- plan was implemented to develop and put into ricultural colonies and cooperatives. Most of the production all “nonproductive” lands, including all Volume 78, Number 2 1991 Dodson & Gentry 285 Biological Extinction in Western Ecuador forested lands in western Ecuador. Many of the penetration roads are not maintained by the M.O.P., and in many cases the Ministry is not even aware of their existence. No estimates of the total kilo- meters of these roads are available, but personal experience indicates that they are now very ex- tensive and form a dense network that provides easy access during the dry season to all forested areas. Such easy access has had a devastating effect on the forest. PRESENT SITUATION Figure 4 (Map 3) displays the extent of relatively undisturbed forest known to remain in western Ecuador today. The only two substantial tracts left are in the provinces of Esmeraldas, Carchi, and Imbabura, all in the extreme north, where high monthly rainfall throughout the year has impeded development and utilization of penetration roads. These two tracts encompass the Awá Reserve (80,000 ha, with about 120,000 ha of uncom- mitted forest surrounding it) and the Cotacache- Cayapas Ecological Reserve (204,420 ha). More than 80% of the two tracts lies above the 900-m line outside of western Ecuador as here defined. e know from personal experience that an un- determined but clearly increasing portion of both reserves has already been cut by colonists and lumber interests. Moreover, the government is cur- rently planning an all-weather highway that will pass through the middle of both reserves. Much of the dry-forest reserve of the Machalilla National Park near Manta is now second growth that has been seriously disturbed by colonists and their goats. South of Guayaquil, the Churute Forest Reserve (35,000 ha) has a fairly substantial cover of somewhat disturbed mangrove but also contains disturbed dry forest on the protruding hills. Several small private reserves exist in western Ecuador, including Capeira (10 ha) in dry forest near Guayaquil; Jauneche (130 ha) in tropical moist forest south of Quevedo; the Rio Palenque Science Center (100 ha) in tropical wet forest between Quevedo and Santo Domingo; and La Perla (30– 50 ha) in premontane wet forest north of Santo Domingo near La Concordia. Extensive mangrove forest once covered the coastal regions that were not directly facing the open sea. The advent of mariculture on a scale that provides for the largest export of shrimp of any country worldwide had reduced by 1984 the mangroves from 2,036 km' to less than 1,881 km? for a reduction of 10.6% (Alvarez et al., 1984). Considerable further conversion has taken place since 1984. In total, we estimate that less than 5,800 km? of reasonably undisturbed forest remain in western Ecuador, that is, less than 8% of the land surface is presently covered by primary forest. The only two relatively large remaining forest tracts are dis- appearing rapidly, and construction of a new all- weather road through them will accelerate this process. The conservation picture can only be de- scribed as grim. FOREST REMNANTS BY LIFE ZONE Figures 5 and 6 illustrate the changes in forest cover that we have observed from 1958 to 1988 in four categories of forest: dry, moist, wet, and pluvial. These categories may each include several life zones. For example, the comparative maps of dry forest in Figure 5 encompass tropical desert scrub, tropical desert, tropical thorn scrub, pre- montane thorn scrub, very dry tropical forest, and premontane dry forest in the Holdridge (1967) system. All of these life zones have been severely altered by humans and by domestic animals (Table 2). We estimate that about 35% (28,000 km?) of western Ecuador consisted of dry forest originally. Probably much less than 1% (about 200 km?) of this category remains undisturbed. The moist forest category, which encompasses tropical moist forest and premontane moist forest in the Holdridge system, was once the most ex- tensive forest cover in western Ecuador (Table 2). Since the soils of western Ecuador are generally quite rich, particularly when compared with Ama- zonia, the moist forests were extremely attractive for agricultural development. They also have a rain-free period of at least four months, allowing easy access. We estimate that nearly 40% (32,000 km?) of western Ecuador was covered forest originally. Surely less than 4% (about 1,500 km") remains. Regeneration of moist forest is prob- ably very slow, and in this case, most of the forest with moist was clear-cut, burned, plowed, and permanently altered. The wet forest category, encompassing tropical wet forest and premontane wet forest in the Hold- ridge system, was originally a band extending from Colombia to Peru. The band was about 20 km wide at the northern extreme and very narrow and bro- ken to the south. The soils are very rich and the forests today are mostly converted to banana, Af- rican palm, and rubber cultivation. We estimate that about 15% (12,000 km") of western Ecuador was covered by this type of forest and that now less than 0.8% (9 Pluvial forest, encompassing premontane pluvial О km”) survives. forest and probably lower-montane pluvial forest 286 Annals of the Missouri Botanical Garden WET FOREST PLUVIAL FOREST FIGURE 6. (even though the latter is not included on the Cana- das & Estrada map as known to occur in western Ecuador), forms the only substantial surviving for- est category. We estimate that about 10% (8,000 km?) of western Ecuador was covered by forests of this category and that about 25% (2,100 km”) remains. The reasons for persistence of forest of Comparison of wet and pluvial forest conversion between 1958 and 1988. this category are probably the notoriously poor soils resulting from leaching, inaccessibility due to high rainfall throughout the year that destroys roads that are not all-weather, and the precipitously bro- ken nature of the terrain. These forests are of particular interest because they probably contain the highest biological diversity in western Ecuador, Volume 78, Number 2 Dodson & Gentry 287 1991 Biological Extinction in Western Ecuador TABLE 2. Areas of forest categories with estimated numbers of species and species endemic to them. Esti- mated Original forest Aboriginal forest Remaining as Remaining as d — cover cover of 1958 of 1988 — а num of en- Forest Total ber of demic categories Km? % Km? = % left Km? %left Km? Ф left ^ species species % Dry 28,000 35 15,000 70 12,000 60 200 1 1,000 190 19 Moist 32,000 40 30,000 70 24,000 60 1,500 «4 1,000 140 15 et 12,000 15 10,000 80 7,000 58 90 «0.8 1,700 340 20 Pluvial 8,000 10 8,000 >90 6,000 75 2,100 25 2,300 590 >25 Total 80,000 100 68,000 77 49,000 63 3,890 3! 6,300 1,260 20 ' This number is based on the total of 80,000 km? of western Ecuador. If the total remaining in aboriginal times is used, the remaining primary forest constitutes 4.4% of the total. comprising part of the Choco region (Gentry, 1982b), which contains the most diverse forests on earth (Gentry, 1986b). Though most of the still- forested lands have been declared National Eco- logical Reserves, they are also the only large ex- tensions of forest left in western Ecuador in an increasingly land- and timber-starved era. CONSERVATION ÉFFORTS GOVERNMENT-ESTABLISHED ECOLOGICAL RESERVES AND NATIONAL PARKS In parts of the world such as North America and Europe, an effective means of conserving for- ested areas is to declare them as national parks, wilderness areas, or ecological reserves. It is as- sumed that once so designated, previously estab- lished governmental agencies will take over man- agement and enforcement of law. However, this assumption is not always true even in developed countries and is much less so in developing coun- tries. More than 20% of the surface area of Ec- uador has been declared to be ecological reserves and national parks. The distinction between the two is somewhat hazy, but in Ecuador only in the case of the Galápagos Islands National Park have conservation practices even approximated effec- tiveness. On the mainland, mining and petroleum development activities tend to take official prece- dence over forest preservation. The management of national parks and ecolog- ical reserves is the responsibility of the Parks Sec- tion of the Forestry Department of the Ministry of Agriculture. The Parks Section has had good in tentions and planning but is woefully understaffed and underfunded. One of the primary thrusts of the Ministry of Agriculture through its Institute for Agrarian Reform (IERAC) is **modernization of agricultural techniques and development of unused while the Parks Section has as its primary but conflicting aim the preservation of those same lands. The Ministry of the Interior, with the ac- quiescence of the Ministry of Agriculture, gives concessions for exploratory petroleum drilling (Мет, 1988) and mining within National Park boundaries. Access roads are necessary where occur. Colonists follow the bulldozers, and lan titles are given to them (50 ha each). With Ecua- dor's burgeoning population, there seems little hope for change in political philosophy in the foreseeable future. At this time, private initiatives may be the only effective means of true forest protection. lands," those activities EXISTENCE OF ““CLANDESTINE FOREST” One of the particularly undesirable results of implementation of land reform laws enacted in the 1960s in Ecuador was their effect on privately owned forests. In an attempt to break up large land holdings, the law allowed confiscation and redistribution to landless peasants of all of the land of any hacienda that included ““unproductive” land. Consequently, it became dangerous for a hacienda owner to retain natural forest on his land, since forested land was interpreted as unproductive and thus subject to seizure. The land reform act also required total conversion to production of all forest on lands deeded to colonists. As a result of this law, many haciendas in Ecuador were invaded by colonists with the support of the land reform agen- cy. In 1971 the Rio Palenque Science Center was officially designated as a privately owned National Forest Preserve by president Jose Maria Velasco Ibarra. In 1975 an attempt was made to invade the Science Center. The 44 prospective colonists began and were repelled with the help of neighboring "clearing" land in the overgrown pasture 288 Annals of the Missouri Botanical Garden TABLE 3. Plant species locally extinct at RPSC. Species Family Dicliptera dodsonii Wassh. Acanthaceae Tabebuia chrysantha (Jacq.) Nichols. Bignoniaceae Quararibea soegenii (Cuatrec.) Robyns & Nilsson о Pseudobombax millei (Standley) Robyns acaceae ohnbergia nidularioides Luther Bromeliaceae Tillandsia dodsonii L. B. Smit Bromeliaceae Burme rachyandra Wimmer C nulaceae Anemia a phyllitidis (L.) Sw. Fer Gloxinia dodsonii Wiehler Gesneriaceae Drymonia turrialvae Har Gesneriaceae Trichantha herthae Mu Wiehler Gesneriaceae Clusia venusta Little Guttiferae Calliandra caracasana (Jacq.) Benth. Leguminosae Rhynchosia precatoria (Willd.) HBK Leguminosae Heliconia wagneriana Wara Musacea Neea ampliflora J. D. Зпи Nyctaginaceae Elleanthus aristatus Garay Orchidaceae Elleanthus zamorensis ^ Orchidaceae Gongora grossa Rchb. Orchidaceae Leucohyle subulata es ) Rchb. f. Orchidaceae Lockhartia serra Rchb. f. Orchidac Maxillaria alba (Hook.) Lindley Orchidaceae Psilotum nudum (L.) P. Beat Psilotaceae Erythrochiton giganteus Kaastra & A. Gentry Rutaceae farmers, resulting in no loss of mature forest (see also Gentry, 197 n the other hd there are a series of laws mandating retention of limited amounts of forest, particularly along watercourses, as well as permit- ting governmentally proteeted private forests with ¥ “redeeming social value." Though a perhaps sub- stantial number of privately owned, large farms still have appreciable amounts of natural forest, they are extremely difficult to find because the owners fear expropriation. An educated guess would be that the total of ** Ec uador might approach 100 km clandestine foreste" in western one considers clandestine —— and other remnants along river courses, hilltops, and areas too broken or steep for efficient cultivation, an interrupted reticulum of existing forests still re- mains over a significant part of western Ecuador. EXTINCTION EVIDENCE FROM FLORULA SITES THE RIO PALENQUE SCIENCE CENTER (RPSC) (FIG. 7) Extensive plant collection began at RPSC in 1973. By 197 1,150 species of vascular plants had been recorded 8, when the Flora was published, (Dodson & Gentry, 1978). As of 1988, 1,294 species are known (Table 5). In 1979, ten 2 x 50-m transects, altogether totaling a 0.1-ha sam- ple. were analyzed for woody plants > 2.5 ст dbh to quantify the structure and floristic composition of the forest (Gentry, 1982a). In 1980 a second 0.1-ha sample of the forest was conducted, this time including all vascular plants in an otherwise similar survey (Gentry & Dodson, а). Ме have also conducted а series of phenological sur- veys of understory plants (Gentry & Emmons, 1987; pers. obs.). provide a baseline for future comparison as well e results of these surveys as a documentation of the relative abundance of particular species. Monthly surveys of the status of rare or more infrequently encountered species have been conducted since 1984 Many Rio Palenque plant species are very rare or locally extinct. One species, Dicliptera dodsonit assh., known only from a single plant at Rio Palenque, disappeared in 1987 and is presumed extinct. At least 24 other species appear to be locally extinct (Table 3). Those species have ap- parently disappeared from RPSC but are known to occur or have occurred at other sites in western Ecuador and include a few widespread species as well as many species endemic to western Ecuador. Volume 78, Number 2 Dodson & Gentry 1991 Biological Extinction in Western Ecuador SANTO DOMINGO Ù LOS COLORADO MT. ILLINIZA 5.704 M © `ОЕВ5ТЕ MT. CARIMUAIRAZO 5.020 M © MT. CHIMBQRAZO 6310 М TUNGURAHUA 5.018 M MT. SANGAY 5230 M e ALAUSI Location of field stations in western Ecuador. FIGURE 7. 290 Annals of Missouri EH Garden Plant species known from a 1 ] di d | at the Rio Palenque Science Center. Species Family Осогеа sodiroana Mez Lauraceae Bunchosia sp. (D6960) Malpighiaceae Trichilia septentrionalis C. DC. Meliaceae Epidendrum jejunum Reichb. f. Orchidaceae Orchidaceae Apeiba membranacea Spruce ex Benth Tiliaceae Six species, mostly trees or epiphytes, are known from a single individual at RPSC. These were prob- ably never more than accidental waifs at RPSC since substantial populations are (or were) found on the nearby Montanas de Па or further south in drier forest (see Table 4). Another 106 species are very rare at Rio Palenque and may be in danger of local extinction. Also worth noting are 67 species now known only from Rio Palenque but with a local population sufficiently large to seem in little danger of extinction so long as the RPSC forest is pre- served. We also have relevant data on some animal groups. Since 1973 (except 1976) the senior au- thor has taken an annual census of the euglossine bee populations at Rio Palenque. Of the potential 44 species present at the site, 32 are attracted to 1969). Five Macphail traps are hung out for six succeeding days (usually in March) of each year, each with a different chem- ical bait, and the drowned bees are identified and counted at the end of each day. Fluctuations in the number and relative abundance of individual chemical baits (Dodson et al., species have occurred from year to year, with one species (Euglossa gibba Dressler) demonstrating much greater population size during two years, but all species were still present in 1988 and in essen- tially the same numbers as in 1973. For butterflies, at least three brassolid species have apparently gone extinct locally but their loss may reflect re- placement of their larval host, the banana plants of the prior plantation by oil palms adjacent to the RPSC forest. In 1971 and 1982, two expeditions to RPSC led by ornithologist Oscar Owre of the University of Miami made baseline collections of birds. At that time the forests of RPSC were still connected to the extensive forests across the river leading up to Centinela on the Montanas de Ila. During the years immediately following, Richard Webster, Dan Cary, and Ken Miyata produced a local bird list of 340 species based on several months of observation. At present 355 bird species have been reported from RPSC. With the exception of three species, e.g., the Harpy Eagle, the Giant Ani, and the Long Tailed Tyrant (all of which were considered to be waifs), the resident species appear still to be present (P. Greenfield & P. Scharf, pers. comm.). THE JAUNECHE FOREST The Jauneche forest consists of 130 ha of trop- ical moist forest owned and operated by the Uni- versity of Guayaquil as a biological research station. It is well protected and represents the last sizable remnant of that forest category in western Ecua- dor. In 1985, when the Flora was published (Dod- son et al., 1985), 728 species of vascular plants were cataloged (Table 5). Transects totaling 0.1 ha were analyzed as at RPSC and compared with those of both RPSC and Capeira (Gentry & Dodson, та). Two principal types of forest exist at Jauneche. The western three-fourths of the reserve are typical upland moist forest, which once covered a sub- stantial portion of western Ecuador. The eastern one-fourth represents a forest type, seasonally in- undated, which was once extensive on the plains inland from Babahoyo and north of Guayaquil. To our knowledge this forest type has been otherwise totally eliminated in western Ecuador by conversion to rice production during the wet season, and corn or soybeans during the dry period. Several species reported by the Ruiz and Pavon expedition (prob- ably collected by Tafalla between 1807 and 1812) TABLE 5. Comparison of the status of vascular plant species at four sites in western Ecuador. RPSC Jauneche Capeira Centinela Vascular plant species reported 1,294 128 772 1,164* Cultivated plant species 215 136 141 177 Native plants 1,079 592 631 987* Species now extirpated at sites 25 ___ l 84* Remaining native plants 1,054 592 630 903* * = Approximate due to conversion of natural forest before completion of survey. Volume 78, Number 2 1991 Dodson & Gentry 291 Biological Extinction in Western Ecuador TABLE 6. Risk situation at four sites in western Ecuador. Species at extreme risk due to small populations: RPSC Јаџпесће Capeira Centinela Endemic to site 18 5 0 +84 Also found at other sites in region 49 12 18 17 Species of broad distribution 39 33 63 ? Species with short-term sustainable populations at site 942 542 559 ? Species at risk on a long-term basis 312 73 187 ? 670 469 372 300 Species not at risk, primarily due to wide distribution had never been recollected until collections for the Flora of Jauneche were made from 1975 to 1983. Species such as Turnera hindsiana Benth., Ficus cf. trigonata L. and Pouteria sp. (D8759) appear to be endemic to this forest and are at extreme risk. Similarly, several species in the upland moist forest of Jauneche are presently unknown from other sites. Four of these are newly discovered species that have never been collected outside Jau- neche: Duguetia peruviana R. E. Fries and Ап- nona hystricoides А. Gentry (both Annonaceae), Inga jaunechensis A. Gentry (Leguminosae), and Aspidospermum јаипесћепзе А. Gentry (Apocy- naceae). Others, such as Erythrochiton carinatus Little (Rutaceae), Erythroxylum patens Ruiz ex O. Schulz (Erythroxylaceae), and Aspidospermum myristicifolium (Markgraf) Woodson (Apocyna- ceae), are local endemics in western Ecuador and probably now only occur at Jauneche. Most of these species, with the exception of Annona hystri- coides, seem to have substantial populations at Jauneche and are not at immediate peril as long as the forest remains intact. Fewer data are available than at RPSC but Ta- bles 5 and 6 provide information on the status of the flora. CAPEIRA The forest reserve at Capeira consists of 10 ha of somewhat disturbed tropical dry forest owned at present by the Capeira Corporation but being operated by the Fundación Natura office in Gua- yaquil. It is anticipated that permanent protection by transfer of title to Fundación Natura will be accomplished soon. Transects surveyed at Capeira as part of the Flora preparation and published in the Flora of Jauneche (1986) make comparison with RPSC possible (Table 5). The area surveyed is on the slopes of low hills (crest at 220 m) in tropical dry forest with possible premontane dry forest elements. e most striking feature of the dry forest near Guayaquil is its explosive exuberance of vegetation at the beginning of the rainy season in late De- cember or early January. During the dry period, from May through the end of December, the veg- etation consists of deciduous trees and brown dried leaves of the herbaceous plants and vines. Massive flowering of trees and perennial vines occurs in December and early January, while annuals and shrubs tend to flower from March through April. Sporadic flowering occurs during the dry season. Selective logging of highly valuable hardwood trees such as Tabebuia сћузатћа (Jacq.) Nichols., Tab- ebuia billbergii (Bur. & Schum.) Standley (both Bignoniaceae), Alseis eggersii Standley (Rubi- aceae), and Myroxylon balsamum (Leguminosae) has put those trees at risk locally, but their broad distributions throughout the dry forest remnants of the coast suggest that they may be salvageable. Only one new species, Prockia pentamera A. Gen- try, represented by a single individual (now de- ceased), was discovered at Capeira; this species has since been discovered in the Cerros de Amotope on the Peru-Ecuador border (Gentry, 1988). THE FORESTS OF CENTINELA Centinela was a florula site consisting of about 100 ha, located along the ridge line at the southern extreme of the Montanas de Ila, about 12 km directly east of RPSC. Most of the site is at ele- vations from 550 to 650 m. The forest cover of the entire length of the Montanas de Ila, from Santo Domingo to Centinela, has now been con- verted to pasture, coffee, or cacao. A few remnants still exist in ravines and areas too steep to cultivate. When collections began at Centinela in 1975 there seemed little reason for concern because the Montanas de Ila were essentially untouched. We were becoming uneasy by 1980 when preparation of the florula began, and by 1985 we were alarmed. By 1988, the site of Centinela had been completely converted from forest to agriculture. The flora of Centinela was extremely diverse, particularly in epiphytes and understory shrubs and herbs (Table 292 Annals of the Missouri Botanical Garden 5). Although approximately 48% of the naturally occurring species at Centinela were shared with RPSC, about 10% of the flora appeared to be strictly endemic to the site (see Gentry, 1986a). Comparison of the floral composition of Centi- nela with that of the area between Lita and Alto Tambo in the province of Esmeraldas near the Colombian border indicates that the forests of Cen- tinela were premontane pluvial on the Holdridge system. Approximately 80% of the species are shared with Lita, and some species extremely rare at Centinela are common at Lita. We had assumed that most of the nearly 100 species endemic to Centinela became extinct when the forest was cut. This conclusion was supported by failure to locate these species in similar habitats north of Santo Domingo. However, in only five relatively short visits to Lita we have now found 11 species that we had considered to be extinct at Centinela (plus several additional new species). Another site at Poza Honda near La Mana in the province of Cotopaxi east of Quevedo, about 70 km due south by air from Centinela, has been found to have six species of those previously considered to be extinct at Centinela. Probably others remain to be found, since we have only made one visit; however, this site is also threatened, so their discovery at Poza Honda is likely only a short reprieve. Nearly all of the refound Centinela plants were epiphytes or understory herbs. On the other hand, further study of the Centinela herbarium collections continues to add additional new, perhaps already extinct, local endemics that were not included on our original ist, more or less balancing the “‘lost’” endemics. Many epiphytic plants and understory herbs at Centinela have survived in the cultivated orchards that replaced the forest. We first encountered this phenomenon at Jauneche where 63 species of epi- phytes were found. Transect studies in the forest indicated population density to be very low (Gentry & Dodson, 1987b), orchard most of the same species (and seven species whereas in a nearby cacao that were not encountered in the forest) were nearly an order of magnitude more dense (Dodson, un- published). At Centinela we now find many of the extremely threatened epiphytes forming substantial populations in cacao trees and shade trees for cof- fee. Even Epidendrum ilense Dodson, which had been given up as extinct in the wild (Dodson, 1981), has been recently encountered as a substantial population in a 13-year-old cacao plantation at Centinela. In summary, we do not know how many endemic species have already gone extinct on Centinela; all are in dire peril. Although all or most of the trees that were endemic to Centinela may already be extinct—e.g., Carapa megistocarpa А. Gentry (Мећасеае), Brownea n. sp. (06846) (Legumi- nosae), and the shrubby tree Erythrochiton gi- ganteus Kaastra & A. Gentry (Rutaceae)— of the herbs and epiphytes have thus far survived. many Discovery of several of the Centinela “endemics” at two additional recently explored sites would somewhat lower our estimate of the original level of endemism on the ridge, but several additional new species have also been recognized. The only solid conclusion may be that we still know so ex- tremely little about the plants of western Ecuador and their distributions that we cannot hope to make educated conservational decisions without a far more extensive floristic data base. NATURAL AND MAN-MADE FRAGMENTATION Western Ecuador has an extremely diverse flo- ra. Аз indicated above, we estimate that there are more than 6,300 species in an area of 80,000 m?. A substantial portion of the diversity found in western Ecuador is probably part of the general syndrome of explosive speciation in the northern Andes (Gentry, 1982a; Gentry & Dodson, 1987b). Тће causal mechanisms for that explosive specia- tion are not well understood, but surely one of the factors is the plethora of microhabitats in the re- gion. Not only are there a large number of micro- habitats in western Ecuador, but they are spatially isolated due to the natural fragmentation of man of the life zones. Of the 12 or 13 life zones, eight to ten often were represented, even originally, by small to very small, largely noncontiguous patches of forest strung out in a linear fashion either along the tops of the coastal mountains or the lower slopes of the Andes. It is in those natural fragments that speciation was presumably most rapid, judging by today's exceptionally high endemism. The forest fragments that now remain after hu- mans have essentially obliterated the forest over the majority of western Ecuador tend to be in the most inaccessible areas. They often are located near the summits of the coastal mountains or along the steep flanks of the Andes. These tend to be humid- to wet-forest fragments with high species diversity and high natural endemism, including some of the above. *clandestine" forest reserves mentioned We suspect that much of the flora of western Ecuador is based on a different set of evolutionary parameters from that of l homogeneous tropical zone hábitats. These include anses of relatively extremely rapid speciation, adaptation to survival with very low population densities, adaptation to Volume 78, Number 2 1991 Dodson & Gentry 293 Biological Extinction in Western Ecuador small habitat patches, and perhaps concurrent rap- id natural extinction rates. Thus, we would submit that conservation of even the tiny habitat fragments that remain in western Ecuador might well con- serve a substantial portion of the estimated original complement of 1,260 species endemic to the area, despite the tremendous deforestation that has taken place. Any such effort must be undertaken very soon, however, since no matter how adept at sur- viving in habitat fragments many of these species may be, they will surely all be lost if the few tiny forest patches that remain today are lost. SYNOPSIS OF DATA AND DISCUSSION 1. The forest cover of a significant but inde- terminate part of western Ecuador had already been somewhat altered prior to the time of the Spanish conquest, perhaps as much as 30%. The catastrophic effect of European diseases apparently reduced the indigenous population drastically, and much natural reforestation occurred before the be- ginning of the twentieth century. 2. At the end of the nineteenth century most of western Ecuador was covered by intact forest (at least 75%) except for regions easily accessible by boat along the rivers. 3. Most forest conversion in western Ecuador occurred between 1960 and 1980, principally as a result of road construction and demographic ex- plosion; today, approximately 4.4% of the aborig- inal primary forest cover remains. The moist- and wet-forest categories were most severely reduced, from 50% to less than 1% and from 15% to 0.1%, p . The population of Ecuador is increasing at an prada rate that shows little indication of slowing. The present population is more than twice that of 1957 and doubling of the 1988 (10.2 million) population is expected by the year 2020. e need for resources to support that population will make for even greater pressure for conversion of natural forests. 5. Western Ecuador was very diverse ecolog- ically with 12 or 13 Holdridge life zones in 80,000 km”. The forests of most of those life zones were distributed in long narrow bands or scattered through pe coastal mountains. . Considerable local endemism occurred in each k. zone, at least partially due to their long, narrow forms and natural fragmentation. 7. Extrapolation from florula data indicates that the flora of lowland western Ecuador probably con- tained about 6,300 species of vascular plants. We estimate that about 1,260 of these are endemic to this region. 8. Island biogeographical theory suggests that the drastic eradication of habitat during the last two decades should have resulted in substantial extinction by now, but the baseline floristic data are insufficient to know whether this has happened. . Florula data suggest that very many species, about 4% of the florula floras, are at extreme risk at this moment. We estimate that perhaps 12% of the flora of western Ecuador is at risk in the near future. Many species have very precariously sur- vived partially because of a network of minute patches of private, in part clandestine, forest re- serves. 10. With the exception of the two threatened forest reserves at the northern extreme (only 80 km? of which lies within western Ecuador as here defined), very little natural forest remains in west- ern Ecuador— са. 4,000 km? or 4.4% of the ађ- original total of 77% of western Ecuador. ederal protection of established national parks and reserves is inadequate. 12. Demographic pressures and political reali- ties project a very grim future for effective con- servation of what remains unless immediate steps are taken. 13. Unless adequate protection is given to the remaining forest fragments, including the tiny pri- vate reserves in the south as well as the two larger ones in the north, a major extinction of perhaps 1,260 endemic plant species can be expected in western Ecuador in the very near future. 14. Unlike Amazonia, the soils of much of west- ern Ecuador tend to be rich and if well managed will sustain agriculture for an indefinite period; therefore, conservational decisions must be based on the comparative values of conservation of for- ests versus the land's potential value when placed in agricultural production. WHAT CAN ВЕ DONE? Though the conservation scenario in western Ecuador appears to be very grim, we feel that several actions could be taken to save a portion of the remaining diversity of plants and animals. As indicated above, western Ecuador is very diverse biotically with 12 to 13 Holdridge life zones. Small portions of each of those life zones still exist as private reserves, clandestine forests, or federal re- serves. We suggest that if even the tiny extant fragments can be effectively preserved there may yet be reason for optimism of preserving a portion of western Ecuador's ca. 1,260 endemic species, especially in view of apparent adaptation of many of these endemic species to an originally small and fragmented habitat. 294 Annals of the Missouri Botanical Garden ACTIONS THAT MIGHT HELP TO SAVE THE REMAINING DIVERSITY: a. Strengthening of federal efforts through private institutions. Governmental efforts to preserve forests that have been declared to be reserves or national parks have had a dismal record of success in Ecuador except in the Galápagos Islands. Until recent times there has been little external or internal interest in conservation in mainland Ecuador. The development of private, conservation-ori- ented Ecuadorian foundations within the last de- cade, such as Fundacion Natura, dedicated to con- servation principles and with surprising internal support, holds hope for a major change in direction. Those foundations have been spectacularly suc- cessful in modifying popular opinion in the country and are making strides in changing attitudes of governmental agencies. А program of support of natural history education at the elementary school level, daily articles in the newspapers, establish- ment of the Pasachoa Nature Reserve by Funda- ción Natura, and ecological activism on all fronts may play an important role in changing conser- vation practices in the country. b. International pressure to save the Awa and Cotacachi-Cayapas reserves. If the Awá and Cotacachi-Cayapas reserves are to be saved for posterity a major effort must occur soon. Plans exist to construct roads that will open the two areas to colonization. А substantial portion of the biological diversity in western Ecuador still persists within those two preserves. They are still sufficient size that all resident flora and fauna can persist. However, without those preserves mas- sive extinction will surely occur. International attention to the plight of the two reserves must be brought to bear now. Though the Galápagos are world renowned, have many endem- ics, and spectacular tourism impact, the two re- serves have much greater diversity, far more en- demics, and extreme potential impact from extinction. The two reserves have an astounding diversity and much of it is endemic; the 1, endemic species of western Ecuador are nearly twice the total flora of the Galapagos. Even the forests of the 1 km? Rio Palenque Science Center have nearly twice as many species of vascular plants as all of the Galápagos Islands, and the nearly 100 endemic species of the 20 km* Cen- tinela Ridge are nearly half the total of endemics that occur on the Galápagos. Survey of remnants and development of small nurseries at established sites for plant species at extreme risk. It is important that survey and identification of forest remnants and fragments be accomplished as soon as possible. This can be done by use of satellite imagery and overflights. Once located, the frag- ments must be explored to determine preserva- tional status of the forest and the presence of endemics. When the endemics are discovered, their population status must be ascertained. The impor- tant forest fragments should be protected as forest reserves. Species at extreme risk could be propa- gated in small, relatively unsophisticated nurseries in similar habitats and reintroduced as necessary. In addition, seeds can be collected for preservation of genotypes in seed banks. Nurseries can be developed at already estab- lished sites with a minimum of expense. For ex- ample, nearly all of the dry-forest species could be maintained at Capeira with minimum care. Moist- forest species can be grown at Jauneche, and wet- or pluvial-forest species at RPS UPDATE ON STATUS OF FORESTS OF WESTERN ECUADOR (10 Feb. 1991) n 1990, Luis Mejia of the Ministry of Agriculture panying that report is a detailed map showing forest types and extent. Mejia ти that we yer not correct in our assessment ч t were even overly ly optimistic. Mejia's study area was much larger, including all of the western slope of the FR and totaling 102,000 km’, whereas we only used up to the 900-m line, and included 80,000 km. Mejia indicates 15.4% of western Ecuador still in primary forest. The fundamental difference in the two studies is that the northern portion of the area above the 900-m level contains the major portion of the 3,000 a reserves, 8 B = 5 EH E- & = = e р 5 а. 5 e = = = aq 2 5 о Е. |=] о a & 5 я sequently, even less than the 4.4% of lowland western Ecuador, still in primary forest predicted by us, existed block lies just west of Volcán Corazón at elevations be- tween 2,000 and 3,000 m. No others occur in the lower two-thirds of western Ecuador. Mejia also ое that of the 102,000 km? of western Ecuador, 15,7 km? of primary — exist, but only 22, м 0 km? ho forest of any so uring the 2 to 3 years since the Landsat photos were taken deforestation чы ассејега!ед. Only the northern reserves and the Chindul mountains in Volume 78, Number 2 1991 Dodson & Gent Biological Extinction in Western Ecuador extreme western Manabi and Esmeraldas provinces re- main forested, and even there the chain saws drown out all other sound from dawn to dusk all of western Ecuador is considered (as did Mejia) using the same criteria (based on the relatively well known orchid data western Ecuador, including more than 3,000 endemic species. LITERATURE CITED ALVAREZ, A., В. VASCONEZ & L. GUERRERO. 1984. Mul- ti- -temporal study of mangrove, shrimp farm and salt Island/ Agency е International Development, Coastal Resources Projec ANONYMOUS. 1980. Information Handbook of the Re- public of Ecuador, Volumes 1-3. E. Romero Palacio (editor), Cientifica Latina Editores, Cuenca, Ecuador ANONYMOUS. 1987. О i la Red Vial Nacional. Ministerio de Obras Public Вокм5-Ватосн, P. & Р. еа lution апа phylogeny in the tribe Thelymitreae hl о Мо еб Pl. Syst. Evol. 159: 1988. Floral evo- Шо, AS, L. EsTRADA. 1978. Ecuador Map Ecológico. PRONAREC- Ecuador, Ministerio de Hen cultura y Ganaderia. Рорзом, C. Н. 1981. Epidendrum ilense —the saving of a truly endangered species. American Orchid Soc. Bull. 50: 1083-1085. . H. GENTRY. 1978. Flora of Rio Palenque. Selbyana 4:1-628. — — — &. А. Н. Gentry. Flora of ios (in Wisi I . M. VALVERDE. 1985. Flora of Jauneche. Selbyana 8: x о В. DRESSLER, Н. Hirs, К. Арлмѕ & М. WIL- s. n Bu. active compounds in or- chid fragrances. Science 164: 1243-12 GENTRY, A. H. 1979. Exti nction а conservation of plant species in tropical Amer a phytogeograph- ical perspective. Pp. 100-126 ink Hedberg (editor), Systematic Botany, Plant Utilization, and nm Conservation. Almqvist & Wiksell, Stockholm 1982a. pege н neotropical plant species diversity Evol. Biol. -84. 1 ее patterns in north- west South America and southern Cen E America as evidence for a Chocó Refugium. Pp. 112-136 гл G. Prance (editor), Biological улей ка in the m Colombia Univ. Press, New York. 198 Neotropical floristic dives: uns Connections America, Pleistocene climatic fluctuations, or an ac- cident of the Ande y orogeny. Ann. Missouri Bot. Gard. 69: 556-59 a. En PE in tropical versus temper- ate plant communities. Pp. 153-181 in M. E. Soulé (editor), Conservation Biology. Sinauer Associates, Sunderland, Massachusetts. Species richness and floristic com- position of the Choco region plant communities. Cal- dasia 15: 71-79. 1988. New species and a new combination for plants from trans- haee South America. Ann. Missouri Bot. Gard. 75: 1429-1439. 1 Northwest “зел Peng (Colombia, Bousdor and Peru). P 91- n D. Campbell H. Hammond (editors) наи eb of Trop- . Ga ical Forests. New York Bot гд., Bronx, New York. & С. Н. Dopson. 1987a. Contribution of non- trees to Qo uy wr ow of tropical rain forest. Bio- tropica 19: 149- P Diversity and biogeogra- phy of ir gage vascular epiphytes. Ann. Missouri Bot. Gard. 74: 205-233. & “E MONS. 1987. Geographical variation in fertility and composition of the understory of neo- tropical forest. Biotropica 19: 216-227 HARLING, G. 1986. Current Scandinavian Botanical Re- search in Ecuador. Rep. Botanical Institute, Univ. Aarhus 15: 9-10, HASSAUREK, F. 1967. Four Years Among the Ecuador- ians (1861-1866). Southern Illinois Univ., Carbon- dale & Edwardsville. [C. H. Gardiner (editor).] МА L. Life Zone Ecology. Tropical Sci- nce Center, San Jose, Costa Rica. Ks. N. J. 86. The systematic e of vas- cular epiphytes: an update. Selbyan MACARTHUR, R. & E. O. WILSON. 1967. The Theory of ашан Biogeography. Princeton Univ. Press, New Jer MEJIA, L E. (in press for publication in 1991). Deter- minación de Areas de Bosques Remanentes en la Región Occidental Ecuatoriana. Fundación Natura, bom Ecuador. Pp. 1-62 with maps and appen dic нес `N. Mass extinction of species: a great creative challenge. е! Lecture in Conservation. iar California 87. The extinction spasm ПИ syn- ed at work. Conservation Biology 1: 14-21. 1988. Threatened biotas: “hotspots” in trop- ical ‘forest. Environmentalist 8: NEILL, D. Oil wells, Indians e E E of upper Amazon. Missouri Bot. Gard. Bull. 76: аса D.S. 1986. Are we on the verge of a mass al rain forests? Pp. 165-180 in liot (editor), Dynamics of Extinction. Wiley 1976. Island verd theory d conservation practice. Science 191: 285-286. Skoc, L. 1978. Flora of Panama, ann Ann. Missouri Bot. bo 65: 783-998. SvENSON, Н. К. 46. Vegetation of the coast of Ec- uador and Peru and its relation to the Galapagos Islands. Amer. J. Bot. 33: 394-498. TERBORG, J. & B. WINTER. 1982. Evolutionary cir- cumstances of species with small ranges. Pp. 587- 600 in G. Prance (editor), Biological Diversification ress, New York. 1979. Cienc. z < Publ. > Н. 1918. The vegetation of the Ga- lapagos Islands, Thesis, State Univ. Utrecht, Neth- erlands. 1983. Species number, area, and habitat di- ade in the Galapagos Islands. Vegetatio 54: 167- 175. STUDIES OF ANGIOSPERM PHYLOGENY USING PROTEIN SEQUENCES: P. G. Martin? and J. M. Dowd? ABSTRACT In previous papers we have reported the N-terminal 40 amino acids of the small subunit of rubisco for samples from four families of gymnosperms, nine families of monocotyledons, and 26 families of dicotyledons. We expanded this list to 122 families of dicots and derived a phylogenetic tree for all 335 species. The main computing program used was HENNIG86, with which a reliable result can be assured with only 17 subsequent testing using the longer sequences from nucleic acids. u species, the nature of the N-terminus of rubisco-SSU, and evidence that natural selection is powerful taxa or less, so a major part of this Topics discussed include amino acid sequence. The rate of evolution has been shown to vary between major taxa, and data suggest that angiosperms originated in the Jurassic. The problems of angiosperm phylogeny are well illustrated by a consideration of the differences between four classifications, all less than a decade old and all by highly respected and experienced authors. The dicotyledons are divided into six sub- classes by Cronquist (1981) and seven by Takh- tajan (1983), while, for the other two authors, the major groupings are superorders, Thorne (1983) having 19 and Dahlgren (1983) 25. The number of dicotyledonous orders recognized is, respective- ly, 58, 72, 41, and 83; these figures alone indicate the resulting diversities of names and content, all of which reflect our comparative ignorance of the course that evolution has taken in the angiosperms. In contrast to this, at the next level down the hierarchy, there is basic agreement about the “core” families to be recognized (Heywood, 1978). Macromolecular sequences provide taxonomic characters whose homology over widely diverse species can be assumed with some confidence. Se- quence data can be analyzed objectively with com- puters. We will probably see in the next decade the publication of nucleic acid sequences long and variable enough to solve some of the problems of angiosperm phylogeny (e.g., Palmer et al., 1988; Zimmer et al., 1989). It is therefore an appropriate time, when nucleic acid sequencing is supplanting protein sequencing, to set out the results of a de- cade of work that has produced 335 partial protein sequences from a wide range of angiosperms. These sequences are shorter than nucleic acid sequences already published and therefore contain less infor- mation and are less able to resolve the sequential divergences of early radiations. Nevertheless, we believe that our phylogenetic trees will indicate likely relationships and profitable working hypoth- eses for future investigations. A SUMMARY OF PUBLISHED INVESTIGATIONS USING PROTEIN SEQUENCES The pioneer of the use in botany of protein sequences for investigating plant phylogeny was D. Boulter of the University of Durham, England. During the 1970s, Boulter, along with his col- leagues and students, published 25 sequences of cytochrome c, 12 complete and 58 partial se- quences of plastocyanin and seven sequences of Gardens (Canberra). We are indebted to Richard Norris order. As indicated in Table 2, we have obtained leaves from man sources, to all of ful; we are » especially grateful to the Adelaide Botanic Gardens, the Missouri Botanical Garden, the National Botanic Gardens (Canberra), and the Royal Botanic Gardens, Kew * Department of Botany, University of Adelaide, Box 498, G.P.O., Adelaide, South Australia 5001. ANN. Missouni Bor. Garp. 78: 296-337. 1991. Volume 78, Number 2 1991 Martin & Dowd 297 Em Phylogeny Using Protein Sequen ferrodoxin. These have been collated, with ref- erences, by Ramshaw (1982), and Scogin (1981) has reviewed the results from the taxonomic point of view. Although this work generated much in- terest, it also gave rise to skepticism, some of which can, with hindsight, be attributed to the inadequa- cies of computing methods that were being devel- oped concurrently. The mostly unfavorable reac- tion of systematists, epitomized by the review of Cronquist (1976), influenced the cessation of re- search in Boulter's laboratory about 1980. Before this, however, partial sequences (up to 25 N-terminal amino acids) of the small subunit of ribulose-1,5-bisphosphate carboxylase/oxygenase (rubisco-SSU) were obtained from six species (Has- lett et al., 1976; Strobaek et al., 1976). This work led to a complete SSU sequence from spinach (Mar- tin, 1979), a forerunner of the work presented here which concerns the N-terminal 40 amino acids of this protein. (The complete sequencing of a protein requires prior purification of several frag- ments and is at least an order of magnitude more time-consuming than the direct sequencing of the N-terminus of the whole protein using an automatic sequencer.) Nucleotide sequences of rubisco-SSU from a few species have been published, and all of them have been studied using our method. The only new data comparable to our 334 species are from two closely related orchids and their hybrid (G. C. Martin et al., 1987). We are unaware of phylogenetically useful sequences of other proteins since those of Grund et al. (1981) and Nakano et al. (1981). Work in our laboratory has proceeded in five phases. In phase 1 species were chosen because Boulter had already published their complete se- quences of cytochrome c and partial sequences of plastocyanin. When a pattern failed to emerge from analyses of these data, we decided to sample each family with sequences from at least two more rep- resentative genera. Thus, the families Apiaceae, Asteraceae, Brassicaceae, Caprifoliaceae, Cheno- podiaceae, Fabaceae, Malvaceae, Poaceae, Polyg- onaceae, Ranunculaceae, and Solanaceae have each been sampled at least three times. These early results were published in a series of papers (Martin et al., 1983; Martin & Dowd, 1984a, b, c). The sequences for rubisco-SSU, cytochrome c, and plastocyanin were analyzed for these families by Martin, Boulter, and Penny (1985) using de- rived estimates of familial node sequences. Anal- yses of data from single macromolecules were not consistent with one another but, for nine of the families, a phylogenetic tree derived from combined data remained consistent when ferrodoxin or 55- ribosomal RNA (available for some of the families) was added. This result indicated the need for longer se- quences and better sampling of families. Although rubisco-SSU was always multiply represented, in 17 of the 33 samples of other macromolecules there was only a single sequence. This situation is precarious because, if the average distance from a familial node to a species is N, then on the average a single sequence will misrepresent the familial node by N. This source of error might be responsible for part of the poor agreement observed. Sampling a family at least twice, preferably from widely divergent representatives should give a better es- timate of the familial node (see phase 5). In phase 2 we sequenced rubisco-SSU from 11 members of Onagraceae (Martin & Dowd, 1986a), 15 monocotyledons (Martin & Dowd, 1986b), and 14 species of Solanum (Martin et al., 1986). We reasoned that the reliability of our methods might be estimated by comparison with taxonomically well understood groups. The results were similar to oth- er taxonomic treatments. Additional species of As- teraceae were also studied and those results will be presented in this paper. To estimate the rate of evolution, Proteaceae, Solanaceae, Fagaceae, and Winteraceae were sam- pled in phase 3 using species whose ancestors are thought to have been separated by continental drift at known times. This led to a preliminary publi- cation (Martin & Dowd, 1984b), and the derivation of a molecular evolutionary clock (Martin & Dowd, 1988), which indicated that on average one nu- cleotide difference arose between two diverging lines once in seven million years. In phase 4 we tested the hypothesis that leghe- moglobin had evolved in plants by lateral transfer from animals. This led to an investigation of all species for which leghemoglobin sequences had been published, and it was shown that the pathway of evolution in those species was closely parallel in hemoglobin and rubisco-SSU (Martin & Dowd, 1986c), suggesting that there was no need to in- voke novel evolutionary processes. А consequence of this study was that we increased the number of species of Fabaceae sequenced to eight (see Group 4 below) and obtained sequences from several additional families. Many of these were too small to be studied in the normal course of this investi- gation but were obtained either because they are known to include nitrogen-fixers or thought to be relatives of the legumes; these include Betulaceae, Chrysobalanaceae, Сопапасеае, Crossosomataceae, Datiscaceae, Elaeagnaceae, Casuarinaceae, Moringaceae, and Myricaceae. 298 Annals of th Missouri Bm Garden In phase 5 we surveyed the dicotyledons which increased the number of families studied from 24 124 А SURVEY OF THE DICOTYLEDONS There are about 250 families of dicots. Because it was impractical to sample all of them, a decision was made to sample about half, i.e., the number from the 24 mentioned above to 124. Three families (Acanthaceae, Loranthaceae, San- talaceae) failed for reasons that will be discussed later. The additional 97 families were chosen pri- marily on the basis of size. The majority of families sampled have more than 20 genera. To cover as wide a range of variation as possible, some small families were also sampled. For example, the order Illiciales has only three genera, so the family Schi- to increase sandraceae (two genera) was chosen to represent it. Only three orders are unrepresented out of Thorne's 41 (two of which are parasitic and devoid of rubisco), 10 out of Cronquist's 58, 19 out of Takhtajan's 72, and 21 out of Dahlgren's 85. It is impractical, mainly because computers are limited in their capacities to analyze large numbers of taxa simultaneously, to contemplate building a phylogenetic tree for 122 families (comprising 310 species) without some subdivision into groups. We have done this by referring to all four current phylogenies. Thorne (1983) and Dahlgren (1983) have superorders as their major groups, the former nominating 19 and the latter 25. If th authors agree that families are in the same super- order then they have been grouped together in our scheme, with one proviso. Takhtajan (1983) and Cronquist (1981) have respectively seven and six subclasses as their major groups, and these two authors have been allowed a veto; if either if them does not also agree that families are in the same subfamily, then they are left ungrouped. In this way we have divided 102 of the studied families into 25 Groups, leaving 20 ungrouped because ese two there is disagreement. We are reluctant to use a formal term like superorder but need to make it clear that our use of Group does have a defined meaning, so we have used a capital С. The Groups are shown in Table 1. It was practicable to sample each new family only twice, and we have done this by choosing two species not only from different genera but, if pos- sible, from different subfamilies or tribes. Some- times this criterion has broken down because fresh leaves have not been available. In Table 2 the 335 species for which sequences are available are arranged by families and Groups, and their sources and sequences are given. BIOCHEMICAL METHODS The methods published by Martin and Jennings (1983) have stood the test of time, so, rather than repeat them here, a general description will be given and the few modifications mentioned. Two methods were described, one for “pungent” leaves with high concentrations of phenolics or other substances that make protein purification difficult, the other for “bland” species whose leaves are much more amenable. The bland method gives better quality protein and is therefore to be pre- ferred. However, because the pungent method works well with bland leaves, but not vice versa, it was preferred when there was doubt or too few leaves for trial extractions. Both procedures started with maceration of about 100 g of leaves from which the midribs were re- moved if practicable. For bland leaves the extract- ing buffer was essentially a reducing, saline tris- НСІ buffer at pH 7.4, while for pungent leaves a reducing, saline borate buffer at pH 8.6 and con- taining the detergent Triton X-100 was used. After crude straining and centrifugation to remove solids, the extract was passed through a succession of two liquid gel columns. A Sephadex G-25 column was used first to remove low molecular weight sub- stances. А Sepharose 6B column was used to re- move remaining low molecular weight substances and high molecular weight nucleic acids and mem- brane fragments. Eluting buffers were different for the two extraction procedures and for the different columns used. The protein was precipitated with ammonium sulfate for the bland method and with acetone for pungent. Procedures after the second column were the same for both types of leaves. The protein was S-carboxymethylated at pH 8.6 to break disulphide bridges between cysteine res- idues and then passed through a long column of Sephadex 6-100 in an eluting buffer containing sodium dodecyl sulfate. This separated the large subunit from the small subunit, which was precip- itated in acetone and dried before sequencing. (A variation of this procedure was to use a column of G75 followed by G-100.) The methods are rather crude but are successful because rubisco is a very large protein and, by a considerable margin, the most abundant protein in leaves. About 5 mg of small subunit (in 0.5 ml of water without polybrene) was sequenced on the Beckman Volume 78, Number 2 1991 Martin & Dowd 299 Angiosperm Phylogeny Using Protein equences TAB Families of dicotyledons grouped because they are placed in the same major taxon by all of Cronquist бан. ты (1983), Takhtajan (1983), and Thorne (1983). GROUP 1 GRouP 4 GRouP 9 Group 12 Group 17 GRouP 21 Magnoli Ulm Dipterocarp Eric Connar ami Winter Mor Elaeocarp Epacrid Sapind Verben Annon Urtic Tili GROUP 13 Anacardi GROUP 22 Myristic GROUP 5 Sterculi Cunoni Simaroub lan Schisandr Hamamelid Bombac os li Convolvul Monimi Betul alv Saxifrag t Polemoni r g Group 10 GROUP 14 GROUP 18 GROUP 23 Aristoloch Casuarin ol Caesalpini Halorag Scrophulari Calycanth ROUP Flacourti imos Rhizophor Gesneri ROUP Dilleni Datisc Papilioni GRouP 19 Bignoni Berberid hea Cucurbit Group 15 Zygophyll edali anuncul Ochn alic Trap rani Group 24 Lardizabal Clusi Cappar Lyth Tropaeoli Valerian Menisperm GROUP 7 Brassic Myrt Malpighi Caprifol Papaver yric Resed Punic GROUP 20 GROUP 25 GROUP 3 Jugland orin nagr Logani Api Cabomb ROUP 8 GROUP 11 Melastomat Gentian Arali Nymphae Caryophyll Sapot Combret Apocyn Nyctagin Styrac Group 16 Asclepiad Amaranth Primul lac Ole Phytolacc Myrsin Celastr Rubi Chenopodi FAMILIES THAT DO NOT FIT INTO ONE OF THE GROUPS Aster lari deni Nelumbon Polygon Thymelae Bux Crossosomat Hydrophyll iper rote Vit Campanul aeagn ythi Plumbagin Rhamn Chrysobalan Euphorbi Loas Note: *'-aceae" omitted from all names. 890C automatic sequencer using Beckman’s stan- dard quadrol program with 50% quadrol buffer. The phenylthiohydantoin (PTH) derivatives of the amino acids were identified using a Waters HPLC instrument with a C-18 radially compressed column and eluted with 0.1M sodium acetate (pH 6.0) and acetonitrile. This did not distinguish two pairs of amino acids and was therefore supplemented with Using these methods, we could, without assis- tance, produce two proteins each week and se- quence two others. FAILURES Although 90% of attempts led to successful se- quences, the remaining 10% deserve brief atten- tion. Unless there was an identified reason for fail- ure that could be corrected, our policy was to try another representative of the family. Faults that could be corrected include the amounts of extraction and elution buffers used. Some plants gave extracts that were mucilaginous to the point of setting solid. Dilution of the extract corrected this. This problem occurred in Onagra- ceae and a few others with small leaves containing a high proportion of veins. Insolubility of the pro- tein, leading to precipitation in columns, could sometimes be corrected by loading a more dilute extract. Plants with C4 photosynthesis, and rubisco tightly bound in bundle sheaths, were avoided if possible. Plants with C3 photosynthesis often occur in the same genera or families and were unlikely to be phylogenetically biased. However, if unavoid- able (e.g., Welwitschia is reported to be C4), spe- cial care was taken during the maceration process. It is suspected that the most common cause of failure was the presence of powerful proteases in the leaves and, in retrospect, it would have been profitable to try correcting this with research early in the project. Species of Ficus, known to have leaf proteases, showed symptoms of this failure. Large amounts of protein traveled where the small subunit should have been on the G-100 column and gave many amino acids at each position when sequenced. Another casualty of this sort was Gne- tum gnemon, which was particularly desired be- cause it is a gymnosperm thought to be close to Annals of the 300 Missouri Botanical Garden алмохмлллхалиахничтличзаазтлаачт1тлх=5 алмоллтллхалиачичмтлоогвотлаачтлх=5 алмозгхиллхалохмчтлооач5таатлх= алмэзбттлаточтобазтаатл GAMOWNMITATAOWNWVTOASSTAdATAS GAMONTTTATIANTTAGOSTASTAS d4AMS5MNTT1T1ASISMITIS3GSÓSIdST1AS алмокцлтлтлтлхахахуичтлозтлтлаатх=з5 d4AMS9NUI1TIT1A4GIGNVTIOSW/ASAdd'T12AS d4AMSRITITAGAS3NVUIMSUSSTIdd'T1AS @IMOSHITAGCIGNYVYTIASSTAATAS атмэзэттлатзячтчя с таатле атмэзчттлаляячтзбазтаатле алмэзтттлаляучтоз<зазтаатл$ алмэзтттлаляячтоо`ззстаатле алмэмтттлатня 11/10 ятлтаатле$ IAMDATTTAOCIANITOASSTAdTIAS алмохулллхалачиттлочлалачачтх= атмохлллхаличичтлаозгзотлаатх= GAMONTITITAGIAYVYTOTALTLIATIAS d4lMO9SIITASASNVTS3S3SS'ILd'T1A4S GIMONTTITTAGTAGAYVYTOAISSTSATAS ттэхаамл ттэхаамл ттэьхаамл 3RHNAddMA NHNAddMA хѕотаамл хукотаамл х 153/Ad dMA X T9I/AL/dd M A *eursoiÁ3 'Х ¿ueydo3dA13 ‘м “SUTTRA ‘л ¿euTUOSIYA ‘Г /euTies { эитиотчзеш тетлаеприт pue OTFTAUSTOS чатеомџошшод='0'М'1'5'0 'pue[?ez MON="Z'N грџетеџеопдертд :Алозтллед UI9YITON="L*N {вөтем ҷапоѕ MeNe'M'S'N ‘Ss zx W *5'g/'qunuj eor3seuop PUTPUEN *£*g эртетэруч/твеэла вт=чэизее stieqieg uud avaovaludaudsid z апочо zz *5'g PIQUE) W /чэтшс *2'V(AxexotA'M' f£) susoseandind etuueusey я -Z'N чалхом/Арчед (тпоеч) езехотоо елэзитморпеза я о" M9N/197S104 тлезитм SÁWTIC NIM IVIDVIAALN IM я о" мод/човттм 3 19PY9Y PIOTITIQNI PIPUPBSTUOS W *9'g KAeqexieg/ieung eoruode( einspey sos SvdovVdaNVSIHOS Я "оў eQiodeburg/*33noH sueibeij еотавтлАн ^| БЫ: ai етоттезет ешәчу W '$'g эртет us TTON snpToq snuned '9'4 И Ee end обтј еттечоти 2 а/'род-'тпоб еиетбиетпоє еттоџмбеџ 'S'g/''1 unzegrdr[n3 чолрџеротлтт ЭМИ IVIDVUITONIVH :*9*4/*TTTH ечеэттеше eosiod *5'g эртетеру/*т STTITqou вплпет AWT uvaovunvI pPIÓ'N/exetH'L'S(sterd) SSUSTTRIISNO umuiedsoTpI '5*g epreiepV/'xXurT('T) xooeead snuqueuouru) *5'g ертетерџ/'1т зпртлотј snuaueoA[eo Two IVIOVHINVOXTIVO *'£'g TINOSSTH/'T esuepeueo UNIPSY *£'g SPTPTSPY/* т езтзэееа PTYDOTOISTIV SUY IVIDWIHOOTOLSIYVV PTO'N/YaTUS*S*T(*TTOSNH TA) тәелооеш етаотАх PTO'N/dosser (· ттеги' 3) snueezeobD souseg NNV AVIDUNONNY zz BEA zx БЕБЕ ZZ 'SNOQGAITALOOD геџтитбле ‘¥ ;еџмтшезптб “б ieurloid ‘а :еџтбеледве ‘М ‘я ‘эитопет ^1 ‘эитвАт ‘y feutonetost ‘т :еџтодтб ‘5 Puls м “д ‘ртое отшезптб ‘я {ртов отзледве ‘а :eurueTe АЧ :=ртое ouTwe ор [euruiea-N оз дәу (€) *pees ротјт3192='5'0 “UOTIPSTUBDID uoreeseu "етрехаепџ uiejseM-"Y'M 'етлозотле'отл еттелзвпү YANOS="W"S :веодеје ueriezasny • (в)џерлеђ (ie)orue3ogae'9'gd :(редеје еетмлецзо євәтип эртетеру UT) USPIPY эзелтха=‘9‘а :ueuroeds əy} jo eoinos ey3 Aq ремоттој ST ҷѕетѕ е 'хоцчапе pue эшеч setoeds 943 19334 (2) 1933197-29143 943 әлтб әм эшеи Аттшеј yore 19339 ° (2861) едем jo чотзетлелаае Т этаег pue зхәз ey ит реџтетдхе se ваполо Aq pebueire әле виорэтАзоэтр UL (T) :8930N 'seouenbes pue perpnis saradg "с IYL Volume 78, Number 2 1991 Martin & Dowd Angiosperm Phylogeny Using Protein Sequences 301 YLLSSGWVP YLINNGWIP YLLSNGWVP LLKGWIP GWVP VWPPIGLKKFETLSYLPPLTQEQLAKEVE VWPPVNLKKFETLSYLPPLSQE QLLKEVOD Q MKVWPPVGLKKFETLSYLPPLSQEALAKEID MKVWPPLGLKKFETLSYLPPLSP QALAKEVDYLLLNGWIP MKVWPP/VLGLKKFETLSYLPPLSPQALAKEVD YL MOVWPPTGLKKFETLSYLPPLTAE QLAKEVDYLLKN мо о m . + i o o o m omn m Ü Ф а Ая 3 Ф с Ф ны і Mo] „Я о @ ч. 4 4 ~ зоны СУ Ф =“ с~ = © om - Ф < "о > *- N зо < . ~ 90 М. = E . ~ . 0 [T о - G сы ас o ам ~ ~ a Qoo Qe o А = 5 ба E а ГИ ЕСЕ а абы ES R=] Om & <= м = ша H n ~ = "700 зс @ S |Ева 5 ? T ае Вая А Eo RR N d ч Poh) s КЕ Oa ы - 53 Q n ya 5 о = с CE 3 a 4 са цао носи о = | да4ч9 2022 7 аза eae залива а Ц V V M M Fumaria officinalis L./P.G. NGWVP LLANGWVP MKVWPPLGLLKFETLSYLPPLSQQALAKEVDY MKVWOPYNNKKFETLSYLPELTPESLAKME ТЕ Aquilegia vulgaris L./P.G. G. Clematis rehderiana Craib/Adelaide B DYLIRKGWVP VWPPLNNLKFETLSYLPPLSQASLAKEVDYLI VKGWIP MKVWNPVGNQKFETLSYLPPLSTESLAKEVE YLIRKGWVP MKVWOPYNNKKFETLSYLPELTES QILKEI MK Helleborus foetidus L./Adelaide B.G. Ranunculus trilobus Desf./P.G. GROUP 3 CAB Brasenia schreberi Gmelin/Nagambie Vic. CABOMBACEAE MAVWPIYN/VLKKFETLSYLPSLSKE QLAKEVEYLLIKKWIP NYM Nuphar advena W.T.Ait QLAKEVDYLLSNKWIP MK VWP Р YN/DVQ/KKYETLSYLPPLSTE a Public garden Ты A Ay а A Ay Ay A HH H > > ~ ~ == = == яа = ја ox = oo № o ом o м № o = HA H mH м ^а "dud а HA H а HA H AH s] 4 нм м Bu м ба ~ в в а ыы а а >> ~ n> H > ы El El a nm a a а а ва а ба за да < « H HH Hd ч HA A HH A H аа а аа а а VWPPLGKKKFETLSYLPNLSSESLAKEV DYLLLKKWIP MQVWPPSGLKKFETLSYLPDLTEE QLIKEVDYLTI К SKL/WI/VP VWPPSGKKFFETLSYLPQLSSESLA MQVWPPSGLKKFETLSYLPPLTEA MKVWPPLGLKKFETLSYLPPLTAE QLAKEVDYLLLKGWIP MOVWPPLGLKKFETLSYLPPLTTE MOVWPPLGKKKFETLSYLPPLSSESLAKEV DYLLLKKWIP MKVWNPYNNIKFETLSYLPPLTDE MKVWPPYNNIKFETLSYLPPLSDE MKVWPPSNVKKFETLSYLPPLSVE QLLKEVEYLLKSGWVEP MKVWPPSGLLKFETLSYLPDLTSE MOVWPPLGKKKFETLSYLPELTIE MOVWPPPGYKKFETLSYLPDLTTE а 2 o Ч са а 3 у E. Ф . @ = A o H ~ Ф © . H sy $ 7 B óo g Ф " . z v в. о m Е . > E мс E o B M Q A om o о . ° са H Ф са - Ф AO H Ф v но av ӯ ~ Ф z < o 0 я чш . © o ~ О ч ow Ф о d [7] . ~ а © d ЗФ Ка] © ~ н . і Ф < © = a On Yo» ao 0 U мя "pr Ф * d Фо оз чк нц v aa > c < Ф М Ф ч a < ~ ~ а ~ м us Ue d я ~ ° one ag <3 28,58 но A ~ ч 52 шт OH +0 HON о HH 2 о ~ ОЗЕ АВЕМЕ BO: Ф Ф ш z dD GH m & = d o ~ ہے‎ о мо 24 ч o о Ф n og d Ф Ф -A nv Hog чо но 8 @ vc са Ф -A ac оа [5] со Ф М "d с d m ON Uo: со 20 zog @ @ oy GA d ui т gx с ~ E Ss оры iln > T 23 3583 ЕЕЕ E © d чыка но $922 я Ф -A - о" on os о U Ho я он B] 2 Bl әл 543 x 3:58 ч до поч wO wm wo оз си о d о @ 4) o Брзина» ооззвзво 3 Eb ORE щы бм 25523973 [7] E зо = - zz SSe2Z5a5h8 sa EPELE 5 KEVDYLLLKKWIP мо L.Johnson) /N.Qld A Gymnostoma (sp. nov FAG Fagus sylvaticus L./P FAGACEAE MKVWPPLGLQKFETLSYLPPLVSENL AKQIEYLLNKGWIP MKVWPPLGLIKFETLSYLPPLSSEIL G. AKEVDYLISSGWMIP Nothofagus aequilateralis (Baum.-Bodenh.)Steenis/ Missouri Botanical Garden Annals of the 302 паа Oy Oy (и >>>>>> =т= EZ oooooo >>>>>> >> о m ыс mmi aad Hd ud ad за Al El n An» а а а а AZZAN naa аа an на аа ын za а а 2 лат HEH E4 E4 E4 E a a A E El El Ba [a [u [u [n [а HH м "9'd/"1 Team unrpodoueuo и W *5'q әртетәрү/з а езехотТоо eueT[TS W :0*8 eprerepv/'T шпеодџешоз unr3sere2 AUD IVIDWITAHIOAYVVO :0*4/'"T тедџебле етвоте2 "5'g ертетерџ/“хоон('Бод) *uerasrieque eesog WHY IVIOVHLNVIVHY 8 апочо Я *£*g әртетәрү/тәєтот eotuueaTAsuued eori4g ЧАЙ JVIOVO IYAN *£*g ертетерм/'1 етбәх sue[bnp '9'4 epretepv/uooyx(* еа втеџеоџтттт eAreo dqvaoxNaNVIOnn Е апомо PTO"N/SITUM*L*D 5Áerreg'4'p xe 'ллед(Леттен'и'а) тАллецо erueoi3suie] “wis '9'8 A3301 aW/'TÁzsAzg STIPTTTXP PTUOPIOS val IVIDVIHL '9'4 SPTRTSPY/*Da eeindandozj3e euuoo PIOÓ'N/ÁeQi9:v ертзти гтебртлџехоела ноо змаомнноо *S'a/BTTo('PTTTM) suepueos er319qqrH *S'g е11әдчед/тттәзлең (оа) езете еэтчеттта тта GWEOVINATIIA *5'g epteTepy/*ury 3 зчБтм unueriexoou шпотледАн ‘L'N чтмлед ONISO/'1 ечезвобиеи eruroieo Гау беј aVGOWISATS 9 dnou5 zx *5'g eprerTepy/ieÁew'2 eotszed етзлоллеа '5'g әртетәрү/'т ептутое2Азв iequeprnbrT WWH IVIDVAITIAVAVE *9S'g/*T 20401 вполепд 'Z'N '9'4 TIPIO/P9I5190 (J"A00H) тлрџетог епбејочзон '9'4 Mey /лхәввелу(хәтрия 3 HStddeog) ortrund snbegoua3oN '9'4 Mey /ределео(летриа 3 brddeog) елезола епбејочзон MSN рчетбил MON/TTOSNA* A техоош епбејочзом "Z'N '9'8 чозбитттем /peasaeo('3'xXooH) rISerzueu snbegou3oN erueusel/*g'xooHu rruunb впбејочзом “Z'N ^5'g8 тлезо/резелео (*J"A00H) eosnz епбејочзом ‘ota Кемзо edeo/pregue3so rrueuburuuno snbegou3oN '5'd Mey/peasieo(leqaTW) sepror[naeq snbezoy on '5'g M9N/P9I51909 (1935104) еотзаолездџе ENÓPJOUYION етџорејтед мәң ышым 5555 Е Е MS z 'penunuo) ‘с зтауј, 303 Martin & Dowd Volume 78, Number 2 1991 Angiosperm Phylogeny Using Protein Sequences = ЖЕ ом оо HH аны >>> > <= < к << mam ae аа ық ааа ыс ad с аа >> х а аа а а > > xz > > >> >>> аа аа xxx >> zx ам а аа аа zm zz ВЕБ zz = Е = > TVS IVIODVIDITYS "98 SPTETOPY/*T етоезит epesey asu яүзочаз$яч ттемен шпзәлодтү чодт/ ‘шет eriegre[o еБитлтоң pul SVS3OVONISHON '9'd 9PT*T9epV/'T eueorqige етлхетәббту '9'8 epreTepy/uoq'q ечтлзветэеэ PIPZY 214 IVIDVILANODWTII PTO"N/*18*Y PIOTITPNU ветешелјеј '9'g8 әртетәрү/°т eutqeuues еовтзед ono 3539411930905 '9'4/'1 esoutds әшоәтә ааз ячзочачаачо “S°O/°T snatqes snueudey 'D'd/'1d'W e[eurorggo UNTIINISEN '5'2/'1 sndeu eorsserg wad dV3oVOISSWVHd '9'8 ортетерџ/.ј'1 емеотлје etuuewzeds '5'g8 әртетәрү/'т вттезиэртоэо гтмело TIL SIWAOVITIL "D'a ертетерч/.ј хоон TS193s*u eAequog 'S9'H ортетерм/“ттепи'4 лотоовтр џозтцоАцоела 315 AVGOWITOOWALS ‘Drd epreTepy/uoq'5('Ae5) eQeTTequn геотелетцаб “D'd/°T елотјтлахед елтеу '5'd/'A*2(*T1) Pesoz езечзату A'IN dV3oVATIVN '9'8 ортетерм/“ттепи '4 stpuez6 епіхеооәтя 9'd елледџер/“ттепи'4 еответехавпе еттедодетлу oTa яүчочаччоозүтя '9'4 eaodeburs/Bury елотјтопеа г91045 '5'g ezodebuts/reuqreey еотдешоле вдоџетеаоАла ода чмаомаммооматата ‘LN о чтилед 04152/'71 впитцзедта orznq ‘Do'd epreTepv/uauny stubtsut ersriouo WOE IVIDVIVENOS 9'dH 9PT?T9pV/'T1 еототр еээетозАца UMON/* TTONWN’ A(*Jseq) SNTTOJTUTIOS sndieoouopoo LHd GWFOVOOWIOLAHG '9'4 ePIeT9pv/'T edetel втттаелти '9'4 epretepv/Astouo ехаетб eet Ttautebnog DAN dVdOVNISVIOAN '5'0/'1 теоедлејо eroeurds 'penunuo) ‘< 3TAVI A‏ ب۵ Ar Ay Missouri Botanical Garden Annals of the oo а As Я N HA с а о а аа аа Aan AH ди 'S'g/'aieg хеш euroÁTS таа IVIDVNOITIAWA W :0*'9 эртетэрч/‘чзчэея Loneq eueeoneT K '9'4 epre[epVv/uoq'5 xe ‘ичио езетлашту етэеэч WIW IVIDUSOHNIH я 09:9 эртетэрч/*7 вочзиеэетлз етёзтрэтЭ я ‘оя эртетэрч/*т епрттте етиозехео 152 SVSoVINIdIVS3VO УТ düOu5 и *с*а/хәдәт< езетпотита еэбиезрАн и ‘09° эртетэрч/еэполет-елТт$ ттартшцов x erueDieg XNS ЗУЗОЧЭЧНАтТХЧ5 W 9:8 эртетерч/*т srunuuoo БпхАд W *'5'g эртетэрч/* be еттојтоттт snunid 34394504 я Kn Я eiQiequeo/':aipuv i MUR етшооттте2 A 54 PI19QUED/*TPUZ unsoursei шптедедоџецаџ NAD AVIDWINONOO ЕТ апочо я ‘wes '9'4 Азјот ЗИ/`Т шпотзиоЯ чолриэророчн Я '9'4 әртетәрү/ `T эицоелрие впзпал тия gwaowoldd И "W'S чоеея S,A9TTIOSS/ASTPUTT snzoTytarzed uobodooneT ‘Y'S ea3noequoq W /'ттепи'4(лерчоб) seprorude4souoo ешотолзаеч odd uVdoValuovda ёт апомо W ‘9°@ TinosstW/*oonz $ Iəqəşs eoruode( хелхАз5 W 0:9 unit. еџмттодео етєәтен УМЧУОЧОЧЧУАЈ5 '9'4 ертетерч и јеллета ('ттепи'4) вттелздепе еттәцоцоцета и '9'8 эртетерч/зээем$ xe 'uunj едлеооџећо sdosnurg LdS dVdoVLOdWS H *'$'d/3euoueai4 seprooe[eu етпштта и :5'4/ ual oe usweTo Ay змаомтоитма Я 'S'g эртетэру/2эи (‘ZG us и eourdey Я '2'8 на A STTTUNY PTSTPIV HVdOVNISHAN TT 40059 '5'd/'1 езелоро PTOTA W Wi *5'g epteTepy/reqysz0g еплотјтшел впзАэттей OIA AVIOWIOIA я / OR '9'g epreTepy/'T eoruo[Aqeq хттеѕ би -9"8 eprelepv/*1 erbyu sntndog 304 'penuguo) ‘с ATV], 305 Martin & Dowd Volume 78, Number 2 1991 Angiosperm Phylogeny Using Protein Sequences а As а As м (м Ay (ш Ay Ay Ay As As а Ay а Ay а Ay м Ay Ay Ay Ay А AY I а а 3 а а A3/II mM QAQ AQ HM A A OA OH A >> >> > > >> a 9 Y Y Y Y Y n El <= AA Оч AA Ке ко кодни И н An из Из из Ио и 4 4 < < H AH AH A HAH за а а ыс аа oo aa mH fH BRL Ol Ol O OSN AAA Ann” оо о ою о 0 YN HEH E4 E4 E4 E4 E4 Ба fu мм Fea [a [xa [ч [xo [ro [ч [їч fu мыммымым Ба fu ба Га ю за м ox Ба [а х x ххэтптаа NTOLSd NT/X9 Ld/Ld Тоза 1А/10 I Ld хтозаа хутотаа хтэзаа ххозаа тлозаа ххозаа хтэчаа хтэзаа утэчаа хтэзаа хтозаа хлолаа J 19 A û ã HY OI a хухосаа YANGI 4HANIdd HYH OT a хлозчада Nux5'71dd N3325'1dd эта NuoIdad Nuo5oIdad NXuyNIGdd хт SA/Id d NTOIdd хлолаа M M M > > >> > >>> >>> >> W 2:4 = 7 eotput елејтбџеџ кеа еа LT апочо я PTSSUOPUI ·9:4 лхобон/ешпта eorueAef етводшол35 я :9'g erodebuts/sieqsepy теоеаџеше sAyorqsouryoo это IVIDVIWIO я ‘D'a әртетәрү/'Т "*oruode( вошАцопя R "5*gd аса а евотттдед eutsseo 34394415135 91 апочэ Я uedep/rexeN еотџоде(·лел'ј sueqeu edezz Vul dVaoVdVul н *S*da/*T-wnqeuez6 eotung копа AIVIDWIINNOd El ‘W'S ebptig Áeaing/'qepeT езотлае етлечзочео я ‘W'S Биетти/челеч (чзипу) septotded етбтмрпт "o'a TINOSSTH /SA0TP99I 3 USARY(UUPUIJTTA) елрџетешев erzedoT '9'8 TINOSSTWH/*9q suebe[e eAney "o'a гртетерџ/.Лело:џ 3 'upebug тлештечритТ eines '9'd/SSOoA ертлалц ersuona 79'4/' јен шпзеттто шптаоттад "9'd зртетерч/џелец(оџезло) unueo шптаоттал :9*d/*TPUTT езетпотпбип етулето '"D'd/STMeT 5 srMeT('TpurT) ерџпотапл етулето '5'g Tanosstw/eTAoy езерлоо геолто XNO dV 3oVH5VNO 'S'q/uepreW(ueprew) едлеоолотш snqdAtTeong '9'H ортетерм/Аллед 9 ·ллеџд(лелтод) ттузтше еџешоџ LUIN dV3oVIHAR 'S'd/'uboo('5d) ечтәтттАлп еџтцопоатј "'"M'S'N чзахом/чод а эчтууе ешоззетея STW SvaoVLINVWOLSVISN я '2'4 SPTETSPY/ZINY(*T) PSODTIONIJ етрхојроом я '9'8 эртетеру/*т етхеоттеѕ UNIYIAT LAT AVIOWIHLAT A ‘O'd TINOSSTA/*T POTPUT sTTENDSTIND Я "D'a эртетерч/`ахоч шплрџеоер unjeiquoo SNO IVIDULIENOS ст 40040 + [u [ч [u [u [ч [u [u fu zx = = *'S'O/XezeTTM(''T) езетрел PUTA '5'0/"'1 е4; ?TOTA = = = 555 5 o M 4 l С + £ q > & = Y v" ч о Ф o а с ~ Е 2 -A Q о a > х o *5'g елледче2/ 'авттес еттоўтзот етрооо 'penunuo) *z 318V[ Annals of the 306 issouri Botanical Garden M хх ЖЕ HA ян [s] aa a аа а сс m [2] ы ш hl a L 21/31 >> >> ммо ма OM zzz А > > а мм W ‘W'S '9'4 A3301 зи/едон еводотб ттетррпа SOT UWAOVINYSOT W 94 epreiepv/ieKew'a3 sueos$ejnig wnpydIo KW "9: а/чоа`о шпџетттеовепл ewogsna N35 AVIOVUNVILNIO я 'S'g/'ig'u esouieo еАон SY IVIDVOIVIATIOSY W '5'g әртетәрү/'т хоитш eourA и '5'g ертетерч/човроом (полеа 3 2778) ехет етттләрити оаї есе рее 02 dnON5 я "M'S'N ртеџелтен/. ттепи 4 шпзетпотде WNT TAUdOSAZ DAZ SY3oVITIAHdODAZ и *9*g ертетерџи/'л snCew unToeedozj] dul AVIDWIOIVIONL W'5'g eprerepv/'ssnp:ipy(-ueT) шпзеттто чоттАЧЯеи6т3$ R :9"8 эртетерч/-т елаетб етчбтатен ати змаомтнотатчи И "5'4 әртетәрү/чозтүч( . 1) шизеттпопо шптџоблетед W 5" E 19H,T шпзецовош шптрола 435 IVIDVINVIIO 6T апочә W ‘LN urMieq/-aiiew('inoT) езетцоела етттеле2 Я “L'N чтмхеа/пон бота езезетдехе елетпбила Я хемедеб 359104 битҷопу/ ·тттея a eeirÁudos TUY змломчаонао тим ''g елледџед W /PIPYDIO A W(* ттепи“ Я) ттџетлепед чолрџеробелотен и '9'4 0'8'М sntTAydoezo впатеэоцоэ 'IMH ЗУЗОЧЭЧНОТЧН 81 düOu5 -5-а эртетэрч/этБитм$ (леттти) ештеетате snuauertv HWS змзаомапомми15 өң D'a әртетәрү/°ттәпи`д хе чочочета етпрџед етттпдлен ‘отл чәдпоу ‘Боер евоовтл тетџород avs AWZOWINI dvs 'S'g эртетерч/Аеттеч euTTTOO PTSISPUTTA "5° 4/леем5 xe AeyoeW eTTeyotnd геллод '£'g epreTepv/'qunuj esuedeo шпхрџеротео LON IVIOVLIOY A '9'8 epreTepv/ssnp''T'V STSUSUTS еТтәлрә2 таи WaOWI TSN W PTO'N/*TTONW' A епадлеооцоџоо eni?uuo) и PTSSUOPUT*D*E 10b08/*119H (SOUPTT) етллеџтла тәетәбү NNO змаомиммноо W ‘95'da epreq[epv/'gseq POTIUETIP етэезета я '5'g эртетерч/`тБиз ттешттм eTXITM Я 'penuguo) ‘с 318V], Volume 78, Number 2 1991 Martin & Dowd Angiosperm Phylogeny Using Protein Sequences 307 SKWIP GWVP MOVWPTEGLKKFETLSYLPELTEAQLLKEVDYLLLN GWIP MKVWPPLGLKKYETLSYLPDLSSEQLLKEVDYLLL SKWIP MKVWPPLGLKKYETLSYLPPLTTEQLIKEIEYLISM GWIP MKVWPPLGLKKFETLSYLPDLTSEQLLKEVDYLLL MQVWPP/TYGKKKYETLSYLPPLTTEQLIKEIDYLIS T = Ou . 0 ~ o. . . . m = о m d . 9 Ф са ov Е о. 39 а ~ © m 9 9 . = UK ч Фо «~ = UU <. ; |a Sse ш“ 3 dm HAMU © AHH OD. = 25758. & не ја = n о М «uo с я: н © ч ә н о Ф О С] = а с oa Фа а оф о 3 ‘ о ма Фа Cv @ Ф с “о E “ ГД о 3 © = EEE gs са яры с ФОФса Е || 333558 ЕЕЕ S8z9z06 GROUP 21 LAM hmeriana Royle/Adelaide B.G Teucrium flavum L./Adelaide B.G. LAMIACEAE MKVWPTEGLKKFETLSYLPPLTSEELLKEVEYLLL KNLIP MKVWPPEGGGGUKFETLSYLPPLSSEQLIK QVEFLL mis cas Phlo LKNLIP MKVWPTEGLKKFETLSYLPDLSTEQLLKEVDFLLL K KLIP MKVWPTEGLKKFETLSYLPPLT ТЕ/КЕ L L K QVEYLLLS G. Citharexylum myrianthum Cham./Adelaide B KLIP Kirk/Adelaide B.G. Vitex lucens GROUP 22 C CONVOLVULACEAE NV Dichondra repens Forster & G.Forster/Adelaide B.G. Ipomoea learii Paxton/P.G. QVWPPVGLKKFETLSYLPPLSTEQ WPPIT/GLKKFETLSYLPDLSTE QV M M LLLEGWVP MQ/KV WP PLGLKK/EFFETLSYLPELT SQ/KELAKEVDY LLKKGWIP MKVWPPLGLKKFETLSYLPPLTDEQLAK QVDY Phlox paniculata L./P.G. SOLANACEAE Cobaea scandens Cav./P.G. y Ay Ay Ay Си Ay nnm Aa Ay ш Qu OF (См ра (и (и e OQ. A Ay A DD DDDDDDDDDDDPD5u»5»5» > ~ > > м ШЖШ ЖЖЖ ЕЕ ШЕЖЕ ЖЕЕ ЕЖЕ ЕЕ ЕЕЕ ЕЕ х ooonooooooooooooooodooouodooco NZAAZAZAZAZASAZANYMANNNAAANSNNNNNHNAHnN AMMMM MMMM ASM AMM MAM MAMA с AHA HHA AAA AAA AA AAA 2 A AAA AHH Ке ۹م 3ر 3ر‎ Ио ۹ہ ۹ہ جر ۹ہ‎ Де جر‎ ИГЕ ДЕ جر کہ کر‎ ДРЕ جم جر جر جر جر جر‎ 94 рч DA DA рч pM DA DA рч рч DA DA DH рч DA DH рч DH DA A М A A پو‎ a Y A AA A Aa El muoaaoaaaoaonoaaoaaaamuana HHHH DP PHD >к EA OE Аз мм > HH a Cx) fe) nl fx] fs] a о A Ио F3] A Ио A [s] [3] O [i] [3] [3] A [3] в AHAHAAAAAAME RC CAAA «6 3 юю 3 RAHA جم جر جر جر جر کر کہ جر جر جر جر جر جر جر جر جر جر جر جر جر جر‎ AAAA 3 а ۹ر جر کہ 3 کہ‎ 3 3 а 3 ۹ 3 4 За аза 9 f] P) Y Y Y p odd A E A A fb A A A A pl pl Ио nl n] ni часа аааааоаааааааа а оҥ о оо ол щл ощ 0 ЕЕ NUNN ло NON ON Ф ON 0 о 0 0 AAA م 3م ر‎ AA Из ИЗ 4 بع بع‎ 1 4 ИТИИ AAAA сосооооооооооонороороряноооон (Си р би р би р Си Си Си Си (м р Си Си Си Си Си Си Си Си р р р (а جر ر جر جم جر جر جر 3م‎ а جم 3+ جر جم 3 3 جر جر جم جر جر چم جر جر کر‎ PM بو بو بو‎ DA بو‎ DA D4 DA DA بر‎ DU м DU DU м DA DA بو‎ DA بو بو‎ м м жоо оо ооо 0 00 NNN Ф 0 Ф (0 0 Ф Ф 0 0 0 0 41 4 4 جر جر ۹ر 4ر 4م‎ а а بع بع‎ 1 3 на کم‎ 6 la بع‎ AA HH Ез Ез Еч е Еч Еч е Еч EM Еч Еч Еч Еч Еч E4 Еч Еч E4 E4 E4 E4 E4 Ca) p p) pl a a A A A A A A A A A Ио fd A И A pon ны PM PM pA PA pA pd pA pA A о [e PM PA PA PA PA DA DA PA DA DA DA PAM Ва oM M Ма x Ма юа ю ю M s n а а а № а а hA bá bá а а x а M ка x x x xx 5 Ма M M Ма Ма MD Ма Ма MG Ба MG n x x x х= ъ=: >>> x ^ ox M $ AS ја Ва ба № 5 €x № д2 22 2 2 2 2 s одо == 2 2 2 ae ZAZAZO FA Ed И ои ok Hi ALAS A HAHAHAHA HA HA A A A A A > H A Ay Ay би р ба р р би A A A A Си Си р (м Си Си Си Си Су Си (а м Са Са Си Си Си Си Oy بم بم‎ As Си Си Си Си Си Си As Си ب۵ ب۵‎ Са (а © шщ ЖЖЖЖ ЕЕЕ ЕЖЕ Е ЖЕЖ Е Ж Ж Е Ж Ж Е ЕЖЕ >>> >>> > }> }> PP }> }> }> ;> рр рр рр р р р БЕБЕБЕБЕЕБЕЕ хх>>>1>>>1>> >> > > EB = но < ' 8 N oo. E 2 с - ач x. e E 4 оо D A чч ~ ~ n an 8 4 2 Hm = g d Q о E à 38 5; M " л 0 - gh эй. 585 ~ . • - . gm H8. o0 Add а 9 5 N ФО 5 а Е. Nad o 3 о. млад ade ~ Ф e EA ‘NOH ON ~ A E . ч. E - بم‎ ‘HO Ффон "uo “Oo + Qo fu SSedis „Фая оо > Фф: · o ач gne 040-049 ч • (им • IOUO . а б " Б . Фо 45 0 BE ب0‎ < Û = 5 ONG AE са AH "CX ВИ “Zea -3 =зно NUON OUO * WUA * INANA чаше јелом а eM RH E 49 © UAA AHAA g Hn DEA amas ee are d re Kal Jagg NSISE і озасо орЕБ OG 83 -:DaGao3tEEd БЕЕН ОВ MUENRM CH GAH OVODO пон о 555 “0E ~4 @ 0 @ بيبا‎ S oH OVA 3 > g ~ VOGRGMEVAaAQO UT С) HOON nH Q- 3 HOM GDMOHOH N U Еч НО 3 2 P» d HO 0H DONG DONUA @ = -A нт CPE EEE EEC 4 - - ‘A - на " q s a а =i dC GM O U g g 2 со 850 рр р n 099999999 ЗЕБЕББЕБЕВБЕБЕБЕ Аница авва ва авва ва о о о зоо >оо»осссссасссаоассссс«св 4400000000020 d d d от d d (d d d d X Gh dduddudz99990902909292929292929209 LKrTEEEEBLEEPEEREERELRILLLLULI GROUP 23 308 Annals of the Missouri Botanical Garden аа Фр ma шщ ааа А A Au м بم‎ Ay Ay Ay Ay Ay Ay Ay Ay а Ay A A HH HH HH >H >>> >> >>> > > > > р> р р> р р> р> > н > > х = эн == =н EEE uox EEE = = х 23 2 2 2 2 2 == == X ox ао Xo ox ооо ш ооо о о ка ка а юа а а ш zo х ox мм n 2 2 ан AZZ 2 2 nnn o o 2021 0 2 2 № 2 oz nn HH HH HH HA (а n) ni HH [A] м № ар ошын р> [^] AH AH HA AA HH HH AHA HA ен A A جر جر رورم‎ +H ы а а ян мо ва AA чыч AA AHA A A AAAAAAAA aad эн “> м > > > “=> нн м “=> “=> »^ м pA بو‎ pA pA pA > ~ м м м м Ыы ыы аа аа ааа аа ааа а а аааааа а аны аа HH “> HH “> >> > HH > > > > > >>> >>> > н > > > ыы ыы мы мы ыыы ын ыыы ы ш a a a A Bl A p] pnl ы ы ы ы мм мм MM мм ммм мым ан м м M, M M M M M M x < м мы HH HH HH ay © < < оо AA < < к 6 НО < н < < < эн dud ad ек| AHH AH А а а d 4 а 3 3 а 3 3 a AHA HA аа а а ы а aa m < а [e Ol ааа а а ннаоаа а an f n ыы ыы ыны ыны < < < ы n ыыы El El a Aaa o w ыы шо т oo ра nan ох ннн н FH A A AAN AA EA na на а нн не QOO не HHH о о E090 000E 00 ни oU AA 4 аа AH AHA dd HHA da A nr An” A dud Mud RA am аа аа а Q0 QA ааа A A пана й аа E4 A Ay Q 0 мм UU ae بم بم‎ паа аА А ааа A А Q Ay Au Au ш As © Ow Qu A са A чы на AH AH “аа AH чыч A A YAA An 74227.” ad AH “> н м н м н н ня “=> мм بو‎ Lal ч PM PA pd pA pA > ~ “~ м м an nn o0 NN ооо nu ooo n Nn Qo ouooououutd oo nn AH AH AH ARH AHH AH AHA A d HAAAAA “а HA AA нын кн HE не ннн не Bee в E E EF E4 E4 E4 E4 E4 E4 не не ыы [CN а а ы ы (о ل‎ f ы El qua [5] Bl (о f Bl nl f mi га [о fF f nl мы ыы ml я fu fu la А ба dn (а fa їч P4 PA Pa A A да ч [а Ба fu Ba [и ка x мм Xx ox > № юа ба а ox мым x = Ва M Ра Ба Ма ба ^ нм а ба Ба Ба mH X за x за ^X oM x ја x ха x ба х Ба а bá 6 а ы ба а мм мым Mud HH “ам dud Muda Е ю AAF A A HAAA MHA 4 nn аҥ оо оо оо оо ооо == zzz o o ن ن ن ن ن ن‎ o оо оо >> A> ны “> HAA HH HH > а A An» Ann A >” ad нн HE E се QM а мым а ш Q Ay A Ay Ay Ay Ay Ay A A A A мм QUO D fu ба بم بم بت بم‎ QOO Aa ш м Q ш ш Е ш б р ao Au Qu == EE = = = = ЖЕ == TEX = = = Е ЕЕЕ Е = == == >> >> >> >> >>> > > >>> > > >>> >>> > >> >> zx да ox аа Ol x ммм X ox Ба ox x x x ха ха а а а ба x аа ам = = zx zx zx = Е = > zzz = = = > 5 5 5 > = = = = = [© P mo [6] . . om m о d 8 Ela: d 99 . [о] . О ~ 74 . . о. d о < ско о о . (9 o С ђ о ke] б . a ° . v m o д < m а м o < [5] ~ . . ~ . os . ~ Ф о 8 3 оо Ф H о со ч я © О . . ks] ‚ 0 . “A а ~ Ф я v = ч m m d оз m 5$ x ț o 3 E. 5 а д о 9 s mo Ф За а $ EE. 33 3 si 3 2 0 ыш O [2] z о. = Ф ч . © © < 35 © ~ ~ ~ ~ ~ ~ * YN 5 я o о ч ч ~ AD ~ со МС н . ч (и ° [7] с . 9 9 N б ~ 9 oz dà Ф ; 8 Ф ~ о ~ Ф [5] o v 3 o м о a4 88 8. 22 E à Se со 265654 2 $3 € мг || ба P3 5. 8B ¿dual? e deidad ot a Е | очаи Замона guidads 88543 B .8233.3358 gis Has Е a4 85858. © чаю Я МС; 5 а азмдлдачназо~ a d зор д оз о © а an = ~ ~ С] нон н a [5] @ са ~. nv oN = uc gs о Ф О о, uc H ~ о с о чо со сс G = 5 dod о = HD HAA La по с D AIIA A-A оо Ф н ом са а зс я О cob в 4 O u A 0 сз y н с d o 83 оо н pe 4383 Bd 93H о Pp ШШ HDH O GAA = а но 406 > сно ч =>» d о О нросесри РАД бз . 3 Q 20 ЕЕ: H obw € © Е "42292 = но 455 ei во 5 Aa» H -A -A no оо о о -d o لډ لډ‎ -Q dd ~ о & осв»оз > E A o A „Ва оон m s 30 Ф Dg H aa 928 С) u AECA ‚я $ Ф i < 0 « 330. A -A Tx OE ос ш но TA 9 4 dace д фо 5 Е: HC OH QO OK YP чс NO 53 ca N эз б = Озанвв»ово 489 Ба o = дюн осно б х BeOS aa оске wu esguoauggoa с zi & ҤНГЕН БЕЛЕНЕ: SAS 3 HBSo230205 she бат HHTEEEHEHEETHHEEHHETEEHHTERBC EEHIHFEHBHIEHEHE позвана а са & 455585 62224 fey n 2285808882%4% mmo 83 309 Martin 4 Dowd Volume 78, Number 2 1991 Angiosperm Phylogeny Using Protein Sequences A ana а 6 «6 9/33 aa 1 T т a fa fu Eu A A Я S DT/Ad NS/19 d І MSDAd d d d d а а W zz = = *9"d/*T ваттоутвпзао хешпм "9" d/*T wnotquodeyz uneyy *65°О/Чочәоң Baga ари unzAdobeg эта SV3OVNOSDA'TOd "O° ертетерџ/'шет езетпотлпе obequntg '5'8 eprerepv/ezauny ттиттөшб unruourTT «та яүзочнточвйпта о әртетәрү/*т unibru zedtg '9'4/ч3ачпу eAx30qATod etwozedeg ата змаомчзата "5'g еее елејтопџ OQUNTON AVAOYNOGNNTIN '5'0/Ае19:џ 3 Аеддој тАетритт eT Tez AUER WOT AVAOVSVYOT ттемен un3ezoqay uoAT/breq TTTUOd sruaAoeT ттемен шпзелодлу uoAT/3eq[qnv srsueuernb езтаполпод хәт HSVd3OVOIHLAOWI 'S'0/'W3ueg еттоутзеэеиез еттәоеца 'S'O/"UIW 3 *X00H ттветхџеш еттчаошем ЧАН AVIDWITAHIOYAAH ‘W'S 1T8798/929Nn1( (Y3TUS) ертате eroAeeog “Ys ‘ов Азјот m езело етџероод AVIDWINICOOS '9'4/"Т stunumoo snutoty "9"8 epreTepv/'biv ·ттепи ттрџештрлеј џотртцоото '9'4 әртетәрү/°Блү ·ттепи euersexyTT^ eudATeoy ana зүяоүтянонапя '9'4/'аапц тәзцәбле vrpieudeus '9'4 epreTepv/'qunuj suebund зпибееэетя ald dV3OVNOVSVIS '9'8 VUY equeS ououegu/'43ünN ?OTUIOJTT?O ешововво219 ОШО IVIDVIYHOSOSSOYO “O° әцлпоатәу/ҷотттем гтеџетедеџ етлетлод wad IVIDWIYVVIVOO ‘L'N чтихед 081852/' ттепи'4 epuou тлеџтлед ттемен шпделодлу uoÁT/*' ооеот snue[eqosÁiuo GHD IVIDVUNWIVEOSAYHO 'penunuo) ‘с Зву], Annals of the 310 Missouri Botanical Garden Qu Pu (м Au Ay Ay м Ay As Ay Ay Ay Ay Ay = f на н са 0 Н 0 ш “ааняачяаанч ۹ر جر 4ر جرم‎ 4 ۹ 2 azaaaaaa >>>>>>н> aa м بص‎ 0 à, à, 0 YN а надо за a a pl p Pl pl n) n Я Га ыы ыыы z= = a аа мм я "94 ортетерм/:1 ноти snberedsy W *5'd/Ə2ULĽH хотоовтр оеочм W ‘)*d/*PITTM етаветеоо PUTTOUWOJ WHO IVIDUNTTIWNOO ‘D'a SPTETSPY и /*Tpuem(Arog) sueose3nT впадеоорттевАлц) я '9'4 эртетерч/човзем $ езеште тецела qu ЯҮЯОМОЯЧЧ я '9'4/330ц425 шпедџебтб џолхрџероттца я p шпзетлоеш univ IVIDVAY W '4'5 1014W/^J3^1 oo NN uo3ebouody 944 FVASWNOLESONOdY W ‘о: ертетеорч/хпецоти *eeupueib етлезатбеб ITV IVIOWIVASITY SNOOUITALODONOHM W '9'4 эртетеруч/*т 91937 TUTA STITA W “59° 4/чоцоџета errojenburnb впеевтооџецзалеа LIA ЗУЗОЧЕТА я "S'd/'xXooH seposÁud еәтәшта я 'S'd/"qunuj ezopo euudeq АНІ IVIDVIVÍAHAHL W *'S'g/'1 воталечзео snuweyy W ‘о: epreT[epv/'u»suosg sniopgrszÁqua3 snuj3ouee2 WHY ЗУЗОЧМИЧНЯ “Z°N ^9'g тлезо W /5565118'8 3 uosuyop'T('uungo'y) пдоз PTUOTOL W '5'8 eaiiequeo2/r[eeuo sTsueebuou eedo[e] я :9:4/'1('1) suedez еэзота •2 4 eaiequeo я /zeusstey Xe AysjoyT әопәйәешецо eruoosied Я '9'g эртетерч/*ттепи*я еттозтихез етшереоеу Я "N'N '9'8 з5хочехем/‘ ха*ч еэзезиер етзешот '5'g eaiequeo /KqxexotA'M'P 3 196814"] еџеоведодде етзешот *5'g зртеторџ/'олпод('шет) шптоехпет џохрџереопет “ZN '9' 8 T1?30/:3g8'N esTeoxe eraubruy == 555555 м о *6'g елледче5/ * тттает gneoties soujueuepy Ladd Есе) еспке): (1 "рапшцио `g 318V[ 311 Martin 4 Dowd Volume 78, Number 2 1991 Angiosperm Phylogeny Using Protein Sequences > HH а а ы ааа а M ox x ама а ммм х 1 x E TI © © a d за IN/53M мам N N = ~ ZA ~ N лаам млаам AANA ахкам 9 94 AT/OW "S*9/*3"X00H етттаелтш етцоватмтем MIM AVIDVIHISLIMIIM АСИ *9'9 әртетәрұ/биәцозпн вертодолаеозаАатб етопЬэзезен A А A A A A о [e] OXL AWGOVIGOXWL И CM'S'N Aeg s,uwuejeg/uosuqop*T sTUnuuoo eruezoioey a29 IVIDVOIVIAD " '9'4/'1 эаотта обхчто ANO IVIDVODANIO и '9'48 эртетерч/*:$5эа втттБелу елрэецая ная змаоммазназ 5иччазонмилоь '5'4 eiQiequeo и /'XooH xe -uunp'v(*ig'u) шпвошАо шптветдоџодто W '9'g VIIƏqULO/'Id'Y SNTTOJTIPT snydərgsng INS IVIDVIVWIINS W *S'9/*T әлебтпл UNITITIL W `5°а/човвоо(°т) шпеэеттти шплечзезата '4'5 SUSITOL 19ATY W /TƏPNƏJS xe *UTIL(*APD) втТелзепе вэзтшбелца я '5'0/"т емџелед шпттот W '5'2/'Тт әхебтпл шперлон vod AVIDVOd UT319W'2'5 ‘ел 99Е/ (*3*quotey) я UNUPTAOT*D X Аөтритт unpunqriorg x шптрташдо очо чүзочатночо я *'9'8 әртетәрү/*т wunTTAydodAy snosny я "9'H эртетеру/*тпот eqeotds едотлтт я ‘D°a зртетерџ/ешпта хотзетә елаетртдеџ 'penunuo) >р ялау, 312 Annals of the Missouri Botanical Garden angiosperm ancestors. All three members of Acan- thaceae that were tried failed with symptoms like these, as did four out of six species from Caesal- piniaceae. Finally, an entirely different sort of failure oc- curred wit the putatively related families Santalaceae and Lo- our species, all hemiparasites from ranthaceae. These species had abnormally high amounts of phenolics, but it seems unlikely that failure can be attributed to them or to any of the other causes mentioned above. The preparations always yielded abnormally high amounts of plas- tocyanin but no trace of rubisco-SSU. Plastocy- anin, a chloroplast protein, has a molecular weight sufficiently close to rubisco-SSU that it occurs, occasionally, as a small contaminant detected dur- ing sequencing. It could be identified by its se- quence but, except in these two families, it was so weak that it disappeared after about seven posi- tions. The strength of the plastocyanin sequence in all four of these hemiparasites suggests that the absence of rubisco-SSU could not be ascribed to some general difficulty like proteases, but might reflect an unusual, perhaps facultative, photosyn- thetic system. GENERAL REMARKS ABOUT THE SEQUENCES OVERALL VARIATION AND INVARIANT SITES A summary of the variation that we have ob- served is given in Table 3; the amino acids most commonly observed are in the top line The rubisco-SSU gene includes two introns, the first of which is inserted before the codon that determines amino acid 3. It determines valine and this, like tryptophan at position 4, is invariant, the two codons carrying the signal to cut the end of the intron (Berry-Lowe et al., 1982). These in- variant residues were useful early signals that the correct protein fraction had been chosen. Within the ћ position 5 and/or 6, at position 19 and/or 20, rst 40 amino acids, proline always occurs at and at position 40. These three regions correspond to bends in the tertiary structure of the molecule. Chapman et al. (1988) have indicated that between the first and second bend there is alpha-helix and thereafter beta-sheet. There is an almost invariant region from amino acids 13 to 18, a region that makes contact with one of the large subunits (Chap- man et al., 1988; Knight et al., 1989). The only variation we have found in this region is the sub- stitution of phenylalanine for leucine at position 15 in five species of Solanum (Martin et al., 1986). These same species also have phenylalanine sub- stituted for leucine at the almost invariant position ~ 21. The simultaneous occurrence of two very rare substitutions indicates a causal connection. Hydro- phobic bonding between the two positions may sta- bilize the bend at position 19 and, because these species are inhabitants of very hot and arid regions, this may have been a factor in natural selection. HETEROGENEITY WITHIN SPECIES The first reports of rubisco-SSU sequences (Stro- baek et al., acids in species of Nicotiana, and these showed heterogeneity at positions 7 and 8 in tobacco. We have also found it at position 30. These hetero- geneities are undoubtedly associated with the am- phidiploid origin of tobacco and led to the expec- tation that heterogeneity would be fairly common, 1976) were for the N-terminal amino not only because about one-third of plant species are polyploid, but also because in diploids gene duplications are frequent. We may not have de- tected some heterogeneities (for example, those by Pichersky et al. (1986) that in tomato there were at least three different DNA messages for rubisco-SSU, acid sequence, suggests that selection acts strongly all with the same N-terminal amino to preserve primary amino acid structure. There are at least eight different genes encoding rubisco- SSU in petunia (Lamb & Fitzmaurice, 1986); for this reason, when we prepared protein from that species, we used a mixture of equal quantities of leaves from four morphologically different varie- ties, with the aim of finding heterogeneities (Martin & Dowd, 1984b). The sequence was of high quality but no heterogeneity was detected. Likewise, we chose to study Rhoeo discolor because it is a complex interchange heterozygote for all chro- mosomes and might therefore be heterozygous for rubisco-SSU, but we detected no heterogeneity. Heterogeneities that were found presented no prob- lem for the computer analysis. INSERTIONS Only two examples of additional amino acids in the N-terminal sequence have been found. Both species of Epacridaceae that we studied had an additional isoleucine between normal positions 9 nd 10. Teucrium flavum (Lamiaceae) had two additional glycines, probably between the same two > positions. These insertions, while clearly of taxo- nomic significance, have been ignored during data processing. Volume 78, Number 2 1991 Martin & Dowd 313 Angiosperm Phylogeny Using Protein Sequences THE N-TERMINUS Haslett et al. (1976) reported that the N-ter- minus of rubisco-SSU was "'frayed," some mole- cules seeming to have methionine in position 1 while others are without it. This is the situation that we have encountered in the vast majority of species, the effect being that at every position two amino acids are recorded, the correct one and the next one. Usually the two signals are approximately equal, especially when the protein is of highest quality. This property is helpful in that it provides a second opportunity for identification and is useful for identifying minor contaminating proteins whose residues appear only once, but probably means that it is more difficult to obtain long sequences because attenuation of the signal occurs earlier. those species for which nucleic acid sequenc- es have been reported were also studied by us and all show fraying. Because the nucleic acid sequenc- es show the N-terminus to be methionine, there is no doubt. The signals we obtained were not typical for either methionine or its sulfone derivative. Whether the derived amino acid is obtained by dansylation (in manual sequencing) and identified LC, or is the PTH derivative from automatic sequencing and identified by HPLC or TLC, the N-terminal amino acid moves differently from me- thionine; therefore, we conclude that it is a modified form of that amino acid. Two exceptions to the above generalization have been encountered. In 10 out of 11 species of the Onagraceae the N-terminus is phenylalanine, the only variations from methionine known, and i these there is no sign of fraying. In six other species the N-terminal amino acid is methionine (and gives the normal signal for PTH-methionine), but there is no sign of fraying, the difference from the ma- jority of species being sharp and unmistakable; these are two members of Papaveraceae (Papaver orientalis and Eschscholtzia californica), two from Pedaliaceae (Sesamum indicum and Ceratotheca triloba), Vitex lucens (Verbenaceae), and Mentze- lia lindleyi (Loasaceae). Any hypothesis to explain fraying must account for these exceptions, and we believe that they exclude artifacts arising. from techniques of protein production or sequencing. Any hypothesis must also account for the modifi- cation of methionine and the equality of the two forms of the protein. We therefore dismiss as un- likely hypotheses relying on inefficient shortening of the protein either as it passes through the chlo- roplast membrane or after entry. It is known that rubisco-SSU forms dimers (Roy et а|., 1978), and we suggest that this may be through formation of disulphide bridges between the N-terminal methionines of two SSU molecules. This might occur in vivo if the enzyme model of Chapman et al. (1988) is correct, but is more likely an in vitro event if the different model of Knight et al. (1989) is correct. If dimers are formed, S-carboxymethylation at pH 8.6 would not break an inter-methionine bond, but we suggest that the dimer does fall apart so that dimethionine is on one chain and no methionine on the other. This hypothesis would account for all phenomena except for non-fraying species, which presumably do not naturally form dimers. METHODS OF DATA ANALYSIS Before computer analysis, amino acid sequences were converted to inferred nucleotide sequences using the genetic code. Usually this could be carried out after inspection of, for example, all the se- quences in a Group so that the most parsimonious choices of codons could be made. A standard was chosen at sites where substitution was silent. Al- though a program was available (Martin et al., 1983), usually the path was obvious and computing unnecessary. Thus the unit of length in phyloge- netic trees is an inferred nucleotide difference (1.n.d.). The number of dichotomizing trees (phyloge- netic Steiner trees) connecting М taxa is | X 5.... (2N-5). The principle of induta 15 ^ the length of every possible tree is calculated and the shortest tree is chosen as the most probable. This agrees with the parsimonious hypothesis that evolution has proceeded by the shortest route. However, because the total number of possible trees increases very rapidly, i.e., when increasing from М-1 to N taxa it increases (2N-5)-fold, it is not always possible to consider every tree. xcept during the final stages of this project, the program that we used was MINTREE, the "branch and bound" program of Hendy and Penny (1982). With a Vax 785 computer the usual limit for simultaneous analysis was 12 taxa. This limit could be extended to about 15 with a supercom- puter (Cyber 205), but the trouble and expense precluded useful work. Although MINTREE has now been superseded, its co-program ANALYZE is still used because it possesses efficient routines for obtaining ancestral sequences and internodal lengths. Most of the analyses have been carried out using HENNIG86 (Farris, 1988) and a personal com- puter (Microbyte 230). This system is about two c o o > © o © о c 25 4 ے‎ 0 O = 5 23 Ф co c= < > 314 "a ¬ $ А 1 $ 1 L Y d I те 1 W N 1 9 5 | € 1 5 0 9 M И 4 1 2 1 Z D Ig 5 но L S I 3 9 V | zZ I I | £ 8 i 2 S 1 3 d S L 1 1 N у а x L 2 z 1 D I а € 6 D Z € I 2 1 1 L У ә S Y N d V А 5 5 JL S 5 5 € 9 I D 1 € € Ol © Iı Z vt I 8 [| I 1 А A 4 о A I I A Ó L d A 1 83 4 Ó N Z € cl 5 D oz € + 9171 € I ге S 1 152 I 1 L T N I N L а л б d N L И А А 3 А 5 1 3 б v Wl € lo? S + 812 1 * I ї © © реј: I 1 тоя © 9 5 9 3 3 N А 50 3 о I V LNI ү 1 € 8 SI € Z zI ст Ol & Z I * I Z + 6 2 VG V 1 8 9 S NS у ч 2 S а а а I 3 I 1 ¥ И А I A 3 1 Z L ТГ I а I Z S IZ I OC LI 222 1 AZ I Z Ll ZIZ 0€ 8 I 6 S N NUL 3 NH JX 17 I 4 V б d Ad 1 1 AI I ма 3 1 V W S8 9 t 19 SE 6b L 9 SS LS 87 6 25 € OS 61 St SL € 956 S re 65 $ OL Lb $5 Z£ 8 € £L 1 А 1 3 MI I I 4 3 I O у 1 1 3 у L L ма S 4 A kK 1 N N A LN 3 d 9£198 OET 001 ZO SB ZIT €£l SEI 001 EOI IZI 821 18 92156 SIL Zt S6 YET BOT CET DEI 921 YET YET 961 921 211 ТЕТ 621 86 #2189 FIT СЕТ 921 921 SL 981 а I мо S 11 1 ха A 3 N у 1083 S S 1d d 1 AS 1 1 3 4 хм N12 1d d мл ди ОР 6€ 8$ LE 92 SE PE се ZE TE OF 65 BZ 15 95 SC vC $5 ZZ IZ OF 61 BI LI 91 SI HI EI Gl 11 OL 6 8 L 9 с " € Z I seirurej ќиеш moy ut ѕәдедтрш proe ошше ue 1эрип лодшти au 'siue[d peas jo saurey 921 шол SS-OOSIANY Jo se»uenbes [еипи1э1-№ jo Атешште ° epe ees sjoquiás proe ошше 0} Кох 104 "punoj 51 и "© 318vL Volume 78, Number 2 1991 Martin & Dowd 315 m Phylogeny Using Protein Seque orders of magnitude faster than the above and has the additional advantage that an analysis can be left running for days or weeks. The principles of its algorithms have not been published, but the time of its release suggests a possible connection with a published letter by Johnson (1987). HENNIG86 offers a number of options, which com- parison with MINTREE using the same sets of data suggest are reliable; in order of preference we have used implicit enumeration (ie*); ie followed by bb (branch swapping); mhennig followed by 66. Be- yond the number of taxa that can be handled by MINTREE or the ie option of HENNIG86, correct solutions cannot be guaranteed. A further advantage of HENNIG86 is that it includes a program for successive weighting, which often reduces the result to one or very few trees. This usually eliminates the need to derive consensus trees, a process that we have found unsatisfactory (Martin & Dowd, 1989). Finally, HENNIG86 de- rives the с: (consistency index) (Carpenter, 1988) and ri (retention index) which we record with each figure of a tree. In a personal communication, Farris defines these as follows: if r and m denote, respectively, the smallest and greatest number of steps that a character can require on any tree, and s denotes the number of steps that character re- quires on a considered tree, then c.i. is r/s and r.i. is (m-s)/(m-r). MINTREE uses data such that each of the four nucleotides is entered as 1, 2, 4, or 8, which allows the counting of ordinary differences and also of heterogeneities; no matter what variation occurs at a site, it can be recorded as a sum that is always different for different combinations. Provided there are no heterogeneities, HENNI can use the same notation (using the nonadditive option); if there are heterogeneities, they must be recorded by inserting additional taxa. This is satisfactory if there is only one variable site within a taxon when only two taxa need to be recorded. Assumptions of linkage must be made, however, if there is more than one variable site but only two taxa are to be recorded. This problem becomes increasingly im- portant as an analysis progresses from using raw data to derived ancestral sequences for families and then Groups because, in these, heterogeneities may be numerous. We have therefore used alter- native strategies. The first is inserting additional taxa as just described and accepting the result if the different versions of a taxon cluster without interruption. e second strategy is to use binary coding for nucleotides; e.g., 1000 for A, 0100 for G, 1100 for A and G, 0010 for C, and so forth. In con- junction with HENNIG86, MINTREE notation is slower than binary notation, which is therefore advantageous. Аз would be expected, binary gives a tree length double that using MINTREE notation but it is seldom exact. If inexact, the length is always less than double, and there is a loose re- lationship between the deficit and the number of heterogeneities. Because the details of HENNIG86 have not been published, we have been cautious about choosing between these alternatives and have done all Group analyses with HENNIG86 using both notations. We structured our investigation such that the majority of families were represented by at least two species from different genera. If we accept that taxonomy is seldom wrong when placing gen- era within families (Heywood, 1978), then we have an empirical way of judging the merits of the two notations. Omitting families that have either a sin- gle representative or are multiply represented, all Groups have been analyzed using both notations and, from the minimal trees recorded, we have chosen the best as judged by pairing of represen- tatives of families. Їп eight Groups both methods had the same best tree, in six binary gave the best, and in eleven MINTREE notation gave the best. In the section **Analyses Within Groups" we have therefore used the taxonomically best minimal tree no matter which notation was used to derive it. However, in later sections we have used binary notation exclusively because it is quicker and more convenient. Among best trees, 79% of families showed cor- rect pairing of its members. When judging this result, it should be remembered that a single mis- placed species will often result in the failure of pairing of representatives of two families. While a few such occurrences may be the result of incorrect taxonomy, the remainder are presumably caused by convergent evolution. The details can be seen in the figures for the Groups. ANALYSES WITHIN GROUPS EXPLANATION OF THE FIGURES The figures are drawn to scale, which is indicated by the length of one inferred nucleotide difference (i.n.d.). Only lengths have meaning, not angles. Usually at least two trees have been derived for each Group, one using sequences of individual spe- cies and the second using derived familial nodes. If the first shows congruent grouping of putative members of families, then the second is not needed. 316 Annals of the Missouri Botanical Garden Disruption of familial grouping is often caused by the sole representative of another family or by a member of a multiply represented family. The sec- ond tree is derived from familial nodes and single- tons, is drawn (1982), which is given in Table 2. When appro- priate, trees of multiply represented families or genera are also given. In later analyses, Group nodes, and often one or two others, will be used; each of these is numbered. The Base of the Angiosperm Tree Before detailed analysis of Groups began, many analyses were done using the five gymnosperms, representatives of the monocotyledons and of Groups 1, 2, and 3 which were most likely, on taxonomic grounds, to be near the root of the dicotyledon tree. The angiosperm family closest to the gymnosperms was Schisandraceae. Figure 1 shows the junction of the gymnosperms, Schisan- draceae, and the other angiosperms. The derived sequence of this node has been used as “Base” in all subsequent Group analyses. It will be noted that Figure 1 is different from, and taxonomically more satisfactory than, the equivalent figure of Martin and Dowd (1989). Since then a sequence of Welwitschia has been obtained and this paired with Ephedra between the angio- sperms and the other gymnosperms. Three tempts to study Gnetum were made but all failed with symptoms suggesting strong leaf protease ac- tivity. WELWITSCHIA WINTERACEAE EPHEDRA CALYCANTHAGEAE METASEQUQIA t N RANUNCULACEAE SCHISANDRACEAE и и МУМРНАЕАСЕАЕ ASE OF DICOTYLEDONS FIGURE 1. = of five angiosperm families from and e ancestral sequence derived for the MASE of Five gymnosperms analyzed with familial Groups 1, he has been used as an outgroup for ana- lyzing the Groups of а with a different scale, and uses only the three-letter familial abbreviation of Weber Group An attempt to study Hedycarya (Monimiaceae) having failed, Peumus was left as a singleton, which was therefore omitted from Fig- ure 2a but included in Figure 2b. Correct pairing and grouping occurs in all families except Aristo- lochiaceae for which a derived familial node is shown in Figure 2b. As indicated above, Schisan- draceae is nearest to Base with a rather long gap to the remainder. Group 2. |n contrast to the straightforward- ness of the previous Group, Group 2 has presented problems that arise from the great intrafamilial variation of Ranunculaceae and Papaveraceae, the trees for which are shown in Figure 3b and c. The two derived familial nodes are shown with the rest of the Group in Figure 3a. We interpret Ranun- culaceae and Papaveraceae to be ancient angio- sperm families, and only with some misgivings have we adhered to our acceptance of current taxonomy at the levels of family and below. This is especially so for Papaveraceae, but splitting off Fumariaceae creates more problems than it solves. ANNONACEAE TRES M i го Xylopia irioden б, Drimys Idiospermum Pseudowintera m himonant = с a WINTERACEAE Tasmannia Aristolo Calycanthus T % z 2 « Kadsura (4) T 2 ma G oc T < < 0 SCALE ci 85 Ri 89 I 1 ind. 9 Schisandra Base SCALE BASE iind Ci 93 Ri 89 FIGURE 2.— (А). Group 1, omitting the single repre- sentative of Monimiaceae. —(B). Tree of family nodes for ;roup 1. Моште 78, Митбег 2 1991 Martin 4 Dowd Angiosperm Phylogeny Using Protein Sequences 317 MENISPERMACEAE Cocculus BERBERIDACE AE r (A) Lardizabala 3 [e] Akebia E E RANUNCULACEAE m е т Decaisnea PAPAVERACEA SCALE | fing, С! 91 Ri 92 Aquilegia Helleborus Papaver auus: eo ci "d Dicentra Clematis viticella RANUNCULACEA е ак поде lematis rehderiana yox 87 Ri 86 л Ci 92 Ri 88 FIGURE 3. (A) Group 2, with family nodes, derived from (B) and (C) for Ranunculaceae and Papaveraceae s.]. Nymphaea “м. “e ^c, S4 CABOMBACEAE Victoria Nuphar Brasenia } SCALE Ci 100 Ri 100 | 1 i.n.d. FIGURE 4. Group 3. Liriope ШЕ. i Е ыи Азрагадиз dino POACEAE ARECACEAE Let COMMELINACEAE p ел T iin m i BASE of MONOCOTYLEDONS SCALE Ci 88 Ri 80 1 ind. Aspidistra idas ARACEAE и Е The monocotyledons that have been stud- ied with some families represented by their nodes. Group 3. Failure with a species of Cabomba has left Brasenia as a singleton that does not separate from the three members of Nymphae- aceae; however, the internode is so short that we draw no significance from it (Fig. 4). Piperaceae, Nelumbonaceae, and the mono- cotyledons. Piperaceae and Nelumbonaceae have not been placed in a Group because in both cases opinions differ among the four phylogenies consid- ered when we nominated Groups. All place them in either our Group 3 or Group 1, so it is appro- priate to carry out a joint analysis with the members of these two Groups, and at the same time to consider the links with the monocotyledons. Al- though no species have been added to those re- ported earlier (Martin & Dowd, 1989), the se- quences have been reanalyzed with HENNIG86. To reduce the number of taxa to be compatible with the ie option, familial nodes have been used for Araceae, Агесасеае, Commelinaceae, Роасеае, and Smilacaceae (Fig. 5). The monocotyledon node has been derived and is included in Figure 6. The result is different from that of Martin and Dowd (1989), and this is presumably because of the new computing program. Annals of the Missouri Botanical Garden 318 pe Pipturus E MAGNOLIACEA LAURACEAE ARACEAE ARISTOLOCHIACEAE CALYCANTHACEAE ANNONACEAE т Other WINTERACEAE tj MONOCOTS 9 i 2 MYRISTICACEAE Párasponia SCALE v, Ci 93 Ri 89 PIPERACEAE Group 3 Group 2— 1 ind. SCHISANDRACEAE FIGURE 7. Group 4. SCALE Ci 88 Ri 88 GYMNOSPERMS FIGURE 6. The base of the angiosperm tree showing зы laa ge of the monocotyledons to members of , 2, and 3 and Nelumbo and Piperaceae. Parottia ve Ш N.aequilateralis uj g Gymnostoma S o ~ [т] > = e < = 3 N.solandri : i e MINE N.cunpinghamii Liquidambar llocasuarina N.antarctica N.menziesii N.procera Nothofagus node (B) SGALE ci 94 Ri 91 1 ind. FIGURE 8. —(A). Group 5. The node for Nothofagus was derived from (В) Volume 78, Number 2 1991 Martin & Dowd Angiosperm Phylogeny Using Protein Sequences 319 Gordonia FIGURE 9. Group 6. Group 4. There is good pairing between mem- bers of Ulmaceae and Urticaceae but not Moraceae (Fig. 7). It was only after failures with two species of Ficus and one of Maclura, almost certainly due to protease activity, that we supplemented with Humulus, knowing that its taxonomic position was not entirely clear. The failure of Humulus and Morus to pair was therefore not surprising. Sub- sequently, Humulus was removed from Group 4 and added to the list of uncertain taxa (see below). Group 5. The tree of Nothofagus (Fig. 8b) is slightly different from that of Martin and Dowd (1988) because it is influenced by the weighting procedure of HENNIG86 and because it includes Fagus, which does not separate. From this tree a node has been derived and used in Figure 8a. While Betulaceae, Casuarinaceae, and Hamamelidaceae have correct grouping, the junction with Base di- vides Nothofagus from Quercus. Group 6. Figure 9 differs from one alread published by Martin and Dowd (1989); the two trees are of the same length but this one is preferred because it shows perfect pairing and reflects the Brackenridgea 3V3OVNH2O SCALE cı gg Ri И cr Ci 66 Ri 64 order when only familial nodes are used. However, the other probably conforms better with taxonomy in that Dilleniaceae is separate from the other three families. Group 7. Figure 10 shows that the two rep- resentatives of Juglandaceae pair leaving Myrica, the sole representative of Myricaceae, separate. Juglans MYRICACEAE Myrica arya 1 ¡SCALE Ci 91 Ri 931 ind. FIGURE 10. Group 7. 320 Annals of the Missouri Botanical Garden NYCTAGINACEAE N es Phytolacca Chenopodium * / 7 ш iy K x О * g < Celosia Y 5 8 * 8 a Q 9 О < 5 g Bosea “Y, О Е S А 5 m pinacia Beta 3 - Codonocarpus 2 3 © tium Y a (A) c G Silene Q pc * SEALE Ci 88 Ri 85 PHT AMA CHN CRY E (в) Son Ci 98 Ri 94 FIGURE 11.—(A). Group 8. — ~ B). Group 8 family nodes. Group 8. Тһе “Centrospermae” is one of the most unsatisfactory groups with representatives of two families, Chenopodiaceae and Amaranthaceae, failing to form pairs (Figure lla). Spinacia and Beta have identical sequences, but these are quite different from Chenopodium. The tree for family nodes is in Figure 11b Group 9. There is a marked difference be- tween the two representatives of Tiliaceae; Grewia is at the bottom of the tree (Fig. 12), while Spar- mannia disrupts the clustering in Malvaceae. How- ever, the remaining four families are satisfactory. Grewia was removed from Group 9 and added to the list of uncertain taxa. (As will be mentioned later, it subsequently rejoined.) Group 10. While Violaceae, Cucurbitaceae, Salicaceae, Brassicaceae, and Flacourtiaceae formed good clusters (Fig. 13a), the two represen- tatives of Datiscaceae (Datisca and Tetrameles) were very different. Attempts to study Capparis having failed, Cleome was left unpaired so we chose Reseda from the putatively related family Rese- daceae. Since these two did group, we did not seek correct partners for them. In addition to these two Chorisia BOMBACACEAE Durio Sparmannia Dombeya < 8; x | Ж. ` ~ p * еи o г phaeralcea Althaea L g VACE q Malva 3 a <=" Dryobalanops Т) < [ө <, Elaeocarpus o X » S Shorea 2 0 а > З id > Ye Aristotelia Grewia | йсй Ci 81 Ri 79 FicuRE 12. Group 9. singletons, Moringa represents a monogeneric family. A tree from family nodes is shown in Figure 13b. Group 11. The representatives of all four fam- ilies form pairs (Fig. 14), Myrsinaceae adjacently and the other three families dichotomously. Group 12. As mentioned earlier, the two spe- cies of Epacridaeae are distinguished by having an additional amino acid inserted in their sequences. Although this could have been used as a character, it was unnecessary because the two species paired separately from the two Ericaceae species (Fig. 1 Group 13. When family nodes are derived for Rosaceae, Cunoniaceae, and Saxifragaceae, they are very close (Fig. 16b), so it is not surprising that there is confusion when individual species are analyzed (Fig. 16a). The representatives of Rosa- ceae pair correctly, however. Volume 78, Number 2 Martin & Dowd 321 1991 Angiosperm Phylogeny Using Protein Sequences el © Raphant Nasturtium ОХ ОЕ Рапспопена NEL вы opP Mimusops Kiggelaria чак се йс i | о“ g 5 [e] (A) о Styrax 2 = > О T | — Azara "m Ш Salix q зы < ә Populus ucurbita с, $ © E 2 ш Melicytus E > ag \ © Primula | —- Карапеа 2 < N i а \ [A ш y = Ardisia т, = Cucumis Ш m Viola a | > [A БАР Seat Ci 89 Ri 90 3 \ = \ x \ Cyclamen SCALE \ ту | T PT. Ci 94 Ri 92 | "T BRA | CUC 5 2 FIGURE 14. Group 11. Мо эсле cies Ri B3 Arbutus FIGURE 13. —(А). Group 10, omitting single represen- tatives of families. —(В). Family nodes of Group 10. y ш cause they did not pair and are sometimes placed i < different families, Datisca and Tetrameles аге e d here. 5 "to, E 4) fry Astroloma Че, Ye eucopogon Group 14. Among minimal trees derived when all legume species are analyzed simultaneously, there are some in which the two Mimosaceae spe cies pair and so do the two Caesalpiniaceae; how ever, the eight Papilionaceae species are confused. We have therefore derived a Papilionaceae node separately (Fig. 17b) and show this with the other two families (Fig. 17а). FIGURE 15. Ci 93 Ri 73 Group 12. Annals of the Missouri Botanical Garden 322 Hydrangea Leucaena 4 CUNONIACEAE ру, us Callicom: a & Prunu ut Ci 96 Ri 92 | m Ind. Ci 98 Ri 94 ROS Glycine Lu SAX B и Oxylobium P Hovea n CUN FIGURE 16.—(A) Group 13.—(B). Family nodes of — NOG Grou зеље 1 ind. FIGURE 17. (A) G resented by a node derived from Lopezia Ludwig ONAGRACEAE = О Ф Ф Ф Tibouchina (A) z ©. EA o 3 D SCALE сЕ Ci 87 Ri 88 MELASTOMATACEAE Ludwigia E.glabellum Ci 90 Ri B4 SCALE лв (С) Наџуа | SCALE Ci 92 Rias В tind FIGURE 18. (А) Group 15 omitting Trapa and Punica, which are included with family nodes in (В E ). гопр 14 жић Раршопасеае гер- (В). (С) Опаргасеае. Volume 78, Number 2 1991 Martin & Dowd 323 Angiosperm Phylogeny Using Protein Sequences Cassine $ Xo ~ ~ > 597 Strombosia % о $ o^ Ochanostachys Euonymus 2 3 1 к=к Ci 96 Ri 97 FIGURE 19. Group 16. Group 15. This Group, which corresponds to the order Myrtales, was discussed by Martin and Dowd (1986a). Since then only Trapa and Punica, both singletons, have been added. When the rep- resentatives of the other five families are analyzed, pairing is good except in Lythraceae (Fig. 18a). The three members of Onagraceae in this tree are from the bottom of the family tree (Fig..18c). When family nodes are analyzed (Fig. 18b), the root of the tree is in a different place from the one pre- viously published; it is uncertain why this is so, but Dodonaea y „у co теа Pistacia а ш Ailanthus 4 praem [9] < а = 2 = < Mangifera | FIGURE 20. it could be due to the inclusion of new families, the earlier choice of inappropriate outgroups, or the new analytical methods. Group 16. As discussed earlier, all represen- tatives of the hemiparasites of the Santalales and Loranthaceae failed to yield protein samples, so this Group is reduced to Olacaceae and Celastra- ceae in which pairings are straightforward (Fig. Group 17. This Group is not very satisfactory possibly because, as indicated in Figure 20b, there has been a rapid radiation. The consequences are that the members of Simaroubaceae and Sapin- daceae do not pair, while Flindersia, sometimes excluded from Rutaceae, does not group with the other two representatives of that family. However, there is good pairing for Connaraceae and Ana cardiaceae (Fig. 20a). Melia having failed, Cadrola is left as the sole representative of Meliaceae. Group 18. The two members of Haloraga- ceae, Gonocarpus and Haloragodendron, are so confounded with the three members of Rhizopho- raceae (Fig. 21) that there was no point in deriving family nodes to derive a Group node. Group 19. Nitraria having failed, Zygo- phyllum is a singleton as is Tropaeolum, for which no partner was available. As shown in Figure 22a, the members of Geraniaceae and Malpighiaceae pair. The family node tree is shown in Figure 22b. Connarus Q (A) RUT SCALE с; ¡ А Ci 89 Ri 76 1Tn.d. о sour Ci 95 Ri 88 | (в) (A) Group 17 omitting Melia, which is included with family nodes in (B). 324 Annals of the Missouri Botanical Garden Anisophyllum . Az, Stigmaphyllon Ke, < r & = Pelargonium 23 aloragodendron Erodiu (A) SCALE ci g7 Rios } SCALE ci 100 Ri 100 i 1 ind. FIGURE 21. Group 18. Group 20. Hoya was left a singleton by failure to extract protein from two other members of clepiadaceae, Asclepias and Cryptostegia. The members of four families showed dichotomous pair- ing while Logania paired alongside Buddleia, which is only possibly a member of Loganiaceae (Fig. 23). Group 21. and Verbenaceae were very similar but there was nevertheless a minimal tree in which congruent pairing occurred (Fig. 24) The representatives of Lamiaceae SCALE, а. Ci 88 Ri 50: FIGURE 22. (A) Group 19 omitting Zygophyllum and (B). Tropaeolum, which are included in Group 22. There have been previous reports of Solanum (Martin et al., 1986) and Nicotiana (Martin & Dowd, 1984b). Using HENNIG86, new Mandevilla < uj RUBIACEAE g enia = © ce Coprosma & e Buddleia y Vinca ЈУ inus Q E РЕА Fraxinus s & > = lea t $ Eustoma SCALE C Ci 75 Ri 82 1 i.n.d. FIGURE 23. Group 20. Volume 78, Number 2 Martin & Dowd 325 Angiosperm Phylogeny Using Protein Sequences Teucrium nodes have been derived for both genera (Fig. 25b, c) and were used, with Anthocercis, to represent Solanaceae in Figure 25a. These group well but there is confusion between the representatives of Convolvulaceae and Polemoniaceae. Group 23. Аз mentioned earlier, all attempts to extract rubisco from Acanthaceae species failed. The representatives of the other four families of this Group pair well, Scrophulariaceae, Gesneri- aceae, and Bignoniaceae dichotomously and Pe- daliaceae adjacently (Fig. 26). А ; Cob SCALE Ci 87 Ri 89 obaea 1 i.n.d. FicURE 24. Group 21. (A) ea e SOLANACEAE © Nicotiana node > o RN > о Anthocercis Solanum node | p Ci 94 Ri 93 S.viridifolium Petunia hybrida N.tabacum2 (=N.tomentosiformis) па is „ы. er SCALE 1 ind. | ју Ci 85 Ri 91 FIGURE 25. (A) Group 22 with the Solanum node derived from (B) and the Nicotiana node from (C). 326 Annals of the Missouri Botanical Garden Digitalis SCROPHULARIACEAE Paulownia Sinningia x $ Q ES Pandorea 2 PEDALIACEAE d Ceratotheca Sesamum Saintpaulia g о) 2. О 2 >, О Jacaranda > acaranda > { oir Ci 89 Ri 84 i 1 ind. FicuRE 26. Group 23. CAPRIFOLIACEAE Group 24. Sambucus The three members of Caprifoli- aceae are substantially different from the two mem bers of Valerianaceae so that correct grouping is observed (Fig. 27). Group 25. Viburnum `N This Group is unusual in that Api- um and Foeniculum of Apiaceae have identical sequences as do Schefflera and Fatsia of Arali- aceae. Consequently, the tree of this Group (Fig. 28) is very simple. e E Centranthus 7, Y, Y, k с oe Pastinaca X С à; < У Valeriana QN iss р Foeniculum/ ARALIACEAE Apium Schefflera Fatsia ISCALE ~, y | SCALE ci 96 Ri 96 a ae нер | 1 ind. {НЕ FIGURE 27. Group 24. FicunE 28. Group 25. Volume 78, Number 2 1991 Martin & Dowd 327 Angiosperm Phylogeny Using Protein equences THE DERIVATION OF A TREE FOR THE GROUPS OF DICOTYLEDONS FIRST STAGE; A TEST OF THE REALITY OF GROUPS Depending on the size and complexity of the Group, one, two, or three nodes have been marked near the bases of each Group tree; altogether there are 58 basal nodes and the ancestral sequence of each has been derived using ANALYZE. These have been used for a test of the reality or integrity of the Groups. If a family does not really belong to a Group, it should usually behave like an out- group and assume the position closest to the base of the tree. Thus, in a simultaneous analysis of all 58 basal nodes, it would be expected that nodes truly belonging to the same Group should cluster together. If a family is misplaced in a Group, the nodes should separate. The only program that can be used with 58 taxa simultaneously is HENNIG86 with the option mhennig followed by bb. This was done three times, each yielding large numbers of trees for which strict consensus trees were derived. Inspection indicated that most Groups behaved as if they were real, but some separation of nodes occurred in Groups 5, 8, 14, 15, 22, and 24 (see below). It is unlikely that this sort of analysis would give a completely reliable result, but our interpretation is that where there is no separation of within-Group nodes, that Group should be accepted as valid. We understand that our test is not infallible, but we are reluctant, at this stage, to attempt another obvious test, viz. the simultaneous analysis of adjoining Groups. This test was used earlier with Groups 1, 2, and 3 and led to considerable mixing of the first two. The amount of convergent evolution between Groups is probably such that, if this test were applied widely, confusion would result. Therefore, even though we understand the limitations, we confine our testing of the integrity of Groups to one sort of analysis. For the six Groups where there was doubt, we applied the test devised by Lake (1987). This is confined to four species, A, B, C, D and uses a chi-square test to decide which is the most probable of the three possible relationships, viz. A + C+DoA+CS€B+DorA + р € В + С. Thus representatives of each part of a divided Group were tested with representatives of the Groups with which they most closely clustered. These tests gave no further grounds for doubting the integrity of Groups 8 and 15 and consequently, in the next stage of the analysis, they were included un- changed. The tests reinforced the doubts about Groups 14, 22, and 24, so their separate parts were added to the list of uncertain families to be incorporated later. These were: from Group 14, Mimosaceae plus Papilionaceae on the one hand and Caesalpiniaceae on the other; from Group 22, Convolvulaceae plus Polemoniaceae on the one hand and Solanaceae on the other; from Group 24, Valerianaceae and Caprifoliaceae. Tests with Group 5 were equivocal, so Hamamelidaceae was removed and added to the list of uncertain families, but the node for the remaining three families was used at the next stage. SECOND STAGE; DERIVING A PRELIMINARY, ABBREVIATED TREE Following the first stage, the basal node was used to represent each of the 22 remaining Groups (though amended in Group 5 after removal of Ham- amelidaceae). Several analyses, using mhennig and bb, were carried out on these nodes. The object was to identify apparently constant associations from which nodes might be derived in order to reduce the number of taxa to 16, a number com- patible with analyses using the reliable ie program at the next stage. The following five pairings were chosen and their nodes derived: Groups 2 and 3; Groups 6 and 16; Groups 7 and 9; Groups 11 and 19; Groups 21 and 23. Group 25 was omitted at this stage because it was small, well-defined by morphology and our own work and could be in- corporated later in the same way as the uncertain taxa. The resulting tree of 16 taxa is shown in Figure 29. THIRD STAGE; INCORPORATING TAXA OF UNCERTAIN AFFINITIES At this point there were 28 taxa on the list of those with uncertain affinities, comprising 18 fam- ilies that were not placed in a Group (see Table 1 but note that Piperaceae and Nelumbonaceae were considered earlier), two genera (Humulus and Grewia) excluded from Groups during their anal- ysis, five families and two pairs of families excluded during the first stage, and Group 25 omitted at the second stage. We wanted to add these into the second stage tree as accurately as possible using the ie program. These analyses with 17 taxa could each be performed in about a day. Although there was some variation, the second stage tree remained reasonably stable during these analyses, and we noted where each uncertain taxon fit. Six joined 328 Annals of the Missouri Botanical Garden in the basal third, 14 in the middle third, and 16 in the distal third. (The nonadditivity reflects that rigid demarcation was not exercised and borderline taxa were placed in two sets.) The members of each of the three sets were then analyzed with the cor- responding members of the second stage tree and possible new or amended Groups were identified. FOURTH STAGE; REDEFINITION OF GROUPS Putative new or amended Groups were tested extensively to ensure that they were real. In this process an important factor in determining the coherence of Groups was the length of the inter- node joining a hitherto uncertain member to the Group. Penny et al. (1987) have emphasized that “long edges attract,’ that the junction of a distantly connected taxon is ” and we have long been aware subject to so much variation that it is scarcely reliable. Thus, we have usually rejected a potential new member of a Group if it joins with a dispro- portionately long internode and have left it as un- certain. 21 & 23 4 SCALE и] 1 in.d. Ci 91 Ri 91 8 URE 29. had dissolved. In three cases, definite hypotheses arising from our work could be tested. In each case the question was whether a taxon belonged to the Group to which it was initially assigned (Table 1) or to the Group indicated by stage 3. This could be answered by considering the lengths of the alternative trees. In two cases, Humulus and Hamamelidaceae, the new grouping was shorter and therefore preferred. For Grewia, the trees were the same length so there was no good reason for preferring the new grouping (with Group 18 As a consequence of these tests, only 15 of the original 25 Groups have the same composition as they had before the first stage of this section. The other ten Groups have been increased, decreased, or merged. Where a nucleus of an original Group remains, the number has been retained but A add- ed. Original Groups 13, 24, and 25 have disap- peared. New Groups 26, 27, 28, and 29 have been formed. Humulus has been removed. Group ТА. BASE(1) The provisional tree of Group nodes abbreviated by combining some Groups and omitting others that Моште 78, Митбег 2 Martin & Dowd 1991 m Phylogeny Using Protein Sequen 329 Group 54. Hamamelidaceae has been re- Group 124. At all stages, Convolvulaceae and moved. Polemoniaceae grouped separately from Solana- Group ВА: Alo фе caga Ga ceae, the other member of Group 22. At the third (““Centrospermae””) remains intact, Lecithydaceae Pa с еее risa 2n Humulus join the same branch of the шее (Fig. re а и the other three families grouped | appropriately. Gouroupita Group 144. The second stage tests suggested that the legumes should be divided between Caesal- piniaceae on the one hand and Mimosaceae and Papilionaceae on the other; Caesalpiniaceae clus- tered with Group 13. A series of Lake tests (see “First stage” above and Martin & Dowd, 1990) was therefore performed. These tests strongly in dicated, first, that Caesalpiniaceae was closer to Rosaceae than to either of the other two legume families and, second, that Mimosaceae and Papilio- naceae were closer to other Groups (e.g., Connar aceae in Group 17, Chrysobalanaceae in Group 18A below) than were Caesalpiniaceae and Rosa- ceae. ۸92س 34354 Lecythis Н p cies мач Other second stage tests had indicated that Pro- teaceae, Coriaria, Crossosoma, and Hamameli- POLEMONIACEAE | | daceae were also linked to the complex of Groups FIGURE 90. Group BA: oe Figure Jor Groups 13 and 14. The tree that resulted when these were Cobaea Arbutus m 2 > O m Сој, м “Oy Dichondra Vu, Phlox Асу. Rhododendron 4 Ipomoea © ш Ғадоругит о, - 4 Astroloma L % Y x ~, S ES б ES < Leucopogon Rheum Q "m Rumex SCALE ci 81 Ri 85 1 i.n.d. FIGURE 31. Group 124A. 330 Annals of the Missouri Botanical Garden SAXIFRAGACEAE sc Ci 91 Ri 88 (A) a HAMAMELIDACEAE PERSOONIOIDEAE soon (B) a t; < € Telopea ^O, о € <, «Ў Leucadendron Darlingia L.arborescens © 9 A Protea .dentata Brabeium Macadamia Adenánthos SCALE lind. Ci BB Ri 85 FIGURE 32. (A) Group 144A. For three legume families see Figure 17, and for Proteaceae see (B). all analyzed together is shown in Figure 32a. The Proteaceae node was derived from Figure 32b, while the Mimosaceae- Papilionaceae node is node 3 of Figure 17a. Group 184. Third stage tests suggested that the Chrysobalanaceae and Vitaceae might cluster with Group 18 and also with Group 25 (Apiaceae and Araliaceae). Incorporation of these (Fig. 33) does nothing to repair the previous (Fig. 21) dis- junction of Haloragaceae while Rhizophoraceae s.l. remain apart from Anisophyllea. Group 224. With the other two families join- ing Group 12А (above), the Solanaceae were left as the sole representative. Group 26. This new Group (Fig. 34a) consists of three families (Campanulaceae, Caprifoliaceae and Goodeniaceae), each with well-paired repre- sentatives. With them is Asteraceae, the node for which is derived from Figure 34b Group 27. This comprises the families Elaeag- naceae and Rhamnaceae, the members of which form pairs (Fig. 35). Group 28. pair with Simmondsia, which is sometimes placed in Buxaceae. While the latter clusters with Eu- phorbiaceae (Fig. 36), Buxus does not. As noted below, Buxus does not Group 29. The species of Hydrophyllaceae, Thymelaeaceae, and Valerianaceae form pairs in this new Group (Fig. 37) TAXA THAT REMAIN UNPLACED There are three families for which we have no acceptable hypothesis. (a) Loasaceae. It was un- RHIZOPHORACEAE s.l. Vitis Gonocarpus e © р AX Anisophyllea ~ Haloragodendron FIGURE 33. Group 18A. Bruguiera Carallia Q Ay, PL 2 Chrysobalanus Ke q inari С Parinari A Pastinaca NC су Apium $ Foeniculum ¥ | —— —Fatsi Schefflera ARALIACEAE SCALE rry Ci 89 Ri 85 Volume 78, Number 2 Martin & Dowd 1 Angiosperm Phylogeny Using Protein Sequences Sambucus E^ % Q Viburnum < Lactuca Q S (B) (A) «© Cichori Senecio L Abelia а Ш : x Helianthus Eupatorium 9 Lobelia 5 д Canarium ASTERACEAE ASTERACEAE node ашна = | & Scaevola Q ea Ci 98 Ri 95 2 ; Z SCALE Goodenia ў 1 in. e Ci 91 Ri 88 ES FIGURE 34. (A) Group 26 with node for Asteraceae from (B). Acalypha NA cem Rhamnus SS e e Ceanothus : SCALE i кс Ci 73 Ri 77 411 FIGURE 35. Group 27. Simmondsia SCALE ci 73 Ri 77 : 1 ind. : FIGURE 36. Group 28. 332 Annals of the Missouri Botanical Garden Nemophila P VALERIANACEAE = Centranthus s 9, кл Valeriana < > © Ta — SCALE 1 ind, Ci 82 Ri 88 FIGURE 37. Group 29. Ex БРЕ CRR MIM peas RO PPL HAM "i BOMDPC ELC MLV STR TIL 9 JUG MYR SCR PED GSN BIG e MRT MLS LYT 15 TRA CMB ONA PUN САВ МУМ 3 RAN PAP LAR 2 BER MNS pes Ci 91 Ri 87 1 ind FIGURE 38. The underline MRS РАМ T STY LAM VRB 21 C SAL 17 fortunate that we failed to obtain a sequence for Eucnides bartonioides because this left Mentzelia as a singleton and therefore with a “long edge g edge" that joined unreliably. (b) Plumbaginaceae. Al- though the two representatives , Limonium and Plumbago, paired well, there remained a very long internode joining the family to the tree, and so we have left it unplaced. (c) Buxaceae. Originally both Buxus and Simmondsia were chosen as represen- tatives of Buxaceae (s.l.), but they proved quite different and, since there was taxonomic opinion to support this, they were treated as such. Where 26 AST CAM CPR GOD HYD THY V AL | 29 N " L— 29 gi ANS CHB ВНЕ VIT 1 ARL HAL 28EUP SMM 16CEL OLC n ВБА MRG CPP R 6 CLU DIL OCH TEA TTM FLC VIO ANA MEL SAP CNN RUT SMR S GER ТАР МІР ZYG GEN ERI EPC 2160 OLE RUB 12A CNV PLG PLM BALCY CRY ee PNT SOL NYC CHN PHT ANN MAG LAU ARS MNM 1SCS CAL MYS WIN overall tree for the dicotyledons. Groups are numbered and their constituent families indicated using the three-letter acronyms of Weber (1982), given in Table 2. Families in which nitrogen-fixation is known Volume 78, Number 2 91 Martin & Dowd 333 Angiosperm Phylogeny Using Protein equences . 30» а 2 — 254 2 ш € 204 | У tee а 154 TERME HHT ш 10- О < | E 0 a 0 1 2 З 22A8A 15 5А 11 17 19 12А 20 23 4 10 6 16 28 14А 9 27 7 18A 29 26 21 GROUPS the X axis in the order that they depart from the trunk of the VER FIGURE 39. Groups are arranged alon tree (Fig. 38). Solid dots are the indicate the range from smallest to greatest. Simmondsia grouped reasonably well with Eu- phorbiaceae, Buxus did not and remains unplaced. FIFTH STAGE; THE SIMULTANEOUS ANALYSIS OF REVISED GROUPS Initially, the nodes of all 26 revised Groups were analyzed using the option mhennig followed by bb, and the resulting tree was divided into a top, middle, and bottom section. Thus, with overlaps, each con- tained 14 taxa, a number that could be analyzed using the ie option. Fortunately, there was no con- fusion at the overlaps, and the three parts were fitted together to give the overall tree (Fig. 3 DISCUSSION THE RATE OF EVOLUTION AND THE AGE OF THE ANGIOSPERMS In Figure 38, which shows the overall tree for the dicotyledons, there is a “trunk” from which branches depart at irregular intervals of up to 5 i.n.d. In Figure 39, we arranged Groups in the order that they branch from the trunk. For every species we measured the number of differences (in i.n.d.) between it and the base of the angiosperm tree (Fig. 1). For each Group we show the mean of these distances and also the range from smallest to greatest. The mean of all Groups is 16.2 i.n.d. We have also analyzed variance and shown that there is significant (P « 0.001) variation between Groups. Thus, although the difference between a slowly evolving Group such as Group 3 (mean 14.1) and a rapidly evolving Group such as Group 21 (19.7) is not great, it is probably real. The age of the dicotyledons can be derived from the product of the mean number of differences of species from base and the rate of evolution. Since mean distances of species of that Group from the angiosperm origin, and bar Figure 6 suggests that the monocotyledons are derived from the dicotyledons this is also the age of the angiosperm. Martin and Dowd (1988) es- timated the rate to be 1 i.n.d. in 14 Ma for a single evolutionary line. However, this estimate was based on members of the Fagaceae, Proteaceae, Sola- naceae, and Winteraceae, all of which belong to Groups that evolve more slowly than average; their mean number of differences from base is 14.7 i.n.d. Thus, the inferred age of the angiosperms is 14 х 14.7 — 205 Ma, that is, at the beginning of the Jurassic. Crane et al. (1989) and Wolfe et al. (1989) have estimated the age of the angiosperms as 200 Ma. If the monocotyledons are indeed de- rived from the dicotyledons, there is good agree- ment. THE RELIABILITY OF OUR TREES The current limitations of computers and com- puting programs make it impossible to conduct a large phylogenetic analysis in a completely objec- tive manner. Our first important deviation from objectivity has been accepting taxonomic opinion that species belong to the same family. Our second has been seeking a consensus in placing these into Groups. The assumption of correct assignment to families is strongly supported by the correct pairing (or formation of clusters of three when appropriate) shown in the objectively derived figures of the fina 26 Groups. Of the 95 families with two or three representatives, only 11 had disjunct representa- tives and, of these, at least four were families sensu lato with taxonomic opinions that they should really be split. These are the separation of Humulus from Morus in Group 4, of Flindersia from other Ru- taceae (Group 17), of Buxus from Simmondsia Annals of the Missouri Botanical Garden (Group 28), and of Anisophyllea from other Rhi- zophoraceae (Group 18). When it is further con- sidered that one aberrant species can disrupt two families, we submit that the high proportion of correct grouping is strong evidence not only for the correctness of other taxonomy at this level but also of the soundness of our approach. If our methods, while not perfect, are good at the level of placing taxa into families, is there any reason why they should not be equally acceptable at higher levels? We have investigated this with the assumption that the probability of errors will increase as internode lengths decrease. From each of the Group trees we have determined that the average length of internodes within families (re- stricting the measurements to families with only two correctly paired representatives) is 5.6 i.n.d., while the average length of internodes between families is 4.7. From the final tree showing the relationships of Groups (Fig. 38), the average length of internodes is 3.0. Thus, if our assumption is valid, the ratio 5.6:4.7:3.0 should reflect the re- liability of arranging species within families, fam- ilies within Groups, and Groups in the final tree. We suggest caution about accepting relationships as the taxonomic level increases. There is no obvious reason why the ratio just reported should not be similar for other macro- molecular sequences. However, with nucleic acid sequencing (see review by Palmer, 1988) the amount of information available might increase by an order of magnitude over that presented here; thus, even if internode lengths at the highest levels are still proportionately small, the probability of errors due to chance when using small numbers should diminish and lead to more decisive phylog- enies. THE VALUE OF THIS STUDY We believe that the demarcation of plant taxa at all levels should be the prerogative of botanists with a broad background in taxonomy and that the same specialists are best suited to compare the results of this study, expressed as phylogenetic trees, with published phylogenies. Because we do not have that background, we resist the temptation to point out the similarities and differences that we perceive and to assess when our trees are likely to be incorrect. Our perceptions are likely to be un- balanced. One difference between this phylogenetic study and most others is that it is repeatable. Without detracting from the value of published angiosperm phylogenies, they do seem to depend on the ac- cumulated wisdom and experience of rare individ- uals whose relevant brain functions are not easily transmitted in entirety. On the other hand, anyone improved analytical procedures it is possible that more acceptable endpoints may be reached. We have avoided the word *'conclusions" be- cause we do not claim that this work is definitive. Rather it has led to new working hypotheses which, we hope, others will test with more extensive sam- pling and more data including much longer se- quences. To such investigators our analytical meth- od, whether perceived as successful or not, may be a useful example. NATURAL SELECTION AND THE EVOLUTION OF RUBISCO-SSU Under “General remarks about the sequences," we discussed heterogeneity within species and quot- ed the evidence of Pichersky et al. (1986) that natural selection acts to keep the amino acid se- quence constant. Below we present other evidence for the importance of natural selection. Under “Methods of Data Analysis,” cussed Lake's test, which is based only on trans- versions (mutations from a purine to a pyrimidine we dis- or vice versa) and ignores transitions (purine to purine or pyrimidine to pyrimidine). Lake (1987) quoted evidence (Brown et al., 1982) that in animal mitochondrial DNAs, transitions occur an order of magnitude more frequently than transversions. Zimmer et al. (1989) have found for higher plant cytoplasmic rRNA that, on average, transitions were twice as frequent as transversions with the lowest ratio in the most invariant regions. We have investigated this in 44 families of Groups 1 to 10 and have scored those amino acid changes within families that can be ascribed unequivocally to trans- versions and transitions. There were 123 transi- tions and 306 transversions, a proportion of 0.28 transitions and therefore quite different from the evidence just quoted. We have considered each of the 61 codons in the genetic code and, assuming that each nucle- otide can change to another with the same prob- Volume 78, Number 2 Martin & Dowd 335 Angiosperm Phylogeny Using Protein Sequences ability, calculated the frequencies of the four pos- sibilities, i.e., transition causing amino acid change, transversion causing amino acid change, no amino acid change, and lethality (stop). Thus, for the two codons that determine phenylalanine the ratio of transitions to transversions is 0.25, for eight of the amino acids it is 0.33, and the ratio varies from 0.14 to 0.34 with an average of 0.2845. Averaging the 31 variable amino acids in the top line of Table 3 gives the ratio 0.268, which may be compared with the observed figure of 0.287. This suggests that, at the nonsilent positions, which are the only ones we are able to consider, there is close to randomness with respect to the occurrence of tran- sitions and transversions. We suggest that the large discrepancy between our result and the expectations from chemistry and nucleic acid sequencing is due partly to our inability to score silent substitutions and partly to the over- whelming importance of natural selection in de- termining the amino acid sequence of an important enzyme. Even though most nucleotide substitutions are presumably transitions, this has little effect on the final outcome, the amino acid sequence, on which natural selection can act. Other evidence of strong natural selection comes from consideration of variation at positions like 8 and 9. At position 8, 84% of species have glycine and 15% asparagine. This substitution requires at least two nucleotide changes so, in the absence of selection, the single-change intermediates serine or lysine would often be expected, though they have not been observed. Similarly, at position 9, 56% of species have leucine and 28% lysine. Again, this is a two-nucleotide change, but the only single- change intermediates found are methionine and isoleucine, and these are much too rare to occur унета: Apparently, glycine and asparagine are "adaptive peaks" at position 8 and leucine and lysine are at position 9. When positions 8 and 9 are considered together, there is a small excess over chance expectations of the combinations gly- cine-leucine and asparagine-lysine; these may be adaptive peaks because both combinations are found within Tiliaceae (Group 9), Papilionaceae (Group 14A), Apocynaceae (Group 20), Proteaceae (Group 14A), and different families of Group 15. Clearly, convergent evolution has occurred. his last evidence suggests that adjacent posi- tions influence one another, which is known. An- other example is probably found in the Onagraceae, the only family with N-terminal phenylalanine and, alongside it, asparagine, again only found in On- agraceae. Solanum species with the same rare substitutions at positions 15 and 21 are examples that the effect can extend further. Another example concerns positions 30 and 39, both of which are almost always either valine (V) or isoleucine (I). The frequencies within species of the four possible combinations (VV, VI, IV, П) indicate that the two positions evolve independently; nevertheless, they are different in the monocotyledons, with 67.5% isoleucine, and the dicotyledons, with 35.4% iso- leucine. It is conceivable that monocotyledons are richer in isoleucine because they have a more ef- ficient synthetic pathway for isoleucine so that, in the absence of other strong selective forces, the substitution of isoleucine for valine may be favored. espite the evidence that natural selection is acting strongly, there are few decisive changes, such as the change from proline to isoleucine at position 6 during the evolution of the monocoty- ledons. At positions 7 and 8, the combination ty- rosine-asparagine occurs in the gymnosperms, Groups 1, 2, and 3, but in no other Groups, sug- gesting that these amino acids are primitive. How- the distinction between primitive and ad- vanced is usually equivocal; for the following example, normal taxonomic criteria have been used ever, to distinguish 58 primitive genera (those in the gymnosperms, Piperaceae, Nelumbonaceae, and Groups 1, 2, 3, and 5) from 67 advanced genera (those in Asteraceae, Campanulaceae, Goodeni- aceae, Hydrophyllaceae, and Groups 10, 20, 21, 22, 23, and 24). At position 12, tyrosine occurred in 5% of primitive and 31% of advanced genera while at position 20 aspartic acid occurred in 5% of primitive and 43% of advanced genera. While admitting that the sampling is not entirely satis- factory, it appears that tyrosine at position 12 and aspartic acid at position 20 are advanced. How- ever, the important point is that the divergence is so indecisive, the primitive amino acids phenylal- anine at position 12 and proline at position 20 still occurring in the majority of genera in all advanced Groups. If it is correct that natural selection acts strongly to determine the amino acid sequence of a protein, this could be important in considering “molecular evolutionary clocks.” If the clock that is considered is derived from nucleic acid sequences, the rare event that is the basis of regression of number of 336 Annals of the Missouri Botanical Garden differences on time is nucleotide substitution, the most common form of mutation and not always subject to natural selection. If, however, the cloc is derived from amino acid sequences, the rare LITERATURE CITED BERRY-LOWE, S., T. D. MCKNIGHT, D. M. SHAH & К. В. MEAGHER. 1982. The nucleotide sequence, pression, and Bona of one member of a шын: ied encoding the small subunit of ribulose-1,5- sphosphate carboxylase i in soybean. J. Molec. Appl. а Й 483-4 Brown, W. M., E. M. PRAG ER, A. WANG & А. C. WILSON. 1982. Mitochondrial sequences of а tempo and mode of evolution. J. Molec. Evol. 22 CARPENTER, J. M 88. Choosing among multiple equally parsimonious cladograms. Cladistics 4: 291- 2 CHAPMAN, M. S., S. W. Sun, P. M. С. Curmi, D. Cascio, W. W. SMITH & D. 5. EISENBERG. 1988. Tertiary structure of plant ei domains and their con- acts. Science 241: CRANE, P. R., M. J. ы Е, Ј. А. E & E. M. 1989. Buc udin ied (London) 342: 131. ae da A. 1976. The taxonomic f the ucture of m putes a classical taxonomist 5 view. Brittoni : 1-27. An Integrated System of Classification of Flowering Plants. Columbia Univ. Press, New York. DAHLGREN, R. 1983. General aspects of angiosperm evo and macrosystematics. Nord. J. Bot. 3: 11 HENNIG86 Reference: Version . Јат . Farris, Port Jefferson Station, New Ya rk. GRUND, C., J. GILROY, T. GLEA U. JENsEN & D. BouLTER. 1981. Sys к o of the d on amino acid sequence data. 5 Phytochemistry 20: 1559-1565. G., A. HASLETT YARWOOD, I. M. Evans & D. BOULTER. 197 Studies on the small subunit of fraction 1 protein from Pisum sativum L. and Vicia faba Г. Biochim. Biophys. Acta 420: 122-132. D. Ремму. 1982. Branch and bound algorithms to determine minimal evolutionary trees. Math. Biosci. be 277-290. 978. Flowering Plants of the World. oe London More approaches to p travelling : 525. сз R саа of the three- dimensional structure of the small subunit of RuBisCO from higher plants. Science 244: 702-705 LAKE, J. A. . A rate-independent technique for analysis of ар“ acid sequences: evolutionary par- . Biol. Evol. 4: 167-191. 1986. в сгор improvement. Nature (London) 324: 41 MARTIN, С. С., J. KEEN, B. V. Fonp- отр & Н. J. NEWBURY. 1987. Isolation and partial amino acid sequencing of the small subunit of ribulose-1,5-bis- event is not mutation alone but its incorporation into the gene pool by natural selection. In both cases clocklike behavior is observed, but the rare event is different. phosphate carboxylase/ doped from three species of orchid. Pl. Sci. 50: 233-237. MARTIN, P. С. 79. Amino acid sequence of the small subunit of deep 1,5- ED EE rid lena from spinach. Austral. J. Pl. Physiol. —— — & J. M. Eine The Study of е phylogeny using a cid sequences of ribu 1,5-bisphosphate cibo de Ш Addition of Mal. vaceae and Ranunculaceae to the phylogenetic tree. ану Ј. Вог. 32: 283-290. — € —————. 19845. The study of plant phy- a using amino a bisphosphate carboxylase. I a- gaceae and the rate of evolution of the small bur Austral. J. Bot. 32: 291-2 & —————. 1984c. The study of plant phy- logeny using amino acid sequences of ribulose-1,5- bisphosphate Cuore V Magnoliaceae, Polygo- naceae and the concept of primitiveness. Austral. J. Bot. 32: 301-309 & 6a. Phylogenetic studies using protein sequences within the order Myrtales. Ann. Missouri Bot. Gard. 73: 442-448. РЕ 19865. A phylogenetic tree for some monocotyledons and gymnosperms derived from protein sequences. Taxon 35: 460-475 & —————. 1986c. Is the gene for Бай globin widespread among dicotyledons? Pp. 81-82 in W. Wallace & S. E. Smith (editors), Proceedings of the Eighth Australian Nitrogen Fixation Confer- ence. dab e и ute of Agricultural еек Parkville, Victor & 8. A molecular л clock for m— Taxon 37 4- & —— ——. 1989. ligated is rm flow- ering plants as derived t rom amino acid sequence data. Pp. n B. Fernholm H. Jornvall (eie. The Hierarchy ol Life. Elsevier, Amsterdam. . А protein sequence study i the dicotyledons and its relevance to the evolution mes and nitrogen fixation. Austral. Syst. Bot. 3: 91-100. & A. C. ne 1983. The study of plant aha ЕЛ sing amino acid sequences of ribulose- 5-biphosphate carboxylase. I Biochemical a d the patterns of variability. Austral. J. Bot. 95- D. "BOUL LTER & D. PENNY. 1985. Angiosperm phylogeny studied using sequences of five m molecules. Taxon 34: 393-400. ‚ 1. M. Down & & S. J. L. d study of plant phylogeny using amino acid sequences of ribulose-1,5-bisphosphate cube. П The analysis of small subunit data to form phylogenetic trees. Austral. J. Bot. 31: 411-419. RRIS & D. E. Зумом. 1986. The study of ‘plant phylogeny using amino acid se- 1983. The Volume 78, Number 2 1991 Martin & Dowd 337 Angiosperm Phylogeny Using Protein Sequences quences of ribulose-1,5-bisphosphate carboxylase. VI е к from different continents. Austral. 87-1 1. NAKANO, Ка T Hase & Н. MATSUBARA. 1981. The complete amino acid sequence of parsley (Petroseri- num sativum) ferredoxin. J. Biochem. 90: 1725- 1730. PALMER, J. D., В. К. JANSEN, Н. J. MicHaELs, М. W. CHASE & J. В. MAHHART. 1988. Chloroplast DNA variation and plant phylogeny. Ann. Missouri Bot. Gard. 75: 1180-1206. Penny, D., M. D. HENpv & I. M. HENDERSON. 1987. Relability of evolutionary trees. Cold Springs Harbour Symposium on Quantitative Biology 52: 857 PICHERSKY, Е., R. BERNATSKY, S. D. TANKSLEY & A. R. Evidence for selection as a subunit of ribulose-1,5- e go pae expe aed ох- ygenase. Proc. Natl. Acad. U.S.A 3880-3884. КАМЗНАУ, J. A. M. 1982. Structures ч plant proteins. 229- 290 in D. Boulter & B. Parthier (editors), Nucleic Acids and Proteins in Plants 1. Encyclopae- dia of Plant Physiology, NS Volume 14А. Springer, Berlin. Кох, H., А. VALERI, D. Н. РОРЕ, L. RUCKERT & К. A. Costa. 1978. Small subunit contacts in ribulose- 1,5-bisphosphate carboxylase. Biochemistry 17: 665- 668. ScocIN, R. 1981. Amino acid sequence studies and plant phylogeny. Pp. 229-290 in D. A. Young & D. eidler (cdi кеи and Angio- sperm Phylogeny. Praeger, New Yor STROBAEK, S., С. C. GIBBONS, В. HASLETT, p. BOULTER & S. С. WILDMAN. 1976. On the nature of the polymorphism of the small subunit of ribulose-1,5- e carboxylase in the amphidiploid tiana tabacum. Carlsberg Res. Commun. No. = TAKHTAJAN, A. The systematic War ie of dicotyledonous families. Pp. 180-201 in C. R. Met- calfe & L. Chalk (editors), Anatomy of the Dicoty- ledons, Vol. II (2nd Edition). Clarendon Press, Ox- for THORNE, Б. Е. 1983. Proposed new ینان‎ ВЫ in ot. 3: А Mnemonic re jns же for the families of vascular plants: a device for more effective herbarium curation. Taxon 31: 74-88. . Соот, Y-W. YANG, Р. M. SHARP & WOLFE, К. H., W-H. Lr. 1989. Date of the monocot-dicot diver- gence estimated from ар DNA sequence data. Proc d 6: 6201- = ZIMMER, E. HA MBY, M. L. ARN D, D. A. LEBLA oy E L THERIOT. 1989. Ribosomal RNA phylogenies and flowering at: evolution. Pp. 205- 214 in B. Fernholm emer . Jornvall (ed- itors), The Herde of Life. Elsevier, Amsterdam. PHYTOGEOGRAPHICAL RELATIONSHIPS OF THE GENERA OF ANGIOSPERMS IN THE FANJING SHAN MOUNTAIN RANGE, NORTHEASTERN GUIZHOU, CHINA! Tsun-shen Ying,” David E. Boufford,’ and Yuling Tu‘ ABSTRACT The Fanjing Shan mountain range, one of six Man and the Biosphere (MAB) sei’ in China, covers an 6 area of ca. plants in 636 g ‚ 23 0 the 20 largest families, which together contain more than half of ra are analyzed and мора апі temperate genera are nearly equally represented, The distributional patterns of the g and some comparisons are made with eight other major mountain regions in south-c Мо contain rich floras of phytogeographical ed Fanjing Jinfo Shan, and Shennongjia regions, areas to the north an affinities of the genera of angiosperms in the one, are located to the south and west. m? in northeastern Guizhou Province, China nd east, than it is to any o Besides the келк and temperate components of the flora, the Fanjing Shan . The flora comprises ca. O species of flowering which are endemic to China or barely extend into bordering + areas. The importance of the valua briefly erg Tn the Жо. Shan region, enera are analyzed, entral and oi China that an is most similar Pss ed to the Wuyi Shan, e others, which, except for region also contains a rich representation of both Sino-Himalayan and Sino-Japanese elements. China is particularly rich in numbers of plant species. The estimated 29,000 species of flowering plants and ferns within the country, including about 7,500 indigenous trees and shrubs (Pei, 1984), represent approximately 8-12% of all vascular plants known worldwide. Part of the richness can be accounted for by the diverse geology, complex climatic patterns, and broad altitudinal and lati- tudinal ranges. Elevational differences in China span more than 9,000 m from the summit of Qomo- langma Feng (Mt. Everest) on the border with Nepal at 8,848 m above sea level to 154 m below sea level in the Turpan Depression in the Xinjiang Uygur Autonomous Region. China also occupies the only place on earth where an unbroken tran- sition of vegetational types exists, ranging from trop- ical rainforests in the south through subtropical, temperate, and boreal forests, to tundra and alpine vegetation in China's southwest and north. This vegetational continuum has resulted in associations of plants not seen in other parts of the world and includes taxa with diverse and widespread phyto- geographical relationships extending around the world. Many of the plants in these associations, or their ancestors, are among those considered to have had a much wider distribution during the Tertiary. Fieldwork in the Fanjing Shan mountain range in northeastern Guizhou by Chinese and American botanists in 1986, and fieldwork in other mountain ranges throughout China by Ying and his collab- thank our colleagues, Bruce Bartholomew, Qianhai Chen, Sizhao Fang, Jingen Qi, Stephan A. Spongberg, 1 We Zhanhuo Та, Peishan and Ying-chien Chien, Yenfeng Fu, Weilian Huang, and Wang, and Yinghai Xiang, who accompanied us in the field in northeastern Guizhou in 1986, Xianpu Wang for logistic arrangements prior to our fieldwork. We are particularly grateful to the National Geographic Society, the Chinese Academy of Sciences, the Academy of Sciences of Guizhou, of Tongren District, Jiangkou Xian, and the Institute of Botany, Beijing, for generous financial support. The local government officials Yinjiang Xian, and Songtao Xian aided us in many ways and deserve special an thanks for their efforts and bornes We are also grateful to two anonymous reviewers for their comments useful information. ? [nstitute of Botany, Chinese Academy of Sciences, 141 Xizhimenwai Dajie, Beijing, People's Republic of China. * Harvard University Herbaria, 22 Divinity Avenue, Car nbridge, Massachusetts 02138, U.S.A * Department of Geography, Guizhou Normal University, Guiyang, Guizhou, People's Republic. of China. ANN. Missouni Bor. GARD. 78: 338-358. 1991. Volume 78, Number 2 1991 Ying et al. 339 Phytogeography of the Fanjing Shan orators, has provided the basis for interpreting the phytogeographical relationships of one particularly rich area, the Fanjing Shan mountain range, lo- cated between subtropical and temperate China. HISTORICAL AND BOTANICAL SETTING Botanically, Guizhou is one of the least known provinces in China. The province, which occupies an area of about 174,000 km? and reaches from ca. 24?50' to ca. 29°10’М and from ca. 103?46' to 109%0'E, is located in a transitional position between the tropics in Guangxi and southern Yun- nan and the subtropical and temperate areas of Hunan, Sichuan, and Hubei. The province also occupies a transitional position between the Sino- Himalayan flora, which extends from Yunnan through the Himalayas, and the Sino-Japanese flora of central and eastern China. As such, Guizhou is vital for the analysis and understanding of vege- tational and floristic patterns and relationships in all of eastern Asia. Yunnan Province, one of only a few provinces in China for which more or less accurate figures are available, and which borders 14,000 species of flowering plants (Wu, 1984), or nearly half of all the angiosperms in China. Si- chuan, to the north, is also considered to be flo- ristically rich, but no accurate figures for the num- bers of indigenous plants there are available. Similarly, Guizhou has a rich flora but, again, exact figures are not known. Until relatively recent times Guizhou was one of the most backward and sparsely populated prov- inces in China (Geelan & Twitchett, 1974). It is largely mountainous with the major ranges in the north-central and south-central parts of the prov- ince. The highest peak, in the Fanjing Shan range in the Wuling Mountains, exceeds 2,550 m ele- vation (Deng, 1982). Extensive areas of flat land are scarce, and except for narrow areas along major streams, much of the land in cultivation has resulted from terracing. The two major roads in Guizhou roughly divide the province into quarters. One road runs from north to south through the center of the province and the other extends from Hunan Province in the east to Yunnan Province in the west. They intersect in the capital at Gui- yang. Minor roads reach most other parts of the province, but travel in Guizhou has always been difficult because of the rugged terrain; even now travel by highway is slow, averaging only about 40 km per hour on the best highways. Because of these factors and the difficulties involved in ex- ploiting the natural resources, the vegetation of Guizhou, although frequently heavily disturbed, is somewhat better preserved than in other parts of China. One exceptionally well conserved region in northeastern Guizhou, the Fanjing Shan mountain range, an area with an unusually rich assemblage of plants and animals, has been designated a Man and the Biosphere Preserve. Few vegetational studies have been conducted in Guizhou. The first was а general discussion on the flora and vegetation of northeastern Guizhou based on observations made in the late summer and fall of 1931 during a collecting trip to the Fanjing Shan region (Steward & Chiao, 1933). Photographs published at the time show the veg- etation to be more intact than at present, but signs of drastic disturbance are already evident in some of the photos. The more recent and most important and ex- tensive vegetational studies in Guizhou were also conducted in the Fanjing Shan mountain region and the findings published in Scientific Survey of the Fanjingshan Mountain Preserve by the Fan- jingshan Mountain Preserve Institute (1982). In- cluded in this publication are discussions on David- ia involucrata Baillon forests (Yang et al., 1982), Quercus stewardiana А. Camus (reported as Cy- clobalanopsis stewardiana (A. Camus) K. M. Lan) forests (Yang, 1983), the general vegetation in the Fanjing Shan region (Huang et al., 1982a), a dis- cussion of soils in the mountains (Zhang & Zhang, 1982), and geomorphology and recent crustal movements (Yang, 1982). Another paper by Huang et al. (1982b) discusses the newly described Abies fanjingshanensis W. L. Huang, Y. L. Tu & S. Z. Fang and its forest associations and phytogeo- graphic relationships. The Davidia forests are composed of a mixture of deciduous trees with some broad-leaved evergreens and conifers at elevations between 1,000 and 1,800 m on the lower parts of moist slopes. These forests have been described as "complex," and from the descriptions are rem- iniscent of forests in which Davidia occurs in the Shennongjia Forest District in western Hubei Prov- ince (Bartholomew et al., 1983). The Davidia for- ests in the Fanjing Shan region have been consid- ered so significant that they have been given government protection. The Abies fanjingshanen- sis forests are among the southernmost Abies for- ests in southeast Asia. According to Huang et al. (1982b) these forests, at 2,100-2,350 m eleva- tion, were first discovered in 1981. The trees grow on north-, northwest-, and northeast-facing slopes with gradients of 50—60°. The species composition and community structure of these forests are some- what similar to other Абгез forests in southwest 340 Annals of the Missouri Botanical Garden China, but are reported to differ in a number of characteristics. The occurrence of five species of Fagus L. in Guizhou Province (Guizhou Flora Committee, 1982) is of particular interest and is perhaps the greatest number of species of the genus growing in close proximity anywhere. Three of these species were the subject of investigation in the early 1970s by Tsien et al. (1975; that they form distinct forests on Fanjing Shan. Guizhou is also located near the center of con- ee discussion below), who found centration in China of many eastern Ásian-eastern orth American disjunct genera. Among the taxa belonging to this association in the Fanjing Shan region are Liriodendron L. (Magnoliaceae), several species of Asarum L. sensu lato, whose affinities appear to be somewhat intermediate between As- arum sensu stricto and the eastern North American Hexastylis Raf. In the Polygonaceae are several species of Polygonum L. sect. Echinocaulon Meis- ner, a section with species in Asia and North Amer- ica and one species, P. sagittatum L., attributed to the eastern part of both continents (Steward, 1930; Park, 1988). Another genus in the Polygo- naceae, Antenoron Raf. (Tovara Adanson in many references), with two species reported for Guizhou, is also frequent and abundant in thickets, moist woods, and along streams. The number of species in the genus is questionable, and the relationship of the Asian taxa to each other and to the single North American species, A. virginianum (L.) Rob- erty & Vautier, is uncertain. Two species of Liq- uidambar L. occur in the Fanjing Shan area, L. formosana Hance and L. acalycina Chang. Small forests and individual trees of either or both of these two species are rather common in north- eastern Guizhou. Liquidambar acalycina is con- sidered to be closely related to the North American L. styraciflua L. and has been placed in the same section by Chang (1979). ost recently, a treatment of the vegetation of B entire province has appeared (Huang et al., 988). This work, which the authors state is aimed at university teachers and students, conservation- ists, geographers, ecologists, foresters, technicians, and farmers, discusses the physical environment, regionalization of the flora, and main vegetation types and their evolution, and utilization of the forest types in conjunction with farming, forestry, and animal husbandry. CONSERVATION Conservation in China has recently been given high priority, and it is particularly noteworthy that the Fanjing Shan mountain range and forests have been designated a Man and the Biosphere Pre- serve. Unfortunately, however, habitat destruction is taking place at a rapid rate throughout China and even Fanjing Shan is not being spared. Trees around the margins of the preserve are actively being cut for conversion to charcoal, which is then transported several kilometers to villages at lower elevations, where it is sold. Other selected trees are being cut throughout much of the easily ac- cessible parts of the preserve to build houses and for other construction purposes. The herbaceous flora suffers from the activities of overzealous dig- gers of medicinal plants who sell their harvests in all of the larger cities and towns surrounding Fan- jing Shan. Tourism is being encouraged, and while most tourists remain close to the trails, their num- bers are becoming so great that they are having a very definite and negative impact on many of the areas they visit. Their presence has become even more noticeable with the introduction and spread of nonbiodegradable food and beverage containers throughout China over the past few years. The lack of adequate sanitary facilities will undoubtedly have a deleterious effect on water resources. Unless these threats to the integrity of the area are ad- dressed in the very near future, the Fanjing Shan region will ultimately resemble much of the rest of rural China, and many species of both plants and animals will be lost locally, if not completely. GEOGRAPHICAL LIMITS, SOIL TYPES AND CLIMATE The Fanjing Shan mountain range is generally latitude and 108?45'55" 108?48'30"E longitude in northeastern Guizhou. The area is deeply cut on all sides by five deep, V-shaped valleys, resulting in a steep topography. The highest peak is about 2,570 m above sea level, and the relative elevation over the surrounding countryside ranges from 1,000 to 2,000 m. Four major soil types have been described in this region: (1) yellow-red earth, below 700 m; (2) yellow earth, (500-)700-1,400 m; (3) yellow-brown earth, 1,400-2,000 m; (4) mountain brown podzolic soil, above 2,200 m. АП exhibit an acid reaction, with pH values lower than 5.0. Zhang & Zhang (1982) also reported a soil type, which they call a dark, mountain thicket soil, different from the mountain yellow brown earth and the soil in shrub meadows, in a zone between 2,000 and 2,300 m. An analysis of the climatic features of the Fan- Volume 78, Number 2 1991 Ying et 341 g et al. Phytogeography of the Fanjing Shan jing Shan mountain area was made by Wang (1982), who classified the vertical climatic zones as middle subtropical, northern subtropical, south- ern temperate, and middle temperate. On the basis of these floristic, vegetational, and environmental studies in the Fanjing Shan moun- tain range, we are now able to draw some conclu- sions on the relationships of its highly diverse and interesting flora and to discuss to some extent the phytogeography of an important segment of the Chinese flora. COMPOSITION OF THE VEGETATION IN THE FANJING SHAN MOUNTAIN RANGE There are five forest zones from the base to the crest of the Fanjing Shan mountain range: (1) evergreen broad-leaved forest zone (below 1,300 m); (2) mixed evergreen-deciduous broad-leaved forest zone (1,300-1,900 т); (3) deciduous broad- leaved forest zone (1,900-2,100 т); (4) subalpine coniferous forest zone (2,100-2,350 m); and (5) subalpine shrub and meadow zone (above 2,350 m). The dominant genera in the forests are Cas- tanopsis (D. Don) Spach, Quercus L., Fagus, Acer L., Rhododendron L., and the gymnosperm gen- era, Tsuga Carr. and Abies Miller, while the dom- inant members of the subalpine shrub-meadow zone are Rhododendron spp., Sinarundinaria chungii (Keng) P. C. Keng, and Arundinaria fangiana C. Camus. Because the zone occupied by the latter association is so narrow, it is not an important floristic component of the Fanjing Shan range. The distribution of the dominant species in the other zones is as follows. EVERGREEN BROAD-LEAVED FORESTS The dominant species of this forest type are Castanopsis tibetana Hance, C. eyrei (Champion) Tutcher, C. fargesii Franchet, C. carlesii Hayata, Quercus stewardiana, Q. englergiana Seemann, and Q. dentata Thunberg var. oxyloba Franchet. Almost all of the dominant species have distribu- tions south of the Chang Jiang (Yangtze River) and are concentrated in central-southeastern China. Castanopsis fargesii, C. carlesii, and C. eyrei extend to Taiwan. DECIDUOUS BROAD-LEAVED FORESTS The deciduous broad-leaved forests contain the primary portion of the Fanjing Shan flora. The dominant members of the forests are Fagus lucida Rehder & Wilson, F. longipetiolata Seemann, F. engleriana Seemann, and Acer flabellatum Reh- der. The range of distribution of Acer flabellatum includes northeastern Yunnan, eastern Sichuan, western Hubei, northern Guangxi, northwestern Jiangxi, and Guizhou, an area where Acer is richly represented. The three species of Fagus were the subject of investigation in the early 1970s by Tsien et al. (1975), who found that they form distinct forests on Fanjing Shan. Each of the species is segregated according to slope exposure, with Fagus lucida on west- and east-facing slopes, F. engleriana on north- facing slopes, and F. longipetiolata on south-fac- ing slopes. Paleopalynological studies reveal that these Fagus communities have been in a contin- uous state of flux in response to changes in climatic conditions and that the three species have occupied various positions in relation to each other as cli- mates differed in the past (Kong et al., 1977). The present distribution of these three species of Fagus in the Fanjing Shan region corresponds to their latitudinal distribution, where F. longipetiolata is a more southern species, F. engleriana is northern, and F. lucida is a more centrally distributed plant. SUBALPINE CONIFEROUS FORESTS Among the important species in this forest type are Tsuga forrestii Downie, T. longibracteata Cheng, and Abies fanjingshanensis, with sporad- ically intermixed deciduous broad-leaved trees. Of the two species of Tsuga, one occurs in southern Sichuan, northwestern Yunnan, and western Hu- bei; the other extends from northeastern Guizhou through southern Hunan, northern Guangxi and Guangdong to Fujian. The distributional areas of these two species overlap in the Fanjing Shan and adjacent regions (Fig. 1). Абгез fanjingshanensis, undoubtedly a relict species, is endemic to Fanjing Shan at elevations between 2,100 and 2,350 m. As with the examples mentioned above, we find that nearly all of the dominant species in the flora of Fanjing Shan are endemics concentrated in cen- tral and southeastern China. SURVEY OF LARGER FAMILIES OF ANGIOSPERMS The largest families of flowering plants in the Fanjing Shan mountain range are the Rosaceae (66 species), Poaceae (54 species), and Fabaceae (51 species) (Table 1). The Rosaceae are a major family in this region and are very characteristic of the flora and vegetation of temperate areas in Chi- na. Poaceae and Fabaceae also exhibit an extraor- dinary frequency in forests and elsewhere in this Annals of the Missouri Botanical Garden EAR FRANZ 7502 8 FIGURE 1. Fanjing Shan r region. The development of the Fabaceae is mainly in the Papilionoideae, which has 41 species, while the Mimosoideae and Caesalpinioideae have only 2 and 8 species, respectively. There are three families with between 40 and 50 species: the Liliaceae (48 species), Asteraceae (45 species), and Lauraceae (41 species). The iaceae and Orchidaceae form the major part of ilis wealth of the petaloid monocots, and together com- prise nearly 9% of the total flora. The Lauraceae, with some 41 native species, comprise an important element of the flora. This family is tropical in nature, and even most genera in temperate regions have their distribution centers in tropical or sub- tropical locations. There are five families with between 30 and 40 species. They are the Orchidaceae (38 species), Lamiaceae (36 species), Fagaceae (33 species). Cyperaceae (33 species), and Ericaceae (30 spe- cies). The 33 species in 6 genera in the Fagaceae are important elements in the flora. Many members of the family, especially species of Castanopsis, Lithocarpus, and Quercus, are the dominant forest species, and Fagus, with five species in China and three on Fanjing Shan, is centered and well de- veloped in this region. Two families, the Ericaceae HUE 0,03 xc МШ Н 1977 9 ЛЕНИ ы of Tsuga forrestii (А) and T. longibracteata (Ф). Note overlap in their ranges in the and Lamiaceae, are relatively poorly represented, but three genera of Lamiaceae endemic to China (Bostrychanthera Bentham, Hanceola Kudo, and Rostrinucula Kudo) occur here. The above 11 families comprise nearly 34.8% of the species of flowering plants in the Fanjing Shan flora. There are nine families with between 20 and 30 species. [n order of size they are: Ranunculaceae (29 species); Apiaceae (29 species); Theaceae (28 species); Saxifragaceae (28 species); Scrophulari- aceae (24 species); Rubiaceae (24 species); Vita- ceae (22 species); Caprifoliaceae (21 species); and Celastraceae (21 species). Except for the Thea- ceae, Rubiaceae, and Vitaceae, these families are north and south ope or only north temper- ate, in distributio e large iain ed of archaic families in the flora of Fanjing Shan should also be mentioned. Their presence may be explained by the antiquity of the flora, although they are always represented by only one or a few species. The polycarpic Magnoliidae, generally considered to comprise the most primitive families and genera of flowering plants (Cronquist, 1981), are distributed mainly in eastern and south- eastern Asia and in North, Central, and northern South America; there are 13 species in 8 genera Volume 78, Number 2 Ying et al. 343 1991 Phytogeography of the Fanjing Shan TABLE 1. Ranking of families in the Fanjing Shan flora based on numbers of species. Number of Number of genera species Number of endemic Number of endemic genera! to China species? to China 1. Rosaceae 14/48 66/855 51 2. Poaceae 41/228 54/1,202 8 3. Fabaceae 33/163 51/1,252 22 22.3% 4. Liliaceae 21/55 48/335 25 5. Asteraceae 41/227 1 45/2,323 8 | 6. Lauraceae 8/20 41/422 24 34.8% 7. Orchidaceae 29/165 38/998 13 8. Lamiaceae 20/99 3 36/808 16 9. Cyperaceae 9/31 33/668 6 10. Fagaceae 6/6 33/281 22 | 11. Епсасеае 7/14 30/718 21 12. Ranunculaceae 12/40 29/736 14 13. Apiaceae 15/95 29/525 11 14. Тћеасеае 6/1 28/397 18 15. Saxifragaceae 14/26 28/440 19 16. Scrophulariaceae 12/60 24/634 10 17. Rubiaceae 19/75 1 24/477 13 18. Vitaceae 6/7 22/109 12 19. Caprifoliaceae 5/12 21/207 17 20. Celastraceae 5/13 21/184 12 Тота! 322/1,398 701/13,571 342 Percentages 50.2% 50.1%/46.73% 48.9%? у ' Nuaber of genera in Fanjing Shan/number in кы umber of species in Fanjing Shan/number in 3 1 Percentage of total species of larger families in Fan anjing Shan region. (Illicium L. (Illiciaceae) Liriodendron L., Mag- RELATIONSHIPS OF THE FANJING nolia L., Manglietia Blume, Michelia Г. (Mag- SHAN FLORA noliaceae), Kadsura Jussieu and Schisandra Mi- chaux (Schisandraceae), and Tetracentron Oliver (Tetracentraceae)), in this category within the Fanjing Shan area. The Hamamelidaceae are also The Fanjing Shan mountain range is very rich thought to be an archaic group; in the Fanjing Shan in genera of flowering plants, with 636, but is not region the family is represented by 13 species in especially rich in numbers of species, with only 6 genera. The main families of the old “Amentifer- 1,431 (2.25 species per genus). Based on the geo- ae," such as the Betulaceae, Fagaceae, Juglan- graphical distribution given for each genus in Wil- daceae, Moraceae, Salicaceae, and Ulmaceae, are lis's (1973) Dictionary of the Flowering Plants found on Fanjing Shan. In addition, three families and Ferns (eighth edition), and Mabberley's (1987) endemic or nearly endemic to China, and also con- The Plant-book, and supplemented by information sidered to be primitive or relictual, the Davidiaceae, gathered from specimens in the herbaria at A, GH, Eucommiaceae, and Sargentodoxaceae (also in НСА (Guizhou Academy of Sciences), Guizhou northern Laos and Vietnam), are also found here. | Normal University, and PE, we were able to classify The Fanjing Shan mountain range therefore har- the genera of flowering plants in the Fanjing Shan bors a great number of archaic or primitive taxa preserve into 14 distribution types (Table 2) as regardless of the criteria on which they are based. follows. The combined families mentioned above contain 701 species, or 51.1% of the total flora, and play an important role in shaping the characteristics of Some genera, mainly aquatics and mesophytes, the forests of the Fanjing Shan mountain range. are so widespread that they can be classified as A. GEOGRAPHICAL AFFINITIES OF THE GENERA 1. Cosmopolitan genera 344 Annals of the Missouri Botanical Garden TABLE 2. Distribution types of genera of seed plants in the Fanjing Shan region. Percent- Number of age of all genera in genera in Percent- China with China with Number of age of all Number of Percentage dis this dis- genera in genera in species in of all species tribution tribution Distribution type the region the region the region in the region pattern pattern Cosmopolitan 56 8.8 162 11.3 107 52.3 Pantropic 109 17.1 29] 20.3 373 29.2 Tropical America and tropical Asia 11 1.7 34 2.4 96 11.5 Old World tropics 36 5.7 63 4.4 162 22.2 Tropical Asia and tropical Australia 29 4.6 48 3.4 152 19.1 Tropical Asia and tropical Africa 23 3.6 31 2.2 160 14.1 Tropical Southeast Asia 60 9.4 114 8.0 570 10.5 North Temperate 112 17.6 332 23.2 297 37.7 Eastern Asia and North America 46 7.2 114 8.0 120 38.3 Old World Temperate 31 4.9 54 3.8 163 19.0 Temperate Asia 6 0.9 13 0.9 58 10.3 Mediterranean, western Ásia and central Asia | 0. 1 0.07 169 0.6 Eastern Asia 93 14.6 149 10.4 294 31.6 Chinese endemics _23 _3.6 25 1.7 214 10.7 Total 636 99.9 1,431 100.07 2,935 21.7 cosmopolitan. Fifty-six genera fall into this cate- ory. These taxa belong to 31 of the 143 native families in the Fanjing Shan region and include 162 species. These cosmopolitans comprise about 8.876 of the flora of Fanjing Shan, and 21.756 of the total families. Many mesophytes such as Senecio L., Aster L., Geranium L., Ranunculus L., Carex L., Polyg- onum L., Galium L., and Cyperus L. are the dominant species of the herbaceous layer in forests and in montane grasslands. Because information derived from the distribution of cosmopolitan gen- era makes it impossible to determine the geograph- ical affinities of the flora of the Fanjing Shan range, they are excluded from the statistical data. 2. Pantropical genera Genera found in all three tropical regions (the Americas, Africa- Madagascar, Asia- Australasia) may be considered pantropical. A typical example is Hypoxis L. Other pantropical genera are con- centrated in the tropics of one continent, but have few species in one or both of the other two regions. As an example, Nertera Banks & Solander ex Gaertner contains about 12 species, one common to New Zealand, South America, and Tristan da Cunha, one on Tristan da Cunha, one in Australia, one in Chile, one on Granada, two in the Philip- pines, one in south China, and eight in New Zealand (van Steenis, 1962; Wu & Li, 1965). Although the distribution of the genus might argue for the existence of Nertera before the breakup of the southern landmass, a more reasonable explanation is that members of the genus have been dispersed more recently by migrating birds that eat the berry fruits. In the Fanjing Shan region 109 genera, or ap- proximately 17.1% of the total genera (excluding the cosmopolitan ones), may be classified as pan- tropics. They contain about 291 species, which is 20.3% of all species in the flora of Fanjing Shan. Of the 109 pantropical genera in Fanjing Shan, about 52 (47.7%) are herbaceous 42 (38.5%) are woody, 4 (3.7%) consist of lianas, and 11 others include both herbaceous and woody species. Al- though the distribution centers of these genera are primarily concentrated in the tropics of both hemi- spheres, a few of them, such as Пех L., Cynan- chum L., Impatiens L., Buddleja L., Celastrus L., Cuscuta L., Euphorbia L., Indigofera L., and Grewia L., extend to subtropical or even temperate regions. 3. Tropical America and tropical Asia This distribution type is represented by propor- tionally few genera, especially when compared to those with a pantropical distribution. Only 11 gen- era in 10 families, representing only 1.7% of the total genera, are included in this category. These genera, all of which are woody, contain 34 species, Volume 78, Number 2 1991 Ying et al. 345 Phytogeography of the Fanjing Shan or 2.476 of the total flora of Fanjing Shan. In Australasia some of these genera may extend to northeast Australia or the southwest Pacific Is- lands, but their distribution centers are always lim- ited to the American and Asian tropics. In China, several genera extend into temperate regions, e.g., Picrasma Blume, Sageretia Brongniart, and Lin- dera Thunberg, the latter to Liaoning province at lower elevations. All three, however, are concen- trated in south-southwest China. Similar examples are found in Litsea Lamarck, but the genus attains its greatest diversity in southern and southwestern China, which has been considered to be its place of origin (Li, 1979). In North America, Lindera extends into the New England states and just reach- es southern Canada; Litsea is found as far north as southeast Virginia and Tennessee. The American members of these two genera may not, therefore, be considered strictly tropical. These genera are common in the Fanjing Shan region and many of their species are the main constituents of the ev- ergreen and deciduous broad-leaved forests. 4. Old World tropics This distribution type is difficult to separate from the subtropical and sometimes even temperate pat- tern. We have segregated these genera mainly on the degree of their concentration in the tropics of the Old World. An example is Dopatrium Bu- chanan-Hamilton ex Bentham, which has about 12 species in the Old World tropics (Mabberley, 1987), but only one species, which extends as far north as Henan province, in China. Within the flora of Fanjing Shan, 36 genera (5.7%) may be classified as Old World tropical. Of these, 14 (38.9%) are woody, 14 (38.9%) are herbaceous, 3 (8.3%) are lianoid and 5 genera have both woody and her- baceous species. Among them, Pittosporum Banks ex Gaertner, Albizia Durazzini, and Mallotus Loureiro are constituents of broad-leaved forests, while others such as Maesa Forsskal and Embelia Burman f. are also rather common, but only in the shrub layers of the forests. 5. Tropical Asia and tropical Australia In one sense this category is a subset of the Old World tropical pattern in that it includes genera with much the same kind of distribution, but with the main areas of concentration in the eastern and southeastern portion of the Old World tropics rath- er than in the west. The western extremity of this distribution type extends to Madagascar, but not to the African mainland. Ап example is found in Balanophora Forster & G. Forster in which 80 species occur in Australia, Polynesia, Malaysia, South China, Japan, and и but none occur on the African continen Twenty-nine genera (4.6% v all genera) and 48 species (3.4% of all species) can be placed in this category. Among the genera, 10 are woody, 4 are herbaceous, and 5 are lianoid; only one genus (Dunbaria Wight & Arn.) has both woody and herbaceous species. The genus Microcarpaea R. Brown, with only one species, is a typical rep- resentative of this distribution type. In China these genera occur predominantly in southern China, but a few of them enter the temperate zone. Some, such as Elaeocarpus L., Toona (Endlicher) M. Roemer, Wikstroemia Endlicher, and Nothopa- nax Miquel, are important constituents of sub- tropical forests and shrublands. Nothopanax has a disjunct distribution in southwest China and New Zealan 6. Tropical Asia and tropical Africa This category is also a subcategory of the Old World tropical pattern in that it includes genera with much the same kind of distribution, but with their centers of concentration in the west and south- west rather than in the east. The eastern extremity of this distribution type extends to Fiji in the South Pacific, but not to Australia. A good example of this pattern is seen in Lecanthus Wedd., where three species occur in tropical Asia, one in tropical Africa, and one in Fiji, but none occur in Australia. A similar example is found in Girardinia Gaudi- chaud-Beaupre, which also extends into subtropical to temperate China. The genus Crassocephalum Moench contains 30 species, 29 of which are re- stricted to Africa and Madagascar, but one ag- gressive species occurs in eastern Asia. In the Fanjing Shan region, 23 genera (3.6% of the total) belong to this distribution type. These genera contain 31 species (2.2% of the total spe- cies). Among the genera, 6 are woody, 10 are herbaceous, 3 are lianoid, and 4 contain both woody and herbaceous members. Except for Debregeasia Gaudichaud-Beaupre and Miscanthus Andersson, all of them are uncommon in this region. The monotypic genus 7oddalia Jussieu is a typical rep- resentative of this distribution type. 7. Tropical southeast Asia The Indo-Malesian region has one of the richest floras in the world. The eastern portion of the Indian region, and the northwestern portion of the Malesian region, through the Malay and Indo- chinese peninsulas, are very closely linked floris- tically with the mountains of south-central China. 346 Annals of the Missouri Botanical Garden Thus, a fairly large number of “typical” Indo- Malesian genera occur in south China. In the Fan- jing Shan region, 60 of the 636 genera (9.4%) are included in this category. These genera contain about 114 species and constitute 8.0% of all the species in the flora. Among these 60 genera, 33 (55%) are woody, 17 (17%) are herbaceous, 6 (10%) are lianoid, and only 4 genera contain both woody and herbaceous species. Some of the woody genera, such as Phoebe Nees, Machilus Nees, Neolitsea (Bentham) Merrill, Camellia L., Schima Reinwardt ex Blume, and Adinandra Jack, Daphniphyllum Blume and Carrierea Franchet are the main constituents of evergreen, deciduous, and mixed broad-leaved forests in the Fanjing Shan area. Ophiorrhiza L. and Dichroa Loureiro are rather common in the herb and shrub layers of these forests. 8. North Temperate Included under this heading are all the widely distributed genera restricted to the temperate regions of Eurasia and America. There are also a few genera that are equally characteristic of North Temperate areas and parts of the tropics, especially in the more mountainous regions of the latter. For example, the genus Vaccinium L. is widespread in temperate Eurasia and America, but some species extend into tropical regions of Asia and America, especially into tropical mountainous regions. Other examples are Juglans L., Cotoneaster Medikus, Artemisia L., Rosa L., Thalictrum L., and Sam- bucus L. In the Fanjing Shan area the genera belonging to this category total 112, or 17.6% of the total genera enumerated. They represent the highest proportion of distribution types in the flora of Fanjing Shan. These genera contain over 330 species, nearly 25% of all species in this area. Among them are 42 woody and 64 herbaceous genera, 8 genera with both woody and herbaceous species, and only 1 lianoid genus. This category forms the nucleus of the flora of the Fanjing Shan mountain range. Among the es- pecially prominent and significant genera in terms of their dominance in the vegetation of the region are Acer, Fagus, Betula L., Carpinus L., Alnus Miller, Cornus L., Castanea Miller, Juglans, Sor- bus L., Prunus L., Tilia L., Rhododendron, Пех, and Fraxinus L. Most of these genera contain only deciduous trees and all of them are typical North Temperate genera. Within this category, however, 17 of the 112 genera occur disjunctly in both North Temperate and South Temperate regions, while Coriaria L., a common shrub genus, shows a dis- continuous distribution in the Mediterranean area, eastern Asia from the Himalayas to Japan, New Zealand, Mexico, and Chile. 9. Eastern Asia and North America Among North Temperate disjunction patterns, the one between eastern Asia and North America is perhaps the most important biogeographically for the amount of interest it has generated. It has been one of the best known since Asa Gray (1846) first drew special attention to it, and it has been much studied and discussed (Boufford & Spong- berg, 1983; Graham, 1972a, b; Hara, 1952, 1956, 1972; Hu, 1935, 1936; Li, 1952; Wu, 1983). In the flora of Fanjing Shan 46 genera, which represent 7.2% of the total genera, belong to this distribution type. These genera contain 114 spe- cies, or 8.0% of all species in the region. Among these genera, 23 (50% of this distribution type) are woody, 17 (37%) are herbaceous, 4 (8.7%) are lianoid, and only 2 genera (4.3%) have both woody and herbaceous species. Although genera with this distribution pattern are concentrated in eastern Asia and North America, some of them (such as /tea L., Nyssa L., Castanopsis, and Lith- ocarpus) extend into the Indo-Malesian region, or occasionally even into Central Asia (Veronicastrum Moench), or Australia (Lespedeza Michaux) in the Old World, and into tropical regions in the New World (Catalpa Scopoli, Muhlenbergia Schreber, and Illicium). Tsuga and Castanopsis are impor- tant constituents of hemlock forests and broad- leaved evergreen forests, respectively. Lithocar- pus, Liquidambar, Hydrangea L., Illicium, Sas- safras Nees & Ebermaier, Photinia Lindley, and Aralia L. are also major constituents of evergreen and broad-leaved deciduous forests in this region. Two additional genera, Абеба R. Brown and Cley- era Thunberg, are also rather common; they are disjunctly distributed in eastern Asia and Mexico. 10. Old World Temperate Discussed here are genera widely distributed over the North Temperate Zone of the Old World. At both ends of their range, and more especially in the west, many of these genera, such as Dip- sacus L., Inula L., Lactuca L., Lotus L., Oenanthe L., and Peucedanum L., often show a tendency to extend southward into Africa. Two additional gen- era, Ligustrum L. and Daphne L., although they range throughout the Old World Temperate Zone, may also be mentioned here because, respectively, they also have one or a few species in Australia and in the Indo-Malesian region. In the flora of Fanjing Shan, 31 genera, or 4.9% of all genera, belong to this distribution type. These Volume 78, Number 2 1991 Ying et al. 347 Phytogeography of the Fanjing Shan genera contain 54 species, or about 3.8% of the total. Among the genera, 6 are woody, 22 are herbaceous, and only 3 have both woody and her- baceous members. The woody genera are not the main constituents of the forests of Fanjing Shan, but the herbaceous genera are the main constitu- ents of the montane grasslands or of the herb layers of the forest. 11. Temperate Asia This type is represented by proportionally few genera, especially when compared with the North Temperate type. Included here are all the genera mainly distributed in Temperate Asia. Their dis- tributional areas sometimes extend southward to the more elevated regions of the subtropical zone. For example, Campylotropis Bunge (Fabaceae) contains about 45 species, of which 29 are dis- tributed in China, and pou all of which are concentrated in the pical regions of the coun- try (Fu, 1987). These species show their temperate nature, however, in their ecologically restricted occurrence to areas with elevations between 1,000 and 3,000 m. We have segregated this group of genera mainly on the basis of the a pref- erence of their subtropical representative n the flora of Fanjing Shan this E type is represented by only 6 genera, or 0.9% o all genera. Together they contain about 13 species, which comprise only 0.9% of all species. With the exception of one woody genus, Campylotropis, all are herbaceous. 12. Mediterranean, western Asia to central Asia and China; North America to Central America In the flora of Fanjing Shan, only Pistacia L. belongs to this distribution type. 13. Eastern Asia This category is a large one calculated to contain about 93 genera (14.6% of the total) and is rather difficult to distinguish from the temperate Asian distribution type. The genera occupy smaller rang- es and are essentially distributed in China and Japan. These genera can be roughly divided into three subgroups based on their distribution centers in relation to Guizhou. One includes genera extending from the Himalayan region across China to Japan; the other two subgroups have eccentric distribution centers, either westward or northeastward. Mem- bers with a westward distribution show a Sino- Himalayan range, from China to the Himalayas; the northeastward group shows a Sino-Japanese pattern, from China to Japan. Among the primarily Eastern Asian genera, about one-fifth belong to the Sino-Himalayan-Japan subgroup, one-fifth to the Sino-Himalayan subgroup, and three-fifths to the Sino-Japanese subgroup. These genera sometimes extend southward into Burma, Indochina, and/or the Malay Archipelago. Among the genera distributed from the Hima- layan region to Japan are: Choerospondias B. L. Burtt & A. W. Hill, Dichocarpum W. T. Wang & Hsiao, Hovenia Thunberg, Меша D. Don, Houttuynia Thunberg, Euscaphis Siebold € Zuc- carini, Pterostyrax Siebold & Zuccarini, Liriope Loureiro, Boenninghausenia Reichenbach ex Meisner, and Helwingia Willdenow. Of the 93 genera in this category in the flora of Fanjing Shan (14.6% of the total genera), 47 are herbaceous, 38 are woody, and 8 are lianoid. The woody genera, such as Cercidiphyllum Sie- bold & Zuccarini, Асатћорапах (Decaisne & Planchon) Miquel, Dendrobenthamia Hutchinson (= Cornus L. subg. Syncarpea (Nakai) 0. Y. Xi- ang) Enkianthus Loureiro, /desia Maximowicz, Platycarya Siebol Zuccarini, Pterocarya Kunth, and Tetracentron Oliver, constituents of the broad-leaved forests, while the herbaceous genera, such as Ophiopogon Ker-Gaw- ler, Liriope, Reineckea Kunth, Tripterospermum Blume, and Ainsliaea DC., play an important role in the herb layer of the same forests. Sinarundi- naria Nakai is the main constituent under trees in the broad-leaved deciduous forests and in subalpine scrublands. are important 14. Endemic genera For the Fanjing Shan mountain range as a whole, 23 of the 636 native genera (3.6%) are endemic to China, but all of them are shared with other provinces. These endemic genera are listed with their ranges and number of species in Table 3. Twelve of these genera are monotypic and nine are oligotypic; the two remaining genera have mul- tiple species. Among the 23 are 11 arborescent genera, of which nine are deciduous and two are evergreen. Eleven genera are herbaceous and only one, Clematoclethra Maxim., is lianoid. Except for three genera, Eurycorymbus Handel-Mazzetti (Sapindaceae: 1 species), Dysosma Woodson (Ber- beridaceae: 7 species), and Cunninghamia R. Brown (Taxodiaceae: 2 species), which occur in Taiwan, all are restricted to central and south- eastern China. Based on the figures cited above we were able to conclude that among the native genera in the flora of Fanjing Shan, 268 (42.1%) are tropical, 348 Annals of the Missouri Botanical Garden TABLE 3. Chinese endemic or near endemic genera in the flora of Fanjing Shan with overall distribution of the genus. The second column indicates the number of species in Fanjing shan/number of species in Chi Asteropyrum (Ranunculaceae 1/2 Guangxi, Guizhou, Hubei, Hunan, Sichuan, Yunnan Bretschneidera (Bretschneideraceae) 1/1 Guangdong, iine о Hunan, Jiangxi, орнай, Tai- an, E Yunnan, Zhejia Bostrychanthera (Lamiaceae) 1/1 m Cannas Guangxi, Guizhou, Hubei, Sichuan, Tai- Chimonanthus (Calycanthaceae) 1/2 bu Fujian, Gansu, Guangxi, A Henan, Hubei, Hun- an, Jiangsu, Jiangxi, Shaanxi, Sichua Camptotheca (Nyssaceae) 1/1 Fujian, Hubei, Hunan, Guangdong, us Guizhou, Jiangxi, Sichua Clematoclethera (Actinidiaceae) 1/4 Guizhou, Hubei, Shaanxi, Sichuan, Yunnan Cyclocarya (Juglandaceae) 1/1 Anhui, Fujian, iq dd Ie Guizhou, Hubei, Hunan, Jiangxi, Sichuan, Zhej Cunninghamia (Taxodiaceae) 1/2 South of the Qinling am Davidia (Nyssasaceae) 1/1 Guizhou, Hubei, Sichuan, N Yunnan. Dysosma (Berberidaceae) 3/7 South of the Qinling range. Dickinsia (Apiaceae) 1/1 Guizhou, Hubei, Hunan, Sichuan, Yun Dipteronia (Aceraceae) 1/2 S Gansu, Guizhou, Henan, W Hubei, эче и Sichuan, SE Yunnan. Eucommia (Eucommiaceae) 1/1 Eto sw China. Еотесоп (Papaveraceae) 1/1 Fujian, Guangxi, Guizhou, Hubei, Hunan, Jiangxi, Sichuan. Emmenopterys (Rubiaceae) 1/1 o SW China (and India). Eurycorymbus (Sapindaceae) 1/1 Sand SW China, Taiwan. Hanceola (Lamiaceae) 1/8 Guangxi, Guizhou, Hunan, Sichuan, Yunnan. Latouchea (Gentianaceae) 1/1 Fujian, on Guizhou. Pteroceltis (Ulmaceae) 1/1 N to S and SW China. Rostrinucula (Lamiaceae) 1/2 Guangxi, Guizhou, Hubei, Hunan, Shaanxi, Sichuan, Yunnan. Sinojohnstonia (Boraginaceae) 1/3 Guizhou, Hubei, Jiangxi, Shaanxi, Sichuan, Zhejian Thyrocarpus (Boraginaceae) 1/3 Anhui, Fujian, Guangdong, Guangxi, Guizhou, Hubei S Jiang- su, Jiangxi, E Yunnan Whytockia (Сезпепасеае) 1/3 Guizhou, Taiwan, SE Yunnan. 289 (45.476) are temperate, and 23 (3.6%) are Chinese endemic genera. It is clear in the numbers and percentages that tropical and temperate genera are of about equal importance in the flora of Fanjing Shan. Taking a more general view, however, and looking for ultimate sources, it appears that the flora of Fanjing Shan can be thought of as com- prising two main elements, a northern and a south- ern one, disbursed, as would be expected, along an altitudinal gradient in five vegetational zones (as discussed above). B. RELATIONSHIPS WITH OTHER MOUNTAINOUS REGIONS IN CHINA To determine the area's floristic affinities within China, eight mountainous regions in the central and south-central part of the country (Fig. 2) were selected for comparison with the Fanjing Shan re- gion. Based on the worldwide distribution of the genera in the nine mountain regions, 15 distribution patterns can be recognized (Table 4). As expected, temperate genera increase and tropical genera de- crease in numbers with increasing latitude. Fanjing Shan, with 268 tropical and 289 temperate genera (excluding the Chinese endemic genera), has about an equal representation of temperate and tropical elements. Taibai Shan, located to the north of Fan- jing Shan in Shaanxi province, has 397 (68.0%) temperate and 105 (18.0%) tropical genera; cos- mopolitan genera (60, or 10.2%) and endemic genera (22, or 3.8%) make up the remaining 14%. Wuzhi Shan has the highest proportion of tropical genera (88.8%) of any of the areas examined, which would be expected from its southerly location on Hainan Island at about 19°N latitude. The num- bers of shared genera and the coefficients of sim- ilarity between these mountainous regions are given Shennongjia (D) (50.6%), and the Fanjing Shan region and Jinfo Shan (C) (50.7%) indicate closer phytogeographical links between those four areas 349 Phytogeography of the Fanjing Shan Ying et al. Volume 78, Number 2 1991 = 9'61/171 8'6/€TI 6'ct/Tt6 6'c€£/60€ vT'SI/09I L'L6/t66 пецс 1421 AA 1 6 S£/026 c 9€/0T€ с Yc/L8TI 8' 66/666 O 26/lYvG чецб enxguo[n X H = O'CE/IPE 8 1G/PLI 875/595 L'66/6S6 чечб reqre p О L'SCc/6L6 Dl'66/60€ 6'I£/19€ емедеием A == 6'86/156 C 9Tv/£TE£ UBUS 1901A d v'9b/81b efSuouueyg ( 0'1P/60€ ueyg ojutf 2 ueys Кам Я ueys Зш(ие y V I H 9 A a d Я “Кулери Jo уџагоцјаоз /елэиэ8 pa1eus jo 1aquinN ви) ut suordo1 urejunour io[eur әши jo Аплејпшв jo әәлдә] ‘S яту, си и — = 0 © —ч со t- + + — из из $5 01 61 SG c8 L8 86 TE IP 01 9 01 61 66 86 Iv OF 15 £I 0c SẸ £I СТ 81 £6 6c 9€ ueys enxZuo[n X ueuc 190€] uw Opa ueys Зи(ие y ¿UN 619098 Ма OL VO ЭМИ УІ ALO волу rre [eo], eıaua3 зјелодшој, eJoua3 [eotido1], uotdo1 urejuno] 9191589 = УЗ *tuejodoursoo = SOD ‘esy [94495 = ту Iseaymog pordon = JSL ‘ецедѕпу [eordog pue esy [eordon = ПУ, uzy pordon pue etsy [eordo1] = Jy L ‘uuy pedon pue esy peoidon = үү, ‘ву ојелодшој yy ЧеоЧоцие4 = N yg ‘dwa ром PIO = ALO eoidon pom PIO = LO ‘duaj, U110N = LN (ту [едиәо pue иләззәм-ивәивләнрәр{ = ОМИ ‘вэмэшу U110N-BISY ulajsea = үм ‘эпиэрие = yq ту уд euyy jo ви01891 urejunoui зиш jo Selo au] ш e.19u93 941 jo sadAy џоцпа дар [ediouud 9u р TIBV L 350 Annals of the Missouri Botanical Garden | HMR — —TEBSA42 || ua gg mme 13 FicunE 2. flora is compared. — A. Fanjing Shan. — B. Wuyi p 110 D MAR 0.03 л SOE HOT 19774 9A 60096 Map of China showing location of Fanjing Shan and eight major mountainous regions with which its han. — =F, . Jinfo Shan. — D. Shennongjia. — E. Miaoer Shan. Nanjiabawa.— С. Taibai Shan. — Н. Yulongxue Shan. — I. Wuzhi Shan. than between Fanjing Shan and any of the other regions. The significance of the coefficient of sim- ilarity between Fanjing Shan and Wuyi Shan on the Fujian-Jiangxi border in eastern China is re- markable because these two ranges are nearly twice as far apart as Fanjing Shan is from either Miaoer Shan (Е) in northeastern Guangxi ог Jinfo Shan (in southeastern Sichuan), and about 25% farther apart than Fanjing Shan and the Shennongjia re- gion. Noteworthy, however, are the especially low coefficients of similarity between Fanjing Shan and Nanjiabawa (F) in Xizang (Tibet), Taibai Shan (C) in the Qinling Mountains in southern Shaanxi, Yu- longxue Shan (Н) in northwestern Yunnan and Wuzhi Shan (I) on Hainan, and the clear contrast between regions А to E and regions F to I. The second column in Table 5 shows that the соећ- cients of similarity between regions B, C, D, and E are still higher than they are with any of the other areas. The coefficients of similarity between I and С (9.8%) and I and Н (12.6%) are surpris- ingly low, attesting to the uniqueness of the Wuzhi Shan fl A comparison of the degrees of similarities with Ora. nine mountainous regions in China, based on the distribution of genera, shows that the floristic af- finities of Fanjing Shan are with Wuyi Shan in southeastern China and secondarily with Jinfo Shan and the Shennongjia region in central China. Phy- togeographically, the Fanjing Shan region is tran- sitional between the Sino-Himalayan and Sino-Jap- anese floras and contains an almost equal number of temperate and tropical genera. LITERATURE CITED BARTHOLOMEW, В., D. E. BOUFFORD, A. L. CHANG, 2. CHENG, T. В. DUDLEY, S. А. HE, У Li, J. L. LUTEYN, 5. А. Згомвенс, 5 C. Sun, Y. C. TANG, J. X. N& T. a н _ 1983. The 1980 E inns n bot to western Hubei Province, People” 8 dide a China. J. Ar- nold Arbor. 64 1- 103. BOUFFORD, D. Е. & S. A. SPONGBERG. 1983. Eastern sian- Eastern North American А relationships —a history from the time of Linn to the twentieth century. Ann. Missouri Bot. Со а. 70: 423-439. CHANG, Н. T. НЕА Liquidambar. Fl. Reipubl. Pop. Sin. 35(2): 54-5 CRONQUIST, А. An Integrated System of Classi- fication of Flowering Plants. Columbia Univ. Press, rk. New Yo Volume 78, Number 2 1991 Ying et al. Phytogeography of the Fanjing Shan DENG, F. L. 982. Survey = ipi Geology and Geo Preserve Institute, Scientific Survey of the Fanjing- shan Mountain Preserve. [In Chinese, English ab- stract. FANJINGSHAN MOUNTAIN PRESERVE INSTITUTE. 1982. papers in Chinese with brief English abstracts.] Ко, P. У. 1 . А study of the vr Campylotropis in China. Bull. Bot. Res. 7(4): English abstract.] GEELAN, P. J. M. & D. C. TwITCHETT (editors). 1974. The Times Atlas of China. Times Newspapers Ltd., 11-55. [In Chinese, London GRAHAM, А. (editor). 1972a. Floristics and Paleofloris- ti a and Eastern North America. Elsevier i ede 1972b. Outline of the origin and historical recognition of floristic affinities FENY Asia and 2: 135, 136. [Reprinted in ree 19725, and Stuckey, 1978.] GUIZHOU FLORA COMMITTEE. 1982. Guizhou Zhiwu Zhi ople’s Press of 956. Contributions to the study of variations in Japanese plants closely eue to those of Europe or North America. Part 1. J. Fac. Sci. Univ. Tokyo III. 6: 29-96. Part 2. Ј. Fac. Sci. Univ. Tokyo III. 6: 343-391. 1972. Corresponding taxa in North America, Japan and the Himalayas. Pp. 61-72 in D. Н. Val- entine (editor), art Phytogeography, and Evolution. Academic Press, London HuH.H. 1 . À pda of itis ligneous flora of China and eastern North America. Bull. Chinese Bot. Soc. 1: 79-97. The characteristics € affinities of Chinese dors. Bull. Chinese Bot. Soc. 2: 67-84. ‚ У. L. То & Г. YANG. 1988. панка of Guizhou. Guizhou People’s Publishing House, Gui- yang. [In Chinese. ] i L. YANG, S. Z. Fanc & J. L. 1982a. Vegetation in ко Fanjingshan Nature od serve. Pp. 93-130 in Fanjingshan Mountain Pre- serve Institute, Scientific Survey of the Fanjingshan Mountain Preserve. [In Chinese, English abstract. H Z. КА YANG & J. L. 1982b. Discovery of the Abies ud. forests and a 1 Mountain Preserve Institute, Scientific Survey of the Fanjingshan Mountain Preserve. [In Chinese, English abstract. Коке, Z. C., J. Hsu & М. С. Du. 1977. On the discovery of paleobotany i in eastern Yunnan and west- ern Guizhou and its significance in botany and ge- ology. Chinese Quarternary, glacial, and geographi- cal reports. Geographical Press, Beijing. [In Chinese.] Li, H. L. 1952. Floristic relationships between eastern Contr. Gray Herb. 88: — —— а. C. Y. NM — а Н. Asia and acr IE America. Trans. Amer. Phi- 42: los. Soc 429. [Reprinted, with Forward and adicional pe ada. as a Morris Arboretum 1971 79. The a distribution of Chi- P Lauraceae plants. idi totax. Sin. 17(3): . [In Chinese, English su a D. J. 1987. The Plant =, Cambridge niv. Press, Cambridge РАКК, C. У. 1988. E of Polygonum section Е жт (Polygonaceae). Mem. New York Bot PEI, s. б pen nil еч in China. Harold L. Lyon Arboretum Lec id ber 13. Univ. of Hawaii Press, Honolulu, Haw SrEENIS, C. G. G. J. vaN. 1962. The land-bridge theory in botany, with particular reference to tropical plants. Blumea 11: 235-372. STEWARD, А. М. 1930. The = of eastern Asia. CHIAO. 1 = Rec ее ех- plorations in Kweichow. The China 1. 18: -143. Bonam R. 1978. Essays on n Aen Plant Geography from tl tury. Агпо Press, New York. TSIEN, C. P., T. S. YING, C. С. Ma, Y. L. Lr, C. S. CHANG & T. L. Mic. 1975. The distribution of beech forests of Mt. Fa E: Shan and its еси in plant geography. Acta Phytotax. Sin. | 5- 18. [In Chinese, ond abstract WANG, 2. D. ев Пра of the F ir: bi mountain area. Pp. 61–78 in Fanjingshan Mou Preserve Institute, PR Survey of the Fa mum shan Mountain Preserve. [In Chinese, English ab- stract. WILLIS, J. С. 1973. A Dictionary of the Flowering Plants and Ferns, 8th edition, revised by H. K. Airy Shaw. Cambridge Univ. Press, Cambridge. Wu, C. Y. 1983. On the significance of Pacific inter- continental discontinuity. Ann. Missouri Bot. Gard 77-5 (editor). 1984. Index Florae Yunnanensis, = umes ] & 2. People's Publishing House, Yunna 1965. Research Report of ir of Tropical and Subtropical Yunnan, Part 1: -146. Science Press, Beijing. oa L. 1983. The structure and dynamics of the лч stewardiana forest of the Fanjing- n Mountain. Acta Phytoecol. Geobot. Sin. 7: 204- vi 1982. Geomorphology and recent crustal movement of the Fanjingshan Mountain area. Pp. 49-60 in Fanjingshan Mountain Preserve Institute, Scientific Survey of the Fanjingshan Mountain Pre- serve. [In Chinese, English abstract.] YANG, Y. Q., Y. Y. Хо & У. Г. Tu. 1982. The Davidia ine rata forests in the Fanjingshan Mountain Pre- serve. Pp. 157-163 in Fanjingshan Mountain Pre- serve и ка на Scientific Survey of the Fanjingshan Mountain Preserve. [In Chinese, English abstract.] ZHANG, Е. Н. & М. ZHANG. 1982. Report of soil survey in the Fanjingshan Mountain area. Pp. 79-92 in Fanjingshan Mountain Preserve Institute, Scientific Survey of the Fanjingshan Mountain Preserve. [In Chinese, English abstract. 352 Annals of the Missouri Botanical Garden DIX. Genera in the Fanjing Shan flora and their APPENDIX. Continued. worldwide distribution. Number of species (approximate- ly): Fanjing Shan/ China/ worldwide. Viola L. 5/120?/500 Xanthium L. 1/5/30 1. COSMOPOLITAN Amaranthus L 1/13/40 . PANTROPICAL GENERA Апетопе 1. 3/52/150 Achyranthes L. 3/3/15 Apium L 1/1/1 Adenostemma Forst. Aster L. 2/130/500 & G. Forst. 1/2/20 Astragalus L. 1/130/2,000 Alchornea Sw. 1/6/70 Bidens L. 1/8/230 Ardisia Sw. 7/69/400 Carex L. 19/400/1,500-2,000 Aristolochia L. 1/51/300 Clematis L. 10/110/250 Arundinella Raddi. 1/11/55 Cyperus L. 2/30/550 Arundo 1/3/12 Digitaria Haller 1/20/380 Bauhinia L. 2/35/300 Eleocharis R. Br 2/25/200 Begonia L. 2/90/900 Eragrostis Wolf 1/35/300 Boehmeria Jacq. 8/35/50 rigeron 1/33/200 Brachiaria Griseb. 1/6/90 Galium L 1/50/400 Buddleja L. 4/45/100 Gentiana L 1/250/500 Bulbophyllum Thouars. 1/36/1,200 Geranium L 2/65/400 Bulbostylis Kunth 1/3/100 Goodyera К. Br 3/15/40 Caesalpinia L 2/16/100 Hippuris L 1/1/2-3 Calanthe R. Br. 4/40/120 Hydrocotyle L 3/15/75 Callicarpa L. 5/42/140 Hyperic 8/55/3170 Calystegia R. Br. 1/6/25 Juncus L 3/65/300 ] 2/22/550 Lemna L. 1/3/15 Celastrus L. 3/20/35 Lobelia L. 1/20/250 Celosia L. 1/3/60 Lysimachia L. 6/90/200 Celtis L. 3/20/60 Mimulus L. 2/5/100 Centipeda Lour. 1/1/6 Nymphoides Hill 1/6/20 issus L. 1/11/350 Oxalis L. 2/10/800 Clerodendrum L. 3/30/400 anicu 1/20/500 Cocculus D 1/2/11 Physalis L 1/5/100 Commelina L 1/7/100 Pimpinella L 2/40/150 nyza Less 1/8/50 Plantago L 1/16/265 Corchorus L 1/3/100 Polygala L 4/40/500-600 Crotalaria L 6/34/600 Polygonum L 12/120/300 ta 2/10/170 Potamogeton L 1/30/100 Cynanchum L. 4/52/200 Pycreus Pal 2/10/100 Cynodon Rich. 1/2/10 Ranunculus L 3/90/400 Dalbergia L. f. 2/25/100 Rhamnus L 4/59/160 Dendropanax Decne. Rorippa Scop. & Planchon 1/16/30 Rubus L. 26/280/600 Derris Lour. 1/20/40 Rumex L. 3/30/200 Desmodium Desv. 2/55/350 Sagittaria L. 2/6/20 Dioscorea L 6/49/600 Salvia L. 3/84/7700 Diospyros L 5/56/500 Sanicula L 3/22/37 Dolichos L 1/7/70 Scirpus L 1/37/300 Eclipta L 1/1/3-4 Scutellaria L. 1/98/300 Eleusine Gaertner 1/2/10 Senecio L 2/160/2,000 Erianthus Michaux 1/8/30 Solanum L 4/39/1,700 Eriocaulon L. 3/45/400 Sophora L. 1/23/50 Erythroxylum P. Spirodela Schleiden 1/2/6 rowne 1/2/250 tac Я 1/18/300 Euonymus L. 14/90/175 Stellaria L. 1/57/120 Eupatorium L. 81. 1/15/1,200 Swertia L 1/50/100 Euphorbia L. 81. 2/60/2,000 Teucrium L. 2/18/300 Ficus L. 10/120/800 Utricularia L. 1/17/120 Fimbristylis Vahl 14/47/200 Volume 78, Number 2 1991 Ying et al. Phytogeography of the Fanjing Shan 353 APPENDIX. Continued. APPENDIX. Continued. Gleditsia L. Glochidion Forst. Gouania Jacq. Haloragis Forst. & G. Forst. Hibiscus L. Hypoxis L. ex L. Impatiens L. Indigofera L. Kyllinga Rottb. Laportea Gaudich. Lasianthus Jack. Leersia Sw Lindernia All. Lycianthes Hassler Mariscus Vahl Millettia Wight & Arn. Murdannia Royle Nertera Banks & Sol. ex т Oplismenus Pal. Ormosia Jackson Ottelia Pers. Paspalum L. Pennisetum Rich. ex Pers. Perrottetia Kunth Phaseolus L. Phytolacca L. Pilea Lindley iper L. Pouzolzia Gaudich. Pratia Gaudich. Psychotria L. Rapanea Aublet Sacciolepis Nash Sapium P. Browne Schefflera Forst. G. Forst. Schoepfia Schreber Setaria Pal. Siegesbeckia L. Smilax L. Sporobolus R. Br. Styrax L. Symplocos Jacq. Ternstroemia Mutis ex L. f. Torenia L Trema Lour. 3/6/16 2/25/300 1/2/20 1/2/26 2/24/200 1/1/100 17/118/400 14/190/850 3/70/7700 1/25/500 3/15/75 1/15/50 1/44/300 1/6/60 2/16/23 4/32/180 1/4/15 3/26/100 1/9/200 1/7/200 4/30/180 1/8/80 3/18/50 1/1/12 1/3/15 1/35/120 1/8/40 1/10/250 1/8/70 1/3/20 1/15/50 1/4/35 6/65/400 1/35/2,000 1/5/50 1/3/35 1/15/700 2/7/200 1/3/30 2/9/120 2/37/200 1/3/25 4/17/140 2/3/6 11/61/300 1/6/150 4/55/130 11/125/350 1/20/100 2/11/50 2/6/30 о Triumfetta L. 1/8/150 Uncaria Schreber 1/13/60 1/4/6 Verbena L. 1/1/250 Vitex L. 1/20/250 Wahlenbergia Schrader ex Roth. 1/1/150 Zanthoxylum L. 5/50/250 Zizyphus Miller 1/13/100 . TROPICAL AMERICA AND TROPICAL ASIA Clethra L. 4/16/70 Gaultheria L. 1/26/210 Lindera Thunb. 8/54/100 Litsea Lam. 7/64/400 Meliosma Blume 4/10/25 Microtropis Wallich x Meissner 2/30/70 Mirabilis L. 1/1/60 Picrasma Blume 1/2/6 Sageretia Brongn. 2/14/35 Sloanea 3/14/100 Turpinia Vent. 1/10/30 . OLD WORLD Tropics Alangium Lam. 3/8/17 Albizia Durazz. 2/17/150 Alpinia Roxb. 2/26/250 Asparagus L. 1/24/300 Blumea DC. 1/30/50 Capillipedium Stapf 1/6/10 Cayratia Juss. 3/13/45 Centotheca Desv. 1/1/4 Clausena Burm. f. 1/10/25 Didymocarpus Wallich 1/47/120 Dopatrium Buch.-Ham ex Bentham 1/1/12 Elatostema Forst. & G. А 2/39/200 Embelia Burm. f. 2/20/130 Emilia Cass. 1/4/30 Eulalia Kunth 1/11/30 Euodia Lam. 3/25/45 Galeola Pour. 1/4/25 Gardenia Ellis 1/6/250 1 1/27/150 Maesa Forssk. 3/27/200 Mallotus Lour. 4/40/142 Monochoria C. Presl 1/5/6 Mussaenda 2/28/200 Osbeckia L. 1/12/100 Pharus P. Browne 1/9/50 Pittosporum Banks ex Gaertner 9/34/150 Pollia Thunb. 1/6/16 Porana Burn. f. 1/14/24 Rostellularia Reichb. 1/3/22 354 Annals of the Missouri Botanical Garden APPENDIX. Continued. APPENDIX. Continued. Smithia Aiton 2/4/30 Myrsine L 2/4/1 Stephania Lour. 3/30/50 Neyraudia Hook. f. 1/2/6 Syzygium = 1/72/500 Peristrophe Nees 1/5/30 Tinospora M 1/7/32 Premna L 2/45/200 Uraria Desv. 1/9/20 Ricinus L 1/1/1 Viscum L. 1/12/100 Strobilanthes Blume 2/20/250 Zehneria Endl. 1/12/30 Taxillus Tieghem 2/15/60 | : Themeda Forssk. 1/4/10 9. RUNE ASIA AND TROPICAL AUSTRALIA Thladiantha Bunge 3/29/30 Aglaia Lour. 1/10/250-300 Toddalia Tues 1/1/1 Ailanthus Desf. 1/5/10 Tricalysia A. Rich. Alyxia R. Br. 1/18/112 1/4/100 Balanophora Forst. ns Pm La | 2 Е 7. TROPICAL SOUTHEAST ASIA Proin ЕР. е Adinandra Jack 3/20/80 innamomum Schaeffer /250 Cleisostoma Blume 1/20/100 Aeginetia 1/3/10 Cudrania Trécul. 1/4-8/12 Aganosma G. Don f. 1/5/15 Dendrobium Sw. 2/63/1,400 OU киш. Dimeria R. Br. 1/4/40 G. Forst. 1/1/2 Dunbaria Wight 4 Агп. 1/7/25 Alniphyllum Matsum. ыл Elaeocarpus L. 2/38/200 Altingia Noronha 4/8/12 Eremochloa Buese. 2/4/10 Amentotaxus Pilger. 33 Eria Lindley 1/36/350 Amesiodendron Hu 1/2/2 Gastrodia R. Br. 1/3/20 Branatia Hook, Helicia Lour. 1/18/90 aan gei о Lagerstroemia L. 2/15/53 Briggsia Стар. - ыа Leptopus Decne 1/8/11 Broussonetia L'Herit. Mazus Lour. 1/22/30 idis ага Microcarpaea R. Br. 1/1/1 Camellia L. 6/190/220 Nothopanax Miq. emend. ampanumoea Blume 2/4/8 Бе 1/3/15 Carrierea Franchet 1/2/3 Кађаоза Hassk. 2/77/100 Chirita Buch.-Ham. Spathoglottis Blume 1/2/40 ex D. 1/40/80 Tetrastigma Planchon 3/45/90 Chlamydoboea Stapt ы Thrixspermum Lour. 1/10/100 Chloranthus Sw 3/15/35 a M. Roemer 2/3/15 Christisonia Gardner 1/1/17 Trichosanthes L. 3/40/55 Cipadessa Blume 1/2/3 Wikstroemia Endl. 1/40/70 oix 1/1/9 Zingiber Boehmer 2/14/85 Collabium Blume 1/2/8 Coptosapelta Korth. 1/1/13 6. TROPICAL ASIA AND TROPICAL AFRICA yclea Arn. ex Wight 1/13/30 Amorphophallus Blume оаа Blume 3/10/24 x Decne 1/25/100 Dichro ur 1/4/13 Arthraxon Pal 1/6/15 Рушт Siebold Asystasiella Lindau 1/1/3 Zuc 3/12/18 Crassocephalum Engelhardia Leschen. Moench 1/1/30 ex Blume 1/6/15 Debregeasia Gaudich. 2/4/5 Exbucklandia R. W. Br. 1/2/3 Dichrocephala L’Herit. Eurya Thunb. 12/60/130 А 1/3/10 Gastrochilus D. Don 2/12/22 Dregea E. Meyer 1/5/12 2/3/10 Girardinia Gaudich. 1/7/8 Gynostemma Blume 1/2/2 Gynura Cass 1/15/100 Hemiboea C. B. Clarke 4/5/8 Lecanthus Wedd. 1/3/5 Hemsleya Cogn 2/2/2 Microstegium Nees 2/10/30 Indocalamus Nakai 1/17/20 Miscanthus Andersson 1/6/20 Kadsura Juss. 1/7/22 Volume 78, Number 2 1991 t al. Phytogeography of the Fanjing Shan 355 APPENDIX. Continued. APPENDIX. Continued. e Lysidice Hance Machilus Nees Manglietia Blume Michelia L. Myrioneuron К. Br. ex Kurz. Neolitsea Merr. Oenanthe L. Ophiorrhiza L. Oreocnide Miq. Paederia L. Pellionia Gaudich. oebe Nees Phyllagathis Blume Pleione D. Don Pueraria DC. Reevesia Lindsey Sabia Colebr. Sarcandra Gardner Sarcococca Lindle Sarcopyramis Wallich Schima Reinw. ex Blume Sindechites Oliver Sycopsis Oliver Ypsilandra al NORTH TEMPERATE Abies Miller cer L. Асопџит L. Adenocaulon Hook. Aesculus L Agrimonia L. Alisma L Angelica L. Aquilegia L. Arenaria L. Arisaema C. Martius Artemisia Asarum L. Berberis L. Capsella: Medikus Carpinus Castanea Miller iaa Rich. Cerastium L. Cercis L. Chrysosplenium L. 1/2/2 7/68/100 1/19/32 2/35/55 1/4/15 4/40/80 3/11/30 2/25/150 1/9/20 2/11/50 1/16/50 3/9/15 1/2/3 1/2/7 1/4/5 1/18/50 12/100/200 3/160/300 1/3/3 1/2/10 1/1/5 1/8/13 1/4/15-20 1/3/10 1/110/700 2/8/40 3/50/100 1/60/100 3/26/80 1/9/100 1/60/250 5/82/150 6/170/350 3/30/70 7/160/450 3/29/100 3/11/70 1/42/55 Cimicifuga Wernisch. caea L. Cirsium Miller Clinopodium L. Coptis Salisb. Coriaria L. Cornus L. Corydalis DC. orylus Cotoneaster Medikus Deyeuxia Clarion ex Pal. Elaeagnus L. Epilobium L. Fagus L. Heracleum L. Tris L. Juglans L. roce iia R. Br. Ligustic Lilium L Listera R. Br. Lonicera L. L. Melandrium Rohl. Mentha L. Monotropa L. Morus L. Myrica L. Orobanche L. Ostrya Scop. inni i: L. Pinus L. унине: Rich. Polygonatum Miller Populus L. Potentilla L. iola L. Rhododendron L. 1/8/15 3/50/200 3/11/20 1/6/15 1/3/15 5/22/45 1/200/320 2/1/15 5/45/50 1/1/5 1/5/13 2/10/50-60 1/23/40 1/40/50-60 3/43/100 5/10/45 8/37/165 3/5/10 2/25/70 1/70/500 1/2/103 1/23/60 1/40/300 1/4/15-20 1/41/50 1/30/60 2/40/80 7/18/30 6/100/180 4/22/25 3/4/40 1/33/100 1/6/25 1/1/5 3/6/10 1/4/50 1/25/150 1/4/10 1/30/300 1/36/50 5/350/600 1/15/65 2/22/95 2/40/200 2/35/55 1/25/35 4/90/500 2/300/500 17/140/200 2/23/25 14/70/450 2/75/90 16/650/800 356 Annals of the Missouri Botanical Garden APPENDIX. Continued. APPENDIX. Continued. Rhus L 3/6/250 Ribes L 3/45/150 Rosa L 2/82/200 Rubia L. 2/11/60 Sabina Miller 1/14/50 Sagina L. 1/4/30 Salix L. 10/200/500 Sambucus L. /20 Saussurea DC. 5/320/403 Saxifraga L 1/180/370 dum 4/125/350 Solidago L 1/1/100 Sorbus L 8/55/85 Spiraea L. 6/50/100 Spiranthes Rich. 1/2/40 Streptopus Michaux 1/5/10 Thalictrum L. 5/67/150 Tilia 1/35/50-80 Trifolium L 1/7/300 Jlmus 1/23/45 Urtica L. 1/15/50 Vaccinium L. 5/47/450 Valeriana L 4/24/200 Veratrum L 1/13/25 Veronica L 1/64/250 Viburnum L 11/74/150 Vicia L 1/40/150 Vitis L. 7/25/60-70 9. EASTERN ASIA AND NORTH AMERICA Abelia R. Br. 2/6/30 Aletris L. 1/13/25 Ampelopsis Michaux 1/1/2 Amphicarpaea Elliot ex Nutt. 1/1/3 Antenoron Raf. 2/2/4 Apios Fabr. 1/6/10 Aralia L. 3/30/35-40 Astilbe Buch.-Ham ex D. Don 1/15/25 Berchemia Necker ex DC. 2/16/22 Cacalia L 1/50/80 Castanopsis Spach. 7/60/122 Caulophyllum Michaux 1/1/3 Cladrastis Raf. /5/6 Cleyera Thunb. 3/11/17 Disporum Salisb. 3/10/20 Hamamelis L. 1/2/6 Hugeria Small Hydrangea L. 15/45/80 Illicium L. 1/21/42 Itea L. 1/12/15 Leibnitzia Cass. 1/3/7 Lespedeza Michaux 4/25/40 Liquidambar L. 2/2/5 Liriodendron L. Lithocarpus Blume Muhlenbergia Schreber Nyssa L. Osmanthus Lour. Parthenocissus Planchon Photinia Lindley Phryma L. Pieris D. Don Pogonia Juss. Sassafras Nees erm. Schisandra Michaux Smilacina Desf. Tiarella L. Toxicodendron Miller Trachelospermum i Lemaire и arriere Veronicastrum Moench . OLD WORLD TEMPERATE Adenophora Fischer Ajuga Anthriscus Pers. Asyneuma Griseb. & Schenk. Carpesium L. Cucubalus L. Daphne L p Я Dendranthema Des Moul Dipsacus L. Elsholtzia Willd. Epimedium L. Fagopyrum Miller Hemerocallis L. Inula L. Lactuca L. Leontice L. Ligularia Cass. Ligustrum L. Melissa L. Nerium L. Origanum L. Paris L Peucedanum L. Pyracantha M. Roemer 1/1/2 9/70/300 3/6/35 4/30/90 4/40/80 3/5/1 2/6/100 1/6/10 2/15/17 1/6/8 4/9/15 4/40/60 1/1/1 1/6/10 1/3/10 1/2/3 1/19/25 1/18/30 1/1/5 3/15/20 6/10/20 3/5/14 1/14/20 2/49/60 1/18/40-50 1/2/20 1/1/51 4/10/20 1/1/1 2/35/70 1/17/30 2/8/15 4/33/35 2/14/21 2/8/15 1/11/15 1/20/200 1/40/118 2/3-5 3/100/150 5/38/51 1/3/100 1/3/4 1/2/4 1/1/15-20 3/16/20 2/30/170 1/43/100 3/7/10 Volume 78, Number 2 Ying et al. 357 1991 Phytogeography of the Fanjing Shan APPENDIX. Continued. APPENDIX. Continued. — — — ~ — о Roegneria K. Koch 1/70/90 Scopolia Jacq. 1/2/5 Tamarix L. 1/18/60 Torilis Adan. 1/2/12 Zelkova Spach 2/4/6-7 . TEMPERATE ASIA Bergenia Moench 1/6/10 о Bunge 1/50/65 Kalimeris Cas 1/7/20 Myriactis Less, 2/5/10 Trigonotis Steven 2/32/50 Tripterospermum Blume 6/20/35 . MEDITERRANEAN, WESTERN TO CENTRAL AND NORTH TO CENTRAL ASIA Pistacia L. 1/2/10 . EASTERN ASIA Acanthopanax Miq. 7/27/35 Actinidia Lindley 4/52/54 Ainsliaea DC. 2/45/70 Akebia Decne. 2/2/5 Amitostigma Schltr. 1/20/23 Aucuba Thunb. 1/3/4 Aulacolepis Hacket 1/2/4 Beesia Balf. f. & Sm. 1/2/2 Belamcanda Adanson 1/1/2 Bletilla Reichb. f 1/4/6 Bothrispermum Bunge 1/5/5 Cardiocrinum Lindley 2/2/3 Caryopteris Bunge 1/12/15 Catenaria Bentham 1/1/1 Cephalotaxus Sieb. 3/7/9 Zucc. Cercidiphyllum Sieb. ucc. 1/1/1 Choerospondias B. L. & A. W. Hill 1/1/1 Codonopsis Wallich 2/39/50 н Sieb. & Zuc 5/20/30 Crenasira Lindley 1/2/7 Decais k. & Thomson 1/1/1 Deinostema Yamazaki 1/1/2 Dendrobenthamia tch. 2/12/12 Dichocarpum W. T. Wang & Hsiao 1/9/18 Disporopsis Hance 1/4/4 Enkianthus Lour. 3/7/10 Euscaphis Sieb. & Zucc. 1/1/1 Gardneria Wallich 1/5/6 ASIA, inkgo L. Helwingia Willd. Hemiphragma Wallich oia LUN Hosta Tra id ‘Thunb. Hovenia Thunb. Idesia Maxim. Kalopanax Miq. Kerria DC Keteleeria Carriére Kummerowia Schindler Liriope Lour. о В. Вг. x Reichb. Ly coris Herbert Maddenia Hook. & Thomso Mosla Buch.-Ham. ex Maxim. Nandina Thunb. Nonocnide Blume Neillia D. Don opogon Ker-Gawler Oreocharis Bentham Patrinia Juss. Paulownia Sieb. ucc. Perilla L. Phtheirospermum isch. eyer Phyllostachys Sieb. Zucc. Pileostegia Hook. & Thomson Platycarya Sieb. ucc. Pogonatherum Pal. Prinsepia Royle Pternopetalum ranchet Pterocarya Kunth Pterostyrax Sieb. & Zucc. Reineckea Kunth Rhaphiolepis Lindley Rhynchospermum Reinw. Rohdea Rot Sargentodoxa Rehd. & Wils. Schizophragma Sieb. & Zucc. Semiaquilegia Makino 1/1/1 2/4/4 1/1/1 2/11/12 1/3/40 1/1/1 1/3/3 1/1/1 1/1/1 1/1/1 1/2/2 1/2/2 2/6/8 1/2/3 2/15/20 1/29/31 1/2/2 2/4/5 3/11/22 1/1/1 1/2/4 1/10/12 1/2/2 7/40/60 2/20/20 3/13/20 2/6/17 1/1/1 2/3/7 3/40/50 1/2/3 2/2/2 1/2/2 1/4/4 3/25/27 1/9/10 2/2/4 1/1/1 1/7/14 1/1/2 1/1/1 1/1/1 3/6/8 1/1/7 358 Annals of the Missouri Botanical Garden APPENDIX. Continued. Serissa Comm. Juss. Sinarundinaria Nakai Spodiopogon Trin Stachyurus Sieb. & Zucc. Stauntonia DC. Stranvaesia Lindley Tetracentron Oliver Toricellia DC. Tricyrtis Wallich Triplostegia Wallich ex DC. Tripterygium Hook. f. Tubocapsicum Makino Tupistra Ker-Gawler Weigela Thunb. Youngia Cass. 14. ENDEMIC TO CHINA Asteropyrum J. R. rumm. & Hutch. Bostrychanthera Bentham Bretschneidera Hemsley Camptotheca Decne. Chimonanthus Lindley Clematoclethra Maxim. unninghamia R. Br. e Minsk. Davidia Baillon Dickinsia Franchet Dipteronia Oliver sma Woodson Emmenopterys Oliver mecon Hance Eucommia Oliver Eurycorymbus Hand.- а22. Hanceola Kudo Latouchea Franchet Pteroceltis Maxim. Rostrinucula Kudo Sinojohnstonia Hu Thyrocarpus Hance Whytockia W. W. Sm. 1/2/3 2/10/23 1/1/1 2/2/2 2/2/3 1/3/3 1/6/10 4/8/10 4/22/25 1/4/5 1/1/1 1/2/3 1/5/11 1/2/2 1/4/4 1/2/2 2/16/25 1/4/12 1/40/50 1/2 REVISION DEL GENERO STREPTOSTACHYS (POACEAE-PANICOIDEAE), SU POSICION SISTEMATICA DENTRO DE LA TRIBU PANICEAE: Osvaldo Morrone? and Fernando Zuloaga? RESUMEN Los límites genéricos de Streptostachys Desv. son discutidos, tomando en consideración caracteres exomorfológicos S. ramosa. Streptostachys acuminata es tratada como Urochloa hus. acuminata y S. robusta como una especie dudosa dentro del género Ichnant ABSTRACT In this work the generic limits of Streptostachys Desv. are discussed, considering exomorphological and anatomical le inicia elt is mainly defined е 7 the thickening of EE Мга с ын е ligule . Thre нр of the anthecium and w boundaries of eae S. asperifolia, B ma crantha, iude a linear hilum, ted within the acuminata is treated ciliate or = aa M and S. r du s Urochloa acuminata and S. robusta is discussed d а as a eres, жен o Ichnanthus El género Streptostachys fue creado por Des- vaux en el año 1810 sobre la base de una especie, S. asperifolia Desv. Streptostachys fue estudiado por varios autores en los ültimos anos, incluyendo el análisis del genero en particular o como parte de floras regionales o de estudios de géneros de gramineas para todo el mundo. Renvoize (1984b) incluyó dentro de Streptosta- chys cinco especies en su tratamiento de las gra- mineas de Bahia y lo caracterizó por incluir plantas anuales o perennes, con espiguillas solitarias o apa- readas sobre inflorescencias racemosas, gluma in- ferior % a ?4 del largo de la espiguilla y separada de la gluma superior por un entrenudo corto. Zuloaga & Soderstrom (1985) trataron en Streptostachys a S. asperifolia, S. macrantha (Trin.) Zuloaga & Soderstrom, y 5. ramosa Zu- loaga & Soderstrom por tener ligulas ciliadas o ausentes, espiguillas engrosadas entre la gluma in- ferior y el antecio superior, cariopsis con hilo linear y antecio superior piloso y con papilas simples & Renvoize (1986) consideraron a Streptostachys como género monotipico, ubicando S. ramosa y S. macrantha en Panicum, no validando en el primer caso la combinación co- rrespondiente, y sin mencionar las especies des- criptas por Renvoize en 1984a, 5. robusta y S. acuminata. Estos autores Lata al género básicamente por poseer raquilla engrosada e in- curva por debajo de la ійѕегсіоп de la espiguilla; citaron la presencia de células fusoides en el género y lo relacionaron con Acroceras por poseer hilo linear. Watson 4 Dallwitz (1988) siguieron el criterio ' Deseamos expresar nuestra gratitud al Gerrit Davidse por la revisión crítica del manuscrito, su valioso aporte acerca de la presencia de eleosoma en S. asperifolia, y sus opiniones sobre la delimitación del género y especies dentro del mismo 2 Instituto de Botánica Darwinion, Casilla de Correo 22, 1642 San Isidro, Argentina. ANN. MISSOURI Вот. GARD. 78: 359-376. 1991. 360 Annals of the Missouri Botanical Garden de Clayton & Renvoize (1986), sin discutir las relaciones de Streptostachys ni mencionar la ubi- cación de las especies excluidas del mismo. Watson & Dallwitz describieron al género con ligulas cilia- das o ausentes, espiguillas comprimidas dorsiven- tralmente, con raquilla conspicua entre la gluma inferior y superior y entre la flor inferior y superior, antecio superior estipitado, glumas subiguales, flor inferior con o sin pálea y masculina o neutra; antecio superior endurecido y cariopsis con hilo linear. Anatómicamente caracterizaron а Strep- tostachys como no-Kranz y con células fusoides. Todos los autores previamente citados no han tomado en consideración la totalidad de las especies descriptas dentro del género para su delimitación en la tribu Paniceae, ni han analizado los caracteres exomorfológicos y anatómicos de las mismas. En el presente trabajo se estudian las entidades hasta ahora incluidas en Streptostachys desde el punto de vista exomorfológico e histofoliar, apor- tándose nuevos datos y discutiendose la delimita- ción de Streptostachys y su ubicación taxonómica dentro de la tribu Paniceae. MATERIALES Y METODOS F! estudio exomorfológico fue realizado sobre la base de materiales de los siguientes herbarios: BAA, CEN, СЕРЕС, Е, С, IAN, IBGE, K, M, MO, NY, P, R, RB, SI, SP, UB, US, Las fotografias de cortes transversales de hoja y epidermis fueron tomadas con un microscopio electrónico de barrido, siendo el aparato utilizado un Jeol JSM-25 SII, de la Facultad de Odontología (Universidad Nacional de Buenos Aires, Argenti- na). Para el estudio de los caracteres histofoliares se empleó la penúltima lámina de la innovación fértil, realizándose los cortes transversales y longitudi- nales en el tercio medio de la lámina. Se utilizó material proveniente de ejemplares de herbario, previo tratamiento con Contrad 70 (Schmid & Turner, 1977). Para la obtención de ambas epi- dermis se siguió el método de Metcalfe (1960). Los cortes fueron realizados a mano alzada y coloreados con azul de metileno al 1% en agua, y safranina al 1% en alcohol 80? o con safranina- fast green y montados en gelatina-glicerina o en bálsamo de Canada. células clorenquimaticas se realizaron macerados siguiendo el método de Jeffrey (Sass, 1940). Para la identificación de los cuerpos de silice y células Para la observación de las suberosas se utilizó respectivamente fenol (Met- calfe, 1960) y Sudán III (Sass, 1940). Para la detección de cristales se siguió el método propuesto por Sánchez & Casabona (1981). La determinación de almidón y su distribución se realizo mediante unas gotas de solución iodo-iodurada (Sass, 1940). La presencia de aceites en la espiguilla se detectó con el reactivo Sudán IV (Jensen, 1962). Para la caracterización de lignina se utilizó floroglucina clorhidrica (Johansen, 1940). Las observaciones y dibujos fueron realizados con un microscopio Wild 20 con cámara de dibujo. Para las descripciones se adoptó la terminologia propuesta por Ellis (1976, 1979). Los ejemplares utilizados para el estudio histofoliar de las especies han sido identificados con un * en la cita del material examinado. REsULTADOS Y DISCUSION De acuerdo con los resultados obtenidos en el presente trabajo se ha hallado una manifiesta va- riabilidad en las cinco especies revisadas. Conse- cuentemente, se han tratado de establecer carac- teres válidos para definir el género y cuales son las especies que pertenecen al mismo. Asi, se considera que el poseer ligula ciliada o ausente; espiguilla comprimida dorsiventralmente, con raquilla engrosada y manifiesta entre las glu- mas, antecio inferior y superior; antecio superior piloso y papiloso, y cariopsis con hilo linear, ocu- pando todo el largo de la cariopsis, son los carac- teres que permiten distinguir a Streptostachys del resto de los géneros afines pertenecientes a la tribu Paniceae. RAQUILLA Zuloaga & Soderstrom (1985), Clayton & Ren- voize (1986), y Watson & Dallwitz (1988), men- cionaron entre los caracteres distintivos de Strep- tostachys la presencia de raquilla engrosada. Este engrosamiento se encuentra tanto en 8. asperifolia como еп S. macrantha у S. ramosa; en las tres especies la raquilla se halla conspicuamente en- durecida y ensanchada entre la gluma inferior y el antecio superior, siendo los bordes inferiores involutos en S. asperifolia y rectos en S. ma- crantha y 5. ramosa. Al efectuar cortes trans- versales de la raquilla en 5. asperifolia se observó que la misma es hueca en la porción distal debajo el ant mientras que en 5. macrantha y 5. ramosa la lis se presenta maciza en toda su extensión a la madurez de la espiguilla (Fig. 1р Se ha detectado un eleosoma en la porción en- grosada de la raquilla en espiguillas maduras de 5. asperifolia, S. macrantha, y S. ramosa. A traves de un ensayo con Sudán IV se hallaron abundantes Volume 78, Number 2 Morrone & Zuloaga 361 1991 Revision del Streptostachys FIGURA l. Streptostachys macrantha. — А. Aspecto general.— B. Detalle de la región ligular. — С. Espiguilla, vista lateral. —D. Corte longitudinal de espiguilla: 1, gluma inferior; 2, gluma superior; 3, lemma inferior; 4, реа inferior; 5, lemma superior; 6, pálea superior; 7, cariopsis. — E. Flor inferior con estambres y lodiculas. — F. Pálea inferior. — C. Pálea superior con lodículas. — Н. Antecio superior, vista ventral. — I. Cariopsis, vista Баг. — J. Cariopsis, vista escutelar. — К. Cariopsis, vista lateral. (De Irwin & Soderstrom 7131.) 362 Annals of the Missouri Botanical Garden gotitas de aceite en la epidermis y tejidos hipoder- micos de la raquilla, principalmente entre la in- serción de la gluma inferior y superior. La presencia de este tejido en diásporas de gra- mineas ha sido extensamente correlacionada con la dispersion a través de hormigas (mirmecofilia), habiendo sido documentada la existencia de eleo- soma en diversos generos de gramineas (Davidse, 1987). Beattie (1983) señalo diversas hipótesis para el establecimiento de la mirmecofilia, entre las que incluye protección contra el fuego, dispersión a distancia, y el hecho de poseer los hormigueros un microhabitat rico en nutrientes. HILO Si bien no existen estudios generales sobre el tipo de hilo y sus relaciones evolutivas en las gra- mineas, en particular en la tribu Paniceae, Fil- gueiras (1986) y Sendulsky et al. (1987) consi- deraron la utilización del mismo de valor taxonómico. layton 4 Renvoize (1986) citaron hilo linear en la tribu Paniceae en los géneros Acroceras Stapf, Homolepis Chase, Louisiella Hubbard $ Leonard, Mesosetum Steudel, Oplismenopsis Pa- rodi, Streptostachys, Tatianyx Zuloaga & So- derstrom y en cinco especies de Panicum. n la presente revisión se considera importante el carácter de hilo linear para la delimitación de géneros dentro de las Paniceas, y se desestima el considerar a Panicum como un género variable en о que a este carácter se refiere. Asi, de las cinco especies citadas por Clayton 4 Renvoize (1986) para el género, las tratadas como Panicum га- mosum y Panicum macranthum pertenecen al genero Streptostachys, Panicum glutinosum Sw., y Panicum villaricense Mez a Homolepis (Zuloaga & Soderstrom, 1985). La restante especie con hilo linear tratado por Clayton & Renvoize (1986) en Panicum, P. pil- gerianum (Schweick.) Clayton, originalmente des- cripta para el genero Acroceras, fue segregada por Launert (1970) en el género Psilochloa, carac- terizandose el género por ser Kranz, del subtipo anatómico PS y con hilo linear. TIPO DE LÍGULA Y ORNAMENTACIÓN DEL ANTECIO SUPERIOR Webster (1988), en su tratamiento de los gé- neros de Paniceas de América del Norte, consideró entre los caracteres genéricos distintivos en la tribu el tipo de ligula, la presencia o ausencia de la misma, y la ornamentación y textura del antecio superior. La ligula es ciliada a nula en S. asperi- folia, S. macrantha, y S. ramosa, mientras que el antecio superior se presenta endurecido y piloso con papilas simples distribuidas regularmente en toda su superficie. ANATOMÍA FOLIAR Los caracteres epidérmicos de las especies ana- lizadas responden al dermotipo panicoide, el que se caracteriza por poseer micropelos bicelulares fusiformes, cuerpos de silice costales halteriformes o menos frecuentemente nodulares, aparatos es- tomáticos romboidales a subdiscoidales, y células cortas solitarias o en pares en la zona intercostal (Metcalfe, 1960; Prat, 1932, 1936; Tateoka et al., 1959). El estudio de los caracteres histofoliares de 5. asperifolia, S. macrantha, у 5. ramosa ha per- mitido observar discrepancias en lo que al sindrome de Kranz se refiere. De acuerdo con los caracteres histofoliares 5. macrantha y S. ramosa son dos especies Kranz, del subtipo anatómico MS por tener: — Presencia de una vaina mestomatica Kranz al- rededor de los haces vasculares, cuyas células poseen cloroplastos especializados. — Ausencia de vaina parenquimática. — Presencia de almidón en la vaina Kranz. — Cuatro tipos de haces vasculares, con haces vasculares de tercer y cuarto orden por debajo e las células buliformes. — Ров a tres células clorenquimáticas entre haces vasculares contiguos. —Células Kranz rectangulares, alargadas longi- tudinalmente en vista paraderma — Relación de isótopos estables del carbono (0'*C). К. Н. Brown (comm. pers.) reportó un д'*С de — 14.39% у 12.199 para S. macrantha y S. ramosa respectivamente. Hattersley & Watson (1976) establecieron un carácter para diferenciar los distintos subtipos fi- siologicos presentes en gramineas, distinguiendo el tipo XYMS., el cual se corresponde con el subtipo MS de Brown (1977) y Ellis (1977) por presentar las células Kranz en contacto directo con los ele- mentos del metaxilema. Se ha observado que este carácter es variable tanto en S. macrantha como en S. ramosa, hallándose, aún dentro de un mismo ejemplar, haces vasculares primarios con o sin се- lulas intermedias entre la vaina Kranz y los еје- mentos de metaxilema. Streptostachys asperifolia es, en cambio, una especie no-Kranz, con doce células clorenquimá- Volume 78, Number 2 Morrone & Zuloaga 363 g Revision del Streptostachys ticas entre haces vasculares contiguos, presencia de almidón en las células clorenquimáticas, haces vasculares con vaina mestomática y paren uimá- tica sin cloroplastos especializados, hallandose en la especie células fusoides y mesófilo laxo. Brown (19777) reportó para esta especie un valor de iso- topos estables del carbono de — 27.0%. El hecho de poseer S. asperifolia un tipo fo- tosintético y anatómico diferente al presente en S. macrantha y S. ramosa, podria implicar la exclu- sion de las dos ültimas especies de Streptostachys. Existen sin embargo en la tribu Paniceae otros géneros, como por ejemplo Alloteropsis y Pani- cum, que incluyen especies C, y C,. Idéntica va- riacion fue reportada por Ellis (1984) para el ge- nero Eragrostis Wolf, el cual incluye hasta el momento una única especie no-Kranz, E. walteri. Ellis (1984) senaló que todos los representantes de la subfamilia Eragrostoideae son Kranz, y que F. walteri podria representar un ejemplo de reversión en la evolución del sindrome de Kranz en la sub- familia. Este autor mencionó que un fenómeno similar puede haber ocurrido dentro de las Pani- coideas C,, senalando: owever, this also implies that caution is called for in emphasizing anatomical characteristics of the Kranz syndrome when mak- ing taxonomic decisions." Consecuentemente, se considera prematuro el establecimiento de un nuevo género sobre la base de caracteres anatomicos, existiendo uniformidad en los caracteres exomorfologicos, al menos hasta la obtención de nuevos datos cariológicos о em- briologicos. POSICION SISTEMATICA DE S. ACUMINATA У S. ROBUSTA Finalmente, y de acuerdo con la delimitación previa del genero Streptostachys, se deben excluir del mismo a S. acuminata y S. robusta por las siguientes razones: Streptostachys acuminata carece de raquilla engrosada, posee ligula membranáceo-ciliada y an- tecio superior transversalmente rugoso y crestado, con papilas verrugosas en toda su superficie. Апа- tómicamente es una planta Kranz, del subtipo PS, con numerosos cloroplastos especializados en la vaina parenquimática. Ellis (1977) y Prendergast & Hattersley (1987) distinguieron dentro de la categoria anatomica PS los subtipos NAD-me y PEP-ck, definiendo al ul- timo por poseer cloroplastos especializados centri- tarn fugos en las células Kranz, vaina K irregular (en particular en los haces vasculares secundarios), y pared tangencial externa marca- damente convexa en las células Kranz. Estos ca- racteres se observan en el material examinado de acuminata, aunque la posición de los cloro- plastos, si bien es aparentemente centrifuga, se debe considerar dudosa por las características del material examinado. Streptostachys robusta es una planta no-Kranz, con mesófilo sin células fusoides. La espiguilla ca- rece del tipico engrosamiento de la raquilla, posee ligula membranácea y laciniada en la porción api- cal, láminas pseudopecioladas, espiguilla compri- mida lateralmente, у gluma inferior % del largo de la espiguilla y con largos pelos hacia el ápice. Desafortunadamente el material en que se baso la nueva especie es inmaduro, por lo que no se pueden apreciar con claridad los caracteres del antecio superior. No obstante, fue posible ver en deter- minadas espiguillas dos pequenos apéndices, o au- riculas, membranosos en la base de la lemma su- perior. Este carácter, sumado al tipo de lámina, compresion de la espiguilla, y largo relativo y pi- losidad de la gluma inferior, permiten suponer que la especie puede pertenecer al género /chnanthus. Sin embargo, y dado I no fue posible relacionar a la misma con algu e las especies tratadas en la monografia de cel (1982, 1987), se ha con- siderado inapropiado efectuar una nueva combi- nación sobre la base de material incompleto, por 5 o que esta especie debe ser excluida del género Streptostachys y su posición considerada momen- táneamente dudosa. Los caracteres distintivos de las cinco especies incluidas hasta el presente en Streptostachys han sido resumidos en la Tabla 1 TRATAMIENTO TAXONOMICO di ig Desv., Nouv. Bull. Sci. Soc. Phi- ris 2: 187-190. 1810. ESPECIE TIPO: S. asperifolia Desv. Plantas perennes, cespitosas, con canas decum- bentes y ramificadas en los nudos inferiores a erec- tas. Hojas con ligulas ciliadas o ausentes; láminas ovado-lanceoladas a lanceoladas, planas, de pilo- sidad variable. Inflorescencias terminales laxas, con espiguillas solitarias o apareadas; panojas axilares presentes o ausentes. Espiguillas desarticulándose por debajo de las glumas, comprimidas dorsiven- tralmente, elipsoides, densamente hirsutas o gla- bras, con raquilla engrosada y manifiesta entre la gluma inferior y el antecio superior, incurva o no en la porción basal; gluma inferior 4-4 del largo de la espiguilla (1)3-5(7)-nervia, gluma superior y lemma inferior subiguales, gluma superior 5-7- 364 Annals of the Missouri Botanical Garden TABLE 1. Caracteres comparativos de las especies tratadas hasta el presente en Streptostachys. S. asperifolia S. acuminata S. macrantha S. ramosa S. robusta Raquilla + = + + - engrosada Hilo linear T ? ? Tipo de ligula ciliada a membranaceo- ciliada a ciliada a embranaceo- ausente ciliada ausente ausente laciniada Ornamentación del liso, piloso y transversal- liso, piloso y liso piloso y liso, sin papilas, antecio superior papiloso mentrugoso, papiloso papiloso glabro labro Auriculas = = = = Anatomia foliar no-Kranz Kranz, PS Kranz, MS Kranz, MS no-Kranz Células fusoides = = - - Pseudopeciolo Ausente Ausente Ausente Ausente Presente Compresión de la Comprimida — Comprimida omprimida Comprimida Comprimida espiguilla dorsiven- dorsiventral- dorsiventral- dorsiventral- lateralmente tralmente mente mente mente nervia, lemma inferior (3)5- 7-nervia; palea inferior conspicua, de tamano similar a la lemma inferior, glabra, a reducida o ausente; flor inferior masculina o ausente; antecio superior elipsoide, endurecido, liso, piloso, con pelos uni- o bicelulares sobre toda la superficie, con papilas simples distribuidas re- gularmente; lodiculas 2, truncadas, conduplicadas, abrazando los bordes inferiores de la рајеа; estam- bres 3; estilos libres, estigmas plumosos; cariopsis con hilo linear ocupando todo el largo de la ca- riopsis, embrión menos de Yê del largo de la ca- riopsis. Genero americano con tres especies, incluyendo, además de la especie tipo, a 5. macrantha y S. ramosa. CLAVE DE LAS ESPECIES la. Panoja terminal y axilares numerosas saliendo en el extremo de las canas. Espi- los bordes inferiores involut 74-*4, del largo de la es piguilla, Planta no-Kra con células fusoides : asperifolia Panojas axilares ausentes. Espiguillas de 5. de largo, raquilla con los bordes it. riores rectos. Gluma inferior Ye % del largo la ig ie Plantas Kranz, MS, células e es ausen . Plantas A 30- 70 ст de alto. Laminas de 7- 7 N £e no hundidos y no cu- diim por papila S. macrantha s de 1.3- 2; m de alto. Láminas de 35- de ancho. Panojas ко E con zonas costales ostales marcadas; estomas abaxiales SNC y cubiertos ~ pa- pilas . S. ramosa l. Streptostachys asperifolia Desv., Nouv. Bull. hilom. Paris 2: 190. 1810. Pan- icum asperifolium (Desv.) A. Hitchc., Contr. U.S. Natl. Herb. 22: 489. 1922. Panicum streptostachys Sprengel, Syst. Veg. 1: 316 1825. Streptostachys hirsuta P.Beauv., Ess. Agrost.: 50, pl. 10, f. 11, 1812. про: *Hab- itat in America (guian.)," Desvaux s.n. (ho- lotipo, P; fragmento, US). А ПЕК. Nees, Agrost. Bras.: 142. 1829. asil. Раш: “Habitat in graminosis in de Serra oin balan innaea asil. one Colares, Poeppis- s.n. (holotipo, n.v.; isotipos, US, W). Panicum кнн Steudel, Syn. Pl. Glumac. 1: 80. : Guyana Francesa: sin localidad y co- lector ibus. P. ан К). Plantas perennes, cortamente rizomatosas, ca- nas decumbentes, pio ue en los nudos inferio- res, luego erectas, de 25-100 cm de alto, rami- ficadas en los nudos superiores; entrenudos huecos, cilindricos; nudos comprimidos, pubescentes o gla- ros. Vainas de 4.5-10 cm de largo, densamente hirsutas o glabras, los márgenes pestanosos; ligulas ciliadas, de 1-1.8 mm de largo, o ausentes; láminas ovado-lanceoladas, de 10-28 cm de largo, 1.8- 4.3 cm de ancho, planas, cordadas y amplexicau- les, densamente hirsutas o glabras, los bordes ba- sales ciliados con largos pelos blanquecinos, tuber- culados. Inflorescencias terminales de 6-16 cm de largo, 1-9 cm de ancho, oblongas a piramidales, largamente pedunculadas; eje principal esparci- damente piloso a glabro, racimos divergentes, al- ternos, ejes de las ramificaciones esparcidamente pilosos a escabrosos, triquetros, espiguillas en pares Volume 78, Number 2 1991 Morrone & Zuloaga 365 Revision del Streptostachys sobre ambos lados del eje; pedicelos cortos, an- gulosos, escabrosos, ocasionalmente con largos pe- los blanquecinos. Panojas axilares numerosas, si- milares a la panoja terminal. Espiguillas elipsoides, de 3.6-5.1 mm de largo, 1.2-1.9 mm de ancho, densamente hirsutas o glabras, con raquilla cons- picua y engrosada entre la gluma inferior y el antecio superior, los bordes inferiores involutos y rodeando el арсе del pedicelo. Gluma inferior oblonga, obtusa, de 3-4.7 mm de largo, %—У del largo de la espiguilla, 5— 7-nervia, con nervios anas- tomosados hacia el ápice. Gluma superior de 2.7- 4.7 mm de largo, dejando libre el ápice del antecio superior, 5-7-nervia, obtusa. Lemma inferior 5-nervia, membranácea a endurecida y cubriendo con los bordes al antecio superior. Palea inferior 1.3-2.4 mm de largo, 0.4-0.7 mm de ancho, castana a blanquecina, hialina, glabra, o ausente; flor inferior ausente. Antecio superior elipsoide, de 3-3.9 mm de largo, 1.3-1.6 mm de ancho, pajizo, liso, con papilas simples distribuidas regularmente, piloso hacia la porción superior; lodiculas abrazan- do los bordes inferiores de la рајеа; estambres 3, anteras de 2.4 mm de largo; estilos libres, jus doni plumosos. Cariopsis de contorno oblon 2.8 mm de largo, 1.2-1.3 mm de visi pajizo a castano; hilo linear, ocupando todo el largo de la cariopsis, embrión menos de %% del largo de la cariopsis. Distribución. Se encuentra desde Trinidad y Venezuela por las Guyanas hasta el norte de Brasil. Material representativo examinado. BRASIL. AMAPÁ: estrada de rodagem Macapá e Clevelandia, entre os kms 106 e 110, Black & Froes 51-12407 (IAN); 2 km from Rio Araguarí, along road from Porto Platon to Macapá, Pires et al. 51084 (K, US); Porto Platón, Mattos 10120 dan 52520 (K, 8 e 14, Black & Froes 51-12330 (IAN); Масара, Black & Froes 51-12279 (IAN); Matapi, Pires & Silva 4799 (IAN, P). AMAZONAS: sin localidad, Martius s.n. (M). BA HIA: Bom Jesüs, Luetzelburg 416 (M, NY), 1913* (NY); Grameal Nanica, Black 54-17938 (IAN); Serra do Aporá, Pinto 256 (IAN); 16 km NW of Lagoinha, on side road to Minas do Mimoso, Harley et al. 16642 (P). CEARA: Granjeiro, Taboleiros, Luetzelburg 26008, 26507 (M); Quixadá, Fazenda Normal, Black 55-18473 (IAN). GOIAS: 2 km М of Araguainia, Irwin et al. 21133 (Е, К, NY, US). MARANHAO: 20 km ao sudeste de Sao Joao dos Patos ao longo da BR-230, Valls et al. 8450 (CEN); Sao Luiz, Granja Barreto, Froes 24312 (IAN, P, SP); Carema, Froes 25860 (IAN); Itapicurü, Black et al. 54-16679 (IAN); Alcantara, Periuasinho, Froes 30794 (IAN). PARA: near Soure, Marajó Island, Goeldi 299 (F, SP, US, W); Maraba, Serra dos Carajas, da 2d 1868 (SP); Marabá, arredores do N5, Secco et al. 467 (K); Caracara River, Marajó Island, Goeldi 99 (W); с ‚ Rio Сиа}ага, Black 54-16163 (IAN); Ilha de Colares, Black 54-17015 (IAN); capoeira de Condeixa, Ilha de Магајо, Black 48-3584 (IAN); camino de Jubim para Condeixa, Ilha de RAS Black 48-3549 (IAN); estrada e Salvaterra para Jub Ilha de Marajo, Black 48-3507 (IAN); estrada = ro- dagem para Vigia, Froes 27819 (IAN); Santarem, Maciel & urs d 335* (NY), Spruce 606, 697 (P), Swallen 3267 (K, P, W) Pires & Black 1137 (IAN), Black & Ledoux 50- 10273 (IAN); Marajo, es et al. 924 (NY). PERNAMBUCO: Paulista, Tenoiro 66/151 (M). PIAUÎ: sin ie Wes Gardner 2341 (P, W); 32.8 km along road uth of Aramante, Еџеп & Eiten 10371 (P); Calverao, Luetzelburg 410 (P). RIO GRANDE DO NORTE: João Pessoa, nto & Assis 750* (NY). GUYANA. Sin localidad, mers 4392 (P), Ile de Cayenne, route de la Chaumiere, de Granville BC.75 (P), Cayenne, route de Rochambeau, Hoock Black & Klein 54-17392* Cayenne, Black & D’Aage 54-18173 (IAN); Route de Sinnamary, Hoock s.n. (IAN 95073); Cayenne, Sagot 1395 (P, W). SURINAM: Zuid Rivier, 25 km above con- fluence with Lucie Rivier, Maguire et al. 53953* (P, SI); Kayserberg airstrip on the Zuid Rivier, Kramer & Hekking 2979* (NY); Kayser airstrip, Irwin et al. 55991* (F, P). VENEZUELA. ANZOATEGUI: bosques cercanos a Playa Colorada, a unos 25 km de Puerto La Cruz, via Cumaná, Aristeguieta & Labbiente 7303 (MO). BOLÍVAR: Isla en el lago de Guri (sector Las Carolinas), 15 km al E de la presa К. Leoni, Aymard et al. 7648 (MO). SUCRE: 7 km E of the Mochima Hwy. intersection along Hwy. 9 between Cumaná and Puerto La Cruz, Davidse 5044 (MO). CARACTERES HISTOFOLIARES EN CORTE TRANSVERSAL (FIGS. 2A, B; 3A) Transcorte plano, de 200-290 um de espesor, zonas costales e intercostales suavemente pronun- ciadas; costilla central con un haz vascular primario asociado a células parenquimaticas incoloras hacia la cara adaxial y solitario o asociado a haces vas- culares secundarios; 4-6 haces vasculares secun- darios entre haces vasculares primarios contiguos, hallandose ambos tipos de haces en corresponden- cia con las zonas costales; haces vasculares ad- yacentes a una distancia de 330-350 um y se- parados por aproximadamente 12 células clorenquimáticas. Haces vasculares primarios de contorno circular, trabados, rodeados por 2 vainas, la interna mestomática y la externa parenquima- tica, ambas sin cloroplastos especializados; vaina parenquimática continua, compuesta por 15-21 células globosas. Haces vasculares secundarios de contorno angular, trabados; vaina parenquimática continua, de 8-10 células; vaina mestomática com- pleta. Clorénquima irregularmente radiado alre- dedor de los haces vasculares, con células tabu- lares, raquimorfas, dispuestas en dos capas hacia la cara adaxial, y con células isodiamétricas hacia la cara abaxial; por debajo de las células buliformes Annals of the Missouri Botanical Garden con células clorenquimáticas tabulares distribuidas regularmente y semejando un tejido en empalizada, con abundantes cloroplastos en relación al resto del tejido clorenquimático; presentes, próximas a la cara abaxial; espacios intercelulares presentes. Células fusoides presentes. Células bu- liformes en grupos regulares de 4-5 células “tipo abanico" presentes en las zonas intercostales ada- “arm cells” xiales. Esclerénquima hipodérmico, en correspon- dencia con los haces vasculares primarios y se- cundarios y el margen foliar. Cristales ausentes. Amiloplastos presentes en las células clorenqui- máticas. CARACTERES EPIDÉRMICOS DE LA LÁMINA EN VISTA PARADERMAL (FIG. 2C) Zonas costales con cuerpos de silice halterifor- mes, alternando con células suberosas rectangu- lares; aguijones ausentes. Zonas intercostales con células largas rectangulares, longitudinalmente Psp i más pei tres veces mas largas que anchas, lada de ; células cortas en pares silico- suberosos. поени estomáticos romboidales a sub- discoidales, de 30—46 um de largo, 22.5-25.5 um e ancho, distribuidos en las zonas intercostales en 2-6 hileras, menos frecuentes en la cara adaxial. Ganchos no vistos. Micropelos 2-celulares, fusifor- mes, de 45-52.5 um de largo, célula basal más corta que la distal, célula distal de paredes delgadas y ápice obtuso. Macropelos unicelulares, de 600— 900 um de largo, de paredes engrosadas, base bulbosa y ápice largamente atenuado, asociados a células epidérmicas sobreelevadas con respecto al resto de la epidermis. Papilas ausentes. Observaciones. Diversos autores, como Met- calfe (1960), Jacques-Félix (1962), y Soderstrom & Ellis (1987) señalaron la presencia en el mesófilo de células fusoides como un carácter distintivo de la subfamilia Bambusoideae. Las células fusoides de las Bambusoideae son, en corte transversal, fusiformes, de lumen amplio y traslúcidas, de posición lateral con respecto al haz vascular, y se hallan en contacto directo con las células de la vaina externa de los haces vas- culares. En corte longitudinal se presentan estre- chas, con escaso lumen celular, ligeramente más amplio hacia ambas caras epidérmicas, y están separadas unas de otras por ase espacios intercelulares (Metcalfe, 1960; Calderon & So- derstrom, 1973; Ellis, 1976). Bod et al. (1987) diferenciaron a las células fusoides de la tribu Phareae por presentarse, cuando maduras, globosas, transversalmente alargadas, y con pa- redes excepcionalmente delgadas, no dejando cons- picuos espacios entre si en corte longitudinal. Watson et al. (1985) y Watson & Dallwitz (1988) mencionaron la presencia de células fusoides en el género Homolepis dentro de la supertribu Pani- canae. Indicaron como posible una evolución pa- ralela o convergencia del carácter en las subfa- milias Bambusoideae y Panicoideae Killeen & Clark like cells" en especies de la sección Laxa del género 986) reconocieron “fusoid- Panicum. De acuerdo con estos autores, estas cé- lulas eventualmente no deberían ser consideradas homólogas a las células fusoides presentes en la subfamilia Bambusoideae. Renvoize (1987) señaló células fusoides en dos géneros de Paniceas, Homolepis y Streptostachys. Al estudiar estas células en S. asperifolia se observó que las mismas difieren de las células fu- soides previamente descriptas para la generalidad de las Bambusoideae, por ser en corte longitudinal más o menos cuadradas, con lumen celular amplio y por presentar escasos espacios intercelulares en- tre células fusoides contiguas. De acuerdo con esta descripción, dichas células coinciden con las men- cionadas por Soderstrom et al. (1987) para la tribu Phareae. En el presente trabajo se mantiene el término de fusoides para las células presentes en S. aspe- rifolia, hasta tanto se efectúen estudios ontoge- néticos comparados entre taxa de las subfamilias Bambusoideae y Panicoideae. 2. Streptostachys macrantha (Trin.) Zuloaga Soderstrom, Smithsonian Contr. Bot. 59: Langsdorff s.n. (holotipo, LE n.v.; isotipo, К; fragmento del tipo, US). Figure 1A-K. Panicum vaginatum var. alpha et beta Nees, Agrost. Bras.: 156. 1829. TIPO: Brasil, sin localidad, Sellow s.n. (holotipo, B n.v.; isotipo, K; fragmento, US). Plantas perennes, cortamente rizomatosas, con rizomas fuertes cubiertos de catafilos vilosos; canas de 30-70 cm de alto, simples, con hojas basales numerosas y manifiestamente disticas, canas flo- riferas de 2-3 nudos y con hojas con laminas reducidas; entrenudos comprimidos, glabros; nudos oscuros, pilosos o glabros. Vainas rala o densamente pilosas, cuello pajizo a violáceo, piloso o glabro. igulas ciliadas o ausentes. Láminas lanceoladas, acuminadas, de 7-14 cm de largo, 0.7-1.3 ст de ancho, verde oscuras, rigidas, atenuadas o cor- dadas. Inflorescencias terminales laxas, de 5-10 cm de largo, 2-7 cm de ancho; eje principal an- Volume 78, Number 2 Morrone & Zuloaga 367 Revision del Streptostachys guloso, glabro, ramificaciones inferiores verticila- das o alternas, las superiores alternas, divergentes, con espiguillas solitarias sobre pedicelos cortos; pe- dicelos glabros, convexos. Espiguillas elipsoides, de -7.4 mm de largo, 2.1-2.5 mm de ancho, verdosas y con tintes violáceos, pilosas sobre el арїсе de la gluma superior y lemma inferior o glabras, con raquilla conspicua y engrosada entre la gluma inferior y el antecio چ‎ don Gluma in- e largo, 6—26 del largo de la ы l- 2s -nervia, nervio medio escabroso hacia el ápice. Сита superior de 5-6.6 mm de largo, dejando libre el ápice del antecio 5—7-пегмта, los nervios anastomosados ferior ovada, superior, hacia el ápice. Lemma inferior de 5.4–6.7 mm de largo, 3-5-nervia. Pálea inferior lanceolada, de 4.3-5.8 mm de largo, 1–1.6 mm de ancho, hialina, glabra; flor inferior presente, lodiculas 2, trunca- das, conduplicadas, estambres 3, anteras de 4.2 mm de largo. Antecio superior ovoide, de 5.1–5.9 mm de largo, 1.8-2.4 mm de ancho, pajizo, liso y lustroso, piloso y papiloso; lodiculas de 0.7 de largo, abrazando los bordes inferiores de la рајеа; estambres 3, anteras violáceas de 3.8 mm de largo; estigmas emergiendo por el ápice del antecio su- perior. Cariopsis fusiforme, de 3.6 mm de largo, 1.7 mm de ancho, hilo linear, ocupando todo el largo de la cariopsis, embrión menos de / del largo de la cariopsis. Distribución. Brasil, en el Distrito Federal y los estados de Goiás y Minas Gerais, hallándose ocasionalmente en Paraguay en la Sierra de Amam- bay. Habita en cerrados sujetos a fuegos periódicos, en suelos pedregosos. Llega hasta aproximada- mente 1,200 m. Florece entre octubre y diciembre. Material о examinado. BRASIL. DIS- TRITO FEDERAL: in the area of the Corrego Сађеса do al. 3. 355 (K, UB); 20 km S of Brasilia on road to Goiania, near Rio Melchior, Irwin et al. 8658* (US); proximidades da Escola Fazendaria, Heringer 18255 (IBGE); Fazenda S of т, on road to Jataí, [eee & Soderstrom US*) Cristalina, Filgueiras & Pereira 923 (IBGE); Serna do Caiapó, 30 km N of Jatai on road to Caiaponia, Irwin & Soderstrom 7307 (К, US); Luziania, Heringer 13963 (IBGE, UB); Chapada dos Vea- deiros, Heringer et al. 2754 (IBGE); sin localidad, 51. Hilaire 660 (P). MINAS GERAIS: 37 km 5 of Prata on highway BR-14, Goodland 3967 (UB); Serra do Cipo, 110 km NE of Belo Horizonte, Chase 9297 * (US); Morro do Cruzeiro, Lagoa Santa, 42 km N of Belo Horizonte, Chase 9018* (NY, US); Paracatü, Filgueiras & Pereira 924, 925 (IBGE); estrada de Cristalina a Paracatú, Heringer & m 17447 (IBGE); Rio Sào Francisco, Pohl 329* (M, ; ad Lagoa Santa, Warming s.n. (M, W); sin МА St. Hilaire 473 (P), Riedel s.n. (P, W), in i a ú pis arenosis Estrella, Besster 10115 (BAA, C). Observaciones. Еп К hay un ejemplar de Bra- sil: Minas Gerais, in campis Lagoa Santa, Riedel 2152, que pertenece a la especie y que ha sido identificado en ese herbario como un isotipo de la misma. Cabe destacar que Renvoize (1978) indicó que determinados ejemplares coleccionados por Riedel en su viaje con Langsdorff por el Brasil deben ser considerados como de la misma colección hecha por el ültimo autor (y sobre la que Trinius basó la mayor parte de sus descripciones en Pa- niceas). CARACTERES HISTOFOLIARES EN CORTE TRANSVERSAL (FIG. 2D, E) Transcorte plano, de 200-350 шт de espesor; zonas costales e intercostales suavemente pronun- ciadas; costilla central no diferenciada; 3—4 haces vasculares secundarios, 4-6 haces vasculares ter- ciarios y 2-3 haces vasculares cuaternarios entre haces vasculares primarios contiguos, haces vas- culares adyacentes a una distancia de 60-15 y separados por 2-3 células clorenquimáticas. Ha- ces vasculares primarios de contorno eliptico, tra- bados, y equidistantes de ambas epidermis, en co- rrespondencia con las zonas costales, rodeados por una vaina mestomática Kranz compuesta por 13 a 20 células; células Kranz globosas en corte trans- versal, de tamano igual o menores que las células clorenquimáticas adyacentes, y rectangulares en vista paradermal, con su eje mayor paralelo a las venas, con cloroplastos especializados de posición centrifuga u ocupando todo el lumen celular, cé- lulas de la vaina de paredes radiales y tangencial externa curvas, con paredes igualmente engrosadas o con las paredes radiales y tangencial interna l hilera de células esclerenquimáticas presente o ausente entre la vai- manifiestamente engrosadas; na mestomática y los elementos del metaxilema. Haces vasculares secundarios de contorno eliptico, trabados y equidistantes de ambas epidermis, con 9-15 células еп la vaina Kranz. Haces vasculares terciarios libres, dispuestos hacia la cara abaxial, de contorno angular, asociados a las zonas inter- costales, con 6-8 células еп la vaina Kranz. Haces vasculares cuaternarios libres, próximos a la cara abaxial, de contorno cuadrado, asociados a las zo- nas intercostales, con 4 células en la vaina Kranz. Clorénquima irregularmente radiado con células 368 Annals of the Missouri Botanical Garden УЗ Y. BEBA 1] \ / 500 дат A ا‎ سے — «2 "rp ¿ua <. quy X re З еа: = S M Ie TA IR ^ eN PET 169. 2 8 6555 nie FIGURA 2. -C. Streptostachys asperifolia. — А. Transcorte de una porción де la semilámina mostrando la costilla central (flechas). — B. Detalle de la semilamina en transcorte, entre un haz vascular primario у secundario. — C. Detalle de la epidermis abaxial. D-F. Streptostachys macrantha: —D. semilámina en transcorte, señalando nervio medio (flechas). — E. Detalle de la semilámina en transcorte. — Е. Detalle de la epidermis abaxial. G-I. Streptostachys ramosa. —G. Aspecto general de un fragmento de semilámina en transcorte. — H. Detalle de la semilámina en transcorte. — I. Detalle de la epidermis abaxial. a, haz vascular primario; b, haz vascular secundario; c, haz vascular terciario; d, haz vascular cuaternario; f, célula fusoide. (A-C de /rwin et al. 55991; D-F de Irwin et al. 8658; G-I de Harley et al. 16876.) Volume 78, Number 2 1991 Morrone & Zuloaga 369 Revision del Streptostachys FIGURA 3. C. Detalle de la epidermis abaxial, con Streptostachys macrantha: — Е. Epidermis abaxial. — Е. rectangulares alrededor de los haces vasculares, y con células isodiamétricas en el resto del mesófilo; espacios intercelulares escasos. Células fusoides au- sentes. Células buliformes en grupos de 4-6 cé- lulas, tipo “abanico” presentes en las zonas inter- costales. Parénquima incoloro distribuido por debajo de las células buliformes. Esclerénquima poco de- sarrollado, hipodérmico, en correspondencia con los haces vasculares primarios y secundarios y el margen foliar. Cristales ausentes. Amiloplastos pre- sentes en las vainas mestomáticas Kranz. aparato estomático, cubierto por pa ў Detalle de la zona costal abaxial. (А de Kramer & Hekking 2979; B-D de Filgueiras 855; E, F de Irwin et al. 8658.) Streptostachys asperifolia. — А. Detalle de una porción de la semilamina en transcorte, con un haz vascular primario y células fusoides (flecha). B-D. Streptostachys ramosa. — B. Detalle de la zona intercostal adaxial. — pilas. — D. Transcorte de C. E, F CARACTERES EPIDÉRMICOS DE LA LÁMINA EN VISTA PARADERMAL (FIGS. 2F; 3E, F) Zonas costales con cuerpos de silice halterifor- mes, menos frecuentemente nodulares; aguijones presentes en la cara adaxial. Zonas intercostales con células largas rectangulares, alargadas longi- tudinalmente, hasta tres veces más largas que an- chas, de paredes ligeramente onduladas; células cortas solitarias o en pares. Aparatos estomáticos romboidales, de 33-36 ит de largo, 22-27 um 370 Annals of the Missouri Botanical Garden de ancho, dispuestos en 2 hileras en las zonas intercostales. Ganchos presentes. Micropelos 2-ce- lulares, fusiformes, de 52.5-72 um de largo, célula distal más larga que la basal, de paredes delgadas y ápice redondeado. Macropelos unicelulares, pre- sentes en la cara adaxial, de 300— rigidos, de paredes engrosadas, base bulbosa y ápi- 00 um de largo, agudo, asociados a células epidérmicas no so- breelevadas. Papilas ausentes. Observaciones. A pesar de hallarse esta es- pecie en un habitat similar al de 5. ramosa, 5. macrantha carece de las adaptaciones xeromór- ficas presentes en esa especie. E Streptostachys ramosa Zuloaga & Soder- rom, Smithsonian Contr. Bot. 59: 52. 1985. TIPO: Brasil. Bahia: 22 km S of the Rio Gal- heirao along highway BR-020, 8 Apr. 1976, 860 m, Davidse et al. 12216 (holotipo, MO; isotipos, NY, SP). Planta perenne, fuertemente rizomatosa; canas e 1.30-2 m d numerosas, cañas floriferas de 2-3 nudos, entre- e alto, simples, con hojas basales nudos cilindricos, glabros, huecos, sólidos en la porción inferior de las canas; nudos oscuros, com- a — Vainas estriadas, densamente vi- abras, los bordes cortamente pilosos, cuello 5 glabro o cortamente piloso. Ligulas ciliadas, de 0.8-2.5 mm de largo, o ausentes. Làminas lan- ceoladas, de 30-60 cm de largo, 1-1.4 cm de ancho, planas o con los bordes involutos, lisos, subcordadas, densamente hirsutas en ambas caras o escabriüsculas en la cara adaxial y glabras en la сага abaxial. Pedúnculos de aproximadamente 70 cm de largo. Inflorescencias terminales laxas, pi- ramidales de 30-45 cm de largo, 15-30 cm de ancho; eje principal cilindrico, glabro, ramificacio- nes inferiores verticiladas, ocasionalmente alternas, las superiores verticiladas a subopuestas, divergen- tes del eje principal, eje de las ramificaciones gla- bro, axilas de las ramificaciones densamente pilo- sas; espiguillas solitarias y dispuestas unilateralmente sobre las ramas, ramificaciones de tercer orden presentes o más comünmente ausentes; pedicelos glabros, convexos. Espiguillas largamente elipsoi- des, de 6 3 mm de largo, 2.1-2.7 ancho, glabras, verdosas y con tintes violáceos, mm de raquilla manifiesta y engrosada entre la gluma in- ferior y el antecio superior. Cluma inferior de 1.5— 4.4 mm long., И—\У del largo de la espiguilla, aguda, 1 -3-nervia, cortamente pilosa en el ápice. Gluma superior obtusa a aguda, 5- 7-nervia. Lem- ma inferior 3-5(7)-nervia, cortamente pilosa en el ápice de la cara interna. Pálea inferior lanceolada, de 5.3-7.9 mm de largo, 1-2.2 mm de ancho, membranácea, glabra; flor inferior presente, lodi- culas 2, truncadas, estambres 3, anteras de 4. mm long. Antecio superior elipsoide, de 5.4-7.9 mm de largo, 1.9-2.4 piloso y papiloso; lodiculas abrazando los bordes mm de ancho, blanquecino, inferiores de la рајеа. Cariopsis fusiforme, de 3.8 mm de largo, 1.7 mm de ancho, hilo linear, ocu- pando todo el largo de la cariopsis, embrión menos de Yê del largo de la cariopsis. , en cerrados de los es- tados de Bahia, Goiás, Minas Gerais, y Mato Gros- so, a 400-1,500 pedregosos. Florece entre enero y abril. Distribución. Brasil m. Crece en suelos arenosos, Material representativo examinado. BRASIL. BAHIA: Ipatá, Harley et al. 24147 (К); Mucugé, rodovía para Palmeira, Hatschbach & Kummrow 48033 (K); Serra do Sincorá, ca. 15 km NW of Mucugé on the road to Gune & Palmeiras, Harley et al. 20997 (К); 15 km S of the intersection of highway BR-020 and the Rio Roda Velha, Davidse et al. 12085 (К, MO, NY, SP, US*); Serra do Curral Feio, 22 km NW of Lagoinha, on side road to Minas i NY, P, US*); Espigão gie ca. 100 k Barreiras, Anderson et al. 36657 (MO*, NY, UB). Golds: Ro d a ВЕ-050, Mun. Campo A ss de Goias, Hatsch- 776 s gem 29-1-1980, Heringer & Riz- zini 17626 (IBGE). M G к, Ки inane Francisco Sa, Serra do = a e 1 195 (SP): ca. 20 km S of Alto do Paraiso, Irwin et al. 24716 (МУ) = 27 (Davidse & Pohl, 1978, bajo Panicum macranthum Trin.). Nümero cromosómico. Observaciones. Streptostachys ramosa se aparta exomorfologicamente de 5. macrantha por el mayor tamano de plantas, láminas y panojas. En el ejemplar Glaziou 20123a (K, P), figura como lugar de colección Rio de Janeiro, Brasil, lo cual obviamente es un error de etiquetado dada la distribución y el hábitat donde crece la especie. CARACTERES HISTOFOLIARES EN CORTE TRANSVERSAL (FIGS. 26, H; JD abierta, involuto, de 360—440 um de espesor; cara adaxial con zonas Transcorte en forma de “U” costales conspicuas, de ápice truncado y zonas intercostales hasta 144 de profundidad en relación al ancho de la lámina; cara abaxial con zonas costales e intercostales poco manifiestas; costilla central conspicua; 2-4 haces vasculares secun- darios, 4—8 haces vasculares terciarios y 2-6 haces vasculares cuaternarios entre haces vasculares pri- Volume 78, Number 2 Morrone & Zuloaga 371 g Revision del Streptostachys marios contiguos; haces vasculares adyacentes a una distancia de 60-160 um y separados por 2- 3 células clorenquimáticas. Haces vasculares pri- marios de contorno eliptico, asociados a las zonas costales, trabados y equidistantes de ambas epi- dermis, rodeados por una vaina mestomática Kranz, compuesta de 13-25 células de tamario igual o menor que las células del mesófilo; células Kranz globosas en corte transversal y rectangulares en vista paradermal, con su eje mayor longitudinal; células Kranz con las paredes radiales y tangencial externa curvas, igualmente engrosadas o con las paredes radiales y tangencial interna fuertemente engrosadas; cloroplastos de posición centrifuga u ocupando todo el lumen celular; 1-3 hileras de células esclerenquimaticas de lumen pequeno y pa- redes igualmente engrosadas presentes o ausentes entre la vaina Kranz y los elementos del metaxi- lema. Haces vasculares secundarios de contorno eliptico, en correspondencia con las zonas costales, trabados a semitrabados y equidistantes de ambas epidermis, con 10-15 células Kranz globosas a reniformes, con las paredes más o menos igual- mente engrosadas, extensiones de la vaina presen- tes o ausentes, cuando presentes compuestas por 1 o 2 células dispuestas hacia ambas caras. Haces vasculares terciarios libres en correspondencia con las zonas intercostales, dispuestos hacia la cara abaxial, de contorno angular, vaina mestomática con 5-8 células Kranz. Haces vasculares cuater- narios libres, de contorno cuadrado, próximos a la cara abaxial y en correspondencia con las zonas intercostales. Clorénquima irregularmente radiado alrededor e o menos rectangulares, el resto del clorénquima e los haces vasculares con células más con células isodiamétricas y con escasos espacios intercelulares. Células fusoides ausentes. Células buliformes en grupos de 2-4 células tipo **abanico" distribuidas en las zonas intercostales adaxiales. Esclerénquima discontinuo, hipodérmico, abundan- te en las zonas costales en correspondencia con los haces vasculares primarios, secundarios y el mar- gen foliar. Cristales ausentes. Amiloplastos presen- tes en las vainas mestomáticas Kranz. CARACTERES EPIDÉRMICOS DE LA LÁMINA EN VISTA PARADERMAL (FIGS. 2I; 3B, C) Zonas costales con cuerpos de silice halterifor- mes; aguijones frecuentes en la cara adaxial. Zonas intercostales con celulas largas rectangulares, alar- gadas longitudinalmente, más de tres veces más largas que anchas, de paredes onduladas, fuerte- mente engrosadas; células cortas solitarias o en pares, rectangulares. Aparatos estomáticos subdis- coidales, de 37-45 um de largo, 33-39 um de ancho, distribuidos en 1-4 hileras en las zonas intercostales, en el mismo nivel que el resto de las células epidérmicas en la cara adaxial y hundidos en la cara abaxial. Ganchos y aguijones presentes en las zonas intercostales de la cara adaxial. Mi- cropelos 2-celulares, fusiformes, de 79.5-94.5 um de largo, con ambas células de aproximadamente igual tamano; célula basal de paredes engrosadas, célula distal de paredes delgadas y apice obtuso. Macropelos unicelulares de 750-1,200 um de lar- go, de paredes engrosadas, base bulbosa y apice agudo, asociados a células epidérmicas en el mismo nivel que el resto de la epidermis. Papilas presentes en las zonas intercostales abaxiales, asociadas a las células largas e interestomáticas, de posición distal y excentrica, cubriendo parcialmente a los estomas. Observaciones. see caracteres predominantemente xeromórficos, como ser la presencia de cuticula gruesa, zonas Streptostachys ramosa po- costales e intercostales adaxiales conspicuas (con zonas intercostales profundas) con numerosos aguijones cubriendo los aparatos estomáticos ada- xiales, y hallándose a su vez los estomas abaxiales hundidos y cubiertos parcialmente por papilas, sien- do este ültimo carácter poco frecuente en la epi- dermis de las gramineas. Eiten (1978) senaló para la region del cerrado (incluyendo a los campos rupestres), una estación seca entre abril y septiembre con aproximadamente 250 mm de precipitación, existiendo en el resto del ano una precipitación promedio de 1,350 mm. Los suelos son comünmente arenosos, bien dre- nados, y generalmente pobres en nutrientes, siendo la radiación solar intensa en este tipo de ambiente. ESPECIE EXCLUIDA El análisis del material tipico de 5. acuminata permitió comprobar que la especie fue casi simul- táneamente descripta por Zuloaga & Soderstrom (1985) bajo el nombre de Brachiaria tatianae. Los caracteres diagnósticos del género Bra- chiaria (Trin.) Griseb. han sido cuestionados en los ultimos anos por autores como Hsu (1965) más recientemente por Webster (1987). Este úl- timo autor consideró a Brachiaria como género monotípico, incluyendo en el mismo únicamente a B. eruciformis (J. E. Smith) Griseb. por tener an- tecio superior liso, punto primario de desarticula- ción por encima de la gluma inferior, lemma mútica y pálea no cubierta por la lemma en su porción apical. Webster (1987) distinguió a Urochloa P. Beauv. por guillas desarticulando por debajo de la gluma ш- oseer antecio superior rugoso, espi- 372 Annals eco PL Garden ferior, lemma mútica, crestada o apiculada y pálea cubierta por la lemma en su apice Siguiendo el criterio sustentado por Webster (1987, 1988), y por coincidir S. acuminata con los caracteres diagnósticos del genero Urochloa, en la presente contribución se trata a la especie en el género Urochloa. Urochloa acuminata (Renvoize) Morrone & Zu- loaga, comb. nov. Streptostachys acuminata Renvoize, Kew Bull. 39: 182. . TIPO: Brasil. Bahia: estrada de Lençois, BR-242, 5 5 km ao N de Lengois, 19 Dec. 1981, André M. de Carvalho, G. P. Lewis & J. L. Hage 1007 bis (holotipo, K; fotografia del tipo y fragmento, Brachiaria tatianae Zuloaga & Soderstrom, Smithsonian ntr. Bot. 59: 7. 1985. [Panicum megasta тё neralium, Martius s.n. (holotipo, M, fragmento, US). Plantas perennes, con canas multinodes de 1.80- 2 m de alto, simples, erectas; entrenudos cilindri- cos, ralamente pilosos hacia la porción superior а glabros; nudos pubescentes, comprimidos, casta- nos. Vainas de aproximadamente 15 cm de largo, pubescentes, los bordes membranáceos, auriculas reducidas, cuello piloso. Ligulas cortamente mem- branáceas en la base, luego largamente ciliadas, de 1.5 mm de largo. Láminas lanceoladas, de 40 cm de largo, 1.5 cm de ancho, planas, sub- ordadas, acuminadas, cortamente pilosas en am- bas caras, los márgenes cartilaginosos. Inflores- cencia terminal laxa, multiflora, piramidal, de 25- 35 cm de largo; eje principal anguloso, | racimos aproximadamente 25, los inferiores ver- ticilados, los superiores alternos a и di- vergentes, ejes de los racimos triquetros, de 5-1 cm de largo, 0.4-0.6 mm de ancho, flexuosos, cortamente pilosos y escabrosos, terminando en una espiguilla desarrollada; axilas de las ramifica- ciones cortamente pilosas; espiguillas en pares, la inferior subsésil, solitarias hacia el apice de los racimos; pedicelos triquetros, de 0.8-4 mm de largo, largamente pilosos cerca de la inserción de la гра Espiguillas largamente elipsoides, de m de largo, 2.2-2.4 m Баса M ápice de glumas y lemma inferior. Gluma inferior ovada, de 3.5-3.8 mm de largo, abraza- dora, poco más de / del largo de la espiguilla, 5 7-nervia, con los nervios anastomosados hacia el ápice, separada de la gluma superior por un corto entrenudo. Gluma superior de 5.2-6.1 mm de lar- mm de ancho, pilosas go, 7–9-пегуја, dejando libre el ápice del antecio superior. Lemma inferior 5- 7-nervia. Palea infe- rior elíptica, de 4.5-5.9 mm de largo, 1.6-2.1 mm de ancho, hialina, los bordes ciliados, glabra en el resto de la superficie; flor inferior presente, lodiculas 2, de 0.5 mm de largo, estambres 3, anteras purpüreas de 2.7 mm de largo. Antecio superior elipsoide, de 4.7-5.9 mm de largo, 1.6– 2.1 mm de ancho, pajizo, endurecido, transver- salmente rugoso, con papilas verrugosas sobre toda la superficie, lemma de ápice cortamente crestado y piloso, raquilla prolongada por encima del antecio superior en forma de un mucrón de tamano varia- ble; lodiculas 2, conduplicadas, abrazando los bor- des inferiores de la pálea, estambres 3, anteras de 3 mm de largo; estilos libres, estigmas plumosos. Cariopsis no visto. Distribución. | Conocida hasta el presente para Bahia, Brasil, creciendo en bordes de bosques. Observaciones. Entre los caracteres más im- portantes de la especie se encuentran los pedicelos largamente pilosos en su porción apical, antecio superior transversalmente rugoso y con raquilla prolongada por encima de la inserción del antecio superior. En el catálogo de colección de Andre M. de Carvalho figura (Mattos Silva, comm. pers.) üni- camente el nümero 1007. А! enviarse material de este botánico en consulta a Kew, Renvoize advirtió que había mezcla dentro de dicha numeración, correspondiendo el numero 1007 a Ichnanthus bambusiflorus (Trin.) Doell, existiendo otro espe- cimen diferente, al que Renvoize dió el nümero 1007 bis y describio como S. acuminata. Cabe destacar que el holotipo de esta ültima especie se encuentra depositado en K y no en CEPEC como consta en la diagnosis original de la especie. CARACTERES HISTOFOLIARES EN CORTE TRANSVERSAL (FIG. 4D, F) Transcorte plano, de 130-200 шт de espesor; zonas costales e intercostales suavemente pronun- ciadas; costilla central no diferenciada; 7-9 haces vasculares secundarios entre haces vasculares pri- marios contiguos, haces primarios y secundarios en correspondencia con las zonas costales; haces vasculares adyacentes a una distancia de 150-1 um y separados por 2-3 células clorenquimáticas. Haces vasculares primarios de contorno circular, trabados, equidistantes de ambas epidermis, ro- deados por 2 vainas, la vaina interna mestomatica y la vaina externa arenquimatica Kranz, de 12- 14 células con cloroplastos especializados de po- Volume 78, Number 2 1991 Morrone & Zuloaga 373 Revision del Streptostachys ST | STE pe FIGURA 4. A-C. Streptostachys robusta. — А. Transcorte de una porción де la semila У E . 19 blade as viewed in transverse section. Bothalia 12: 65-109. 1977. Distribution of the Kranz syndrome the southern African Eragrostoideae and PME according to bundle sheath anatomy and citology. ntae 9: 73-110. 79. А procedure for standardizing com- parative leaf anatomy in the Poaceae. e epi- ermis as seen in surface view. Bothalia 12: 641- Eragrostis walteri—a first record of deae (P bie Т. mineas e seu uso па taxonomia da familia. Рена Agropecuaria Brasileira 21(2): 93-100. HATTERSLEY, P. У. & Г. Watson. 1976. C, grasses: an anatomical criterion for distinguishing between NADP-Malic enzyme species and PCK or NAD-malic enzyme species. Austral. J. Bot. 24: Ls у Hsu, С. С. 19 The classification of Panicum (Gra- mineae) and its allies, with special еи to the ac. raminees (Poaceae) "Afrique tropicale, 1: Generalités, classifcation, de- scription des genres. Institut de Recherc Agro- omiques Tropicales et des Culture lea Bulletin Scientifique (Paris), 8 JENSEN, W.A. 1962. ean =, Ргїп- ciples and Practice. W. n, San Francisco. o D. A. 1940. Plant ы к McGraw- , New York & London. нь T. & Г. CLARK. 1986. Fusoid-like cells іп pop sect. Laxa (Poaceae: Panicoideae). Amer t. 73(5): 771-772. 0. Mi Ре taxa of Gramineae outhwest Africa and a nt areas. Mitt. Bot ет München 8: 136-158. МЕТСАТЕЕ, C. R. 1960. Anatomy of the Monocotyle- ons. I. Gramine d. Prat, Н. 1932. L'epiderme des Graminées, etude ana- tomique et systematique. Ann. Sci. Nat. Bot. 14: 17-324. Lun E fro 1 La systematique des Graminées. Ann. Sci. Nat. (Paris) XVIII: 165-258. PRENDERGAST, H. D. У. € P. W. HarrERSLEY. 19 Australian C, grasses (Poaceae): leaf blade anutomicil features in relation to C, acid decarboxylation types. Austral. J. Bot. 35: 355-382. RENVOIZE, S. А. 1978. The Pen Panicum group Lorea (Gramineae). Kew Bull. 32: -428. . 1984a. New grasses с Bahia. Kew Bull. 39: 179-183. . 1984b. The Grasses of Bahia. Royal Botanic Gardens, Kew. —— — —. 1987. А survey of leaf-blade с іп grass- s. XI. Paniceae. Kew Bull. 42: 68. T E. & M. I. CASABONA. . Presencia de cristales en hojas de Paniceae (Gramineae). Bol. Soc Argent. Bot. 20: 83-90. Sass, J. E. 1940. Elements of Botanical Microtech- nique. McGraw-Hill, New York & London SCHMID, К. & M. D. TURNER. 1977. Contrad 70, an effective softener of quien material for anatom- ical study. Taxon 26: -552. SENDULSKY, T., T. S. a & A. G. BURMAN. Annals of the Missouri Botanical Garden 1987. Fruits, embryos and seedlings. Pp. 31-36 in T. R. Soderstrom, K. W. Hilu, C. S. Campbell & M. E. Barkworth (editors), Grass Systematics and Evolution. Smithsonian Institution Press, Washing ton, D.C. SODERSTROM, T. В. & R. Р. ELLis. 1987. The position У У erst ; W. Hilu, С. 5. Campbell & M. E. Barkworth (editors), Grass Systematics and Evolution. Smithsonian Insti- tution Press, Washington, D.C. & E. J. Jupziewicz. 1987. The Phareae and Streptogyneae (Poaceae) of Sri Lanka: a е anatomical study. Smithsonian Contr. Bot. 65: 27 pp. STIEBER, M. T. Revision of Ichnanthus sect. Ichnanthus анк Panicoideae). Syst. Bot. 7: 85-115. 1987. Revision of Ichnanthus sect. Foveo- latus (Gramineae: Panicoideae). Syst. Bot. 12: 187- 216. ТАТЕОКА, T., S. Inou & S. Kawano. 1959. Notes on some grasses IX. Systematic кш of 2-cellular mucronate d leaf epidermis. Bot z. (Crawfords- ville) 121 l. WATSON, L. & 4 n DarLwiTZ. 1988. Grass Genera classification, interactive retrieval. Research School of Biological Sciences, The Australian National n ‚ Н. T. Си р & М. J. DaLLwitz. 1985. Тће ава игин prs Poaceae: и and su- pertribes. Austral. J. Bot. 33: 433-484. WEBSTER, В. D 87. The пена Miri (Poa- и ү Cramer, Stuttgart, bien n . Genera of the North шы са Pani- ceae а Panicoideae). Syst Bot. 13(4): 576– 609. ZULOAGA, F. O. & T. В. SODERSTROM. 1985. Classifi- cation of the outlying species of New World Panicum (Poaceae: Рашсеае). Smithsonian Contr. Bot. 59: 46-55. A KEY TO THE GENERA OF LAURACEAE IN THE NEW WORLD! Henk van der Werf? ABSTRACT A key to the twenty-nine genera of Lauraceae naturally occurring in the western hemisphere is presented. The key is based on floral and vegetative characters on ly. A brief statement on distribution and size o the genera is provided, together with a note on cupule shape. A Spanish translation of the key is included. Lauraceae have, in general, the reputation of being difficult to identify. After having studied the neotropical members of this family for the last six years, I fully agree that the reputation is well deserved. There are various reasons for this. In the first place, many collections are sterile or fruit- ing and lack the floral characters needed for iden- tification. To date, there are no keys for the generic identification of nonflowering specimens, and it is unlikely such a key will become available in the future. Sterile specimens may be recognized to species by a botanist familiar with the family, but if such instantaneous recognition does not occur, the specimen is likely to remain unidentified even to genus and therefore inaccessible for a long time. Fruiting specimens stand a better chance of being identified because the cupule (or its absence) con- tains useful information, especially when old sta- mens remain on the cupule. However, 1 am not able to construct a workable key based on fruit and vegetative characters. From a taxonomic point of view, fruiting specimens are valuable, even though difficult to identify, whereas sterile specimens are of little or no use. For botanists making forest inventories or studying species diversity, sterile voucher specimens are useful, hence collected, and end up on the desk of a specialist, where they often remain unidentified. Even with good flowering material at hand, iden- tification to genus is often difficult. In my opinion, the main reason for this problem is that no workable keys exist. The published keys are partly obsolete (such as Mez, 1889), are of regional scope (Mac- Bride, 1938; Kostermans, 1936; Allen, 1945), use both floral and fruit characters (Kostermans, 1957; Hutchinson, 1964), or share a combination of these problems. Keys that are based on floral and fruit characters create special problems. As a rule, both floral and fruit characters are necessary for generic identification, but a single specimen rarely has both flowers and fruits. Typically, fruits need several months for development, by which time the flowers have long fallen off. Rarely, fruits from an earlier flowering season persist long enough to be collected with fresh flowers, but that is an exception. In such cases, flowers and fruits do not occur on the same inflorescence. Occasionally, one finds flowers and " on the same inflorescence. In my expe- are usually diseased and de- "fruits rience, such “fruits” formed. Thus, because flowering specimens almost never have fruits, a key that requires both flowers and fruits is not workable and will frustrate its user. However, it is possible to construct a generic key for neotropical Lauraceae based solely on floral and vegetative characters. | present such a key below, which includes all American genera. most important floral characters used in this key, as well as in earlier keys, are number of stamens, number of anther cells per stamen, sexual condition of the flowers, and tepals equal or un- equal. These characters, although readily visible, are often variable within a genus, and this variation has rarely been incorporated in earlier keys. For example, Persea is usually presented as having nine 4-celled stamens, whereas the neotropical spe- cies can have nine or six stamens, these being all 4-celled, all 2-celled, or some 4- and some 2-celled. Moreover, the tepals can be equal or unequal and "А. Gentry, M. Gra W. D. Stevens kindly iod the key into Spanish. J. ? Missouri Botanical Garden, P.O. Box 2 um, J. Rohwer, and W. D. Stevens critically read the е and suggested improvement. Myers made the illustration 299, St. Louis, Missouri 63166-0299, U. s. А. ANN. MISSOURI Bor. GARD. 78: 377-387. 1991. 378 Annals of the Missouri Botanical Garden deciduous or persistent in fruit. In order to accom- modate these permutations of characters, it is at times necessary to have a genus appear several times in the key. While constructing this key, I have completely ignored any phylogenetic relationships among the genera, mainly because phylogenetic schemes o Lauraceae are partially based on fruit characters (Kostermans, 1957) or are based on wood and bark anatomy (Richter, 1981). These supposedly phylogenetic classifications (which are quite differ- ent from each other) cannot readily be followed in the construction of a key based on floral characters. Instead, my aim is to provide a workable key for the identification of flowering Lauraceae. Before presenting the key, I will briefly describe the flower structure of Lauraceae. Examples are based on American genera, and unusual conditions of palaeotropical genera (such as dimerous flowers in Potameia) are not discussed. Typically, each flower has two whorls of three tepals each. The whorls are usually equal in size and shape, but there are exceptions. Tepals are unequal in all species of Caryodaphnopsis, in many species of Persea (Fig. 1c), in Anaueria, and in a few species of Aniba and Licaria. When tepals are unequal, the outer three are smaller, except in Licaria, where the outer three tepals can be larger (Fig. 14). “Chlorocardium” tepals in whorls of four or irregular. The stamens are arranged in four whorls of three each (with the exception of “Chlorocardium,” which has stamens in whorls of four) and are opposite the tepals. The whorls are counted from the outside to the center of the flower. Whorl I is opposite the outer tepals, whorl II is facing the inner tepals, whorl III is is unusual in having the opposite whorl I, and whorl IV opposite whorl II. With the exception of Litsea and “Chlorocar- dium," whorl IV is always staminodial or lacking. The stamens of whorl III nearly always have two globose glands at their base, the exceptions being Mezilaurus and Williamodendron, in which these glands are absent. In Phyllostemonodaphne and Urbanodendron all stamens have glands at their base. In Pleurothyrium the glands are strongly enlarged and may become fused, completely sur- rounding all stamens. In general, the stamens of whorls I and П are (nearly) identical and seemingly form one whorl of six stamens. The exceptions are Dicypellium and Phyllostemonodaphne, where the stamens of whorl I have become tepaloid and sterile, with only whorl II and III fertile. Stamens of whorl III differ occasionally from those of whorls I and II in number of anther cells. If different, whorl III has usually fewer anther cells (whorls I and II 4-celled, III 2-celled or sterile; whorls I and П 2-celled, whorl Ш sterile), but Endlicheria anomala has whorls I and II 2-celled and whorl Ш 4-celled. A further variation is that any of whorls I, II, and Ш can be sterile. (In some species only whorl I is fertile, or only I and II are fertile, or П and III are fertile, or only Ш is fertile; I have not yet seen a species with only whorl II fertile.) Table 1 presents the various androecial combinations and their distributions among neo- tropical Lauraceae, ““Gamanthera”’ excluded. If a key were to include all known combinations of number of stamens, their position and number of anther cells in each whorl, one would end up with a very long and tedious (but workable) key. How- ever, I found that a shorter and more practical key can be made by excluding the condition of the stamens of whorl III, as has first been done by Mez (1889). Exceptional species or groups of species, such as the 2-celled species of Persea and Сагу- odaphnopsis and the Aiouea species with three stamens are keyed out separately. Following the key, the distribution, approximate number of species and fruit type are briefly de- scribed. When an unusual leaf position occurs in a genus, this is also mentioned. A good magnifying glass, or preferably a dis- secting microscope, is a necessity for identifying Lauraceae. Included in the key are several taxa which are not yet published (“Chlorocardium,” “Сатап- thera,” and “Рагала”). Their inclusion in the key is absolutely not НЕ as their publication, but is done to prevent the key from being obsolete at the moment of publication. Key TO New WORLD GENERA OF LAURACEAE 1 Parasitic leafless vine о Cassytha 1 Shrubs or trees with green leaves ______ 2(1) Bod "lLbyc n cna fusion of 3 sta- nens *Gamanthera" 2 айна more than 1 3 3(2) Flowers bisexual or staminate __________ 4 3 Flowers pistillate 42 4(3) Stamens 3 5 4 Stamens more than 3 5)4) Stamens with 4 anther cells Williamodendron 5 Stamens with 2 anther се о 6(5) Leaves clustered near tips of branches; an- ther cells extrorse; inflorescence a panicle, Mezilaurus anther cells introrse); ый panicu liform, ultimate division cymose ............................ Only stamens of Whorl I fertile, the cells Volume 78, Number 2 van der Werff 379 Key to New World Lauraceae FIGURE 1. Flowers of: — А. Nectandra grandiflora Nees (Gentry et al. 59163); — B. Ocotea atirrensis Mez € Donn. Smith inus 6888): — C. Persea liebmannii Мет (Ishiki 1614); —D. Licaria bracteata van der Werff 56). (Kunkel 5 lateral and large, or lateral basal and small; 9(8) Flowers unisexual 10 inflorescences and flowers glabrous ....... Aiouea Flowers bisexual 11 7 Only stamens of Whorl III fertile, the cells 10(9) Leaves a re in winter; flowers appear- extrorse, introrse ог + apical, not lateral; ing before leave Lindera inflorescences and/or flowers usually with 10 Plants а flowers and leaves present varying amounts of pubescence ..................... Licaria at - same time Endlicheria 12 8(4) Outer 6 stamens each with 2 anther cells __ 9 11(9) Leaves opposite „i 8 Outer 6 stamens each with 4 anther cells ___ 23 11 pre alternate or clustered... 15 380 Annals of the Missouri Botanical Garden TABLE 1. Androecial configurations of American AE Lauraceae. Whorls I, II, & III 4-celled: Caryodaphnopsis pro parte, m pro parte, Litsea, Nec- tandra, Ocotea, Рагала, Persea pro parte, Pleuroth- yrium, Povedodaphne, E Sas- safras, Umbellularia, Urbanodendron pro par Whorls I, П, & III 2-celled: Aiouea pro parte, pt" pro parte, Beilschmiedia, Caryodaphnopsis pro parte, Cassytha, Cryptocarya, Endlicheria pro parte, Ku- bitzkia pro D um Persea pro parte, Urban- Chlorocardium, Cinnamomu odendron pro Whorls I & II 4- alod, III 2-celled: Cinnamomum pro parte, Kubitzkia pro parte, Persea pro parte Whorls I & II 4-celled, III O-celled: Persea pro parte Whorls I & II 2-celled, III 4-celled; Endlicheria anomala Whorls I & П 2-celled, III O-celled: Aiouea pro parte, naueria, Aniba pro parte, Caryodaphnopsis pro parte Whorl I 2-celled, II € III 0-celled: Aiouea pro parte Whorl I 0-celled, П & III 4-celled; Dicypellium Whorl I I II & Ш 2-celled: Phyllostemono- daphn Whorl I & II 0-celled, Ш 4-celled: Williamodendron Whorl I & II O-celled, III 2-celled: Licaria, Mezi 12(11) Stamens 6, triangular, dark red, the fila- ments ма, than the anthers, united at the uter tepals about У as long as the inner ones Anaueria 12 Stamens 6 or 9, filaments free, narrower than the anthers; stamens never dark red; tepals equal or strongly unequal, outer ones < the size of inner ones а. 13(12) Outer tepals =% the size of inner tepals Ca ies 14 13 Tepals equal 14(13) Floral tube deep, about as long as the tepals ry ptocarya 14 Жыз tube shallow, much shorter than the vals Beilschmiedia 15(11) Filamenta of stamens of Whorl II united; flowers red Pod 15 Filaments free; flower color variable, cluding red in some species ccoo 16 16(15) со of Whorl I sterile, tepaloid; fertile stamens 6, representing Whorls II and III Pn HM 16 Stamens of Whorl I fertile; fertile sta mens 9 or 6; if 6, representing Whorl I and Urbanodendron 17 NN stamens 6 or 9, only those of Whorl I Fertile s stamens 9, each with 2 glands at the bas Ш with glands at their base ...................................... 18 18(17) Outer tepals ca. Уг as long as inner ones (Fig. 1с . 19 18 Tepals equal or nearly so (Fig. la and b) . 20 NE Staminodia (whorl IV) well developed, with sagittate or cordate apices; floral tube very $429. МА vd A з à й > ‚> m «v И, 2А ۸ E RE 2. Stamens (whorl I or II) of: — Fic A. Nectandra retic ulata (К. & veraguensis a Mez UT 6476 = pulchella Mart. (Goetzke 210); — Е. “жо ан daphne kunthiana (Nees) Rohwer (Perry 5.п.). short, much shorter pud =e Andean species, 1,000-2,500 m Persea 19 Staminodia (whorl IV) RB floral ici well developed; lowland Amazonian speci uet Aniba canelilla ог А. bu Floral M very shallow, scarcely visible at anthes Beilschmiedia 20 Floral bé well developed, about as long as the tepals 21 21(20) Leaves with a fine, raised renculahon floral staminodia well develope C py plot arya 21 Leaves with lax reticulation or this not a ap- Volume 78, Number 2 van der Werff Key to New World Lauraceae 22(21) 26(25) 26 27(23) 200 28(27) 28 29(28 29 30(29) 30 31(28) 31 32(31) 32 33(32) 33 34(33) — — parent; floral tube shallow or deep; if deep, tepals erect and flowers about as wide as floral tube; filaments frequently poorly dif- f tiated fi thers; staminodi present or absent о pido tomentellous, tomentose with ap sed pubescenc e, but ||| s deredh nain poorly differentiated from an- thers; leaves alternate or clustere ments E d from anthers; leaves alternate . pom 24 Flowers unisexual Flowers bisexual Young inflorescences enclosed in decussate bracts; flowers arranged in pseudo-umbels . Litsea Young inflorescences not enclosed in de- cussate bracts; flowers arranged in panicles or racemes Deciduous trees; flowers appearing before or with young leaves, leaves often lobed ... Sassafras Pyerpnen plants with mature leaves when flowering; leaves never lobed Anther p arranged in a low arch (Fig. 2e) anthers poorly differentiated from fil aments Rhodostemonodaphne Anther cells arranged in 2 rows (Fig. 2c and d); anthers clearly differentiated from ij much narrower filament (rarely fila- nts very short) e Flowers arranged in pseudo-umbel, this, w tea young, covered by bracts ______ Umbellularia m arranged in paniculiform inflores- ences, these never enclosed by bracts ...... 2 Кее opposite 2 eaves alternate or whorled .. 31 Outer tepals < М the size of their inner ones; eaves often strongly triveined „u Car а Tepals equal; я pinnately veined Stamens 1 m n Stamens 9; leaves ferruginous tomentose or tomentellous on lower surface ectandra чен н cells on the flat пр of colum men Po vedadaphne Anther cells not apical; stamens not colum m stamens with 2 glands at their base, the glands free Urbanodendron Only stamens of Whorl III with glands (in ied es dpt glands greatly enlarged and metimes зора of Whorl III with fused filaments; flowers re Kubitzkia Stamens of Whorl III with free filaments; if filaments seemingly fused (in some Ocotea qon then flowers white to yellow, never Stamens of Whorl I devoid of anther cells, tepaloid Dicypellium mn 35 36(35) 37(35) 41(40) Stamens of Whorl I fertile, not tepaloid ES 35 о tepals about Y2 the size of the inner 36 Te Жр equal or nearly so 37 Stamens of Whorls I and II with a sterile, triangular tip; leaf apex rounded or obtuse; leaves whorled; Guyanas and adjacent Brazil and Venezuel Ocotea rubra Stamens of Whorls I and II without a sterile, bissl p, the anther cells occupying the entire anther; leaf apex various; leaves al- who rled; widespread, A. to Brazil and Chile ersea e outer 6 s I and II with at least 2 ae s leurothyrium Staminal glands not enlarged, not protruding between outer stamens; free; outer stamens with introrse cells Staminodia representing Whorl IV well de- veloped, with a cordate or sagittate apex; filaments of stamens as long as anthers or longer Staminodia representing Whorl IV small or tip, but these have stamens with a very short filament) Leaves usually tripliveined, alternate, fre- quently tufts of hairs present in the axils of the € veins. Zo ок ünnamomum with a pubescent base, otherwise glabrous “Paraia” Leaves alternate; twigs without clusters of scars from fallen bracts; tepals not roundish, not hyaline, glabrous or with different dis- tribution of pubescence Athe cells arranged in 2 vertical rows (Fig. and d); stamens and inner face of tepals Rim rous or variously pubescent, rarely pa- individually in old flowers; a few old sometimes present on cupule of young fruits; tepals at anthesis erect (Fig. 1b) or spreading Ocotea Anther cells arranged in an arc (Fig. 2a and b); stamens and inner face of tepals papil- lose; tepals united at the base, usually falling as a unit (together with stamens) in old flow- ers; stamens rarely present on cupule of young fruits; tepals at anthesis spreading (Fig. la Nectandra Plants cold-season deciduous and/or inflo- rescences subumbellate and young flowers Plants evergreen ana inflorescences panic- ulate, the in involucrate 382 Annals of the Missouri Botanical Garden 43(42) Inflorescences racemose or paniculate; leaves trilobe Sassafras ise sn subumbellate; leaves never 43 trilo 44(43) Plants deciduous; pseudo-umbels sessile or nearly so indera 44 S e or evergreen; pseudo-um- bels peduncula Litsea 45(42) Staminodes Кар, the filaments about as wide as the anther; traces of locelli four each a odixtemondaphn 45 Staminodes club- -shaped, ‘the filaments than the anthers 46 46(45) Anthers with 4 remnants of locelli; leaves alternate cotea үне with 2 remnants of locelli; leaves и с ог whorled Endlicheria 46 CLAVE PARA LOS GÉNEROS DE LAURACEAE DEL NUEVO MUNDO 1 Trepadoras parásitas у áfilas ______ E 1 Arbustos o árboles con hojas verdes ___ 2 2 Estambre e 1 por fusión com- pleta de *, 3 estambres “Gamanthera’ 2 Estambres ae más de 1 3 Flores bisexuales o estaminadas 3 Flores pistiladas 4 Estambres 3 5 Estambres más de 3 5 Anteras con 4 tecas ....................... 5 Anteras con 2 tec 6 diy oe en n 10s extremos de las ra- mas; tecas extro inflorescencia una pa- nícula con las тене ша ültimas racemosas Mezilaurus 8 Williamodendron 6 6 Hojas alternas u opuestas, no agrupada orientacion de las tecas variabl n L. can nella hojas algo agrupadas, pero entonces las tecas introrsas); inflorescencia a paniculi- forme, ا‎ Últimas cimosas ow... 7(6) Solamente los estambres del Verticilo I fér- tiles, tecas laterales y grandes o lateral-ba- sales y pequeñas; inflorescencias y flores labra Aiouea 7 сецкани los estambres del Verticilo Ш fértiles, tecas extrorsas, introrsas o +apica- les, nunca laterales; inflorescencias y/o flo- res usualmente pubescentes co Licaria 8(4) Ante e los 6 estambres exteriores con 2 tecas 8 Anteras de los 6 estambres exteriores con 4 tecas 9(8) Flores unisexuales 10 9 Flores bisexuals 11 10(9 ANT flores produci- das antes que las hojas „un Lindera 10 Hojas no deciduas, flores y hojas presentes vez Endlic ^ 11(9) Hojas opuestas Hojas alternas o agrupada: 12(11) Estambres 6, rojo- -obscuros, « con a filamentos más anchos que las anteras y unidos en la base; tépalos exteriores cerca de Jê del largo de los interiores Anaueria 12 13(12) 13 14(13) 19(18) N 2(21) N N 23(8) 23 Estambres 6 o 9, nunca rojo-obscuros, con r | exteriores 1⁄4 o menos del largo de los eriores Tépalos e exteriores 1⁄4 o menos del largo de los interi e Caryodaphnopsis Tépalos i iie Miss 14 Tubo floral profundo, tan largo como los tépalos Cryptocarya d floral ms profundo, mucho más corto e los tépalos s Beilschmiedia Posee del bea III con filamentos unidos; flores rojas ss ubitzkia celos con or aren libres; flores rara- nte ro Estambres “del Verticilo I estériles, tepa- Estambres del Verticilo I fértiles; estam fertiles 9 ó 6, si 6 entonces ий sido los Verticilos Ту II Estambres fértiles 9, cada uno con 2 g dulas en la base cc. banodendron Estambres fértiles 6 ó 9, a los del Verticilo I con glándulas en la bas Tépalos exteriores cerca c % del largo de los interiores (Fig. Tépalos iguales o casi кине (Fig. Ta: y b) ere bien desarrollados, con ápices agitados o cordados; tubo floral mu y corto, stam nodi sidani tubo floral bien de- mani especies amazonicas de eleva ciones bajas — Aniba canelilla o А. parviflora Tubo floral Aone profundo, escasamente vis ible en la а eilschmiedia ancho que el tubo; filamen o más angostos que las anteras; estaminodios bien desarrollados sl ryptocarya ojas con una reticulación laxa o inconspi- teras; estaminodios presentes o ausentes Flores menudamente t tosas o aplicado ubescentes, nunca pruinosas; filamentos con frecuencia densamente pubescentes, poco diferenciados de las anteras; hojas alternas о agrupadas niba Flores usualmente glabras, infrecuente- 5 ferenciados de las anteras; hojas alternas Aiouea 24 Flores unisexuales Flores bisexuales 27 Volume 78, Number 2 1991 van der Werff Key to New World Lauraceae 383 24(23) 24 25(24) 25 26(25) 26 27(23) 29 30(29) 30 31(28) 31 32(31) ор jóvenes envueltas en brác- s decusadas; flores en pseudo-umbelas .. Litsea Inflorescencias jóvenes no envueltas en — decusadas; flores еп paniculas о cimos Arboles deciduos, flores producidas antes o nto con las hojas; hojas frecuentemente lobadas Sassafras Arboles o arbustos perennifolios, flores prod- ucidas con las hojas maduras; hojas nunca lobadas Tecas dispuestas en un arco bajo; anteras poco diferenciadas de los filamentos (Fig. 2e) Rho ondes ETA Tecas a gii en 2 hileras; anteras clara- mente más anchas que los Alam ntos (rara- mente a muy cortos) (Fig. 2c y d) О otea Flores en pseudo-u о éstas cubiertas OT nes Um bellularia а о sur d 9; hojas — tomen nele Nectandra жиды lia Tecas e en nel apice plano de m estambres columnares Povedadaphne Tecas no apicales, estambres no columnares Todo los estambres con 2 glándulas en la MES Urbanodendron ауе (en Pleurothyrium las glándulas y agrandadas y a veces fusionadas) Pisces del а ПІ con filamentos fusionados; flores ro Kubitzkia flor nunca rojas Estambres del Verticilo 1 estériles y tepa- loides i м Estambres del Verticilo I fértiles у по te- paloides Tépalos exteriores cerca de Y2 del tamaño de los interiores Tépalos ias o casi iguales Estambres de los Verticilos I y II con ápices triangulares y estériles; hojas con y re- dondeados u obtusos; Guayanas y áreas yacentes de Brasil y Venezuela _ - Ocoten rubra Estambres de Verticilos I y Il: о a a rasil hast P 37(35) Glandulas estaminales muy agrandadas, em- 41(40) > = 44(43) 44 45(42) pujando entre los 6 estambres es inpea HE a veces fusionadas en una masa almohada; estambres de los Verticilos b y 11 con al menos 2 de las tecas laterales leurothyrium Glandulas estaminales no agrandadas ni fu- arde — exteriores con todas las as intro коол representando el Verticilo IV bien desarrollados, con ápices cordados o sagitados; санах iguales о más largos que las a Estaminodios Bi eem sun Verticilo IV au- sentes о si presentes pequ eños y mayor- mente sin apices cordados о sagitados; fi- (unas pocas especies de Ocotea tienen es- cuentemente con fasciculos de tricomas pre- sentes en las axilas de los nervios basales _ лппатотит Нојаѕ pinnatinervias, alternas o agrupadas, l omas axilares ____ ers sin fascículos redondos, hialinos, pubescentes solamente en la base “Paraia” Hojas alternas; ramitas sin fascículos de ci- catrices; пен ем s ni redondos ni hialinos, gla- m on la pubescencia distribuida de otra rma Tecas dispuestas en 2 hileras (Fig. 2c y d); ta erectos o patentes en la antesis (Fig. Tecas dispuestas en un arco (Fig. 2a y b); estambres Plantas deciduas en el invierno y/o inflo- ы || Lal 1 A e Ј vueltas por brácteas involucrales Plantas perennifolias con inflorescencias paniculadas, fl involucrales 5 ас саз ћо- jas frecuentemente trilobadas Sassafras © florescencias subumbeladas; hojas nunca rilobadas Plantas deciduas; pseudo-umbelas sésiles o asi sésiles Lindera is deciduas o perennifolias; pseudo- mbelas pedunculadas Litsea Puls ligulados, filamentos igual de 384 Annals of the Missouri Botanical Garden ancho que ya anteras, cada antera con ves tigios de 4 tecas ______ odostemonodaphne 45 cad. lair filamentos delgados que las anter 4 46(45) ам de los indt con vestigios de 4 tecas; hojas alternas cotea 46 Anteras de los estaminodios con vestigios de 2 tecas; hojas alternas o verticiladas ............ Endlicheria Aiouea Aublet Probably a polyphyletic genus in the currently accepted sense, as has been noted by several au- thors (Burger, 1988; Rohwer et al., in press; van der Werff, 1987b, 1988). The genus has been monographed by Renner (1982) and consists of about 20 species, ranging from southern Mexico to southern Brazil and Paraguay. M the fruit seated on a shallow cupule with a thickened pedicel. 051 species ha ve г Апаиегіа Kostermans А monotypic genus known from Amazonian for- ests near the border of Brazil and Peru. The seeds are said to be edible after roasting. A cupule is lacking in fruit. Flowers have green tepals and dark red stamens, very unusual in Lauraceae. Aniba Aublet Recently revised by Kubitzki (1982). Aniba in- cludes 40-50 species, many with pleasantly aro- matic yellow wood (rosewood oil is distilled from Aniba rosaeodora). It is distributed from Costa Rica to southern Brazil and Bolivia. Most species occur in the lowlands, but it has also been collected in the Andes up to 1,800 m. The fruit is seated in a rather deep cupule, which is often lenticellate or warty. Several species have clustered leaves. Beilschmiedia Nees In the Neotropics a poorly understood genus with 10-20 species, last revised by Kostermans (1938), who accepted 15 species. It has been col- lected from Mexico to southern Brazil and in the West Indies. The fruit lacks a cupule. In the Pa- laeotropics, Beilschmiedia is equally poorly un- derstood, but is represented by many more species. Caryodaphnopsis Airy Shaw Only recently reported from the Neotropics (van der Werff & Richter, 1985), this genus includes five published neotropical species, with several more awaiting publication. The strongly unequal tepals and opposite leaves are diagnostic. Fruits (in neo- tropical species) are round or pearshaped, and seat- ed on a naked pedicel. It occurs in lowland forests from Costa Rica to Peru and Brazil. The seven palaeotropical species occur in Indochina, with one species extending to New Guinea (Kostermans, 1974 Cassytha L. One species in the Neotropics, from Mexico to Brazil, often found near the coast, but also locally common in Mauritia swamps. More species occur in Australia and South Africa. The genus has been placed in its own family because of its parasitic habit, but is here retained in the Lauraceae. “Chlorocardium” A genus with two species, one from Guyana and Surinam, the other from Amazonian Ecuador and adjacent Colombia, previously included in Ocotea, but differing in characters of flowers, wood, and leaf position. The Ecuadorian/Colombian species is incompletely known. Fruits d as large (to 7 x 5 cm), with a normal or very large cupule. Cinnamomum Schaeffer A large, poorly understood genus occurring in Asia, Australia, and the Americas. Possibly with more than 50 species in the Neotropics, known from Mexico and the West Indies to southern Brazil and Paraguay. The fruit is subtended by a small cupule, this often crowned with persistent tepals. The neotropical species have often been treated in Phoebe, but seem better placed in Cinnamomum. Many Asian species, including C. verum, which yields cinnamon, have opposite leaves, but all neo- tropical species have alternate, mostly tripliveined leaves. Cinnamomum, as accepted here, is possibly polyphyletic. Cryptocarya R. Brown A large, poorly understood pantropical genus, with ten or fewer neotropical species, mostly in southern Brazil and Chile, but also known from French Guyana and adjacent Brazil, Andean Ven- ezuela, Ecuador, and Peru. e fruit is almost completely enclosed in the cupule, which has only a small, apical pore. Dicypellium Nees & Martius A small genus of two rarely collected species, restricted to Amazonian Brazil, reviewed by Roh- wer (1988). The cupule is cupshaped, and has a double margin. The tepals are persistent in fruit. One species, Dicypellium caryophyllaceum, has been heavily exploited because of its aromatic bark. Volume 78, Number 2 1991 van der Werff 385 Key to New World Lauraceae Endlicheria Nees A neotropical genus of about 40 species, re- ported from Costa Rica, the Lesser Antilles south to Paraguay, and southern Brazil. The genus is polyphyletic (Rohwer et al., in press). The cupule is variable; most species have a thick, fleshy and smooth cupule. Several species have whorled leaves. “Gamanthera” A monotypic genus recently discovered in Costa Rica, characterized by the presence of only one actually a synandrium formed by com- plete fusion of three stamens (Endress, pers. comm.). The number of locelli of the synandrium can be 3, 2, or 1. The cupule is rather deeply cupshaped, with a double margin and persistent, reflexed te- pals. The only species is monoecious. Kubitzkia van der Werff A neotropical genus of one, possibly two, species, known from Guyana, Surinam, Venezuela, and Brazil, reviewed by Rohwer (1988), who accepts Systemonodaphne as the valid name for this genus. The cupule is cupshaped, with a double margin and persistent, reflexed tepals. Arguments for re- jecting the name Systemonodaphne are discussed by van der Werff (1986) Licaria Aublet A neotropical genus of about 40 species, revised by Kurz (1983). It occurs from southern Florida and Mexico to southern Brazil and Bolivia. The cupule is cupshaped and has a double margin; however, in some species the double margin is scarcely noticeable. A few species have opposite leaves. Lindera Thunberg Lindera is represented by three species in east- ern North America and an additional hundred in Asia. The fruit is seated on a small, platelike cupule. Two of the American species are rare; one was only recently described (Wofford, 1983). Twigs, leaves, and fruits of the common American species are used to prepare a fragrant tea. Litsea Lamarck In the New World Litsea is represented by about five species, known from the ., the mountains of Mexico, and Costa Rica. Most species (several hundred) occur in Asia south to Australia and the Pacific Islands. The fruit is subtended by a cup- shaped cupule. Leaves of Litsea glaucescens are used as a spice, similar to bay leaves. Mezilaurus Taubert A neotropical genus of 16 species occurring in the Amazon basin (including Bolivia, Peru, Colom- bia, Venezuela, and the three Guyanas), revised by van der Werff (1987a). The fruit is seated on a small, platelike cupule. The flowers are very similar to those of Licaria, but the two genera differ in cupule shape, leaf position, inflorescence type and wood anatomy. All species have leaves clustered at tips of branches. Two species with 4- celled anthers were recently transferred to Wil- liamodendron (Kubitzki & Richter, 1987). Nectandra Rol. ex Rottb. A large, neotropical genus of about 120 species, currently under revision by J. Rohwer. It occurs from southern Florida and Mexico to Argentina. Fruits are seated in a cupshaped (sometimes small) cupule. Tepals are spreading at anthesis. Ocotea Aublet The largest genus of Lauraceae in the Neotrop- ics, with at least 300 species. It is also known from Madagascar and tropical Africa. The genus is very variable and serves as a dumping ground for species that cannot be readily accommodated in other gen- era. Its distribution in the Neotropics is from Mex- ico and southern Florida to Argentina. Shape and size of cupule is variable and ranges from small and platelike to cupshaped, sometimes with a dou- ble margin or with persistent tepals. Tepals are erect or spreading at anthesis. " Paraia" A monotypic genus known from Amazonian Bra- zil. The cupule has a double margin with persistent tepals. Tepals are erect at anthesis. Persea Miller An incompletely understood genus with at least 80 neotropical species, but better represented in Asia and with one species in the Canary Islands. The fruit can be seated on a naked pedicel (as in Persea americana, the type species) or 15 sub- tended by persistent, indurate tepals (as in the great majority of species). A cupule is never present. A worldwide survey is needed to determine if Machi- us belongs in Persea (as is accepted here) or should be recognized as a distinct genus. The neotropical species have been revised by Kopp (1966). Persea ~ 386 Annals of the Missouri Botanical Garden americana is widely cultivated for its edible fruit. A few neotropical species have clustered leaves. Phoebe Nees Phoebe is here considered a palaeotropical ge- nus. Neotropical species formerly placed in Phoebe are, for the larger part, included in Cinnamomum; a few belong to Ocotea. Phyllostemonodaphne Kostermans An infrequently collected, monotypic genus from the Atlantic rainforests in southern Brazil. The fruit is seated in a double-rimmed cupule. The genus was revised by Rohwer (1988). Pleurothyrium Nees A neotropical genus with close to 40 species, known from Guatemala south to Peru, Bolivia, and Brazil. The fruit is seated in a rather deep, cup- shaped cupule, which is often warty or lenticellate. Several species have a pronounced marginal vein, an unusual feature in Lauraceae, or clustered leaves. The genus is being revised by van der Werff. Povedadaphne Burger A monotypic genus apparently restricted to Cos- ta Rica. The rather large fruit is subtended by a small, platelike cupule. The fruiting pedicel is swol- len. Distinctive are the nine columnar stamens, each with four apical locelli. Rhodostemonodaphne Rohwer & Kubitzki A poorly known genus of about a dozen species, two of which are rather common, while several species are in need of description. The genus has been reported from Costa Rica to Brazil and Peru. Тће cupule is deeply cupshaped and rather large. Sassafras Presl A small genus with three species, one widespread in the eastern U.S.A., Leaves are frequently lobed. the other two in China. The cupule is cup- shaped, often crowned with remnants of the tepals. Bark and root have been used to prepare tea or root beer. Systemonodaphne See Kubitzkia. Umbellularia Nuttall A monotypic genus We e in California and southern Oregon (U.S by a small, platelike iie The leaves are use e fruit is subtende as a spice; the wood is valued for cabinetmaking. Urbanodendron Mez A neotropical genus of three species, restricted to southern Brazil. The cupules are known in two species: double-rimmed, with more or less persis- tent tepals. The genus has been revised by Rohwer (1988 Williamodendron Kubitzki & Richter A neotropical genus with two species, infre- quently collected, but known from Costa Rica, northern Colombia, Amazonia, and southern Brazil. The number of species will likely increase as more collections become available. The cupule is small and platelike, resembling that of Mezilaurus. Leaves are clustered at the tips of branches. LITERATURE CITED ALLEN, C. K. 1945. Studies in Lauraceae VI. e- semen overview of the Im and Central pue es. J. Arnold Arbor. 8 1988. Е пем Tenue n Lauraceae from Costa Rica, with c n problems of generic and фе ышып within the family. Brittonia 275-282. Hurcninson, a 1964. The Genera of Flowering Plants, Volume I. Clarendon Press, Oxford. Kopp, L 966. А taxonomic revision of the genus Persea in the КЕ Hemisphere. Mem. New York Bot. Gard. 14: E KOSTERMANS, А. J. e 1936. Lauraceae. /n: Flora of ата: ics Kolon. Inst. Amsterdam, Afd. Handelsmus. Volume II. 244-337 1938. wee a the Тангнга У. Кес. ae Neerl. 35: 1. 957 со UE Forest Res. Inst. dw в А сог ~ yy A. Жылма Ни 9: 12 кас К. yn б», 8 Neotropica 31: 1-84 & . RICHTER. 1987. Williamodendron Kubitzki & Кеш. а new genus of neotropical Lau- raceae. Bot. Jahrb. Syst. 109: 49-58. Kurz, H. Fortpflanzungsbiologie einiger Gattun- gen neotropischer Lauraceen und Revision der Gat- oe Ph.D. Thesis, Univ. of urg, Ham MacBrive, F. irs Todes In: Flora of Peru. Field Mus. Nat. Hist. Bot. Ser. 13(2): 819-931. МЕХ, C. 1889. елш чыр Americanae. Jahrb. Kónigl. Bot. Gart. Berlin 5: 1-556. one In: Flora Neotropica 31: -116. RICHTER, Н. С. 1981. lems und der Rinde der а Зопдегђ. wiss. Vereins Hamburg 5: The genera Dicypellun. deis ‚ Systemonodap and Urban dendron Torino Bot. jab. Syst. 110: 157. 171. Anatomie des sekundáren Xy- b. Natur- — ——-, Н. С. RICHTER & H. VAN DER Кези, Two new genera Lol neotropical Lai Jae ritical Моште 78, Митбег 2 1991 van der Werff Key to New World Lauraceae remarks on the PU: delimitation. Ann. Missouri ot. Gard. 78: 388-4 тап, a VAN DER. 1986. Kubitzkia van der Werff, e for : us of neotropical Lauraceae. Тахоп 35: 164- 1987a. me" а (Lauraceae). -182 98 kan, Missouri Bot. Gard. 74 198 ix new species о ~ мы Laura- ceae. Апп. Missouri Bot. Gard. 74: 401-412. . 1988. Eight new species of neotropical Laura- ceae. Ann. | Вог. Саг І . RicHTER. 1985. Caryodaphnopsis Airy Shaw caved a Ps new to tropics. Syst. pps à 166- WOFFORD, E. B. . А new ARTA (Lauraceae) from North с Ј. Агпоја Агђог. 64: 325-331. the Neo TWO NEW GENERA OF NEOTROPICAL LAURACEAE AND CRITICAL REMARKS ON THE GENERIC DELIMITATION! Jens С. Rohwer,’ Hans Georg Richter,’ and Henk van der Werf? ABSTRACT Two genera, Chlorocardium and Paraia, are described, illustrated, and discussed. Chlorocardium includes two na and Surina m (C. rodiei), and Ecuad or (C. venenosum). The exceptions, are circumscribed by a combination of genera of the Lauraceae, with very few characters rather than by single exclusive char- acters. However, Kostermans (1957) noted that “Apparently the combination of characters is more or less indefinite, and almost all combinations are represented.” Although this statement must not be taken literally, because in strictly mathematical terms millions of combinations would be possible with the given characters, it does emphasize the reticulate variation within the family. Naturally, not all combinations are equally frequent. The more frequent combinations correspond to the larger genera, while rare combinations have been treated in two different ways. They either have been placed in a genus of their own (e.g., Anaueria, Dicypel- lium, Phyllostemonodaphne, Urbanodendron), or they have been included as aberrant species in one of the larger genera. In the present paper we de- scribe two new genera based on rare combinations of characters, and we discuss some of the discor- dant elements in larger genera, as well as several cases of unsatisfactory generic delimitation. Тће first of our new genera has the most unusual flower structure. Ironically, it is based on a species that normally would be considered to be among the better known Lauraceae. Chlorocardium Rohwer, Richter & van der gen. nov. HOLOTYPE: Chlorocardium erff, rodiei (Schomb.) Rohwer, Richter & van der Werff Ceteris Lauracearum generibus differt foliis oppositis, qualibus, omnibus quadrilocellatis latrorso-extrorsis, ses- silibus, basi biglandulosis, ovario pubescent Chlorocardium rodiei (Schomb.) Rohwer, Richter & van der Werff, comb. nov. BASIONYM: Nectandra rodiei Schomb. in Hook., London J. Bot. 3: 616. 1844. TYPE: Guyana. 20-50 mi. inland, along the rivers Essequibo, Cuyuni, Demerara, Pomeroon, Berbice, Schomburgk 1004 (holotype, K; iso- types, G, L; probable isotypes with no. 1703 in B, GZU). Nectandra leucantha var. rodiaei (Schomb.) Griseb., Fl. Brit. W. I. 282. 1860. Ocotea rodiei (Schomb.) Mez, Jahrb. Königl. Bot. Gart. Berlin 5: 236. 1889. Chlorocardium rodiei is a species of consider- able economic importance. Its wood (“Стееп- heart," hence the name Chlorocardium) is among ! We tha nk the curators of F, INPA, and NY for the loan of specimens and/or for providing duplicates. L. E. Skog (US) provided additional flowers of Chlorocardium rodiei, J. sim (NY) provided the wood sample of Paraia bracteata, and John M yers made the habit drawing of dde bractea > Institut für Allgemeine Botanik, Ohnhorststrasse 18, D-2 000 Hamburg 52, Federal Republic of German * Institut für Holzbiologie und Holzschutz, Bundesforschungsanstalt für Forst- und Holzwirtschaft, Leuschnerstrasse 91, D-2050 Hamburg 80, Federal Republic of Germany. * Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166-0299, U.S.A. ANN. MISSOURI Bor. GARD. 78: 388-400. 1991. Volume 78, Number 2 Rohwer et al. 389 1991 New Genera of Neotropical Lauraceae FIGURE 1. stamens each. — B. Flower, nearly all tepals remove of the first whorl, seen from side (note placemen seen from outside. — F. Stamen of the fourth whorl, seen from outside. — G. the most important commercial timbers in the fam- ily. Although the species has been known for more than 200 years, it is still poorly represented in the major herbaria, and very few flowering collections are known. This scarcity of flowers permitted a misinterpretation perpetuated until today. Schom- burgk mentioned in the original description that most flowers appeared to be abnormal, with more than the usual number of parts. Similar statements can be found in all subsequent accounts on the species. Until today everyone who dissected a flow- er (including one of the authors) believed that he or she had picked a monstrosity when the flower turned out to be tetramerous or irregular, yet no one wanted to dissect a second flower because there was so little material. As it turns out now, with more material available, the only regular flowers are indeed tetramerous (Fig. 1A), and irregular flowers are common in all specimens. Bentley & Trimen (1880) described nt on the mde. of the floral tu id Chlorocardium rodiei. — А. Floral е of a sd flower with four whorls of stamens with xed —C. Sta n of the first whorl, seen outside. — D. Sta ).— E. Sta uat the third hal vary. anoles Pipoly 7530. the flower as trimerous, with nine fertile stamens. We could not find this condition. Of the 14 flowers that we examined only one appeared to be trim- erous, but it had 12 fertile stamens. Eight flowers were clearly tetramerous, of these we found: 1 with 4 tepals and 4 whorls of 4 stamens each 2 with 8 tepals and 3 whorls of 4 stamens each 3 with 8 tepals and 4 whorls of 4 stamens each 2 with 8 tepals and 5 whorls of 4 stamens each. The remaining five flowers were irregular, usu- ally with less than the full number of stamens in the inner whorls. One of the irregular flowers had only six tepals, and another had eight tepals but four whorls of five stamens each. Tepals with one or two apparently fully developed pollen sacs were found in both a regular and an irregular flower. The stamens (Fig. 1C-F) are almost equal, broadly sessile, and more or less tongue-shaped, each with two rather inconspicuous glands at the base and 390 Annals of the Missouri Botanical Garden with four locules, of which the upper ones are latrorse and the lower ones latrorse-extrorse. To- ward the center of the flower the stamens become slightly smaller, and the inner whorls are inserted at successively lower levels within the floral tube. This configuration is reminiscent of the closely related Monimiaceae, but without parallel in the Lauraceae. It may be interesting to note that also the native name in Guyana is shared with a Monimi- aceae. It is reported to be “Bebearu,” “Bibiru,” “Sipeira,” “Sipiri,” or “рига” —undoubtedly the same root from which Aublet (1775) derived the name Siparuna. Schomburgk reported that in times of food short- age the seeds of the Greenheart tree were used by the natives in a way similar to cassava, i.e., they were grated, and the starch was repeatedly washed to deprive it of the bitter alkaloids. There is some uncertainty concerning the spell- ing of the specific epithet. The species is named for “Dr.” described the medicinal uses of the plant. In the original publication the epithet is spelled **rodioei," with the oe printed as one character. Sandwith (1939) suggested that this was indistinguishable from ae, and that the correct spelling should be because the name Rodie would be latin- odie, a Royal Navy physician who first "rodiaei," ized as “Rodiaeus.” However, from the following text of the description it is obvious that oe and ae are indeed different characters. Since the spelling with oe lacks any foundation, we suggest that [in accordance with Art. 73.10. of the Code (Greuter, 1988) | it should be corrected to “rodiei,” been done by Mez (1889) and most subsequent as has authors. The second species of the genus Chlorocardium is poorly known: Chlorocardium venenosum (Kosterm. & Pink- ley) Rohwer, Richter & van der Werff, comb. nov. BASIONYM: Ocotea venenosa Kosterm. & Pinkley, Bot. Mus. Leafl. Harvard Univ. 22: 241. 1969. TYPE: Colombia. Putumayo: Rio Guamúes, Santa Rosa, alt. 1,060 ft., 77905", 00?19'N, 26 Nov. 1966, Pinkley 555 (ho- lotype, СН; isotypes, ECON, S, U). In the were od ription of this species the authors state t the flower buds avail- able for огеш уеге на as a result of fungal attack." In fact, they are so few and im- mature on the type specimen that it does not look worthwhile to dissect any of them. The general aspect of the plant agrees very well with Chloro- cardium rodiei, and the wood structure described leaves no doubt that the two species are extremely close. Chlorocardium venenosum is known only from the border region of Colombia and Ecuador, while we have seen material of Chlorocardium rodiei only from Guyana. It also occurs in Surinam, but reports of a much wider distribution (as in Mell & Brush, 1913) appear to have no foundation. The structure of the secondary xylem of the two species described here under Chlorocardium is highly unusual in two aspects: The interconnecting pit pairs between vessels as well as between vessels and parenchyma cells (rays and axial parenchyma) are extremely small (3-4 um diam.), a condition absent not only from Nectandra and Ocotea, but from all Lauraceae with the exception of the small Asiatic genus of Neocinnamomum. More impor- tantly, parenchyma cells are connected to adjacent vessels through unilaterally compound pitting with large, simple pits subtending two or more of the small bordered pits in the vessel elements. These large pits in the parenchyma cells are often tra- versed by branched or simple filiform protrusions of cell wall material. [In addition, vessel elements in both species bear at the ends a partial or com- plete ring of large, windowlike pits that give the арн of a crown or corona, hence the term used by Kostermans et al. (1969) to describe this phenomenon quite А *coronated vessel elements” unique in xylem anatomy. The other aspect considered here is the complete lack of secretory cells in the secondary xylem. What might have been a chance observation for Chlorocardium venenosum described from only one specimen so far (Kostermans et al., 1969) becomes significant for C. rodiei: The countless descriptions dealing with the secondary xylem of this taxon have not once revealed the presence of secretory cells. Thus their absence, although gen- erally not diagnostic in Lauraceae, might justly be used in conjunction with other characters to seri- ously question the traditional taxonomic afhnities of these two species. These findings are corroborated by the equally unusual structure of the secondary phloem of C. rodiei (for detailed description see Richter, 1981), which is characterized by special features different from the prevailing structural patterns in Laura- ceae, among them the complete lack of secretory cells (sic). Thus, the wood and bark structure supports a rather isolated position for Chlorocardium, but unlike the flower structure it does not suggest an affinity with the Monimiaceae. Fruits of Chlorocardium rodiei have been de- scribed as large (6-7 cm long, 5 cm diam.) and Volume 78, Number 2 Rohwer et al. 391 New Genera of Neotropical Lauraceae FIGURE 2. — А. Flowe stamens removed. — D. Undescribed new genus. F. Stamen of the first whorl, seen from side (curved inward H. Stamen of the second whorl, seen from inside ellipsoid to ovoid-globose, seated in a shallow, sin- gle-margined cupule, 2-2.5 cm diam. (Mez, Kostermans, 1936). A recently collected specimen (Pipoly 8616, MO) shows a not very deep, len- ticellate cupule, ca. 1.5 cm exserted, ovoid fruit, ca. 1.5 cm long. Fruits and cupule are attached to the twigs. The tree was still iam. and a clearly flowering at the time of collection, and it is possible that the fruits were immature. Fruits of C. nosum were described as depressed-globose, 4.5 cm high and 5.5 cm diam., enclosed by the cupule. А recent collection with fallen fruits agrees very well with this description; the fruits are up to 6 cm high and 7 cm the cupule largely encloses the fruit, leaving an vene- almost completely iam., orifice of about 2.5 cm diam. Fruits are used by the Kofan Indians to make arrow poison. The very large cupule of C. venenosum readily separates this species from C. rodiei with its much smaller, — B. Flower, tepals removed. . Stamen of the first whorl, seen from inside. — — (С. Stamen of the second whorl, seen from outside. — .— I. Stamen of the second whorl, seen from side.—J. Ovary, its position in the receptacle schematically shown. Voucher: Vásquez 6020. Stamen of the first whorl, seen iver outside. — — С. Receptacle with ovary, outer shallow cupule. With the exception of the recently described Ocotea megacarpa, fruit size of the two Chlorocardium species far exceeds that of any Ocotea or Nectandra species. Recently we found another species with opposite leaves and variable staminal configuration that in some characters resembles Chlorocardium but also shows marked differences (Fig. 2). Like in all other neotropical 1 ( pt Chlorocardium) the flowers are trimerous, invariably with six tepals and not more than three whorls of fertile stamens. The fourth whorl, which in many Lauraceae is stami- nodial, seemed to be represented only in the frag- ments of one broken flower. In 25 reasonably intact flowers examined we found the following configu- rations: 10 flowers with whorl I and II 4-locular, whorl III 2-locular 392 Annals of th Missouri ا‎ Garden 8 flowers with whorl I and II 4-locular, whorl III absent 4 flowers with all three whorls 4-locular 2 flowers irregular In the 2-locular stamens it is the upper pair of locules that is lost. The lower pair, like in Chlo- rocardium, is latrorse-extrorse in all stamens; the upper pair can be latrorse or latrorse-introrse. Like in Pleurothyrium the flaps open toward the abaxial side of the stamens. The outer stamens (especially whorl I, Fig. 20-Е) are bent inward, overarching the inner stamens. All stamens have distinct, broad filaments, and there is no trace of glands. The large ovary is inserted in a deep, cup-shaped receptacle. All parts of the flower, including the ovary and the inside of the receptacle, are covered with papillae. This combination of characters is unique within the family, and we have no doubt that our material represents a new genus. However, as our obser- vations are based on only one collection (К. Vás- quez 6020), we postpone the formal description until more material becomes availa The second new genus that we describe in this paper is far less unusual in its staminal configu- ration. Yet its combination of characters does not fit any of the known genera. Paraia Rohwer, Richter & van der Werff, gen. nov. HOLOTYPE: Paraia bracteata Rohwer, Richter & van der Werff Ceteris Lauracearum generibus differt foliis in apice ramulorum айран и ШАН congestis, flori bus subglobosis racteis conspicu subtentis, tepalis margine hyalinis interne epapillosis glabrisque, staminibus novem, antheris | locellis arcuatim juxta- ovario bene evolut tubo profundo, cupula du- ae nos longiore et perianthii segmentis auctis sexlobato. Paraia bracteata Rohwer, Richter & van der , Sp. nov. TYPE: Brazil. Pará: Mun. Orixi- miná, Rio Trombetas, right bank, Porto Trom- betas, road to Міпегас̧ао Rio Norte, km 60, beyond Bauxite mine, 29 Aug. 1980 (fl, fr), Cid et al. 1890 (holotype, MO; isotypes, HBC, INPA 96058 not seen, NY). Figures 3, 4. Frutex vel arbor ad 8 m (20 m?) alta. Ramuli teretes, hornotini omnia apic ta vel oblonga, adulta 14-35 cm longa, 5 apice ex acuto vel obtuso manifeste pata, basi cu- neata vel rarius acuta (basi ipso interdum obtusa), in sicco + olivacea, adulta supra opaca vel nitidula glabra, subtus peropaca subglabra, margine plano, nervis secundariis utroque costae latere 6-11 e nervo medio sub angulo 40-80? prodientibus. Inflorescentiae e gemmae terminali m axillis prodientes, anguste thyrsoides, quam subglobosum, haud patens, suborbiculares, basi crassa, margine hyalino, externe basi et linea media tomento ferrugineo, interne epapillosa gla- braque. Stamina novem, ca. 0.8-1.5 mm longa, epapil- losa, quadrilocellata vel serie prima raro bilocellata, sex uds eglandulosa filamentis gracilibus locellis subap- calibus arcuatim hd ines unica марш, tria interiora crassimsendia lateralibus. ee sein parva, subulata, dense tomentosa. Receptaculi tubus profundus, saltem longitudine stami- num, angustus, intus dense tomentosus. Pistillum glabrum, elongatum, ca. 2-2.4 mm longum, stylo ovarium ae- quante vel eo paulo longiore. Cupula obconica vel sub- 5 mm diametro, 6-8 mm profunda, , margine exteriore longiore et perianthii lobis incrassatis manifeste sexlobato. Bacca ellipsoidea vel ooo ovoidea, ca. 14-24 mm longa et 10-12 mm diam Shrub or small tree, usually to ca. 8 m tall (only Silva 2220 from a tree 20 m tall). Twigs round, the youngest with a dense cover of long (ca. 0.4— mm long), straight to curled, + erect, relatively thick, reddish brown hairs, quickly becoming felt- like, easily worn off. Terminal buds covered by relatively large (4-7 mm) scales, somewhat spread- ing, with the same kind of indument as the twigs on outside (though not always dense at the mar- gins), glabrous on inside. Leaves drying with a characteristic grayish green to olive-brown color, alternate but with few exceptions clustered at the tip of the branchlets (sometimes a second, older whorl still present), 14-35 cm long, 5-12 wide, ca. 2.3-4(-4.5) times longer than wide, ob- lanceolate, obovate or oblong, widest in the middle or mostly slightly above, toward the apex acute to obtuse, the tip itself with a distinct acumen, base mostly long, + cuneate or sometimes acute (at the very base occasionally abruptly obtuse), lateral veins 6-11 on each side of the midrib, diverging at 40— 80°, midrib and secondaries above = slightly im- pressed but the vein itself convex, below as tertiary reticulation above alrhost invisible to pressed and slightly convex, below slightly raised to prominent. Indument of leaves consisting of the same kind of hairs as on the twigs, even in the youngest leaves above sparse, sometimes dense on midrib, below moderately sparse and highly un- evenly distributed, often dense on midrib, or some- times subglabrous from the beginning; mature leaves entirely glabrous or below often with rudiments of the indument. Inflorescences springing from the Моште 78, Митбег 2 1991 Rohwer et al. 393 New Genera of Neotropical Lauraceae Para .— А. Habit. —B. Fruit. aia bracteata scales of the terminal bud, 2-8(-11) cm long with a peduncle of 1.5-5.5(-8) cm, narrowly thyrsoid, the lateral branches not more than 1 cm long (at anthesis; often enlarged afterward). Flowers ap- parently bisexual (but see below), sessile or with a short pedicel (to 1 mm), clustered at the tip of the inflorescence branches, subtended by relatively large Fic Г —C. Сирше, сш open, showing the double margin. — D. Detail " "leaf base. Vouchers: A, D, Cid 1890; B, C, A. S. Silva et al. 72. bracts, which are often more numerous than the flowers but (at least partly) fall off during anthesis. Perianth subglobose, ca. 2-3.5 mm diam., openin only slightly during anthesis. Tepals 6, suborbicu- lar, fleshy at the base, hyalinous toward the mar- gins, only on the outside at the base and in the middle with reddish hairs, otherwise glabrous, with- 394 Annals of th Missouri Eod Garden Paraia brac teata. — stamens, cut open, tepals and outer stamens removed. — Н. Flower, elongate form. — Н), 2-locular form, with a rudimentary third locule, seen from inside. — J GN == <= کے Pl‏ > = == ая = B» PT, i a E 52.2 . Flower, globular form. — B. Flower, tepals removed. — C. Stamen of the first Stamen of the first whorl, seen from side.— E. Stamen of the second whorl, seen Receptacle with ovary, staminode and two inner n of the first whorl (of . Stamen of the second whorl, seen from inside. — K. Stamen of the second whorl, seen from side. Vouchers: A-G, Cid 1890; H-K, Ducke s.n. out any papillae. Stamens 9, 4-thecous or in the first whorl rarely 2-thecous, without papillae, the outer six 0.8-1.4 mm long, their slender filaments 0.5-0.8 mm long, at least at the base and on adaxial side with some hairs, the anthers wider than long, thecae subapical, introrse, arranged in an arc or (especially in the first whorl) almost in a horizontal line, the inner three stamens 0.8-1.6 mm long, their filaments thick, ж slightly nar- rower than the anther, m long, at least laterally hairy, each with a “к of conspicuous glands at its base, the thecae subapical in two + lateral pairs, almost at the same level (the adaxial ones slightly higher than the abaxial ones). Stami- nodia three, 0.3-0.6 mm long, subulate, very hairy. Receptacular tube 1.2-2 mm deep, narrow, dense- ly hairy inside. Pistil slender, ca. 2-3.4 mm long, style as long as the ovary or slightly longer. Cupule broadly obconical to subhemispherical, 12-15 mm diam., 6 -8 mm deep, distinctly double-rimmed the outer margin longer, crowned by enlarged te- pals (which, however, seem to break off easily). Berry ellipsoid to ovoid, 14-24 mm long, 10-12 mm diam., often with truncate tip. Additional specimens S ae BRAZIL. AMAZONAS: Manaus-ltacoatiara Highway, Colonia Santo Antonio, 27 Oct. 1966 (fr), Allen & L. Coétho 330 (HBG, NY); Rio Cuieiras- Manaus trail, km 1, 6 Apr. 1974 (fl), Campbell et al. in Prance 21901 (HBG, MO, NY); Manaus, Estrada de Flores (Rosa de Maio), Feb. 1974 (st), D. Coêlho s.n. (HBG, INPA 46534), near Igarapé Mindú, 12 July 1929 (8), Ducke s.n. (U), along road to Aleixo, 12 Aug.- 1 Sep. 1936 (fr), rid 7996 (F, МУ); Rio Cuieiras, 2 km below mouth of Rio Brancinho, 11 Sep. 1973 (fl), Prance et al. 17749 (HBG, NY); Manaus, Igarapé do Втда, 19 y = 346 Oct. 1961 (fr), Rodrigues & agas 26 = HBG, INPA 10016, МУ); Manaus-Caracarai road, kr 19, 20 1961 (fr), Rodrigues & L. Coelho 3317 (N R itut eva ico do Norte, 3 O Rio at Lago Moura, 22 Aug. 1980 (fl), Cid et al. 1825 (НСВ, МО); Гаро Сисап, мен де Santarém, area of forest о ђу ТАМ, ЗРУЕА, апа FAO, 15 Арг. 1955 (fl), Fróe $ 31668 (МУ); Мосатро, EMBRAPA Forest Reserve, са. 10 km from Belém, 13 Volume 78, Number 2 1991 Rohwer et al. 395 New Genera of Neotropical Lauraceae Nov. 1984 (st), Gentry 48956 (MO), 10 Oct. 1966 (fr), Pires & N. T. Silva 10245 (HBG), 28 Sep. 1967 (fr imm.), Pires & N. T. Silva 10947 (NY), 2 Sep. 1967 (fr), Pires & N. T. Silva 10949 (NY), 3 Aug. 1973 (fl), Pires 13219 (HBG, NY); Mun. Almerim, Estrada do Редгао, km 40, 12 Nov. 1978 (fr), Santos 284 (Е); Serra dos Carajás, 13 km from AMZA — road to sawmill, 18 Oct. 1977 (fr), A. S. Silva et al. 7 MO); Santarém, Estrada do Palhao, km 35, near мне. Curupira camp, 24 Aug. 1969 (fr), М. Silva & Souza 2402 (HBG, MO, NY); Rio Jari, Monte Dourado, Planalto B., between Pilao and Repartimento, 28 Oct. 1968 (fr), N. T. Silva 1328 (HBG, NY); Jari, road from Bandeira to Pilao, km 39, 18 June 1969 (fl, fr imm.), N. T. Silva 2220 (HBG, NY). WOOD DESCRIPTION The wood and bark descriptions are based on the type collection only. The specimen comes from tall. The wood sample has an average diameter of 7 cm, covered by a grayish brown bark ca. 3 mm thick. Wood and bark are slightly aromatic. a small tree ca. 7 m General. Wood uniformly yellowish brown; heartwood probably dark brown as indicated by the presence of darker colored patches around a zone damaged by insect larvae; evenly textured, dense and heavy (specific gravity ca. 0.85/cm?). Anatomy. Wood diffuse porous throughout, pores evenly distributed, thin-walled and angular in outline; solitary and radial multiples of 2-5 in variable proportions; largest pores 100-150 um diam., frequency 15-22/mm”. Vessel perforations exclusively simple; intervascular pits invariably al- ternate, mostly circular in outline and not crowded, pit apertures included; pits 9-11 um diam. Tyloses common, thin-walled to sporadically sclerotic; no gum deposits observed. Fibers libriform, of medium wall thickness; regularly septate; generally ca. 15 ит diam., rarely up to 20 um; pits very small and inconspicuous, mostly confined to radial walls; growth increments distinct microscopically, marked y 1-2-seriate tangential bands of wide-bodied fi- bers (up to 30 um diam.); no imperforate tracheary elements observed. Parenchyma basically paratra- cheal sparse to irregularly vasicentric; rather in- nspicuous unless interspersed with large idio- dun (oil cells) as part of the vasicentric sheath; vertical strands composed of 4-6 individual cells; pits to vessels enlarged, oval to gashlike and ex- tended horizontally. Wood rays predominantly bi- seriate, heterogeneous with 1-2, rarely up to 4, marginal rows of upright and/or square cells; height of largest rays 0.5-0.8 mm; vessel ray pits en- larged, oval to irregular or gashlike; reddish brown gum deposits sometimes present. Crystalline de- posits and silica absent. Oil cells common, associ- ated with axial and ray parenchyma; very large with diameters up to 70 um and axial extensions up to 0.35 mm (associated with marginal ray cells) and 0.70 mm (associated with parenchyma strands). BARK DESCRIPTION Secondary phloem ground tissue composed of conducting (sieve and companion) cells and phloem parenchyma. Interspersed are numerous oval or diamond-shaped, rarely tangentially extended scler- eid islands consisting of vertically elongate, colum- nar sclereids. Secondary phloem fibers are absent. Toward the periphery the secondary phloem is delimited by a continuous ring of small, irregularly shaped and largely isodiametric sclereids enclosing few and widely spaced groups of primary phloem bers. The sclereid ring is followed by cortical parenchyma gradually phasing into several rows of rectangular, radially oriented phelloderm cells (replete with dark colored gum deposits or scleri- fied) and a narrow (2—3 rows of similar but empty cells) terminal phellem. Just inside the sclereid ring few isolated radially flattened sclereids are present. Phloem rays resemble wood rays in dimensions and composition. Their radial course through the sec- ondary phloem is ragged due to the spatial rear- rangement of other constituents following the early collapse of conducting tissues. Ray cells become sclerified when traversing sclereid islands. Oil cells are scattered throughout the secondary phloem and the cortical parenchyma. Crystalline inclusions are deposited in ray and axial parenchyma in consid- erable quantities, usually as minute acicular crys- s. Note. The sclereid island constituents (colum- nar sclereids) are more or less rectangular in cross section, with the larger side extending in radial direction. The narrow tangential faces have pointed ends, thus resembling fibers in tangential view. The polylamellate structure of the cell wall, the distinct pit canals, and the type of lignification exposed b differential staining, however, betray their probably parenchyma-derived sclerenchymatic nature. araia bracteata was first recognized as a new species by Beulah Coe-Teixeira. She annotated sev- eral specimens as Местапага paraensis Coe-Tei- xeira, but she never published the name. While working toward a revision of the genus Nectandra, J. Rohwer found that this taxon has nothing in common with /Vectandra except the technical character of “four locules of the anther arranged n an arc." There are no other species of /Vectan- dra with the leaves distinctly crowded at the tip 396 Annals of the Missouri Botanical Garden of the branchlets, with hyalinous tepals entirely lacking papillae, or with clearly double-rimmed cu- pules. Also, the general aspect of the plant, its flowers or flower parts, do not suggest /Vectandra. Sterile material is often annotated as Aniba, which can be quite similar. Vegetatively and even in its inflorescences Paraia bracteata strikingly resem- es Rhodostemonodaphne miranda (Sandw.) Rohwer, but this similarity does not extend to the inner organization of the flower. After discussing the possible relationships of this new species, we concluded that we would have to describe it as a new genus. For every known genus we could find at least three important reasons why it should not be included there. udged by its rather commonplace lauraceous wood structure alone, Paraia bracteata cannot be safely differentiated from all but a few neotropical Lauraceae, e.g., Mezilaurus, Clinostemon, Car- yodaphnopsis. Nearly all positive and negative affirmative characters circumscribing the prevail- ing structural patterns of Lauraceae are present. In combination with the very specific bark struc- ture, however, most neotropical genera of this fam- ily can be excluded as possible matches. Very good structural agreement can be found only in a small group of taxa comprising most species of Aniba and Licaria, and a few species currently ascribed to Ocotea. When taking into account the less di- agnostic but nevertheless helpful general wood characters such as color and density, the scope can be narrowed down further to the few species constituting subg. Canella of Licaria (Kurz, 1982; Richter, 1985), and two strikingly similar species of Aniba (A. canelilla (HBK) Mez, A. ferrea Ku- bitzki). Even without supporting evidence from ex- omorphic characters, an attribution to the genus Nectandra, as initially proposed by Coe-Teixeira, can definitely be ruled out. Following the traditional *number's game" of generic delimitation, it would seem that Paraia was not at all close to Licaria, which has only three fertile stamens (whorl III), each with two considering additional floral characters (which have received little attention up to now), a connection between Paraia and Licaria does not seem that far off. In Licaria it is quite common that the tepals are short and scarcely spreading at anthesis, and that only the style protrudes beyond the nearly globular perianth. Also the thickish, columnar sta- mens of the third whorl with their subapical thecae are somewhat similar to Licaria. The fact that the stamens of the first whorl are sometimes only 2-thecous in Paraia indicates a tendency toward reduction. If the thecae were entirely reduced in the outer whorls, the resulting staminodia would be quite similar to those of several Licara species, with a slender filament and a flat, widened apex. Thus, the possibility that Paraia represents a more primitive offshoot of the line leading to Licaria cannot be ruled out. When looked at individually, the flowers of Paraia bracteata appear to be invariably bisexual. However, when flowers from several different spec- imens are dissected, it seems that there are two types of flowers, differing in proportions. In the first type (Fig. 4A—G), the flowers are nearly glob- ular, the receptacular tube not being apparent from the outside. The length of the stamens is approx- imately equivalent to the depth of the receptacle. In the second type (Fig. 4H-L) the stamens, al- though apparently fertile, are noticeably smaller. The receptacular tube is deeper, almost twice as long as the stamens, visible also from the outside. The ovary, however, is more or less the same size in both types. It is possible that the different pro- portions indicate an incipient separation of the sex- es. It may, on the other hand, be just an expression of geographic differentiation. Such a differentiation is obvious in leaf shape: specimens from the sur- roundings of Manaus have relatively much nar- rower leaves than those from the state of Pará. DISCUSSION OF GENERIC DELIMITATION IN THE LAURACEAE OCOTEA RUBRA A highly unusual combination of characters is also found in a species known as Ocotea rubra Mez. This species has more or less obovate, gla- rous leaves, crowded at the tip of the branches. It therefore resembles Mezilaurus rather than Ocotea in its vegetative characters. The flowers, however, are unlike any other known species of Lauraceae. The tepals are shorter than the floral tube and usually more or less unequal, the outer ones smaller than the inner ones. The stamens are tongue-shaped and extremely papillose. Their four thecae are introrse and arranged in an arc on the stamens of the first and second whorl, but latrorse and in two pairs above each other in the third whorl. The staminodia of whorl four are almost as large as the inner stamens, but unlike very large staminodia in other genera they are not glandular. The ovary has a very slender style, as, e.g., in Licaria or Mezilaurus. The fruit is almost com- pletely surrounded by a deep cupule during most Volume 78, Number 2 Rohwer et al. 397 New Genera of Neotropical Lauraceae of its development, but about half exserted at ma- turity. There is no other species of Ocotea with unequal tepals, and most of the other characters of O. rubra are very rare. Anatomical data (Richter, 1981) further underscore the isolated position of this spe- cies. Obviously, there is no reason to retain it in Ocotea other than reluctance to further increase the number of monotypic genera. In several char- acters (unequal tepals, broadly sessile stamens, red- dish brown wood with very large cross-field and intervascular pits), O. rubra resembles the Asian pule, but only a somewhat enlarged pedicel with the tepals persistent under the fruit. TWO-LOCULAR VS. FOUR-LOCULAR ANTHERS All of the taxa discussed above go beyond the current delimitation of the genera simply because they represent rare combinations of characters. They are therefore systematically isolated, but do not jeopardize the delimitation as such. There is, however, also an increasing number of cases that lead to serious doubts about the value of the current system. Traditionally, one of the most important characters used for the delimitation of genera is the number of thecae per anther, two vs. four. Although this character is indeed constant for the overwhelming majority of species within well-de- fined natural groups, its indiscriminate application may lead to artificial groupings. Several neotropical species have the staminal configuration of one genus, yet they can be rec- ognized as belonging to another genus based on all other characters. The current system is inconsis- tent in treating these species. Some are placed with their allies despite a staminal configuration that makes it impossible to key them out to the correct genus, whereas others are placed within the genus to which they would key out, regardless of their obvious affinity to species of a different genus. Thus, Persea cuneata Meissn., P. bilocularis Kopp, and Caryodaphnopsis inaequalis (A. C. Smith) van der Werff & Richter, e.g., have 2-thecous anthers, although their respective genera are de- fined by 4-thecous anthers, and Beilschmiedia sul- cata (Ruiz & Pavon) Kosterm. is the only species with 4-thecous anthers in this otherwise 2-thecous genus. Examples of species placed according to their staminal configuration despite obvious ties to spe- cies of a different genus are all Central American species currently placed in Aiouea (compare van der Werff, 1987, 1988). In nearly all of its char- acters, including the wood structure, A. costari- censis (Mez) Kosterm. is clearly a member of the Ocotea insularis group in the sense of Rohwer (1986). It is so close to Ocotea skutchii C. K. Allen as to make it very difficult to separate the two in fruit. Aiouea lundelliana C. K. en, a member of the same group, is even more obviously intermediate between the generic concepts in that its staminal configuration is variable within the same inflorescence (cf. van der Werff, 1984; Roh- wer, 1986). In other Central American species of Aiouea the ties to a particular species or species group of Ocotea are not as clear, yet there is no doubt that they are closer to that genus than to the South American species of Aiouea. In South America, Aiouea is very similar to species that used to be ascribed to the genus Phoe- be. Again, the main difference is in the number of thecae, but it has long been known that in several of the “Phoebe” species only the outer anthers are 4-thecous, whereas the inner ones are 2-the- cous. “Phoebe” is placed in quotation marks here because in the Neotropics it is anything but a good genus. The type of Phoebe is from Asia, and in all of the palaeotropical species the fruit is subtended by erect, enlarged, lignified tepals on a scarcely thickened pedicel, clasping the base of the fruit. uch a structure does not occur in any of the neotropical species. In some of them there are persistent tepals, but they are not enlarged, and either erect on the margin of a well-developed cupule or spreading from a swollen pedicel or a small, disklike cupule. Others have a normal, more or less bowl-shaped cupule, as in most species of Ocotea. Many of the neotropical species currently ascribed to Phoebe, in particular many Central American and Andean species, probably belong to Cinnamomum, as suggested by Kostermans (1961). However, transferring all of them indiscriminately to Cinnamomum seems rather precipitant. Some are clearly Ocotea, several may go to Aiouea after a revision of the generic delimitation, and still oth- ers may represent a genus of their own. A third example of a probably polyphyletic genus in the New World is Endlicheria. It is defined by the combination of two characters, which indicate a common level of advancement rather than a common origin: (1) anthers 2-thecous, and (2) flow- ers unisexual. Some species of Endlicheria are extremely close to certain species of Rhodoste- monodaphne. Endlicheria metallica Kosterm. and R. grandis (Mez) Rohwer can only be separated by counting the number of thecae per anther. It is questionable whether they should even be treated 398 Annals of the Missouri Botanical Garden as different species. Another striking similarity is found between E. vinotincta C. К. Allen and К. celiana (Allen) Rohwer. These two, although un- doubtedly separate species, agree in many char- acters, from their reclining habit via their indument and the structure of their slender, pendent inflo- rescences to the development of peculiar, mem- branous, possibly glandular outgrowths of the fil- ament on both sides below the anther. Especially this last-mentioned character makes it highly un- likely that the overall similarity is a mere conver- gence. Other species of Endlicheria, e.g., E. dysodan- tha (Ruiz & Pavón) Мег and E. pyriformis (Nees) Mez, are much more similar to Ocotea, although they cannot be linked to a particular species. In E. anomala (Nees) Mez only the outer whorls are 2-thecous, whereas the third whorl is 4-thecous. Its fruit has a very small cupule that is lobed by the persistent bases of the tepals. Such a fruit structure is otherwise unknown in Endlicheria, but a rather similar form occurs in Ocotea campho- romoea Коћмег. This supports the view that Е. anomala has been derived from Ocotea indepen- dently from other species of Endlicheria. he Endlicheria species that form the core of subg. bracteata Mez, Е. glomerata Mez, Е. macro- phylla (Meissn.) Mez, E. multiflora (Miq.) Mez, and E. verticillata Мег) need further investigation. Ampelodaphne (Е. arunciflora Мег, Е. With their crowded, nearly verticillate leaves and their numerous, minute flowers in profusely branched, bracteate inflorescences, they look not only unlike any other species of Endlicheria, but unlike any other neotropical Lauraceae at all. It is possible that they, too, have been derived from Rhodostemonodaphne, but independently from the species mentioned above. How the ancestral forms of this group may have looked is illustrated by Е. loretensis O. C. Schmidt, which has been described in Endlicheria despite its 4-thecous anthers, based on its overall similarity to this genus. | of the above cases suggest that the distinction between 2-thecous and 4-thecous anthers shoul be dropped for the purpose of generic delimitation. What would this mean for the system of the Laura- ceae? Actually, the change would only be mod- erate. Some genera already comprise species with 2-thecous and with 4-thecous anthers (see above), and sometimes the number of thecae is variable within otherwise very homogeneous species (e.g., in Urbanodendron verrucosum (Nees) Мег, see Rohwer, 1988). In the New World only the three genera discussed above would need to be (partly) dismembered, and this would make the classifica- tion more natural. On the other hand, it would become technically more difficult, but not impos- sible, to describe the difference between Ocotea and Aniba. In the Old World the change would involve only a few genera, but a large number of species. Alseodaphne and Dehaasia, with together some 80-90 species, would have to be merged, and new generic definitions would have to be found in the complex formed by Lindera, Litsea, and Actinodaphne, with together some 600-800 spe- cies. Within this complex the involucrum envel- oping whole or partial inflorescences is sometimes derived from budscales at the base of the inflores- cence and sometimes from floral bracts. There are also marked differences in the inflorescence struc- ture (Kostermans, pers. comm.), which do not par- allel the 2-thecous/ 4-th distinction, but which may correspond to structurally different groups in wood and bark characters (Richter, 1981). Only the sheer size of this complex precludes an effective revision. All other genera with 2-thecous anthers, whether palaeotropical or neotropical, are either definable by additional floral or fruit characters, or their species had already been placed within a genus with 4-thecous anthers by a previous author. OTHER GENERIC CHARACTERS But what about the value of other characters? Are they really more reliable than the number of thecae? The answer is more complicated than the question implies. No single character has exclu- sively one well-defined state in every genus. Some- where in the family we always find characters not so clearly expressed, or intermediate states, or ex- ceptions to the prevailing state of a genus. This statement, however, is only possible because we still can define genera, as groups with a syndrome of certain characters that are constant or nearly constant for this particular group, but may be variable in other groups defined by other charac- ters. Staminodia, for example, are consistently ab- sent in Rhodostemonodaphne and consistently large and glandular in Persea, but in Ocotea they can be absent, or small and nonglandular, or small with a glandular tip, or with a relatively large apical gland. This means that on the one hand any char- acter will be misleading when given too much em- phasis, and on the other hand any character (in- cluding the number of thecae) can be useful in conjunction with others. Especially the fruit characters deserve more at- tention. With few exceptions (which need further investigation) they are remarkably constant within Volume 78, Number 2 Rohwer et al. 399 New Genera of Neotropical Lauraceae natural groups as defined by other, often less ob- vious characters. Of course, the fruit characters are not diverse enough to differentiate all genera, but when they do show a clear-cut discontinuity, this mostly warrants a generic distinction. Unfor- tunately, nearly all conceivable intermediates are present in the large genus Ocotea, which is defined only negatively, i.e., by the absence of the diag- nostic characters of more distinctive genera. good example of the variation of important characters (including fruit characters) within a ge- nus is found in the neotropical species of Persea. In Persea subg. Persea the tepals are equal, widely patent to reflexed at anthesis, and lost in fruit, whereas in subg. Eriodaphne they are unequal, almost erect to slightly spreading at anthesis, and persistent, indurate, and widely patent in fruit. This seems to be a better distinction than found between many genera, but the gap is bridged by a few Central American species (P. pallida, P. rigens, P. steyermarkii, P. sylvatica), and by the Asian subg. (or genus) Machilus. These species have the equal tepals of subg. Persea, but they are persistent in fruit as in subg. Eriodaphne (although usually not incrassate in Machilus). Representatives of this group also occur in Venezuela, Ecuador, and Peru, but lack of good collections has hampered their correct placement. Recently, a Persea from Su- rinam was described with yet another combination of characters. Its tepals are unequal as in Frio- daphne, but deciduous, as in Persea (van der Werff, 1989) The number of thecae is also variable in the neotropical Persea species. Most commonly one finds nine 4-celled stamens, but some species have nine 2-celled stamens, six 4-celled and three 2- celled stamens, or six 4-celled stamens and the third whorl staminodial. Needless to say, our un- derstanding of this genus is far from complete, and it is equally not clear how to differentiate Persea from its closest relatives, especially in Asia. ile Machilus and Phoebe are clear-cut gen- era in Asia, the distinction between South American Persea subg. Eriodaphne and Asian Phoebe can be difficult. As mentioned above, the tepals are unequal and more or less patent in fruit in the former group, whereas in the latter they are (near- ly) equal and erect. However, in some species of subg. Eriodaphne the “slightly unequal” tepals are less different from each other than the “nearly equal" tepals in some species of Phoebe, and in some species (e.g., Persea cuneata Meissn., see fig. 3 in Kopp, 1966) they can be nearly erect. If P. cuneata was an Old World species, it would be placed in Phoebe without much hesitation, or in Apollonias, if this is kept separate from Phoebe, based on its 2-thecous anthers. What can we expect from future studies on neotropical Lauraceae? As more and better сој- lections become available, the number of species will increase and the distribution of the taxa will become better known. Some of the new species will possess such unusual combinations of characters that additional genera will have to be described. Two genera have recently been described, Wil. liamodendron (Kubitzki & Richter, 1987), initially only known from Amazonian Brazil, northern Co- lombia, and Costa Rica, but now also known from Peru and southern Brazil, and Povedadaphne (Burger, 1988), only known from Costa Rica. We have mentioned earlier several taxa that probably represent additional genera. We estimate that 25- 35% of the neotropical species of Lauraceae have not yet been described, and even under optimum conditions, collecting plus descriptive studies would require at least a decade before an inventory of neotropical Lauraceae will be reasonably complete. LITERATURE CITED AUBLET, F. 1775. Histoire des plantes de la Guiane francoise, Volume 2. Didot, London, Paris BENTLEY, E & H. TRIMEN. 1880. Medicinal Plants, Volume 3. J. & A. Churchill, London. BURGER, M C. 1988. A new i ien of oor n Е Costa Rica, with co n problems of generic mme = specie delimitation within the family. aes International Code of Bo- 40: к Y. а 1988. ta = Ko- omenclature. Koeltz Scientific Books, йт. im "1966. A taxonomic revision of the genus Persea in a western Perses Mem. New York Bot. Gard. 14: 1-11 KOSTERMANS, А. L 6. 1936. Lauraceae. /n: A. Pulle (editor), Flora of Surinam, Volume 2. Koninklijke — Indisch Instituut, yox rdam 1957. Lauraceae. Commun. Forest Res. Inst. 57: ря 64. . 1961. The New World species " poe mum Trew (Lauraceae). Reinwardtia 6: 24. . V. PINKLEY & W. Г. STEARN. "1969. A new Amazonian arrow poison: Ocotea venenosa. Bot. . Leafl. Harvard Univ. 22: 241-252. Шш К. € H.G 1987. Williamoden- dron Kubitzki & Richter, a new genus of neotropical Lauraceae. Bot. Jahrb. Syst. 109: 49-58. Kurz, Н. 1982. E einiger Gattun- gen neotropischer Lauraceen und Revision der Gat- tung Licaria мина Ph.D. Dissertation, Uni- versity of Ham MELL C. D. & W. p. Brush. 1913. Greenheart. U. 8. е к of Agriculture, Forest Service, Сіг- cular 211. MEZ, C. 1889. 5 Americanae. Jahrb. Konigl. Bot. Gart. Berlin 5: 1-556 400 Annals of the Missouri Botanical Garden n H. G. 1981. inia Ver. Hamburg 985. Wood and bark anatomy of Lauraceae, II, Licaria. IAWA Bulletin, n.s. 6: 187-199. ROHWER, J. С. 19 Prodromus einer Monographie der Gattung Ocotea Aubl. о sensu lato. Mitt. Inst. Allg. Bot. Hamburg 20: 3-278. —————. 1988. The genera Diepelium Phylloste- monodaphne, Systemonodaphne and Urbanoden- dron prag Bot. Jahrb. m 110: 157-171. SANDWITH, N. Y 39. сени names of Greenheart and Balata. Trop. Woods 24. Anatomie des sekundáren Xy- lems und der Rinde der Lauraceae. Sonderbd. Na- EE H. Ne. DER. 1984. eae Notes on кр Laura- nn. Missouri Bot. Gard. 71: 1180-1183. . Six new species of nm Laurace- ae. Ànn. Missouri Bot. Gard. 74: 40 1988. Eight new bs ы one new com- bination of neotropical Lauraceae. Ann. Missouri Bot. Gard. 75: 402-41 1 А new species of Persea (Lauraceae) from Surinam, with a discussion of its position within the genus. Ann. Missouri Bot. Gard. 76: 939-941. GAMANTHERA (LAURACEAE), A NEW GENUS FROM COSTA RICA! jp van der Wep and Peter K. Endre ABSTRACT new genus of the Lauraceae, Gamanthera, is described from Costa Rica. It is distinguished from all other Lauraceae by its completely fused stamens, forming a synandrium, with only three, two, or one locellus, and by its monoecious nature During the past few years, intensive collecting efforts in Costa Rica have resulted in a great num- ber of additions to the Flora Costaricensis, as well as in a number of undescribed taxa. Several new species of Lauraceae have recently been published (Hammel, 1986; Burger, 1988; van der Werff, 1988; Zamora et al., 1988), and 14 species were published in the treatment of the Lauraceae for the Flora Costaricensis (Burger & van der Werff, 1990). Burger (1988) published a new genus, Po- vedadaphne, from Costa Rica and noted that re- cently five mostly South American genera had been added to the known flora of Costa Rica. Recent fieldwork organized under the auspices of the Man- ual Flora of Costa Rica project, organized by the Missouri Botanical Garden, has yielded another Lauracea from Costa Rica, which is here describe as a new, monotypic genus. The surge of new taxa of Lauraceae from Costa Rica can be attributed to the unusual richness of its flora, but perhaps a more likely explanation for these discoveries is the excellence of the Costa Rican botanists participat- ing in the botanical exploration of their country. It is, therefore, with pleasure that we name the new species after Gerardo Herrera, who discovered it and whose excellent collections of Lauraceae have given us much pleasure to work with. Gamanthera herrerae van der Werff, gen. et sp. nov. TYPE: Costa Rica. Alajuela, Upala, Colonia Puntarenas, along Rio Chimuria, tree, 8 m, flowers yellowish, fruits green, cupule pink, 11 Nov. 1987, Herrera 1228 (holotype, MO; isotypes, BM, CR, HBG, US). Figure 1. Arbores monoeciae. Folia alterna, venatione pinnata. r er- 1 -locellatum. Fructus cupulae duplimarginatae aidons, Tree, 8 m tall. Twigs terete, densely brown- tomentellous, with occasional whorls of scars from fallen bracts. Axillary and terminal buds conspicu- ous, covered by several whorls of bracts, the small basal bracts brown-pubescent outside, the large inner bracts glabrous or nearly so, but with ciliate margins; inner surface of bracts glabrous. Leaves alternate, chartaceous, 12- —8 cm, elliptic, margin undulate, the base cuneate to broadly cu- neate, the apex shortly acuminate, the acumen to ca. 1 cm long; upper surface of mature leaves with midrib tomentellous, especially near the base, oth- erwise glabrous, young leaves with a rather dense indument of short (0.2-0.3 mm), erect hairs; lower surface with midrib and lateral veins brown-to- mentellous, the lamina with erect pubescence, hairs ca. 0.3 mm long; lateral veins 5-8 on each side, arching upward near the margin, not or very weak- ly loop-connected, venation immersed or nearly so on upper surface, midrib and lateral veins promi- nently raised on lower surface, tertiary venation also raised. Petioles terete, brown-tomentellous, 1— 1.5 cm (rarely to 2 cm) long, adaxially flattened, abaxially terete. Inflorescences axillary, brown-to- mentose, paniculate, to 7 cm long, with bracts or scars of fallen bracts at the base, flowers clustered near the tips of the secondary or very short tertiary axes, rarely solitary, not cymosely arranged; axes ' We thank Mary Merello for help with the SEM work in St. Louis and John Myers for his skillfull drawing. 299, St. ? Missouri Botanical Garden, P.O. Box Louis, Missouri 63166-0299, U.S.A. ‘Institut für systematische Botanik der Universitat Zürich, Zollikerstrasse 107, CH-8008, Zürich, Switzerland. ANN. MISSOURI Вот. GARD. 78: 401-408. 1991. 402 Annals of the Missouri Botanical Garden VA | Wi y А ~ \ y M N іл f, N Wakes ANN curn ias gi ANG, i, ift 2, Nar E FIGURE 1. Gamanthera herrerae.—A. Habit.—B. Flower.—C. Young and old fruits, the young ones enclosed in the cupule, the old one exserted. — D. Detail of inflorescence. — E. Base of leaf, showing indument. — F. Detail of stem, showing linear scars from fallen bracts and circular scars from fallen buds. —G. Base of axillary inflorescence, showing linear scars from fallen bracts. and flowers subtended by bracts, these broadly to nearly so, unisexual, 4-4.5 mm long, including the narrowly ovate-elliptic, to 2.5 mm long or, when narrowed base; staminate and pistillate flowers ex- fallen off, by a scar. Inflorescences with either — ternally very similar; tepals 6, erect, broadly tri- staminate or pistillate flowers. Flowers sessile or angular, pubescent outside, the free part to 1.5 Volume 78, Number 2 1991 van der Werff & Endress 403 Gamanthera from Costa Rica mm long, inner 3 tepals shorter than outer 3. Staminate flowers with a central, narrowly conical synandrium formed by fusion of 3 stamens (Fig. 2), the fusion complete and individual stamens not recognizable, the synandrium 2 mm long, the gla- brous anther-bearing part and the densely pubes- cent fused filaments each 1 mm long; locelli 3, 2, or infrequently 1; staminodia and pistillode lacking. ras a globose, inaperturate, spinulose (Fig. 3). ual flowers with a one-sided(?) synan- кен ovary situated beside(?) the synandrium, glabrous, narrowly ellipsoid, gradually narrowed into a slender, long style (Fig. 4). Receptacle gla- brous inside. Stigma not seen. Fruit ellipsoid, ca. 1.5 ст long, the exserted part 1 cm long, cupule deeply cupshaped, double-rimmed, the undulate- lobed outer margin strongly reflexed. Paratype. Collected from same tree as type speci- men, Grayum et al. 9039 (CR, F, LE, MEXU, МО, 5, J, 2). ANATOMY OF THE STAMINATE FLOWER MATERIAL AND METHODS Figures 5 and 6 were prepared by P. Endress. Pickled flowers of the collection Grayum 9039 were examined with the scanning electron micro- scope (SEM) after critical point drying and sputter coating with Au/Pd. Other samples were used for microtome serial sections and stained with safranin and astra blue. Figures 2-4 were prepared by Н. van der Werff. Dried flowers were dissected without rehydration and sputter coated with Au. RESULTS The perianth consists of two trimerous whorls of tepals (Fig. 8). The androecium in the center of the flower consists of a massive conical body which bears one to three pollen sacs at its top (Figs. 7, 9, 10, 11). Nectaries are lacking. At anthesis the flower does not expand and retains the globular shape of the bud (Fig. 5). Only the upper part of the androecium with the pollen sacs protrudes through the small opening of the flower and is exposed (Fig. 6). Each pollen sac opens by a flap, which is hinged at the top and recurves toward the top (Figs. 5, 6). The flap is shorter than the pollen sac and does not extend to its base (Fig. 1 1C). The diagrammatic position of the pollen sacs is opposite the inner tepals (Fig. 8). The androecium is always served by three collateral vascular bundles irre- spective of the number of pollen sacs (Fig. 11 C). In androecia with three pollen sacs the vascular bundles alternate with the pollen sacs (Fig. 110). The bundles are situated toward the center of the androecium and terminate somewhat above the lower end of the pollen sacs (Fig. 11C) DISCUSSION — INTERPRETATION OF THE ANDROECIUM How is this highly unusual androecium to be interpreted? Normal stamens of Lauraceae are served by a single vascular bundle situated in the center of the stamen behind the two pollen sacs (if the anther is bilocular), as seen in a transverse section. Here, in Grayum 9039, there are three vascular bundles, which suggest three stamens. The three vascular bundles alternate with the tepals of the inner whorl. In lauraceous taxa, where the androecium is reduced to a single whorl, e.g., in Licaria (cf. Mez, 1889, as Misanteca; Koster- mans, 1957) or Mezilaurus (van der Werff, 1987), the stamens alternate with the tepals of the inner whorl. From this it follows that the conical androecium consists of three stamens that are completely con- genitally fused into a compact synandrium. If three pollen sacs are differentiated, each “anther” sists of two “half” pollen sacs in that the two con- neighboring pollen sacs of two stamens have merged into one. Merger of two pollen sacs of different thecae ( or at least of their flaps) into one functional unit also occasionally occurs in other lauralian groups; e.g., in Siparuna the two flaps of an anther may merge into one (here, two thecae of the same anther) (cf. Endress & Hufford, 1989). The male flowers of Grayum 9039, therefore, represent an extremely reduced stage previously unknown in the Lauraceae. In other taxa with bisexual flowers reduction of the androecium to one stamen whorl, which is congenitally fused into a tube surrounding the gynoecium, is the ultimate stage so far known. This is the case in species of Licaria (Kostermans, 1957; as Misanteca in Mez, 1889) and Mezilaurus (van der Werff, 1987). In Kubitzkia there are still three staminal whorls, whereby the third is connate into a tube (as Sys- temonodaphne їп Kostermans, 1957). Fusion of the filaments of whorls II and III occurs in an undescribed species of Rhodostemonodaphne (van der Werff, pers. obs.). However, the anthers re- main free in all of these cases. In Grayum 9039, congenital fusion of the re- maining three stamens into a massive (compact) central body and fusion of neighboring pollen sacs of neighboring stamens, reduction of pollen sac Annals of the Missouri Botanical Garden Volume 78, Number 2 1991 van der Werff & Endress 405 Gamanthera from Costa Rica FIGURES 5-10. flower from the side, two open pollen sacs visible. — 6. Upper half of open flower, three open pollen sacs аан Synandrium from the side, three closed pollen sacs with the dehiscence lines of the flaps visible. — 8. SEM micrographs of flowers and pollen of Grayum 9039. Scale bars — 0.1 m —5. Open Open flower from above, the six tepals and the synandrium with the three open pollen sacs visible. — 9. Synandrium from above, three open pollen sacs visible; the flap of the uppermost pollen sac has been removed.— 10. Synandrium from above, three closed pollen sacs visible number to two or one in some flowers, and loss of the nectary appendages, represent a further step in this trend of reduction within the Lauraceae. Synandry is relatively frequent in Magnoliidae. Because the shape of the stamens is more or less bulky, synandry is relatively easy to establish from a morphogenetical point of view, mainly in uni- sexual flowers, where a gynoecium is lacking as a potential obstacle in the floral center (Endress, 9) DISCUSSION Gamanthera herrerae is currently known only from the vicinity of Upala in the Alajuela Province, Costa Rica. The two collections were made in con- є<— IGURES 2-4. synandrium. — 3. Pollen grain. — SEM puri doge of flowers and pollen of Grayum 9039. — 2. Dissected male flower, showing the . Dissected female flower, showing glabrous ovary below the he зува and the slender style; part of the 5 removed to show style. Scale bars in Figures 2 and 4 E .005 mm. 5 mm; in Figure 3 406 Annals of th Missouri Bed Garden FIGURE 11. — 0.5 mm. Endothecium of the pollen sacs шен In oo bundles re black, phloem stippled. — A. Synandrium with three pollen sacs. — o pollen Microtome serial transverse sections of synandria of thre Фо e male flowers of Grayum 9039. Scale bar s. —C. Synandrium with one pollen sac. (a) Sections at ynandrium upper end of pollen sacs. (b) эин а! uds level of in sacs. (c) Sections at base of pollen sacs. (d) Sections below pollen sacs. (e) Sections at base of synandrium. secutive years from the same tree. A search in the vicinity yielded only two additional trees, one ster- ile, the other juvenile. The trees were not large, but hard to find; it is likely that, if one would be willing to spend a week searching, more trees could be located. Several inflorescences contain deformed flowers infested by insect larvae. Such flowers have basal, spreading tepals and a central, depressed-globose, densely pubescent growth. This growth probably represents the deformed synandrium, as is indi- cated by the densely pubescent surface. The locelli- bearing part of the synandrium is usually present, but always more or less lateral, never at the tip of the growth. The deformed flowers have a hollow center, which usually contain an insect larva. Such flowers were only found in the male inflorescences and never contained a trace of an ovary. Infruc- tescences with “young fruits" are part of all spec- imens. These “young fruits" are almost completely included in a broad, flat-topped cupule (the two righthand “fruits” in Fig. ІС). Upon dissection these “fruits”? show an absence of a well-defined outer wall and a seed; instead they are filled with nondifferentiated tissue which contains a few, ir- regular cavities. Although we did not find larvae ruits," probably caused by insect or mite activity. The lefthand fruit in Figure 1С is not abnormal; it has in these “ we consider them abnormalities, the distinct outer wall found in other Lauraceae and contains a shrivelled, dried seed. The two col- lections of Gamanthera include only one normal fruit but several dozen of the abnormal **young fruits.” This new genus possesses several features that are either unknown or very rare in neotropical Lauraceae. The most striking of these is the an- droecium, which consists of three completely fused stamens, as is described above. Staminodia have not been found, nor are there glands associated with the synandrium. A complete fusion of stamens, including anthers, has not previously been reported for Lauraceae. This synandrium, which is located in the center of the flower, appears to be a single stamen and only a study of the vascularization showed its true origin. The number of locelli on the synandrium can be three, two, or one; one synandrium with four locelli was found with the locelli arranged in two horizontal pairs, the septum separating the two locelli of one pair much thinner than the walls separating the pairs. A second unique feature in Gamanthera (at least for Lauraceae) is the presence of two flower types on the same tree. Lauraceae have either bisexual or unisexual flowers; if unisexual flowers are pres- ent, the species are dioecious. In most cases, the unisexual character of the flowers can be readily seen, but in some species (for instance, in Ocotea, Volume 78, Number 2 van der Werff & Endress 407 Gamanthera from Costa Rica van der Werff, pers. obs.) the reduction of the nonfunctional organs in the flower is not obvious, and it can be difficult to decide whether flowers are unisexual or bisexual. Howard (1981) reported that some species of Осогеа s.l. were found to be polygamodioecious, with perfect and functionally unisexual flowers present in the same inflorescence, an observation we have not been able to confirm. The first collection of Gamanthera included inflo- rescences and young fruits, sometimes with the inflorescence and fruit attached to the same twig. The inflorescences consisted of only staminate flow- ers, without even a trace of a gynoecium. After much searching, pistillate flowers were found on the infructescences as old, dried flowers. Seven flowers of different infructescences were dissected and these were all pistillate. The pistillate flowers are, without dissection, indistinguishable from the staminate flowers and possess a similar synandrium. rhe synandrium often persists on young fruits. We did not find pollen grains in the synandria of the pistillate flowers, but have not found pollen grains in synandria with opened valves of staminate flow- ers either. Pollen was only found in flowers with the synandrium still included in the flower and valves closed. Thus, it is possible that the flowers described as pistillate are bisexual and that Ga- manthera has unisexual (staminate) and bisexual flowers on separate inflorescences on the same plant, rather than pistillate and staminate flowers on sep- arate inflorescences on the same plant. Either way, this is a situation not previously recorded for Laura- ceae. It is not yet clear how pollinization takes e staminate flowers, the synandrium of the pistillate flowers is hollow and the style ascends through the hollow core (Fig. 4). The style can be followed up to the level of the locelli, where it seems to dis- appear. A pore in the synandrium through which a stigma can penetrate has not been found. Such a pore is likely present in younger flowers, but might become unrecognizable in old flowers. The pistillate flowers seen are only present on infruc- tescences and are likely several months old. On the twig that carries a staminate inflorescence and an infructescence, the infructescence is attached much lower on the twig than the inflorescence. In Gamanthera it is difficult to estimate the age of the flowers, as there are, apart from the elongation of the synandrium, no changes in size or shape. The general trend in Lauraceae is that staminate flowers fall off soon after shedding their pollen. Because young staminate flowers (in which syn- andria are still enclosed in the flowers and locelli have not opened) are present on some specimens, the staminate flowers are probably not very old. On the other hand, the development of the cupules on the infructescences suggests that these pistillate inflorescences are much older. Therefore, it seems likely that the pistillate and staminate flowers on the available specimens have quite different ages, and it is not certain that staminate and pistillate flowers are simultaneously functional on the same tree. If this is not the case, Gamanthera could be an interesting example of dichogamy in a mon- oecious species with unisexual flowers. Dichogamy has been reported for Lauraceae in genera with perfect flowers (Stout, 1927; Kubitzki & Kurz, 1984) and seems widespread in the family. A third interesting feature of Gamanthera is the series of bracts surrounding the inflorescence and vegetative buds. Once inflorescences and twigs grow out of the buds, the bracts fall off and leave clusters of linear scars at the base of the inflorescence and young twigs. Clusters of scars on twigs occur in several genera of neotropical Lauraceae (Aniba, Endlicheria, Mezilaurus, Pleurothyrium, van der Werff, pers. obs.). Species in these genera with bracts all have leaves clustered near the tips of their branches, and the bracts are narrowly elliptic with an acute tip and frequently with a rudimentary midrib, quite unlike the broad, scalelike bracts found in Gamanthera. In Persea schiedeana, another species with clustered leaves, the bracts surround- ing the vegetative buds are broad and scalelike, but this species lacks bracts at the base of the inflorescences. The only genus in the Neotropics with scalelike bracts surrounding the vegetative uds and with alternate leaves (apart from Ga- manthera) is Litsea, which occurs at high eleva- tions from Mexico to Costa Rica. In Litsea the flowers are arranged in pseudo-umbels, which are subtended by scalelike bracts. It is likely that these bracts represent the floral bracts, which, like the flowers, have become greatly condensed. In Ga- manthera, floral bracts are present in addition to the bracts at the base of the inflorescence, and these latter bracts seem unique among neotropical Lauraceae. RELATIONSHIPS Infrafamilial classification of Lauraceae (Kos- termans, 1957; Hutchinson, 1964) is based on the following characters: presence or absence of bracts (or decussate bracts surrounding the subumbellate characters found in Gamanthera do not agree with 408 Annals of the Missouri Botanical Garden any of the tribes as defined by these authors. Mon- oecious trees with inflorescences composed of flow- ers of one sex only had not been reported before in Lauraceae. The bracts at the base of the inflo- rescences suggest a relationship with the Laureae, but against this argue the facts that in the Laureae the bracts subtend the (mostly subumbellate) flow- ers, not the common û peduncles as in сатаан that in the La or racemose, not paniculate as in Gamanthera; and that in Laureae the cupule is never double- margined. The other tribes are defined by cupule and androecial characters. The highly reduced an- droecium of Gamanthera precludes placement in any of these tribes. Probably the closest is the subtribe Anibineae of Kostermans (1957), which includes genera with double-margined cupules and some fusion of stamens (but never to the degree found in Gamanthera). Against the placement of Gamanthera in the Anibineae are the bracts sur- rounding the inflorescences and vegetative buds, and the monoecious condition. In favor of a rela- tionship with Anibineae is the fact that Anibineae are well represented in the Neotropics, whereas Laureae include only one neotropical species. The pollen of Gamanthera (Fig. 3) is globose, inaperturate, and spinulose, and agrees with the general pollen characteristics of neotropical Laura- ceae (Raj & van der Werff, 1988). Diameter of the grains is about 20 и and the spinules are seated on cushionlike bases, these without granular pro- cesses. These characters agree well with the Ani- bineae and do not indicate a relationship with the Laureae. H. G. Richter (pers. comm.) reported that the wood of Gamanthera is nondescript and typical of Lauraceae; the bark corresponds with Richter's (1981) Licaria type 1, Aniba, and Ocotea pro parte. These conclusions rule out a relationship of Gamanthera with the Laureae and place it some- where in the Anibineae. А close relationship with Mezilaurus, which includes some species with fused filaments of the three stamens, is not likely because Mezilaurus differs in its wood characters (Richter, 1981), clustered leaves, and small, platelike cu- pules. Instead, Gamanthera might be closer related to Licaria, although we consider Gamanthera dis- tinct from Licaria because of its synandrium and clusters of bracts along twigs and at base of inflo- rescences. The relationships between genera of Lauraceae are presently not well understood and are in need of much further study. LITERATURE CITED BuncER, W. C. 1988. A new ои of — from Costa Rica with comme n problems of generic and ае delimitation perra the family. Bins 40: о м DER WERFF. 1990. Lauraceae. In: W.C. Burger io) Flora Costaricensis. Fieldiana, ot., n.s. 23: 1- ENDRESS, P. К. io ae of reproductive struc- tures and functions in primitive angiosperms (Mag- noliidae). Mem. New York Вог. Gard. 55: 5-34. Enpress, P. К. & L. D. Hurronp. 1989. The am of stamen structures and dehiscence patterns among Magnoliidae. Bot. J. Linn. Soc. 100: 45-85. HAMMEL, B. E. New species and notes on Laur- aceae from the Caribbean lowlands. J. Arnold. Arbor. 67: 123-136 HOWARD, К. A. Nomenclatural notes on the Lauraceae of the Lesser Antilles. J. Arnold Arbor. : 4 HUTCHINSON, L 1964. The Genera of Flowering Plants. Clarendon Press, Oxford KOSTERMANS, A. J. С 1957. Lauraceae. Commun. Kusitzki, К. & Synchronized di- chogamy and dioecy in neotropical Lauraceae. PI. Syst. Evol. 147: 253-266. Mez, C. 1889. e Americanae. Jahrb. Königl. Bot. Gart. Berlin 5: 1-556. Raj, B. & Н. VAN DER Wenrr. 1988. A а to the pollen morphology of RD Lauraceae Ann. Missouri Bot. Gard. 75: 7, RICHTER, Н. С. 1981. lems und der Rinde дег Lauraceae. Son wiss. Vereins Hamburg 5: 1-148 The flower behavior of Avocados. 03 Anatomie г is Xy- erb. Natur- A revision of Mezilaurus 4: 153-182. WERFF, H. VAN DER. 1987 (Lauraceae). Ann. Missouri Bot. Gard. 74 19 Fight new species and one om- bination a doe Lauraceae. Ann. Missouri Bot. Gard. New species of Lauraceae from Ес- uador i Peru. Ann. Missouri Bot. Gard. 78: 409- 423 ZAMORA, N. L. Povepa & E. CANESSA. 1988. Una de Caryodaphnopsis Airy Shaw (Laur- aceae) para la ropen neotropical. Ann. Missouri Bot. Gard. 75: 1160-1166. NEW SPECIES OF LAURACEAE FROM ECUADOR AND PERU' Henk van der Werff? ABSTRACT Seven species of Lauraceae from Ecuador and Peru are described and their affinities are discussed. The new species belong to the genera Caryodaphnopsis (C. tomen (Е. citriodora), Ocotea (О. alata, tosa), Cinnamomum ( O. infrafoveolata, and O. rotundata), and Rhodostemonodaphne (R. synandra). C. napoense and C. palaciosii), Endlicheria Intensive fieldwork during the past few years in Peru (by A. Gentry, R. Vàsquez, and N. Jaramillo) and Ecuador (mostly by D. Neill, C. Сегоп, and W. Palacios) has resulted in a large number of excellent collections of Lauraceae. Several of these represent undescribed species and seven of the most striking ones are described in this contribu- tion. I expect that as long as these collecting efforts continue, more undescribed species will become known and articles such as this one, describing these novelties, will appear with some regularity. Caryodaphnopsis tomentosa van der Werff, Ogica Jatun Sacha, tree, 25 m ta green, 7 Aug. 1989, Palacios 4337 (holo- type, MO; isotypes, AAU, BOG, CR, F, GH, HBG, К, KUN, L, LE, QAME, ОСЕ, QRS, NY, P, S, US). Figure 1. Ab speciebus ceteris Caryodaphnopsis ramulis et foliis subtus tomentosis, staminibus fertilibus 6, bilocellatis, re- cedit Tree, 25 m tall, the trunk with buttresses 1 m high. Twigs terete or slightly quadrangular, brown- tomentose when young, becoming glabrous. Ter- minal bud brown-tomentose. Leaves opposite, co- 7 cm, elliptic, the base obtuse or acute, the tip acute, triveined, the 2 lateral veins riaceous, 8-18 x 3- beginning at the very leaf base and ending in the apex, impressed on the upper surface and promi- nently raised on the lower surface, the perpendic- ular tertiary veins raised on the lower surface, the lateral veins with up to 10 branchlets toward the margin, these strongly arching upward and loop- connected; lamina glabrous above or with traces of pubescence at the bases of the major veins, brown-tomentose on lower surface. Petioles to 1 cm long, tomentose. Inflorescences clustered in the axils of leaves, paniculate, with opposite or alter- nate branching, to 7 cm long, whitish tomentellous in sicco, flowers grouped in cymes. Bracts of in- florescences persistent, linear, the larger ones ca. 2 mm long, bracts subtending flowers only half as long. Pedicels 2-3 mm long. Flowers green, with- out odor (fide collector). Tepals 6, strongly unequal, the outer three broadly ovate to roundfish, ca. 1 mm long, pubescent outside, glabrous inside, spreading; inner tepals erect, ca. 3 mm long, el- liptic, pubescent outside, inside with ttered hairs mostly near the margin. Stamens of whorls and II fertile, 2-celled, 1.5 mm long, the filament 1 mm long, dorsally pubescent, otherwise glabrous; stamens of whorl III 0.7 mm long, sterile or with 2 minute, vestigial locelli, the filament with two basal glands, the anther well developed and clearly differentiated from the filament; staminodia (whorl IV) ca. 0.7 mm long, with a pubescent filament and glabrous, cordate tip. Ovary glabrous, ellipsoid, gradually narrowed into the slender style. Recep- tacle shallow, pubescent inside. Caryodaphnopsis tomentosa is a very distinc- tive species by virtue of its tomentose pubescence on twigs and leaves, its pubescent inflorescences and flowers, the presence of only six fertile two- celled stamens, and its coriaceous leaves with strongly raised venation. Although its androecium is unique in Caryodaphnopsis, it possesses all other characters of the genus, such as opposite leaves, strongly unequal tepals, and a long, slender style. ' Luther Raechal kindly photographed the type specimens. Drawings were made by John Myers. I thank the National Geographic Society for financial support of fieldwork in Peru and Ecuador (grant numbers 3480-86 and 3821-88) 2 Missoni Botanical Garden, P.O.Box 299, St. Louis, Missouri 63166-0299, U.S.A. ANN. Missouni Bor. Garb. 78: 409-423. 1991. 410 Annals of the Missouri Botanical Garden FIGURE 1. Caryodaphnopsis tomentosa. — А. Habit. —B. Detail of inflorescence. — С. Pistil. — D. Stamen of whorl ni (sterile). — E. Staminode of whorl IV. —F. Stamen of whorl I/II ventral view. — (С. Larva inside flower. Volume 78, Number 2 1991 van der Werff 411 New Lauraceae from Ecuador and Peru Although Caryodaphnopsis tomentosa is only known with certainty from the type collection, three sterile collections from Peru probably represent the same species. These are: Gentry et al. 42 MO) and Vasquez & Jaramillo 6387 (MO), both from Yanamono, Explorama Tourist Camp, Maynas, Loreto, and Gentry et al. 55831 (MO), from the Explorama Inn, Maynas, Loreto. In venation and indument these collections are a good match for the type collection. Although fruits have not yet been collected, they were described by the collector as small and round, a fruit type which is to be expected in this group of Caryodaphnopsis species with strongly tri- veined leaves (van der Werff, 1988). The collector noted that small white larvae were frequently pres- ent in the flowers and, indeed, larvae were found in most dissected flowers. Cinnamomum Schaeffer The group of species to which Cinnamomum napoense and C. palaciosii belong has been in- cluded in Phoebe by earlier authors (Mez, 1889, for instance). Because Mez was the last person to survey all neotropical Lauraceae, his concepts of species and genera have long remained unchal- lenged. Until now I have, with some hesitation, accepted Mez's concept (van der Werff, 1 А 1988). Kostermans (1961) argued for the transfer of neotropical Phoebe species to Cinnamomum and made the necessary new combinations. Because various neotropical species of other genera have been erroneously described in Phoebe, I felt that these problems needed to be sorted out before accepting the remaining species in Cinnamomum. I have now changed my mind on this; it makes no sense to describe new species in a genus to which they do not belong. Although the problems with the delimitation of Cinnamomum are not solved and generic concepts may change again in the future, I think the neotropical species formerly treated in Phoebe are, with some exceptions, more closely related to Cinnamomum than to any other genus, and І accept Cinnamomum as their generic name. Wood anatomical characters also support the transfer of neotropical Phoebe species to Cin- namomum (Richter, 1981). The neotropical Cinnamomum species have two areas of species richness: northern Central America and southern Brazil. The genus is poorly repre- sented in other neotropical areas. Most collections from northern South America belong to the C. cinnamomifolia complex, which usually occurs in montane habitats (mostly 1,200-2,000 m). Al- though Mez (1889) reported the genus from Co- lombia and Peru, it has apparently not been re- corded from Ecuador. Recent fieldwork has shown that it is represented by at least three, and probably more species in Ecuador. Further collecting may demonstrate that the paucity of known Cinna- momum species in northern South America reflects incomplete collecting and that the genus is rep- resented by more species than previously thought. А possibly useful character in the identification of Cinnamomum is the odor of the flowers. The species of Cinnamomum I have collected (C. cos- taricanum (Mez & Pitt.) Kostermans, C. cinna- momifolium (HBK) Kostermans s.l. in Venezuela and Ecuador, and C. palaciosii van der Werff) all had flowers with a similar odor of urine. As a rule, Lauraceae have pleasantly fragrant flowers, and therefore I think this peculiar odor is indicative for the neotropical Cinnamomum species. Cinnamomum napoense van der Werff, sp. nov. TYPE: Ecuador. Маро, Estación Biologica Jatun Sacha, tree, 7 m, 2-5 Sep. 1988 (fl), Palacios 2886 (holotype, MO; isotypes, QCE, QAME, AAU, HBG, NY, US). Figures 2B, 3A-F. b ceteris speciebus subgeneris Heterantherae Mez foliis majoribus, petiolis s op et costarum axillis subtus non-barbellatis differ Trees, to 15 m tall. Twigs terete, glabrous or minutely puberulous when young, with a soft pith and sometimes fistulose. Terminal bud densely pu- berulous. Leaves alternate, somewhat clustered near the tips of branches, chartaceous, 18-30 x 6-12 cm, elliptic, the base acute or rarely obtuse, the tip acute, upper surface glabrous, lower surface glabrous and densely gland-dotted, sometimes with minute, appressed hairs along the major veins, lateral veins 5-8 on each side, arching upward near the margin adm somewhat loop-connected in the upper Y4 of the leaf, midrib and lateral veins impressed on upper surface and prominently raised on lower surface, the tertiary venation less raised and drying lighter than the lamina. Basal lateral veins stronger developed than the upper ones. Tufts of hair absent from the axils of the lateral veins. Petioles glabrous, canaliculate, 3-4.5 cm long. In- florescences clustered on very short, leafless shoots in the axils of normal leaves or along the twigs, minutely puberulous, 5-15 cm long, paniculate, the ultimate division(s) usually cymose, but occa- sionally racemose. Flowers yellow, pedicels 3-4 mm long. Tepals 6, erect at anthesis, ovate, 1.5- 2 mm long, on both surfaces with some appressed hairs, united at the base and falling off together 412 Annals of the Missouri Botanical Garden wssoum BOTANICAL GARDEN виззоли HERBARIUM BOTANICAL GARDEN HERBARIUM. № 3691605 № 3691607 MISSOURI BOTANICAL GARDEN ШЕР HERBARIUM. n BOTANA. caron dy MUBARUM № 3691606 № 3395404 - С а. 2 FIGURE 2. — А. Cinnamomum palaciosii. —B.Cinnamomum пароепзе. —C.Endlicheria citriodora. —D. Ocotea alata. Volume 78, Number 2 1991 van der Werff New Lauraceae from Ecuador and Peru 413 Fi -F. Cinnamomum napoense. — A. Floral detail. 4-celled stamens of whorl I/II, a 2-celled stamen of whorl III with basal glands and a оа — D. Pistil. Stamen of whorl III, the basal glands lacking, dorsal view. of inflorescence. — Н. Flower with several tepals removed. with the stamens in old flowers. Stamens 9, the outer six 4-celled, attached near the base of the tepals and slightly shorter than them, anthers ca. 0.8 mm long, cells opening introrsely and arranged in 2 horizontal pairs, the filament gradually widened into the anther, with some short hairs, especially dorsally; the inner 3 and staminodia attached on a ridge which is bent inward and largely covers the ovary, inner 3 stamens 2-celled, slightly shorter than the stamens of the outer whorls, the cells — B. Flower, outside. — C. Three fallen tepals, with —E —F. Cupule with fruit. G-J. Ocotea rotundata. —G. Part — I. Stamen of whorl I/II, ded view. —J. Cupules. opening extrorsely, the sparsely pubescent filament about as long as the anther; glands small, attached at the base of the stamens of the whorl III; stamino- dia 3, ca. 0.6 mm long, with a very short, pubescent filament and a large, sagittate tip. Ovary glabrous, ellipsoid, included in the receptacle, style slender, about as long as the ovary, the entire pistil ca. 1.8 mm long; receptacle with scattered hairs inside. Cupule ca. 1 cm tall, the rim slightly lobed, the young fruit included, older fruits greatly exserted, 414 Annals of the Missouri Botanical Garden FIGURE 4. F. Cinnamomum palaciosii. —C. Flower, with s A, B. Endlicheria iir oe Cross section dida ыш pistillate flower. — B. Staminate flower. C- me tepals removed. — D. Sta amen of whorl I/II, ventral view. — E. Stamen of whorl III, dorsal view. ЈЕ. о of whorl I. ca. 2 cm long; margin of cupule often with rem- nants of stamens. Paratypes. ECUADOR, NAPO: Rio Wai si ayá, Brand- byge et al. 36207 (fl) (AAU); San Pablo де los Secoyas, Brandbyge et al. 36229 (В) (AAU); km 32 carretera Hollin- Loreto, Neill et al. 8605, (8) (MO, QAME); Re- m Pto. Misahualli, Palacios AME); Anangu, SEF 8807 (fr) (NY); Reserva Floristica El Chuncho, Zaruma 611 (MO, QAME). Flowers: August-September; fruits (mature): January. By virtue of its 2-celled anthers of the whorl III of stamens, Cinnamomum napoense be- longs to subg. Heteranthera Mez. This subgenus consists of two Peruvian and several Brazilian spe- cies. The Brazilian species differ in their smaller leaves (less than 10 cm long), shorter petioles (less than 15 mm long) and often dense pubescence on the lower side of the leaves. Of the two Peruvian es, C. subsessile (Meissner) Kostermans has ele pod with a cordate base and C. het- eranthera (R Kostermans has small "Wi 9 cm long) a with tufts of hair in the axils of the lateral veins. Thus, C. napoense differs from the other species of subg. Heteranthera by its large leaves, long petioles, and lack of axillary tufts of hair. In older flowers the tepals break off just above their base and the margin of the cupule is slightly undulating due to the presence of the short tepal bases. Often a few stamens also persist on the margin of the cupule. Volume 78, Number 2 1991 van der Werff 415 New Lauraceae from Ecuador and Peru Cinnamomum palaciosii van der Werff, sp. nov. TYPE: Ecuador. Imbabura: forest remnant along road Apuela-Otovalo near Apuela, elev. 2,000 m, 14 1989, van der Werff & Palacios 10575 (holotype, MO; isotypes, AAU, QAME, QCE, NY, K, US). Figures 2A, 4C-F. congeneribus folis magnis, inflorescentiis tomen- "s et floribus pedicellisque glabris differt Tree, 25 m, with divided buttresses. Twigs an- gular, solid, when very young yellow-brown to- mentellous, the indument becoming gray and spars- er on old twigs. Terminal bud densely yellow-brown pubescent. Leaves alternate, chartaceous, 15-25 X 8-15 cm, broadly elliptic or elliptic-oblong,the tip shortly acuminate, the acumen ca. 1 cm long, the base cuneate or narrowly acute, midrib and lateral veins impressed on upper surface, promi- nently raised on lower surface, the tertiary venation less raised; lateral veins 3—5 on each side, the basal pair much more strongly developed than the upper ones, thus the leaves subtripliveined, the second pair of lateral veins near or above the middle of the lamina; upper surface glabrous with the ex- ception of some pubescence on midrib and basal lateral veins, lower surface sparsely (occasionally densely) appressed pubescent with the hairs not all oriented in the same direction, barbellate in the axils of the lateral veins. Petioles 2-3 cm long, canaliculate, with similar indument as twigs. Inflo- rescences single or a few together on short, leafless spurs, 4-14 cm long, paniculate, with the ultimate divisions cymose or sometimes appearing racemose through loss of the lateral flowers of a cyme, dense- ly yellow-brown tomentellous, especially along the main axis; bracts frequently present, tomentellous- pilose on outside, glabrous inside, 1-3 mm long, ovate to (very) narrowly ovate; pedicels and flowers c Ф un © Ф 5 = 5 = © =} Ф Uu о Ф 5 о Ф - Ф 5. о Ф, Ф = © " Ф = Ф = © л = long, the lateral flowers of a cyme often with shorter pedicels. Flowers green, smelling of urine. Tepals 6, equal, erect at anthesis, ovate, 3-3.5 mm long at anthesis, glabrous outside, the 3 inner ones pu- bescent inside, especially toward the base. Stamens 9, all 4-celled, the outer 6 with the introrse cells arranged in 2 horizontal rows; the anther ca. 1 mm long, about as long as the pubescent filament; inner 3 stamens with narrow anthers, the cells lateral-extrorse, filaments with 2 large glands near the base and slightly longer than the anthers; stami- nodia 3, 1.5 mm long, the cordate tip ca. 0.8 mm long, filament pubescent; ovary glabrous, globose, style slender, ca. 1.5 mm long. Receptacle pubes- cent inside. Fruit unknown. innamomum palaciosii, known only from the type collection, belongs to Phoebe subg. Phoebe sensu Mez. Within this subgenus, it keys to Cin- namomum salicifolium (Nees) Kostermans, a poorly known Mexican species with much smaller leaves and smaller flowers. The new species can be readily recognized by its large, tripliveined leaves with an angustate or cuneate base, and the rather large, glabrous flowers contrasting with the tomen- tellous inflorescence. It is a pleasure to name this species after Ing. Walter Palacios, who accompanied me on several field trips in Ecuador and who has collected several undescribed species of Lauraceae. Endlicheria citriodora van der Werff, sp. nov. TYPE: Peru. Loreto: Maynas, Jenaro Herrera, roadside near the arboretum, 7 Aug. 1988 (fl), van der Werf, Vasquez & Jaramillo 9991, 8 (holotype, MO; isotypes, AMAZ, HBG). Figures 2C, 4A, B. congeneribus foliis subtus glaucis, indumento га- mulorum brunneo et ramulis citriodoris diversa est. Dioecious tree, to 25 m tall. Twigs longitudinally ridged, roundish, densely brown-tomentellous with the odor of lemon and anise, the lemon scent dom- inating. Terminal bud brown-tomentellous. Leaves alternate, firmly chartaceous, 15-30 x 7-14 cm, elliptic, oblong or elliptic-ovate, the base obtuse, tip sharply acute to acuminate, the upper surface glabrous except for the tomentellous midvein, the lower surface glaucous and with varying amounts of small, appressed hairs, these never obscuring the surface; lateral veins 7-10 on each side, arch- ing upward near the margin but not becoming loop- connected, immersed on the upper surface, raised on the lower surface, the midvein prominently raised below, tertiary venation faintly visible on both sur- faces. Petioles 2-3.5 cm long, with similar indu- ment as twigs, the adaxial side flat, the abaxial side rounded. Inflorescences reddish, in axils of regular leaves or cataphylls, 10-15 cm long, brown-to- mentellous, 2-3 times branched, the ultimate branching cymose; in staminate plants the ultimate cymes usually sessile, the flowers thus in dense groups of 6-7 (rarely 9 or 13) flowers. Pedicels 1.5-2.5 mm long and flowers densely, minutely pubescent, the indument covering the surface en- tirely. Receptacle deep, ca. 0.7 mm long, slightly constricted near the tip, densely pubescent inside. Tepals 6, equal, spreading at anthesis, the tips bent inward, ca. 0.7 mm long, ovate-elliptic, with some 416 Annals of th Missouri Botanical Garden hairs on the inner surface. Male flowers: tepals cous lower leaf surface, the indument of twigs and reddish and stamens yellow, at anthesis the tepals spreading and the stamens fully exposed. Stamens 9, all 2-celled, the outer 6 ca. 0.5 mm long, almost sessile, the filament narrower than the anther, the cells large, the connective extending beyond the cells, filament with some hairs at the base; inner 3 stamens with extrorse cells, 0.7 mm long, the pubescent filament about as long as and wider than the anther; no glands seen on the filaments. Pis- tillode threadlike, glabrous, without a stigma, not surpassing the stamens. Female flowers: tepals at anthesis erect, the tips slightly incurved. Stami- nodes 0.3-0.5 mm long, the filaments broad, pu- bescent, anthers 2-celled, the cells opening. Stam- inodes surpassed by the style and more or less hidden under the stigma. Pistil 1.2-1.5 mm long, glabrous, ovary about as long and wide as the style, stigma large, peltate. Cupule shallowly cupshaped, red in vivo, smooth, 1.1 cm; in sicco black, this contrasting shergly with the tan-pubescent pëd. icel. Fruit ellipsoid, ca. 2.5 x 1 Paratypes. PERU. LORETO: Requena, Jenaro Herrera, A. Castillo 13, 6 (В) (MO), tree 4-478, A., Castillo 34, 8 (fl) (MO), Vasquez & Jaramillo 9593, 8 (fl) (MO, AMAZ); Maynas, Puerto Almendras van der Werff et al. 9776 (st) (МО, AMAZ), Maynas, Mishana, Gentry et al. 25983 (st) (MO) Gentry et S 39301 (fr) (MO), van der Werff et al. 10187, е (A ) (MO, AMAZ); Maynas, Alpa- huayo, field station of IIAP, van der Werff et al. 10242, 8 (fl) (MO, AMAZ). Flowers: August-September; fruits January. Common names: Limón-moena, Anis moena. Endlicheria citriodora has been collected in the vicinity of Iquitos and Jenaro Herrera in Amazo- nian Peru. It is restricted to a forest type locally called **varillal," a ca. 25-30-m-tall, closed forest on a substrate of predominantly white sand with a small amount of clay. The glaucous lower leaf surface and the lemon scent are very good field characters; the scent is less pronounced in dried material, and specimens preserved in alcohol lose the glaucous cast. The two Castillo collections have retained the striking glaucous lower leaf surfaces. The deep floral tube, constricted near the apex, resembles the shape of Aniba flowers. However, the unisexual flowers, anthers with large locelli and the smooth, fleshy cupule are characteristic of End- licheria. This genus was last revised by Koster- mans (1937) and is in need of an updated treat- ment. The closest relatives of Ё. citriodora are in the group of species with a deep receptacle and spreading tepals (E. klugii О. C. Schmidt, Е. browniana Mez, Е. pyriformis (Nees) Mez, E metallica Kostermans, among others). The glau- ty inflorescence and the lemon odor make identifi- cation easy, however. An unusual character of the staminate inflores- cences 15 the frequent suppression of the peduncles of the ultimate cymes. Thus, the flowers appear in pseudo-umbellate clusters, mostly consisting of six flowers (two cymes along an axis), or seven flowers (a terminal cyme with the two lateral divisions again cymosely branched). Occasionally, when the ulti- mate part of the main axis also becomes sup- pressed, the pseudo-umbels may consist of nine (one terminal and two lateral cymes condensed) or 13 (two lateral and a branched terminal cyme) flowers. Frequently, the number of flowers is small- er, because the central flowers of each cyme de- velop faster than the laterals and fall off before the laterals. Ocotea alata van der Werff, sp. nov. TYPE: Peru. Loreto: Maynas, Mishana, Rio Nanay halfway between Iquitos and Santa Maria de Nanay, upland forest on white sand, 12-13 Jan. 1983, 9 (8), Gentry et al. 39447 (holotype, MO). Figure 2D. Ab Ocotea glomerata et O. magnifica ramulis alatis et folis basi rotundatis vel subcordatis recedit. Dioecious tree to 20 m tall, but mostly less than 10 m. Twigs winged, the wings to 5 mm broad, young twigs densely tomentellous, glabrescent with age. Terminal bud densely tomentellous. Leaves alternate, chartaceous, 35-65 x 10-20 cm, (nar- rowly) obovate or elliptic, the base abruptly round- ed to subcordate, the tip acute to acuminate, the upper surface glabrous, the lower surface with short, erect hairs ca. 0.2 mm long, the pubescence spars- er and shorter on old leaves, the surface always visible; venation immersed on upper surface, mid- vein prominently raised on lower surface, the lat- eral veins (15-20 on each side) raised and tertiary venation slightly raised; lateral veins arching up- ward near the margin and becoming weakly loop- connected in the upper half of the lamina. Petioles 1–1.5 cm long, to 6 mm thick, canaliculate, with the same indument as twigs. Inflorescences in axils of leaves or cataphylls, 15-25 cm long, densely tomentellous, the lateral branches once or twice cymosely branched, many-flowered. Flowers sessile or nearly so, white, the tepals erect at anthesis. Tepals 6, equal, m long, densely ap- pressed pubescent outside, sparsely pubescent in- side, broadly ovate-elliptic. Male flowers: stamens 9, all 4-celled, glabrous, the outer 6 with introrse Volume 78, Number 2 1991 van der Werff 417 New Lauraceae from Ecuador and Peru cells, ca. 1 mm long, the filaments distinct, 0.4 m long; inner 3 stamens fused, ca. 1 mm long, the cells extrorse; glands small, globose, attached at the base of the inner stamens; pistillode linear, enclosed by the inner stamens, with a small stigma, glabrous, the receptacle pubescent inside. Female flowers: staminodes 9, traces of anther cells visible but locelli not opening, inner 3 staminodes free; stigma platelike, raised above the staminodes, ova- ry globose, glabrous, enclosed in the glabrous re- ceptacle. Tepals initially persisting in the fruiting stage, but finally deciduous. Immature fruit en- closed in receptacle, mature fruit ovoid, ca. 15 х 9 mm, largely exserted, the cupule deeply cup- shaped. Distribution: Amazonian Peru, from Jenaro Her- rera downstream to the Brazilian border. Paratypes. PERU. LORETO: Prov. Requena, Jenaro Her- rera, Gentry et al. 21199, 8 (fl) (MO); Spichiger & "ncarnación 1167, tree 6/125, 8 (fl) (MO); peg ba : p deos 232. 8 (8) (МО); van der Werff et al. 10 (st) (МО); collector b cae tree 6/125, 8 (8) a Prov. Loreto, Nauta, uez & Jaramillo 5059, 8 (fl) (MO) Пеш 925 es fr) (MO), 1640, 8 (fl) (МО); Rimachi 3745 (fr) (MO); Vasquez & Jaramillo 4483, ê (Я) (MO), 8697, (young fr) (MO), 8918, (fr) (MO); Prov. Maynas, border post on Rio Javeri, Gentry & Revilla 20897 (fr) (MO). Flowers year round (January, April, June, July, October, November, December); fruits collected in March, July, and November. Ocotea alata belongs to the O. guianensis group of Rohwer (1986) because of its nearly sessile flowers, the fused stamens of whorl III, the per- sistent tepals of the young fruit, and the cupule shape. Within this group, it is most closely related to О. glomerata (Nees) Mez s.l., the only species with erect pubescence on the lower leaf surface. Part of O. glomerata s.l. is O. magnifica O.C. Schmidt (isotype, NY). This segregate has larger leaves with an inrolled base and a typical indument on the lower leaf surface (a dense, tomentellous pubescence and longer erect hairs), and is only known to me from the type collected at the middle Ucuyali. Two recent collections from the Tingo Maria area have the pubescence characters of O. magnifica, but resemble O. glomerata in leaf size and absence of an inrolled base. More collections are needed to determine the status of O. magnifica. Ocotea alata resembles O. magnifica in leaf size, but can be easily recognized by its winged stems and rounded leaf base. In the past, I have identified all cited specimens as O. magnifica and duplicates have been distributed under that name. Spichiger et al. (1989) also used the name О. magnifica for specimens from Jenaro Herrera; their illustration does not show the abruptly rounded leaf base, al- though they correctly describe this character. Oco- tea alata occurs in varillal (see previous species). Ocotea infrafoveolata van der Werff, sp. nov. TYPE: Ecuador. Zamora—Chinchipe: at the pass dersson 21968 (holotype, GB; isotype, MO). Figure 5. Arbor parva, ramulis crassis, juvenalibus ferrugineo- tomentosis, vetustioribus glabris; foliis ellipticis, subtus ferrugineo-tomentosis, valde foveola а E tiis ax- illaribus, paniculatis; floribus herm са. 1 cm diametro; staminibus 9, А осећа; LN nullis, cu- pula parva. Shrub or small tree, 6(-10) m tall. Twigs an- gular, ferruginous tomentose when young, becom- ing glabrous, often with conspicuous scars of fallen 6-15 2.5-6 cm, strongly coriaceous, elliptic to elliptic-oblong, the leaves. Leaves alternate, tip acute, the base gradually narrowed, at the very base revolute and decurrent on the petiole; the upper surface of very young leaves tomentose, this tomentum soon disappearing and the surface shiny and glabrous except for traces of tomentum on the main veins and the margin; lower surface with a dense, ferruginous tomentum, the midvein, sec- ondary and tertiary veins with a lanate ferruginous pubescence, the surface between veins covered with a shorter, lighter-colored tomentose pubescence, or rarely the tomentum absent, both kinds of pu- bescence thinner and lighter colored with age; lat- eral veins 10-15 pairs, midvein and lateral veins immersed on upper surfaces, but, together with tertiary venation, prominently raised on the lower surface, which is strongly pitted by the raised re- ticulation; margin of the leaves revolute, more strongly so toward the base of the leaves. Petioles broad, ca. 5 mm wide, and to 1 cm long, but mostly poorly defined because of the decurrent laminae; pubescence on young petioles a ferruginous to- supporting the basal branchlets of large inflores- cences occasionally leaflike and persisting; inflo- rescence ferruginous tomentose near base, becom- ing ferruginous lanate in the distal part. Flowers hermaphrodite, greenish yellow, sessile or nearly so, subtended by a narrowly ovate, ferruginous tomentose bract, to 5 mm long. Tepals 6, equal, spreading ovate, 4 mm long, 3 mm wide, ferru- Annals of the Missouri Botanical Garden DO Y È SER Gi NS Yu Ni sJ MAE 4 pes PTS. * е pe f ۶ E LN " E и UNO кыш bán ES FIGURE 5. Ocotea | Ps Habit. — B. Flower. — C. Pistil. — D. Stamen of whorl I/II, ventral view. — ule and fruit. — C. oe of pubescence along midrib on lower leaf surface E. Stamen whorl III, dorsal view. — F. (lower half) and upper leaf Е ош. half). — Н. Leaf bas Volume 78, Number 2 1991 van der Werff 419 New Lauraceae from Ecuador and Peru ginous tomentose outside, the pubescence more appressed on the inner face of the tepals; fertile stamens 9, all 4-celled, the outer 6 with introrse cells, the filaments glabrous or with few hairs, ca. 0.5 mm wide and 1 mm long, the anthers glabrous, ca. 0.8 mm wide and 1.2 mm long; inner 3 stamens with extrorse cells, the filaments with 2 glands near the base, filaments and anthers each ca. 1 mm long; staminodia lacking; ovary glabrous, enclosed in the cupshaped, pubescent floral tube, globose, style ca. 2 mm long, the stigma platelike, ca. 0.7 mm diam. Infructescence woody, glabrous, the cupule a shallow, small cup, ca. 1.5 cm wide and 0.7 cm tall (outside) with few, rather large lenticels. Fruit a 1-seeded berry, broadly el- lipsoid, to 3 cm long and 2.5 cm wide when dry; young cupule with weakly lobed margin, margin becoming straight at maturity. са. 1 mm diam.; Paratypes. COLOMBIA. DEPTO. CAUCA: Cuadros 359 (MO), Cuatrecasas 18942 (MO, r DEPTO. NARIÑO: Soe jarto 1999 (F). DEPTO. = VALLE: Cuatrecasas 17910 , Cuatrecasas 20418 (F), CD 20588 (F), Cuatrecasas 21906 (F Devia 649 CUADOR. PROVINCE CARCH 2620 (AAU, NY), Holm-Nielsen et al. 29858 (AAU), Jaramillo 7686 (GB). PROVINCE AZUAY: Camp E-4808 ( Y), Harling & Andersson 21177 (GB), Jaramillo 7149 (GB), Knight 218 (WIS), Steyermark 53424, 53425 (F). PROVINCE ГОЈА: van der Werff & Palacios 9426 (AAU, HBG, MO, QAME). PROVINCE UNKNOWN: Espinosa 1087 (F, NY). Flowers and fruits year-round. Ocotea. infrafoveolata is a high-elevation (be- tween 2,600 and 3,400 m) shrub or small tree, only known from southern Colombia and adjacent northern Ecuador and south of Cuenca in Ecuador, with one collection in central Ecuador. Its affinities in Ocotea, a largely neotropical genus of probably more than 300 species, are not clear. The shape of the inflorescence, the cupule, leaf shape, and decurrent leaf bases suggest an affinity to O. cal- ophylla Mez, O. sericea HBK, and O. micans Mez, all densely pubescent, high-elevation species of northern South America and Costa Rica. How- ever, none of these species have a foveolate lower leaf surface and all have unisexual flowers. Some collections of O. infrafoveolata have slightly small- er flowers than others, but locelli of both smaller and larger flowers contain pollen, even on speci- mens with old fruits. Because the unisexual char- acter of O. calophylla and O. sericea is quite bvious, it is unlikely that the smaller and large flowers of O. infrafoveolata represent different sex- es. Another difference is the abse lines on the leaves of O. infrafoveolata, while these nce of vernation are conspicuous on the leaves of O. calophylla and related species In spite of these: differences, O. infrafoveolata seems to be related to the group of O. calophylla. There are other species of Lauraceae restricted to high elevations in the Andes and with a similar pubescence, but these species (in Ocotea and Per- sea) differ clearly in floral and cupule characters, and the ferruginous pubescence is more likely an adaptation to cold, wet habitats than an indication of taxonomic relationship. Ocotea rotundata van der Werff, sp. nov. TYPE: Ecuador. Prov. Loja: Parque Nacional Podo- carpus, along road from Yangana to radio tower on Cerro Toledo, cloud forest, 2,900 m, 29 Арг. 1987, van der Werff £ Palacios 9192 (holotype, MO; isotypes, AAU, HBG, QAME). Figures 6, 3G—J. 0 + ++: ы УМ "ЕР s + dd , , petiolis brevioribus et pubescentia longiore differt. Shrub or small tree, to 10 m tall. Twigs slightly angular and rufous tomentose when young, becom- ing terete and glabrous with age. Terminal bud ca. 9 mm long, rufous tomentose. Leaves alternate, 5-8 x 4.5-6 cm, broadly elliptic, the smallest leaves often roundish, coriaceous, the base round- ed, the tip rounded or with a very short acumen, the upper surface glabrous (young leaves with some rufous pubescence along costa and veins, but this wearing off very quickly) and with costa and veins immersed, the lower surface rufous tomentose, col- or of the tomentum fading on old leaves and the tomentum wearing off, but leaves very rarely be- coming glabrous, costa and veins elevated on lower surface, but tertiary venation not visible; margin plane, not thickened; veins 4-6 pairs, the lower 2 or 3 close together near the base of the leaf. Petioles 3-6 mm long, to 4 mm thick, with similar type and density of pubescence as the twigs. Inflores- cences axillary, 6-12 cm long, rufous tomentose, paniculate or, when small, racemose, the flowers usually arranged in cymes. Flowers perfect, pale yellow or creamy, the floral tube deep and gradually narrowed in the very short pedicel. Pedicels, floral tubes, and the outside of the tepals ferruginous tomentose (inner tepals with glabrous margin and tip. Tepals 6, ovate, ca. 3 mm long, the inner surface with some papillose pubescence near the apex, this best visible in young flowers, otherwise glabrous. Stamens 9, all 4-celled (but many flowers with 1 or 2 cells apparently sterile; the position of the sterile cells variable), the cells arranged in 2 horizontal rows; filaments of the outer stamens ca. 420 Annals of the Missouri Botanical Garden MISSOURI — Ne 3691604 A " FIGURE 6. Ocotea rotundata. 0.8 mm long, dorsally with a narrow row of hairs; anthers of the outer stamens curved inward, their cells introrse; inner stamens erect, ca. 1.6 mm long, the pubescent filament as long as the anther, glands large, attached near the base of the filament; upper pair of anther cells lateral, the lower pair extrorse; staminodia ca. 1 mm long, the lower part pubescent, the tip glabrous and swollen. Pistil ca. .9 mm long, glabrous; ovary enclosed in the deep floral tube, floral tube densely pubescent inside. Fruit ellipsoid, 2 x 1.5 cm, cupule funnelshaped, with a weakly developed double margin. Paratypes ECUADOR. PROVINCE LOJA: Parque Na- cional Podocarpus, Molau & Eriksen 3217 (GB), Larsen & Dall 229 (GB), Stein & D'Alessandro 2718 (MO), van der Werff & Palacios M 9195, 9283, 9285, and 9321 (all MO, HBG, QAM Ocotea rotundata is known only from the Par- que Nacional Podocarpus, to the south of Loja, at elevations of ca. 3,000 m. It only grows in wet habitats; it has not been collected on the drier slopes to the north and west of Loja, such as Cerro Villona- co. The new species belongs to the Ocotea aci- phylla (Nees) Мег group, as circumscribed by Rohwer (1986). Characteristics for this group are hermaphroditic flowers, a deep floral tube (and large cupule in fruit), rather large staminodia, and basal lateral veins that are close together. Although this group includes mostly lowland species, several Andean species (O. arnottiana (Nees) van der Werff, O. heterochroma Mez & Sodiro, O. jelskii Mez, and O. stuebelii Mez) belong here as well. These species, which are not well known, differ from O. rotundata in having narrower leaves with acute tips and bases and longer petioles, as well as in details of the pubescence. In leaf shape, O. rotundata is very similar to O. julianii van der Werff (Phoebe steyermarkii Allen; van der Werff, 1989). The latter is restricted to the summit areas of Chimantá-tepui, Edo. Bolivar, Venezuela, and differs in its shorter, often darker tomentum on the Volume 78, Number 2 van der Werff 421 1991 New Lauraceae from Ecuador and Peru y. И dA i. ДИ RH |; У Uer li NOTAE . И. ПАЈА И d D RAE | у AY И 7 I FIGURE 7. Rhodostemonodaphne synandra. — А. Habit. — B. Staminate flower, tepals and stamens of whorl I D. Тера] with stamen of whorl I.—E. Pistillate flower, some tepals and stamens of whorl I removed.—F. Pistil. — G. Detail of pubescence and venation on lower leaf surface (lower half) and upper leaf surface (upper half). 422 Annals of the Missouri Botanical Garden leaves, shorter tomentum on the twigs, a much larger cupule and fruit, and its distribution. Rhodostemonodaphne synandra van der erff, sp. nov. TYPE: Ecuador. Napo: Estación Biologica Jatun Sacha, 8 km downstream from Puerto Misahualli, tree, 30 m, 8 flowers green, Neill & Palacios 7129 (holotype, MO; iso- types, QAME, HBC). Figure 7. Ab R. solaris жа ` pubescentia cinna- momea, floribus glomeratis, filamentis staminum serierum II and Ш 25 connatis, et t glandulis inb late differt. Dioecious tree, to 30 m. Twigs hollow, slightly angled to round, densely yellow-brown tomentellous when young, the indumentum diminishing with age. Terminal bud densely tomentellous. Leaves alter- nate, elliptic or oblong, 20-30 x 7-11 cm, stiffly chartaceous, the base acute or obtuse, the apex acuminate; upper surface glabrous (except some pubescence on the midrib), often shiny, midrib and lateral veins immersed, but tertiary venation raised; lower surface with raised midrib, lateral veins and tertiary venation, the entire venation system laxly tomentellous; lateral veins 10-15 pairs. Petioles -3 ст long, brown-tomentellous, canaliculate. Staminate inflorescences axillary, to 10 cm long, yellow-brown tomentellous; the lower flowers of the inflorescence clustered near the tips of lateral axes, these axes becoming shorter toward the tips of the inflorescence; pistillate inflorescences to 7 cm long, unbranched, the flowers clustered along the main axis. Flowers gray-tomentellous, sessile or nearly so, campanulate, 3 (dry) to 5 (after boiling) mm long. Staminate flowers: tepals 6, iie spreading to reflexed at anthesis, tomentellous pubescent and slightly papillate inside, ca. 2 x 1.2 mm. Stamens 9, all 4-celle sparsely ; outer 3 stamens at- lateral; stamens of whorl II and whorl III ca. 3 mm long, their filaments fused for the lower 2 mm and free for the upper 1 mm, forming a central column; anthers of whorl II with introrse cells, of whorl III with extrorse cells; stamens of whorl III with 2 glands attached at the base of the free part of the filament; staminodia (whorl IV) and pistillode not seen. Floral tube pubescent inside. Pistillate flowers: staminodia of whorl I ca. 1.5 mm long, free, the thick, fleshy filament pubescent, the tip glabrous, cells rudimentary, introrse; filaments of whorls II and III partly connate, only upper part free; glands of whorl III attached at the base of the free part of the stamens, glabrous, small; ovary ellipsoid, ca. 2 mm long, glabrous; style distinct, ca. 1.5 mm long; floral tube glabrous inside. Fruits unknown. atype PERU. HUANUCO: Fundo Sinchona, km 209 carretera Huanuco-Pucallpo, 1,200 m, Hodge & Les- cano 1, 9 (fl) (F); same locality, Lescano 2 9 (fl) (MAD). Rhodostemonodaphne synandra is only known from the three cited collections. The two Peruvian collections carry exactly the same label informa- tion, were both collected on 9 October, and are both pistillate; it is probable that they are the same collection and that the label erroneously states that one specimen (MAD) was collected in 1945 and the other (F) specimen in Vegetatively, Riodasiemonadaphne synandra is quite similar to the widespread К. kunthiana (Mez) Rohwer. Rhodostemonodaphne synandra differs in its yellow-brown pubescence (ferruginous brown in R. kunthiana) and its narrower, more oblong leaves. The inflorescence of R. synandra is smaller, with clustered, nearly sessile flowers. The flowers themselves are highly diagnostic. They are larger than in R. kunthiana, and the stamens are attached near the bottom of the floral tube, not near the margin, as in R. kunthiana. The filaments are, for the genus, very long, but even more remarkable is that the filaments of the sta- mens of whorls II and III are fused for % of their length. Usually, when there are differences between the staminal whorls in Lauraceae flowers, whorls I and П are similar and differ from whorl III, for instance in Kubitzkia, where the filaments of whorl III are fused, but those whorls I and II are free. Thus, the situation in R. synandra, where whorl I is separate from the fused whorls II and III, is highly unusual. An additional difference between R. synandra and R. kunthiana is that thiana has large, staminal glands which are visible . kun- between the outer stamens at anthesis, whereas in R. synandra the glands are small and only visible in dissected flowers. Meissner (1864) coined the name Synandrodaphne because the very bases of the stamens were fused into a ring like structure. I use the epithet synandra for a different reason, namely the fusion of the filaments of the stamens of whorls II and III Rhodostemonodaphne is a small neotropical ge- nus of about ten species, recently named by Rohwer & Kubitzki (1985), who raised Местапага subg. Synandrodaphne to generic level. Because Syn- androdaphne Gilg is conserved over Synandro- daphne Meissner, they proposed the name Rho- dostemonodaphne. With two exceptions (А. grandis (Mez) Rohwer and К. kunthiana (Nees) Volume 78, Number 2 1991 van der Werff 423 New Lauraceae from Ecuador and Peru Rohwer), the species placed in Rhodostemono- daphne by Rohwer (1986) are very poorly known. Rhodostemonodaphne is characterized by the fol- lowing combination of characters: flowers unisex- ual; anthers 4-celled, the cells arranged in an arc; filaments poorly (or not at all) differentiated from the anther; staminodia lacking; and pubescence usually dense. The exclusion of Rhodostemono- daphne from Nectandra strengthens the unity of Nectandra and seems fully justified. LITERATURE CITED ms А. J. С. H. 1937. Revision of the Laur- The genera Endlicheria, ши баш аи species) and Licaria. Recueil Trav. Bot Neerl. 34: 500-609. — The New World species of ин w. (Lauraceae). Reinwardtia 6: 17-24 Мыне С. T 864. Lauraceae. In: А. М enc (editor), enor es" oo Naturalis Regni Ve- getabilis 15(1): 1 МЕР, С. 1889. er Americanae. Jahrb. Kónigl. Bot. Gart. Berlin 5: 1-556. RICHTER, Н. С. 1981. ан des sekundaeren Xy- lems und der Rinde der Lauraceae. Sonderb. Natur- j Я Prodromus einer Monographie der gw Had Aubl. Mitt. Inst. Allg. Bot. Ham- burg 20: & BITZK Entwicklungslinien im е “Komplex alc Bot. Jahrb. Syst. 107: 129- о ^ J. Меког, P. Loizeau & L. Stutz DE ORTEGA. im Los árboles del н Јепаго me 1. Boissiera 43: 359. WERFF, Н. VAN DER. 1987. Six new ss cies of neo- tropical Lauraceae. Ann. Missouri Bot. Gard. 74: 401- EU . Eight new species and a new combi- nation ч neotropical Lauraceae. Ann. Missouri Bot. Gard. 75: 402-419 1989. Novelties in the Lauraceae from Ven- me Guayana. Ann. Missouri Bot. Gard. 76: 462- 475. REVISIONARY STUDIES IN THE ERIOCAULACEAE OF VENEZUELA! Nancy Hensold? ABSTRACT Preparation of the Eriocaulaceae treatment for Julian Steyermark's Flora of the Venezuelan Guayana has brought to light a new achene-bearing subgenus, Paepalanthus s and one variety: Eriocaulon spongiola, Paep S. duidae var. humilis, some instances of synonymy are discuss Monosperma, as well as the following six new species subg. alanthus chimantensis, P. fasciculoides, Syngonanthus acephalus, 5. fenestratus, and S setifolius. In addition, 14 new names or combinations are made, and sed. The Guayana Highland, particularly in Vene- zuela, is one of the two major centers of distribution for the Eriocaulaceae in the New World, the second area being the Brazilian Central Plateau. Of the approximately 100 species of Eriocaulaceae oc- curring in Venezuela, about 85 occur in Guayana, 38 of which are endemic there. In preparation of the Eriocaulaceae treatment for the Flora of the Venezuelan Guayana (Julian Steyermark, editor), several new taxa of the region have come to light, and many revisionary changes have been found necessary. Of special interest are: Eriocaulon spongiola, a species with floating ribbonlike pe- duncles and spongy hygroscopic involucral bracts; a dimerous species of Syngonanthus, with excep- tionally reduced inflorescences lacking peduncles and involucres; and a new subgenus of Paepalan- thus producing single-seeded indehiscent fruits, the only known departure in the family from 3-locular capsules ERIOCAULON Eriocaulon spongiola Hensold, sp. nov. TYPE: Colombia. Vaupés: Rio Apaporis, Rio Pira- paraná, Cano Teemeena, between 0?15'S, 70°30'W and 0?25'N, 70?30'W, Schultes & Cabrera 17314 (holotype, US; isotype, MO). Figures 1, 2. Planta aquatica caulescens. Folia fasciaria, membrana- cea, valde fenestrata, fluitantia, usque 30 cm longa et 5- 1: | > Gh f, 7 mm lata. P fasciarii, fenestra 2-3 mm lati, folia simulantes. Capitula . Bracteae involucrales pallide dissi apicibus songs hygroscopicis auctae. Flores trimeri; sepala libera vel in spatha connata; petala da. antherae albidae Plants aquatics in shallow streams, the base root- ed, the stems, leaves, and peduncles submerged or floating. Roots relatively fine, ca. 0.5 mm diam. Stems up to 25-30 cm long, ca. 1-2.5 mm wide, glabrous. Leaves linear, ribbonlike, up to 30 cm or longer and 5-7 mm wide, membranous, gla- brous, the tips often broken off, conspicuously fe- nestrate from base to apex, the longitudinal and transverse striations of about equal thickness. In- florescence arrangement unknown (peduncles not attached to the stem in the available material). Peduncles 40-50 cm long, filiform at base, broad- ening and ribbonlike toward the apex, up to 2.5- 4 mm wide, membranous, glabrous, fenestrate like the leaves, not evidently costate. Capitula 5-6 mm diam., hemispheric to ovoid, with an elongate re- ceptacle. Involucral bracts in only 1-2 series, or- bicular or somewhat tapered toward the irregularly rounded apex, ca. 1.5-2.2 mm long and nearly as wide, spongy, pale brownish, glabrous. Floral bracts similar in size and color to the involucral bracts, the base broad, carinate, and membranous, held perpendicular to the elongate receptacle and cup- ping the flowers, the apex abruptly differentiated from the base, broadly rounded to truncate or ! I thank the curators of F, LL, MER, MO, MY, NY, PORT, US, and VEN for loan of specimens. The illustrations are by Bru uno Manara through the Flora of Venezuelan Guayana Project. This work was completed in part during a ica Fellowship supported by the Julian Steyermark Fund.‏ ا Department of Botany, Field Museum of Natural History, Roosevelt Road at Lakeshore Drive, Chicago, Illinois‏ 2 60605-2496, U.S.A. ANN. Missouni Bor. GARD. 78: 424-440. 1991. 425 Hensold Volume 78, Number 2 1991 Eriocaulaceae of Venezuela Eriocaulon spongiola Hensold (Schultes & Cabrera 17314 US). — A. Habit. — B. Leaf. — C. Inflo- FIGURE 1. гезсепсе. — D. Involucral bract. 426 Annals of the Missouri Botanical Garden FIGURE 2. ry 746 (VEN). into a truncate spat lyx.— С Епосашоп spongiola Hensold. A-C. Ber — А. я minate flower, the calyx fused tes & ا‎ , the calyx only shal- istillate flower. — E. Sepal, abaxial view. — Е. Petal, adaxial view. Scale bar = 1.5 mm triangular-obtuse (progressively smaller distally), spongy, hygroscopic, expanding and folding up- ward against the surface of the capitulum to enclose the flowers when wet, pilose abaxially with white, clavate hairs in a transverse band below the apex. Staminate flowers. Sepals 3, free, oblong and rounded (Colombia) or completely fused into a cu- neate, perfectly truncate adaxial spathe (Venezue- la), in either case ca. 1.9-2.4 mm long, light brown, membranous, densely ciliate with white clavate hairs diam. Petals free, equal or subequal, ca. 0.4 mm long, pilose at the upper margin and adaxially at the apex, glandular (Ven- ezuela) or not (Colombia). Stamens with filaments about equaling the petals; anthers white, ca. 0.25 mm long. Pistillate flowers. Sepals 3, free, spatulate (Colombia) or completely fused into a spathe as in the staminate flowers (Venezuela), in either case ca. 1.5-1.6 mm long, the same color, texture, and pubescence as in the staminate flowers. Petals free, broadly spatulate, rounded, about equaling the se- pals, tufted with white clavate hairs adaxially at apex, and also pubescent with curly brownish fil- amentous hairs adaxially below apex, glandular (Venezuela) or not (Colombia). nos specimen examined. VENEZUELA. AS: in caño ca. 4 km S of Macuruco on road to Santa Barbara del Orinoco, Berry 746 (VEN) (inflores- cences only). This species had previously been mistaken for Е. spruceanum Koern., which also may have float- ing, membranous, ribbonlike leaves, and which may occur in the same general area (though it is not known from Venezuelan Guayana). Eriocaulon spongiola is easily distinguished by its elongate stem, its very conspicuously fenestrate leaves, and particularly by its fenestrate, membranous pedun- cles and the spongy, hygroscopic apices of the floral bracts. Eriocaulon spruceanum also has fenestrate leaves, but these often must be held up to the light for the transverse striations to be clearly seen, and these tend to be finer and less conspicuous than the longitudinal striations. In addition, E. spru- ceanum almost always has firm, erect, conspicu- ously costate peduncles, which are never fenes- trate Because floral morphology, particularly sepal form and fusion, seems to be a very plastic char- acter in Eriocaulon, I have not recognized the Venezuelan material, with sepals fused into a cu- neate spathe, as distinct from the Colombian ma- terial, with the sepals completely free. PAEPALANTHUS Paepalanthus subg. Monosperma Hensold, subg. nov. TYPE: Paepalanthus squamulifer- us (Mo Herba perennis caulescens vel rosulata rhizomatosa. Radices veteres saepe vi styli fere aequantes. in pg bifida. Ovarium semen unum efferens; fructus achen Caulescent or rhizomatous rosulate perennial herbs. Roots firm, usually whitish, the epidermis even of older roots, often persistently villous on exposed aboveground portions. Leaves linear to ligulate or lanceolate, chartaceous to coriaceous, smooth above and striate below or the upper sur- face 1-3-sulcate, variously pubescent with fila- ps oni асле solitary to numerous, the hs foliaceous with an acute apex. a 4-16 mm diam., hemispheric to globose or fungiform. Involucral bracts mostly dark black- ish brown, occasionally paler and then dull yellow- brown or copper-colored, + ovate, the apex obtuse to acute, acuminate or apiculate, enclosing but never surpassing the head, glabrous to ciliate and abaxially sericeous. Receptacle pilose. Floral bracts oblanceolate to oblong or squarish, subequaling the flowers, dark brown, barbate abaxially at apex, the Volume 78, Number 2 1991 Hensold 427 Eriocaulaceae of Venezuela trichomes linear-obtuse to clavate, granular within. Flowers 3-merous, unisexual, actinomorphic, short- pedicellate. Staminate flowers. Sepals free or briefly connate at base, oblanceolate to obovate, acute to rounded, dark brown, barbate abaxially. Corolla tubular, borne on an androphore, the androphore commonly stout and about equal to corolla tube in length, though sometimes much shorter, the tube rather narrowly infundibular, membranous, the lobes 3 and sharply triangular-acute or sometimes low, rounded, and irregular. Filaments free, the anthers exsert, white to brown. Pistillodes large, simulating appendages of gynoecium. Pistillate flowers. Sepals free, obovate to sometimes ovate or oblong, similar in color and pubescence to those of staminate flowers, the midrib area thickened and pale in fruit. Petals oblong to narrowly elliptic, lanceolate or oblanceolate, usually much narrower than sepals, blackish streaked, and usually densely pubescent with very long cilia, especially toward the base, rarely (P. chimantensis) subglabrous, thickening throughout and the tips recurving at maturity, dispersed with the fruit. Staminodes pres- ent at base of petals, small and scalelike and ap- pressed against the ovary. Ovary dark blackish brown, producing a single seed, the two abortive locules represented by ridges in fruit; style fre- quently equaling or exceeding the style branches in length, never less than J2 the length of the style branches, robust, often somewhat swollen above, easily breaking just above the ovary, not dispersed with fruit; appendages large, stalked, fusiform to clavate, with the main body smooth and usually blackish, white-papillate only at the upper rim, about % to equal the length of the style branches; style branches shallowly bifid. Fruit an indehiscent single-seeded asymmetrically fusiform achene. This subgenus comprises 23 species, all of which are endemic to the Guayana Shield of Venezuela, Amazonian Brazil, and Guyana. Most of the species are narrow endemics to the summits of one or few tepui formations. They are closely related to each other, distinguished by subtle, reticulately varying characters, and show some evidence of hybridiza- tion. This variation pattern is similar to that seen in certain eriocaul taxa of the Brazilian highlands (e.g., Paepalanthus subg. Xeractis and Aphoro- caulon, and Syngonanthus sect. Eulepis in the Serra do Espinhaco of Minas Gerais). Although the topography of the Brazilian and Venezuelan high- ands is very different, both areas represent sea- sonally wet high-altitude formations with thin, nu- trient-poor soils. Paepalanthus subg. Monosperma is the onl subgenus of Paepalanthus to have undergone ra- diation on the tepui summits of Venezuela. Mem- bers of a few other alliances occasionally reach tepui summits (P. dichotomus Klotzsch ex Koern., P. tortilis (Bong.) Koern., and P. formosus Mold.), but the main distribution of these species and their allies is elsewhere. This subgenus is best characterized by the ap- parently consistent production of single-seeded in- dehiscent fruits, in contrast to the capsular fruits produced by all other taxa known in the family. The achenes are dark-colored and thin-walled, not differing greatly in appearance from fruits of other Paepalanthus in which by chance only a single ovule has matured. At anthesis, the ovaries of subg. Monosperma are to my eye indistinguishable from those of other subgenera. It remains to be shown whether only one functional ovule is produced, or whether all three ovules are functional, with the development of more than one seed chemically inhibited. Thomas Stützel (pers. comm.) has pointed out that a consistent position of the fertile locule would provide evidence for the first possibility. Un- fortunately, because the position of the "fertile locule" cannot be readily distinguished before seed set, and seed set occurs concomitant with abscission of floral parts from the axis, I have not been able to answer this question with the material at hand. Whatever the mechanism, the character of sin- gle seed production appears to be constant through- out the subgenus. The regular abscission of the style from the apex of the ovary is further evidence that the character is genetically fixed. Such ab- scission is not observed in the capsular-fruited spe- cies, in which the seed, not the fruit, is the dispersal unit. Similarly, the sclerification of the petal bases in fruit also may be interpreted as evolutionarily correlated with indehiscent fruit production, be- cause in capsular fruits this sclerification would impede seed dispersal. Characters other than fruit type useful in dis- tinguishing the subgenus are: (1) the villous in- dumentum on exposed portions of the root; (2) the sclerification of both perianth whorls of the pistillate flower in fruit (and not merely the sepals); (3) the relative weakness of sclerification of the sepals of pistillate flowers, and their failure to recoil hygro- scopically upon drying, being left behind in the capitulum after fruit dispersal; (4) the extreme hairiness of the petals of the pistillate flowers in all but one species (P. chimantensis); (5) the very fleshy elongate style with long darkly pigmented nectaries papillate only at the upper rim (see flower of P. chimantensis, Fig. 4); and (6) the regular abscission of the style at fruiting. It is not possible at this time to make a rigorous 428 Annals of the Missouri Botanical Garden comparison with other subgenera because the ge- nus as a whole is poorly known, especially with respect to floral characters. The most complete description possible is provided in order to facilitate future comparisons with other as yet poorly known groups. Characters 1, 2, 5, and 6, above are to my knowledge rather exceptional in the genus. Characters 3 and 4 are useful in identification but are also found in other taxa. The affinities of this subgenus are not yet clear, though it bears some similarity to a complex of capsular-fruited species native to the West Indies, represented by Paepalanthus seslerioides Griseb. and P. retusus Wright. A list of accepted taxa and synonyms to be placed in the subgenus follows. apacarensis Mold. auyantepuiensis Mold. cardonae Mold. chimantensis Hensold, sp. nov. convexus Gleason mo ee ee E TUCUCUCUCUCU CU v vv vv vv m T d Q 5 = © 2. perplexans var. steyermarkii Mold. (syn.) gleasonii Mold. robustus Gleason (syn.) convexus var. strigosus Mold. (syn.) major (Mold.) Hensold, comb. nov. (see below 13. P. parvicephalus (Mold.) Hensold, comb. et stat. nov. (see below 13a. P. parvicephalus var. parvicephalus 13b. P. parvicephalus var. wurdackii (Mold.) Hensold, comb. nov. (see below 14. P. phelpsae Mold. 15. P. roraimensis Mold. Rondonanthus micropetalus Mold. (syn.) 16. P. schomburgkii Klotzsch ex Koern. P. perplexans Mold. (syn.) P. pendulus Mold. (syn.) P. macrocaulon var. venamensis Mold. (syn.) P. scopulorum Mold. P. scopulorum var. auyantepuiensis Mold. (syn. . P. squamuliferus Mold. P. fraternus var. spathulatus Mold. (syn.) . P. stegolepoides Mold. P. fraternus var. chimantensis Mold. (syn.) . P. subcaulescens N. E. Br 21. P. sulcatus Hensold P. stegolepoides var. acutalis Mold. (syn.) — “з — со — ~ ~ o 22. P. turbinatus (Gleason) Hensold, comb. nov. (see below) 23. P. venustus Mold. P. venustoides Mold. (syn.) Paepalanthus chimantensis Hensold, sp. nov. TYPE: Venezuela. Bolivar: Apácara-tepui, Chi- manta Massif, 2,200 m, 30 Jan.-1 Feb. 1983, Huber & Steyermark 6971 (holotype, MO; isotype, VEN). Figures 3F, C, 4. Herba perennis rosulata. Folia 3-4.5 cm longa, 3-8 mm lata, valde mucronata, subtus striata superne laevia nitida ubique glabra. Pedunculi ca. 13-30 cm longi basin versus dm ad apicem sericeo- кшш. Capitula 6.5- 9 mm lat t ovatae um rumque breviter acuminatae, а dis- tales ais Fructus achenium noncte Rosulate rhizomatous perennials forming small clumps. Roots pale, fibrous. Leaves chartaceous, glabrous, often bluish green, 3-4.5 cm long, 3-8 mm wide at middle, tapered, usually rounded to- ward the apex, distinctly mucronate, the upper surface lustrous, especially shiny toward the mar- gins, the lower surface striate. Inflorescences 1-6 per cohort, synchronous; peduncle sheaths equal- ing to exceeding leaves by up to 3.0 cm, glabrous, the apex acute to acuminate; peduncles ca. 13- 30 cm long, 4—5-costate, lightly tomentose at least when young with mixed retrorse and antrorse hairs, usually glabrate toward base and more densely and persistently sericeo-tomentose distally. Heads 6.5- 9 mm diam., hemispheric to subglobose at matu- rity; involucral bracts gray-brown, lustrous, ciliate in bud but early glabrate, lance-ovate, usually aris- tate or short-acuminate, the lower bracts slightly narrower than but usually as long as the upper, and usually with a slightly thickened reddish midrib. Floral bracts about equaling the flowers, obovate to linear-subspatulate, fuscous, subhyaline, ciliate at upper margin and tufted along midvein abaxially with opaque white clavate hairs. Staminate flow- ers. Pedicels ca. 1.0 mm long. Sepals 3, free, elliptic to narrowly obovate, acute, ca. 1.9-2.5 mm long, 0.7-1.2 mm wide, fuscous, pubescent as floral bracts; corolla tube borne on an andro- phore ca. 0.5-0.75 mm long, the tube itself 1.0— 1.5 mm long, narrowly infundibular, the lobes 3, narrowly triangular, acute; anthers cream to brownish, strongly exsert; pistillodes 3, simulating gynoecial appendages and nearly the length of the tube. Pistillate flowers. Sepals 3, free, broadly obovate-elliptic, acute, carinate, ca. 2.8 mm long, 1.6 mm wide, fuscous at least toward the apex, ubescent as in the staminate flowers, the midvein area thickening and expanding at maturity toward Volume 78, Number 2 Hensold 429 1991 Eriocaulaceae of Venezuela 2.6 mm SSS > کے === Д S > EN МИЛА DEP F WH „4 МИ ES ~ E E» / 5 ст FIGURE 3. Habits of Syngonanthus setifolius Hensold, Кысас fasciculoides Hensold, Р. M Hensold, S. acephalus Hensold, and S. fenestratus Hensold. A, B. S. setifolius (Liesner 3914 MO).— A. Нађи. — B. yT in bud. C-E. P. fasciculoides (Silva et al. 1756 MO).—C, D. Habit.—E. E F, G. P. chimantensis (Huber 11598 MO). —F. Habit. — С. Inflorescence. Н, І. S. acephalus (Vareschi & Magdefrau 6789 N) Нађи. —1. Fertile branch with terminal flower. J-L. S. fenestratus (Vareschi & Magdefrau 6612 VEN). —J. Habit. — K. Inflorescence. — L. Leaf. Scale bar at hae applies to all habit illustrations. 430 Annals of the Missouri Botanical Garden F FIGURE 4. Paepalanthus chimantensis Hensold sepals removed. D- F. Petal.—F. Gynoecium. Scale scale bar C = m. Pistillate flower. — le bar A-B, D-F = 3 mm; the base, sepals remaining after dispersal of fruit; petals oblong to elliptic, rounded at apex, ca. 2.7 mm long, 0.8 mm wide, glabrous or sparingly ciliate along upper margins, expanding and thickening at maturity and dispersed with the fruit. Ovary ca. 0.75 mm long, blackish, with one functional locule, the style 0.9 mm long, 3-angled and expanded above, broadest just below divergence of the ap- pendages, where it about equals the ovary in di- ameter; appendages 3, ca. 0.9 mm long, fusiform- terete, smooth and blackish except for the pale papillate apex; style branches pidan slightly but distinctly above the appendages, ca. 1.0 mm long, shallowly bifid, slightly exsert. Fruit an achene. Additional specimens examined. VENEZUELA. BOLÍVAR: Chimanta Massif, Apácara-tepui, 2,100-2,200 m, Huber & Steyermark 7035 (VEN), Huber et al. 8817 (NY), Huber et al. 8823 (NY), Steyermark 128376 (LL, MO); Chimanta Massif, C. section, 2,120 m, Steyermark & Wurdack 750 (F, NY), Steyermark & Wurdack 764 (F, NY); Chimanta Massif, Murey- (Eruoda-) tepui, 2,600 m, Huber 11598 (MO) This distinctive Chimanta Massif endemic is a member of subg. Monosperma. It has previously been confused with Paepalanthus kunhardtii old., P. schomburgkii Klotzsch ex Koern., and Leiothrix turbinata Gleason (see P. turbinatus below). It is best distinguished by its rosulate habit, glossy glabrous sharp-mucronate leaves, more or less oblong grayish involucral bracts, and the petals of the pistillate flowers which equal the sepals and are subglabrous. It has been collected in flower from January to March. This species is unrelated to the unpublished her- barium name *'Paepalanthus chimantensis," which Moldenke (1957) mentions as a synonym of Carp- totepala jenmanii (Gleason) Mold. Paepalanthus dichotomus pore ex Koern in C. Martius, Fl. Bras. 3(1): 8. 1863. Dupatya dichotoma cei ex Koern.) Kuntze, Revis. Gen. Pl. 2: 745. 1891. TYPE: Venezuela. Bolivar: Rué-imeru falls, Schom- burgk 899 (holotype, B not seen; F photo neg. #10597). e guyanensis Klotzsch ex Koern. in C. Mar- us, Fl. Bras. 3(1): 347. 1863. Dupatya guyanen- sis (Klotzsch ex Koern.) Kuntze, arb Gen. Pl. 2: 745. 1891. TYPE: Venezuela. Bolivar: Rué-imeru € Schomburgk s.n. (holotype, B not en) Syn. no ры ibe Mold., Bull. Torrey Bot. Club 948. Paepa laihia savannarum (Mold.) Mold., Phytologia 49: ios 1981. E: Guyana | ге & ечен 23280 dia steyerm markii 125. 1951. TYPE: Venezuela. Bolívar: tween Pins еп and Rio Karuai, 1,220 m, Stey- UR 59369 (holotype, NY; isotype, MO). Syn. Syngonanthus | Mold., Fieldiana, Bot. 28: . TYPE: Venezuela. Bolivar: Gran Sabana, а ween a Kavanayé én and Rio Karuai, 1,220 m, Stey- d abl n (holotype, NY; isotypes, F, MO). yn. Syngonanthus savannarum var. glabrescens Mold., Mem. New York Bot. Gard. 9: 412. 1957. Paepalanthus savannarum var. glabrescens (Mold.) Mold., Phy- tologia 49: 293. 1981. : Venezuela. Воћ Chimantá Massif, C. mu of upper falls of Rio Тиса, Steyermark Wurdack 539 (holotype, NY; isotype, F). Syn. no Paepalanthus apacarensis var. humilis Mold., Phyto- logi JT E Hus Chi- 128164 (holotype, LL aa seen). Syn. Paepalanthus dichotomus, endemic to the Gran Sabana of Venezuela and adjacent Guyana, has previously been recognized under several names, probably owing to the large variation found in pu- bescence, leaf texture, and peduncle elongation. However, these characters are found to vary con- tinuously and independently of each other, and the segregation of distinct taxa within this complex is not supportable. That some of the taxa were de- Volume 78, Number 2 1991 Hensold 431 Eriocaulaceae of Venezuela scribed in Syngonanthus probably owes to the tight coherence of the apices of the pistillate flower pet- als, which are nonetheless not truly fused. In other respects (fibrous roots, sepals tufted with trichomes, trichomes often tuberculate with ornamented in- ternal walls, sepals of the pistillate flowers hygro- scopic and thickening and recoiling to eject fruit at maturity), this species is definitely allied to Pae- palanthus. Paepalanthus fasciculoides Hensold, sp. nov. TYPE: Brazil. Рага: Marabá, Serra dos Carajas, 700-750 m, А. S. L. da Silva et al. 1756 (holotype, MO; isotype, NY). Figures 3C-E, 5. Herba annua erecta caulescens. Folia linearia ca. 5- 10 mm longa, 0.5-1 mm lata attenuata, superne hispida. Inflorescentiae multae terminaliter fasciculatae. Capitula ca. 2 mm lata, globosa. Bracteae involucrales fuscae late oblongo-obovatae, ciliatae; bracteae florales eis similes, rotundatae dense ciliatae, infra apicem ane batae Sepal a asi attenuata. Plants annuals with leafy stems up to ca. 5 cm long. Leaves dark green, ca. 5-10 mm long, 0.5- 1 mm wide, longer toward stem apex, attenuate, not ampliate at base, spreading, the upper surface hispid (with short, stiff hairs). Inflorescences nu- merous and clustered at stem apex. Peduncle sheaths with the apex attenuate-acute and often deflexed at right angles. Peduncles ca. 2-5 cm long, pilose with long spreading hairs, green, ob- scurely 3-costate. Heads 2-3 mm diam., globose. Involucral bracts dark fuscous, broadly oblong- obovate, obtuse, concave to navicular, ca. 1.5 mm long, densely ciliate at upper margin and sparsely pilose in medial area on abaxial side. Floral bracts similar in size to involucral bracts, also fus- cous, broadly spatulate and rounded or obtuse at apex, navicular, distinctly narrowed toward base, ciliate along upper margin and densely bearded in a V-shaped medial band below apex. Staminate flowers. Pedicels ca. 0.3-0.5 mm long. Sepals 3, connate at base, oblanceolate, acute, ca. 1.2 mm long, fuscous at least apically, ciliate at upper mar- gin, enough narrowed at base so that the andro- phore is visible without teasing the bases apart. Corolla inserted on a very long androphore nearly equaling the sepals, the tube itself at least partially exsert from the calyx; anthers white, ca. 0.3 mm long. Pistillate flowers. Sepals 3, connate very brief- ly at base, linear-spatulate, ca. 1.0 mm long at anthesis, cream-colored, sparsely ciliate at upper margin, thickening, elongating and recoiling hy- groscopically at fruiting. Petals hyaline, oblong, btuse to acute, ciliate along the distal margin, (а) (А-В) р (С-Е) FicURE 5. Paepalanthus Eu —— i va et al. 1756 МО). —А. Involucral br ract. epal D.. E u^ taminate flower. —G. Seed. Scale C-F E. Gynoecium. — К. A-B — ] mm; scale bar G bar А-В = 1 mm; scale bar about equaling the sepals. Gynoecium with style 2—3 times the length of the style branches; style branches simple, about twice the length of the filiform appendages. Seeds oblong, 0.6 mm long and less than half as wide, yellow-brown, the testa with prominent longitudinal striations. Additional specimens examined. VENEZUELA. BoLívAR: Cerro Cotorra (El Vigía), south side of Rio Para- gua between La Paragua a Pedro de Las Bocas, 600 m, 5 Aug. 1960, Steyermark 86888 (VEN). BRAZIL. PARÁ: Marabá, Serra dos Carajás, solo com cascalho de pedras, М.С. Silva 2920 (NY); Marabá, Serra Norte, Carajás, N1, on canga outcrop, May 1986, Marli et al. 70 (SPF). p ] ah £ я Ig i dics AN І 1 from its closest relative, Paepalanthus Sut by the dark brown obovate to spatulate involucral and floral bracts with entire upper margins, the dense subapical bearding of the floral bracts, and the narrow oblanceolate sepals. In P. fasciculatus, the bracts are buff-colored, usually conspicuously mu- cronate or apiculate, and at least the floral bracts are usually rather broadly obtrullate with a flattish upper margin. Instead of being densely bearded, they are generally sparsely pilose in one medial or two submedial bands. The sepals of the staminate flowers in P. fasciculatus are more or less broadly elliptic, the lateral margins overlapping each other and concealing the basal part of the corolla. The two species may also be ecologically distinct, the rarer P. fasciculoides being foun mountains of Amazonian only on rocky soils and canga in low Brazil and Venezuelan Guayana, while P. fascicu- 432 Annals of the Missouri Botanical Garden latus is a common species of sand savannas at low altitudes. Paepalanthus major (Mold.) Hensold, comb. et stat. nov. Paepalanthus convexus var. major Mold., Phytologia 15: 463. 1968. TYPE: Bra- zil. Amazonas: Serra Pirapucú, Rio Negro, Rio Cauaburi, Rio Maturacá, Silva & Brazdo 60926 (Da ae NY; isotypes, NY (2)). Paepalanthus fisicas var. zuloagensis Mold., Phyto- logia 23: 211. 1972. TYPE: аи Bolivar: Cer- ro de la Ne re ina, Planicie de Zuloaga, Rio Titirico, oo 103839 (holotype, NY: isotype, NY). Syn. no Paepalanthus major of Cerro de la Neblina is distinguished from P. convexus Mold. of Cerro Duida by its acaulescent to short-caulescent habit, its larger leaves with a smooth (not sulcate) upper a Ane its solitary robust inflorescences with ths far surpassing the leaves. Its clos- tric Р. gleasonii Mold., the latter od Palas are the sympa and probably P. squamuliferus Mold., of tepui summits in the Gran Sabana. It is distinguished (imperfectly) from P. glea- sonii by its usually shorter stems, more sharply cuspidate leaves, and especially by the sericeous upper leaf surface, lack of well-defined cilia, and the slightly revolute margins. In P. gleasonii, the leaves are bluntly cuspidate, densely short-ciliate, and occasionally also hirsute adaxially at the apex, but otherwise early glabrate on both surfaces, and the margins flat in drying. It differs from P. squamuliferus also by the sharply acute leaf apices, as well as by the acute to acuminate involucral bract apices. In P. squa- muliferus, the leaves are subacute to rounded, but never sharp-cuspidate, and the involucral bracts are usually obtuse to rounded. * Paepalanthus fulgidus var. zuloagensis,” also from Neblina and known only from the type, ap- pears to be a dwarf form of P. major, with smaller leaves and heads. 1 am provisionally placing it in synonymy pending the availability of further ma- terial. Paepalanthus convexus, with its numerous small inflorescences, peduncle sheaths approximately equaling the leaves, and leaves strongly sulcate- ribbed on the upper surface, is a more distant relative probably derived from Paepalanthus sul- catus Hensold (originally described from Cerro de la Neblina as P. stegolepoides var. acutalis Mold.). Paepalanthus parvicephalus (Mold.) Hensold, comb. et stat. nov. Paepalanthus convexus var. parvicephalus Mold., Phytologia 52: 19. 1982. ТҮРЕ: Venezuela. Amazonas: Cerro Marahuaca-Fhuif, 2,330-2,460 m, Steyer- mark et al. 126100 (holotype, LL not seen). Although both P. parvicephalus of Cerro Mara- huaca and P. convexus of nearby Cerro Duida share a long-stemmed, clambering habit, they differ strongly in leaf form and inflorescence arrange- ment. the veins inscribed on the upper surface, and the whole leaf strongly reflexed at maturity, while in P. parvicephalus, the leaves are obtuse, the veins n P. convexus, the leaves are acute, with obscurely if at all inscribed above, and the leaves divergent, not strongly reflexed at maturity. In addition, P. convexus is unique for its numerous small, short-pedunculate inflorescences. Although the two taxa appear closely related, P. sulcatus appears on morphological and geographic grounds to be a possible common ancestor to both, even to the extent of intergrading with P. parvicephalus at the southern end of its range. Paepalanthus parvicephalus is therefore elevated to specific sta- tus in order to avoid probable misrepresentation of relationships. Paepalanthus parvicephalus var. wurdackii (Mold.) Hensold, comb. nov. Paepalanthus perplexans var. wurdackii Mold., Mem. York Bot. Gard. 8: 98. 1953. ТҮРЕ: Vene- zuela. Amazonas: Serrania Parü, Rio Parü, Cano Asisa, 2,000 m, Cowan & Wurdack 31141 (holotype, NY). This taxon, native to the Serrania Parü, appears closely related to and probably derived from P. parvicephalus var. parvicephalus of Cerro Duida. It differs from this variety primarily in its smaller heads (4.5- 2 mm diam. vs. 6.5-8 / ). It is distinguished from P. schom- e “Mold. ві the Gran Sabana by the EE characteristics: leaves ligulate, i.e., linear and su mm in var. acute to obtuse, cuspidate, and densely Жека at base, the venation not pellucid; involucral bracts of the lower series thickened and red-brown at base. Paepalanthus schomburgkii is characterized by broader leaves straight-tapered to an acute apex, not distinctly cuspidate, often long-ciliate, but the cilia not densely crowded at the base, and the venation pellucid. The involucres of P. schom- burgkii are generally of a finer, herbaceous tex- ture, and the bracts of the lower series not distinctly thickened as above. Paepalanthus tortilis (Bong.) Koern. in C. Mar- tius, Fl. Bras. 3(1): 354. 1863. Eriocaulon Volume 78, Number 2 1991 Hensold Eriocaulaceae of Venezuela tortile Bong., Mém. Acad. Imp. Sci. St.-Pe- tersbourg, Sér. 6, Sci. Math. 1: 624, t. 49. 1831. Dupatya tortilis (Bong.) Kuntze, Ке- vis. Gen. Pl. 2: 746. 1891. TYPE: Brazil. Ba- hia: in arenosis humidis propre Ilheos, Riedel s.n. (LE not seen). Paepalanthus salticola Herzog. Feddes Repert. Spec. V ov. Regni Veg. 29: 207, pl. 121, figs. i-l. 1931. SYNTYPES: Brazil. Alto Amazonas: Tiquié, Uira Poco, Uferwald, Luetzelburg 23906 (M); Papori, Trini dade, am Fall, . Syn. nov. Paepalanthus tatei Mold., Brittonia 3: 158. 1939. TYPE: enezuela. Bolivar: Auyán-tepui, 2,200 m, Tate 1326 (holotype, . Syn. nov. Paepalanthus killipii Mold., Bull. Torrey Bot. Club 68: 67. ТУРЕ: Colombia. Santander Sur: E cor- dillera, Mesa de los Santos, Killip & Vei aid (holotype, US not seen; isotype, F). Syn. Paepalanthus maguirei Mold., Bull. Torrey Bot. Club 1948. TYPE: Surinam. Tafelberg, North Ridge Савска Savanna I, se 24670 (holo- type, NY; isotype, F). Syn. Luetzelburg 23912 (M, F photo neg. Paepalanthus tortilis owes its lengthy synon- ymy to wide variation in vegetative size and habit of flowering specimens. Ruhland’s (1903) descrip- tion of the species contains no measurements, but indicates a short or scarcely elongate stem with leaves clustered toward the apex. This is generally true of the material from eastern and Amazonian Brazil, as well as more depauperate collections from Guayana. However, much of the more recent col- lections from the Andes, as well as from tepui formations in Venezuela, are much more robust, often with long, uniformly leafy stems, but even in these areas, all gradations in size are found to aking division into discrete varieties or mm wide, and peduncles of 5 cm, to clambering long-stemmed herbs with leaves 6 cm long and 8 mm wide, and peduncles up to 25 cm. According to Stützel (pers. comm.), greenhouse-grown plants in Ulm, all propagated from a single plant of Auyán- tepui, Venezuela, tended to flower in December regardless of plant age and size, resulting in flow- ering specimens of widely varying sizes, from 2 to 25 cm tall, and with leaf sizes varying widely as in herbarium material. Inflorescence size varies by a smaller factor, with heads ranging from 2.5-6 mm diam. I am not entirely comfortable explaining all the asis of age and environmental factors. For example, although large variation in Р. tortilis on the caulescent forms are known from eastern Brazil, they seem to be much rarer there and also localized. aL 7 24 Г ( ===> Y N Y A ZEN tp, 2 222 2222 о 3 gum FIGURE З. Rondonanthus duidae (Gleason) Hensold € Giulietti (Steyermark 58319 MO).—A. Habit. — B. Inflorescence. — C, D. Involucral bracts, lower and upper. E-G. Staminate flower. — E. Flower. — F. Flower with sepals spread apart.— С. Corolla, opened to show stamens and pistillodes. H-L. Pistillate flower. — Н. Flower.—1. Flower with perianth spread apart. — J. Petal, adaxial view. — К. Sepal, abaxial view. — L. Gynoecium, shown with subtending hairs and a staminode. — M, N. Trichomes of floral bract apex. long, bifid. Floral trichomes subacute, lacking gran- ular thickening of the internal wall. mm long, the style 0.3 mm long, the appendages ca. 0.4 mm long, the stalk rather thick, the papil- lose portion pale orange, the style branches di- verging slightly above the appendages, 2.5 mm Phenology. Collected in flower in February. Volume 78, Number 2 1991 Hensold & Giulietti 453 Revision of Rondonanthus Habitat and distribution. Known only from the type; no habitat information given. This species appears more similar to К. aco- panensis of the Chimantá Massif than to “Р. jauensis" (here included in R. duidae), with which it occurs sympatrically at Aprada-tepui, and of which it was originally described as a variety. It is allied with К. acopanensis by the acuminate, col- orless apices of the involucral bracts and perianth, and by the more densely woolly leaf bases which are broadest at the base. Rondonanthus duidae (Gleason) Hensold & Giulietti, comb. nov. Paepalanthus duidae Gleason, Bull. Torrey Bot. Club 58: 329. 1931. TYPE: Venezuela. Amazonas: summit of Mount Рида, 5,500-6,000 ft., Tate 456 (holotype, NY). Figure 3. Пра phelpsae var. pilosus Mold., Mem. New York . Gard. 9: 282. 1957. TYPE: Venezuela. о. Cerro de la Neblina, W escarpment 1,850-1,900 m, Pii cedi et al. 37308 (holotype, NY; isotype, F). S жш ан var. "oiridis Mold., в 26. 1971. TYPE: Venezuela. Amazonas: Сегг га la ү blina, Planicie de Zuloaga, 2,300 m, CHEN н 103840 (holotype, NY; na LL). Syn. Мике jauensis Mold., Mem. New York Вог. Gard. 23: 850, fig. 5. 1972. TYPE: Venezuela. Bolivar: Cerro Jaua, cumbre de la porción centro-oeste, 1, EE Rs 100 m, Steyermark 97906 (holotype, LL). Syn РЕ ен duidae var. p Mold., Phytologia 983. TYPE: Venezuela. Amazonas: Cerro Матаһчаса, 2,750 m, Tillett et al. 752- 340 (ho- lotype, NY). Syn. nov Plants rosulate or short-stemmed cespitose pe- rennials, rarely branching above the base. Main 5-1.0 mm diam., sparingly branched, cream cm long, occa- roots 0. to pale orange. Stems up to 10 sionally up to 40 cm long but then sterile, ca. 2- 5 mm diam., densely leafy to the base, woolly with light brown hairs. Leaves spirally arranged, 1.5- 15 cm long, 0.3-1.5 mm wide; the base slightly dilated just below the blade and then usually con- stricted proximally, pubescent with long matted feltlike hairs as on the stem; the blades linear to d pene usually sharp-cuspidate, subterete and rigid o flat and chartaceous, glabrous to densely pu- E with retrorse malpighian hairs, usually with the midvein on the lower surface prominently thick- ened in contrast to the two narrow lateral veins (occasionally leaves merely acute and midvein sim- ilar in thickness to laterals). Inflorescences about 1—4 per age class, the old ones often persistent on stem. Peduncle sheaths 2-6.5 cm long, twisted, pubescent like the leaves, the apex acuminate, scarious. Peduncles (6-)1 1-39 cm long, 6-costate, pubescent with retrorse malpighian hairs to nearly glabrous, usually with long filamentous hairs fring- ing the base of the capitulum. Capitula 6-14 mm diam., hemispheric at maturity. Involucral bracts in 4—6 series, the outer ovate to narrowly triangular or lanceolate, subacute, often recurved, 0.5-4.0 mm long, 0.4-1.2 i bracts truly oblanceolate though appearing lance- linear to lanceolate in their exposed portions, ta- pered to an acute or narrowly rounded apex, erect or somewhat spreading, 2.5-8.5 mm long, 0.6– 1.5 mm wide, often surpassing the head, sometimes by as much as 3-4 mm; variable in color, ranging from uniform deep brown with a dark reddish mid- vein to blackened only along the distal margins of the inner bracts with the outer bracts and medial areas of the inner bracts a strongly contrasting pale gold; glabrous or often ciliate, and then often hir- sute abaxially (rarely also adaxially) at apex. Floral bracts commonly equaling or slightly surpassed by the flowers, linear to linear-oblanceolate, acute to subacute, blackened distally, bearded with long hairs abaxially at apex. Staminate flowers. Pedicels 0.5— 1.5 mm long. Sepals broadly spatulate, rounded to acute, and sometimes the upper margin lacerate, 1.5-2.7 mm long, 0.6–1.2 mm wide, uniformly blackened distally, white below, rarely dark brown distally, grading into a lighter brown (Cerro Jaua and Sarisarinama), bearded abaxially with long erect trichomes. Corolla connate into a 3-lobed tube, or the petals very rarely free from each other, the androphore lacking, the tube 2.0-3.0 mm long, the lobes rounded and sometimes (Aprada-tepui) the sinuses rather deep, almost equaling the tube in length, the upper margin and abaxial lobe apices densely pilose, involute after anthesis. Stamens with filaments adnate to corolla at base, usually exsert; anthers ca. 0.2-0.3 mm long. Pistillate flowers. Pedicels 0.4-1.5 mm long. Sepals broadly elliptic long, mm wide at base, the inner to ovate-elliptic, concave, acute, 2.0-4.2 mm 1 mm wide, cream-colored or sometimes pigmented as in staminate flowers toward the apex though usually much less so, ciliate for most of their length and bearded at the apex as in the staminate flowers, thickening at base at maturity. Petals mostly free, rarely briefly connate distally, oblong to subspatulate, obtuse to acute, 2.0-3.5 mm long, 0.5-1.5 mm wide, densely ciliate at upper margin and frequently also pilose abaxially at apex or in submarginal bands. Staminodes linear, 0.2-2.0 mm long. Ovary ringed by hairs at base, 0.5-0.8 mm long, the style 0.2-0.6 mm long, the appendages са. 0.6-1.0 mm long, the stalk usually 454 Annals of the Missouri Botanical Garden long and narrow, up to 25 the total length, the papillose portion white, sometimes collapsing after anthesis and appearing nearly filiform, the style branches diverging slightly above the appendages, 1.5-2.5 mm long, bifid, though in many cases the two segments cohering. Floral trichomes subacute to rounded, lacking granular thickenings of the internal wa Phenology. Collected in flower mostly from September to February, rarely in March and June. Habitat and distribution. Locally abundant on open plateaus with shallow, rocky, wet soils, at 1,600-2,800 m. Venezuela. Amazonas: summits of Cerros Marahuaca, Duida, and Neblina; Bolivar: summits of Cerros Jaua, Sarisarinama, and Aprada- tepui. Additional specimens examined. VENEZUELA. AMAZONAS: summit of Cerro Duida, Steyermark 58319 (F, MO, NY, VEN), Steyermark 58361 (F, NY); Cerro Duida, summit of Peak 7, Tate 625 (МУ); Cerro Duida, Caño Sapo, Tate 691 (NY); Cerro Marahuaca, SE section, Maguire et al. 65637 (MO, NY), Maguire et al. 65638 (MO, NY, VEN), Maguire et al. 65639 (MO, 2. = Y 5 ыр: я 3 & ~ = % ~ в ~ ~ ~ A R с © 2 < = e O 4 e | o = دو‎ = > = = © о E extreme SE section, Steyermark 129109(LL, MO); Cerro Marahuaca, extreme NE section, Steyermark 12 (MO, VEN), Steyermark & Delascio 129247 (LL); Le Marahuaca, central part of the SE Meseta, Steyermark 129475 (LL, MO); Cerro Marahuaca-huha, Steyermark 125932 (LL, VEN); Cerro Marahuaca-Fhuif, Steyer- mark et al. 126099 (LL); Cerro Marahuaca- Atuhua- atua, Maguire et al. 36930 (NY), Maguire 37123 (NY, VEN), Bec pd et al. 42261 (Е, NY, VEN), Maguire et al. 4 NY, УЕМ); NW plateau, ИА 16028 (МО), od 16046 (MO); Planicie de , Rio Titirico, Steyer- mark 103878 (NY), nione 103920 (NY). BOLÍVAR: Macizo de Sarisariñama, Fernandez 4753 (MO, PORT); Aprada-tepui, Steyermark 115885 (F, MO, VEN); S sec- tion, 30 km E of Urimán, Huber 9539 (MO). The amount of variation in Rondonanthus dui- dae is very high, with some of the variation geo- graphically distributed, and some probably envi- ronmental, such as leaf length and width, which can vary considerably within a single population. Involucral bract color varies geographically. The plants of Cerro Marahuaca, Aprada-tepui, and some of Cerro de la Neblina have uniformly black in- volucres, while plants of Duida and some of Neblina have highly contrasting gold and black or almost entirely gold bracts. Plants of Jaua and Sarisari- nama have more rounded bract apices, and lack characteristic black color to bracts and perianth, having a dull brown color instead. Plants of Aprada- tepui are distinguished by very long black invo- lucral bracts and by Syngonanthus-like fusion of the petals of the pistillate flower corolla. This species is closely related to Rondonanthus roraimae, from which it can only be satisfactorily distinguished by reference to flower morphology. In general, the leaves of R. duidae also differ from those of R. roraimae by the prominent midvein and aristate tip, but this character is lacking in some collections (Aprada-tepui and some of Nebli- Rondonanthus duidae is also habitally similar to 5. duidae Mold., with which it is sympatric. These species are so similar in general aspect that two collections of R. duidae have been described as new varieties of 5. phelpsae (now placed in S. duidae). At Neblina some plants look intermediate with 5. duidae, with rather short, unevenly black- ened involucral bracts, with the sepal apices sharply acute-lacerate and bearded below the apex, but not at the apex (as in S. duidae), the corolla of the staminate flower developing rather irregularly, and the gynoecium intermediate with a rather long com- mon style and style branches apparently fine and simple, not bifid. This is particularly true of the type of S. phelpsae var. pilosus Mold. It is possible that hybridization is occurring at Neblina, although 5S. duidae has never been collected south of Cerro Aracamuni. Another possible hybrid requiring more study is the type of Paepalanthus fraternus var. radiatus Mold. (Phytologia 49: 383-386. 1981) from Cerro Marahuaca. The leaves of this plant are dull green and rounded at the apex rather like the sympatric species Paepalanthus parvicephalus (Mold.) Hen- sold, but have a cartilaginous texture, and are rosulate. The peduncles are of variable lengths, and the lanceolate radiating involucral bracts and soft heads which flatten upon pressing suggest R. duidae. The flowers are intermediate between the two taxa, with some developmental irregularities, particularly in the gynoecium. Rondonanthus flabelliformis (Mold.) Hensold & Giulietti, comb. nov. Wurdackia flabelli- formis Mold., Mem. New York Bot. Gard. 9: 413. 1957. TYPE: Venezuela. Bolivar, Chi- manta Massif, Torono-tepui, 2,165-2,180 m, Steyermark & Wurdack 671 (holotype, NY; isotypes, F, K). Figure 4. Plants caulescent perennials. Main roots ca. 1– 1.5 mm diam., sparingly branched, pale orange- brown. Stems up to 16 cm long, densely leafy to the base, woolly with grayish hairs. Leaves disti- chous, 9-10 cm long, 5-8 mm wide; the bases Volume 78, Number 2 1991 Hensold & Giulietti 455 Revision of Rondonanthus slightly dilated linear or broader below, densely woolly with matted whitish hairs; the blades ligulate- falcate, obtuse to rounded, rigid, shiny, glabrous, ca. 30-veined, the veins of equal size, the margins pale and slightly revolute. Inflorescences up to 4 per age class. Peduncle sheaths ca. 6 cm long, not twisted, glabrous, the apex acute, green or some- what scarious. Peduncles 13-16 cm long, 7-8- costate, pubescent with retrorse malpighian hairs between costae, and long filamentous hairs a the base of the capitulum. Capitula 15-19 m diam., hemispheric to subglobose at maturity. In. volucral bracts in 5 series, the outer triangular, acute, ca. 4 mm long, ca. 1.0 mm wide at base, the inner bracts spatulate, rounded and minutely apiculate, са. 6 mm long, 1.5 mm wide, about equaling the head; the outer bracts gold-brown, coriaceous, the inner black, chartaceous, except for the gold-brown midvein; the outer bracts gla- brous, the inner bearded abaxially at the apex, glabrate with age. Floral bracts about equaling the flowers, spatulate, obtuse, the distal half blackened, bearded abaxially at apex. Staminate flowers. Ped- icels not measured. Sepals oblong-obovate, rounded and apiculate, ca. 2.7-2.9 mm long, ca. 0.7-0. ] d distally, bearded abaxially with long, erect trichomes. Corolla connate into a 3-lobed tube at first, the segments later separating at least at the base, the androphore lacking; the tube 3.3 mm long, the lobes rounded, the upper margin and abaxial lobe apices ciliate-pilose, involute after an- thesis. Stamens with filaments adnate to corolla, included; anthers ca. 0.3-0.35 mm long. Bisexual flowers. Pedicels ca. 2 mm long. Sepals ovate, concave, obtuse, ca. 4 mm long, 1.3 mm wide, cream-colored, bearded as in the staminate flowers, thickening at base at maturity. Petals connate into a tube except at base, the segments spatulate, + rounded at apex, ca. 3.5 mm long, 1.0-1.2 mm wide, densely long-ciliate at upper margin. Stamens mm wide, bl present, opposite the petals, similar to those of the staminate flowers, but with the filaments slightly shorter. Ovary with no hairs at the base, ca. 0.5 mm long, the style 0-0.3 mm long, the a ca. 0.5 mm long, the stalk thick, the papillose portion brownish, the style branches diverging slightly above the appendages, ca. 2 mm long, conspicuously bifid. Floral trichomes subacute to rounded, lacking granular thickenings of the in- ternal wall. endages Phenology. Collected in flower in February. Habitat and distribution. Known only from the type, reported as locally frequent in a zanjon (large crevice), in moist places among rocks. Rondonanthus flabelliformis is clearly closely related to R. acopanensis, differing in the persis- tence of the leaves along the stem, in the very large size of the leaves and capitula, and in the occurrence of stamens in the morphologically pis- tillate flowers. It would be of great interest to de- termine the relationship between these two species. It is possible that R. flabelliformis has been derived from R. acopanensis by macromutation, a sus- picion supported by the occurrence in the latter species of discontinuous size classes. Alternatively, it may be postulated as a primitive relict, with the persistent leaves and bisexual flowers representing an ancestral condition. Rondonanthus roraimae (Oliver) Herzog, Feddes Repert. Spec. Nov. Regni Veg. 29: 210. 1931. Paepalanthus roraimae Oliver, Trans. Linn. Soc. London, Bot. 2: 286. 1887. TYPE: Guyana. Mount Roraima, 15 Dec. 1884, Thurn 294 (holotype, K). Figure 5. Plants rosulate or caulescent cespitose peren- nials. Main roots ca. 0.3-0.7 mm diam., sparingly branched, cream to pale orange. Stems up to 10 cm long, 2-3 mm diam., leafy to the base, woolly with light brown hairs. Leaves spirally arranged, 1.5-4 cm long, 0.5-1.3 mm wide; the base slightly dilated just below the blade and then usually con- stricted below, pubescent with long matted feltlike hairs as on the stem; the blades linear to acicular, the apex rounded or bluntly cuspidate, subterete and rigid to flat and chartaceous, glabrous or spar- ingly pubescent with retrorse malpighian hairs, es- pecially abaxially, and then usually glabrate, 3-veined, the veins of equal size. Inflorescences 1(-5) per age class. Peduncle sheaths 3-4(-6) ст long, twisted, pubescent like leaves, the apex usu- ally erose-toothed, scarious, eventually lacerate. Peduncles 5-28 cm long, 6—10-costate, pubescent with retrorse malpighian hairs to nearly glabrate, usually with long filamentous hairs fringing the base of the capitulum. Capitula 8-13 mm diam., hemi- spheric at maturity. Involucral bracts in 4—5 series, the outer ovate-triangular to lanceolate, subacute, often recurved, 2.3-3.7 mm long, 0.8-1.2 mm wide at base, the inner bracts lanceolate to oblan- ceolate and then appearing lance-linear to lance- olate in their exposed portions, tapered to an acute or subacute apex, erect or somewhat spreading, 4.5- 7.0 mm long, 1.0-1.5 mm wide, surpassing the head by 1-2.5 mm; the outer bracts mostly castaneous, or sometimes fuscous along the mar- gins, progressively darker above, the innermost bracts commonly deep fuscous with only the mid- 456 Annals of the Missouri Botanical Garden / \ (OMM | НИД у nih AN [/ Uy, MU |; Volume 78, Number 2 Hensold & Giulietti 457 1991 Revision of Rondonanthus FIGURE 5. Rondonanthus roraimae (Oliver) Herzog (Liesner 23239 MO, unless stated otherwise). — A. Нађи. — В. Inflorescence. — C. Involucral bract, adaxial view. D, E. Staminate flower. — D. Flower with bract.— E. Corolla, nodes. — H. Gynoecium, with staminodes and subtending hairs. I-K. Large-flowered morph (Аутага 2480 H= I, J. Petals with staminodes.—K. Gynoecium, with staminodes and hairs.—L, M. Trichomes of floral bract apex. <— FIGURE 4. MM m (Mold.) Hensold & Giulietti واچ‎ & Wurdack 671 МУ).— Habit. B-E. Bisexual flower. — B. Flower. — C. Flower with sepals spread apart. — D. Corolla opened D show а stamens and e Ж со: F-I. Staminate flower. TF. Flower (pedicel broken). — С. Sepal, adaxial view. — H. Corolla, opened to show stamens and pistillodes. — I. Corolla, entire. — Ј. Trichome of floral bract apex. 458 Annals of the Missouri Botanical Garden q t ubacute, blackened distally, bearded abaxially with long hairs. Sta- minate flowers. Pedicels 1.0-2.5 mm long. Sepals broadly oblanceolate to obovate, rounded to acute, 1.5-2.5 mm long, 0.5-1.3 mm wide, uniformly blackened distally, white below, bearded abaxially with long erect trichomes. Corolla connate at first into a 3-lobed tube, but the segments separating before anthesis, the androphore lacking; the petal segments 1.9-2.5 mm long, 0.9-1.5 mm wide, obovate, rounded to subacute, glabrous. Stamens with filaments adnate to corolla at base, included; anthers ca. 0.3-0.4 mm long. Pistillate flowers. Pedicels 0.9-1.3 mm long. Sepals obovate to spat- ulate, concave, acute to obtuse, 1.5-2.1 mm long, 0.6-1.8 mm wide, of the same color and pubes- cence as in the staminate flowers, thickening at base at maturity. Petals free, + obovate, rounded to apiculate, 1.5-2.1 mm long, 0.6-1.2 mm wide, sparingly ciliate along lateral margins below apex, and frequently also pilose abaxially in submarginal bands. Staminodes linear, 0.5-1.2 mm long. Ovary ringed by hairs at base, 0.5-0.7 mm long, the style ca. 0.3 mm long, the appendages ca. 0.4-0.5 mm long, the stalk narrow, the papillose portion white, the style branches diverging slightly above the ap- pendages, ca. 1.0-1.2 mm long, bifid though the two segments often failing to separate. Floral tri- chomes rounded to clavate, lacking granular thick- enings of the internal wall, though sometimes lightly flecked. Phenology. Collected in flower mostly June to September, occasionally April and March. Habitat and distribution. In open rocky or boggy, often partially inundated areas over thin soils or rock, 2,300-2,800 m. Venezuela. Bolivar: summits of Mount Roraima, Kukenán-tepui, Yu- ruani-tepui, and Ilà-tepui. Guyana: summit of Mount Roraima. Additional specimens examined. VENEZUELA. BOLÍVAR: Ilú-tepui, NW sector (Tramén-tepui), Huber 10065 (NY); Yuruani-tepui, Huber 11813 (MO); Ku- kenan-tepui, Huber & Alarcon 10526 (MO), Liesner » NY, VEN), Stey- (F, NY), Steyermark 112608 (VEN). it of Mount Roraima, Abbensetto 8 (К), Connell & Quelch 660 (K, NY), Lies- 9 (MO), Liesner 23316 Luetzelburg 21605 (NY), Quelch & Connell 102 (K, NY), Quelch Rondonanthus roraimae, as the type of the genus, was originally described as dioecious and having free petals in both pistillate and staminate flowers. Though the trend to dioecy is pronounced, monoecious plants are occasionally found. In ad- dition, petal fusion in staminate flowers is found to occur early i lor t, although i p anthesis flowers, which are easier to dissect and interpret, the petals are invariably free. e species, endemic to the Roraima formation of eastern Venezuela, is very similar to R. duidae, found to the west. Rondonanthus roraimae can be distinguished by the glabrous, ephemerally fused corolla segments of the staminate flowers, and also usually by leaf characters (see R. duidae). Other characters distinguishing R. roraimae are the in- cluded stamens with fleshy filaments and larger anthers, as well as the much shorter (but not nar- rower) petals, style branches, and appendages in the pistillate flowers. These characters, together with the more swollen floral trichome apices, all ] er, post- d, 1988). In addition, R. roraimae blooms earlier in the year than R. duidae. EXCLUDED SPECIES Rondonanthus micropetalus Mold., Fieldiana, Bot. 51. TYPE: Venezuela. Bolivar: Mount Roraima, SW -facing side, Steyermark 28777 (holotype, NY; isotype, F). This species is an abnormal form of Paepalan- thus roraimensis, with dwarf peduncles and ab- normal floral development. In the material ob- served, the staminate corollas were abortive and underdeveloped, represented only by small lobes of tissue on the floral axis, which were apparently interpreted as free petals. LITERATURE CITED CiULIETTI, А. M. 1984. Estudos Taxonómicos ` rad nero Leiothrix Ruhl. (Eriocaulaceae). Tese de Livr . HENSOLD. : А Ж genera Comanthera and Carptotepala ма gonanthus (Eriocaulaceae). Ann. Missouri Bot. Lar“: 78: 460-464. GLEASON, Н . A. 1931. Botanical results of the Tyler ida expedition. Bull. Torrey Bot. Club 5 8: 277- : E HensoLD, N. 1988. Morphology and Syne : aepalanthus subg. Xeractis (Eriocaulaceae). y Bot. Revisionary studies in the Erie ceae of Venezuela. Ann. Missouri Bot. Gard. 424-440, plants. VI. Phytologia 3: 35-4 pe 19 Volume 78, Number 2 1991 Hensold & Giulietti 459 Revision of Rondonanthus Ss Е Herzoc, T. 1931. 9 E pr nera of Eriocaulaceae in x Southeastern United States. J. Arnold Arbor. 131-1 MOLDENKE, H. N. 48. Notes on new and noteworthy 46. 5l. Eriocaulaceae. /n: J. St teyermark, эч tanical exploration in iie Fieldiana, Bot. 2 114-129. 1957. E In: B. Maguire et al., Botany of the Chimantá Massif. I. Gran Sabana, Venezuela. Mem. New York Bot. Gard. 9: 408-414 RUHLAND, W. 1903. Eriocaulaceae. /n: А. Engler, Das ami gioi IV. 30. (Heft 13). Wilhelm Engel- = uA “ genus of Епосашасеае. Lbs Herb. 117: 38. тутак. T. 1984. Botas und Infloreszenzmorpholo- зена Untersuchungen zur pose ik der Eri lac е botanicae 71. Cram Die раније prod die Gattung zo (Eriocaulaceae). Flora 177: 335-344. tersuchungen zur Wurzelanatomie der оя Flora 180: 223-239. SYNONYMIZATION OF THE GENERA COMANTHERA AND SYNGO TH (ERIOCAULACEAE)! Ana Maria Giulietti? and Nancy Hensold* ABSTRACT Hos systematic posee of the monotypic Guayanan genera Carptotepala Mold. and Comanthera Lyman B Smit d. Carpto Lo e Mold. is here placed in Syngonanthus sect. Thysanocephalus near _ xeranthemoides, under the e S. jenmanii (Gleason) Ciulietti & Hensold. Comanthera kegeliana (Koern.) Mold. is ids returned to а sect. Eulepis, as S. kegelianus (Koern.) Ruhl. Synonymies аге given for each specie The Eriocaulaceae, as currently treated, com- prise 13 genera and about 1,200 species (Kral, 1989). The family has a pantropical distribution but is much more diverse and species-rich in South America, where two disjunct centers of diversity occur. These are the Cadeia do Езршћасо in Bahia and Minas Gerais, Brazil, and the Guayana High- land, especially in Venezuela. Although the largest concentration of species is found in the Brazilian center, the largest number of genera (10) is re- ported from Guayana. Generic delimitations in the Eriocaulaceae are the subject of some controversy. The bulk of the species are contained in the large genera Eriocau- lon L., Paepalanthus Kunth, Syngonanthus Ruhl., and Leiothrix Ruhl.; the remaining nine genera together contain only about 35 species. Of the five genera described since Ruhland's classic treatment (1903) (Carptotepala Mold., Comanthera Lyman B. Smith, Moldenkeanthus Morat, Rondonanthus Herzog, and Wurdackia Mold.), none were de- scribed with more than two species, and all but Moldenkeanthus of the Madagascar are endemic to the Guayana Highland. (Moldenkeanthus has recently been placed by Stützel (1987) in Pae- palanthus. The systematic positions of Comanthera and Carptotepala have never been adequately inves- tigated in publication, although some workers have informally expressed doubts about their validity (Moldenke, 1968, citing Górts-van Rijn; Giulietti, 1984; Stützel, 1985, in a footnote). Like all genera in the family, they are distinguished on the basis of floral characters, which often prove difficult to interpret in the extremely small flowers. We shall demonstrate here that these monotypic genera are indeed very closely allied with certain sections of Syngonanthus and that they lack any specialized characters to adequately distinguish them as genera in their own right. I. CARPTOTEPALA Moldenke (1951) described the genus Carpto- tepala based on C. insolita from Venezuela. Later he transferred Paepalanthus jenmanii Gleason, a species of British Guiana, to Carptotepala pad denke, 1957), and in 1968 he recognized С. solita as a synonym of С. jenmanii. He distin. guished the genus by the following characteristics: (1) free sepals and petals in both pistillate and staminate flowers; (2) prolongation of the floral 'We thank the curators of F, K, МО, NY, and VEN for the loan of specimens. T was a visiting research scientist at Kew Botanic Gardens, senior author he work was initiated while the e unded by the British Council Fellowship Program, and completed while a visiting research scientist а the Missouri Botanical Garden Postgraduate Fellowship Program, supported by the Jessie Smith Noyes Foundatio ? Departamento de Бшш. Instituto de Biociéncias, Universidade de Sao Paulo, C.P. 11.461-05499, Sao Paulo, SP, Brazil. * Department of n Field Museum of Natural History, Roosevelt Road at Lakeshore Drive, Chicago, Illinois 60605-2496, U.S.A ANN. Missouni Вот. GARD. 78: 460-464. 1991. Volume 78, Number 2 Giulietti & Hensold 461 1991 Synonymization in Syngonanthus FIGURE 1. Flowers of Syngonanthus jenmanii (Gleason) Giulietti & Hensold and Syngonanthus "i dn (Koern.) Ruhl. A-D. 5. jenmanii (Jenman 7486 К). A, B. Staminate flower. — А. ole flower. — B. with perianth parts ойы open to show stamens and pistillodes. C, D. Pistillate flower.—C. Whole flower at indc the style tips barely protruding from the involute petal apices. —D. Flower with perianth parts spread open to show 1 ) ] ] Д К). 1 r.—E. me ower. ; gynoecium. E- ianus (Lan Lindeman 2984 , F. Staminate flower. — E | —F Flower with perianth dissected open to show stamens pis bs stillodes. c- A нне ари flower. — Flower with sepals spread apart, the petals remaining connivent arou tyle.—I. Flower after e “of seeds, the sepals deciduous, the ovary locules open.—J. Same, pen the sei е амау. receptacle above the calyx in the staminate flowers; — Thysanocephalus (Koern.) Ruhl. and Eulepis and (3) presence of caudate appendages at the пр (Bong.) Ruhl., and may be compared as follows. of the petals in the pistillate flowers. In Carptotepala jenmanii, the staminate flow- In fact, the flowers of C. jenmanii are extremely ers are found at the periphery of the capitulum similar to those of Syngonanthus, especially sects. and are much more abundant than the pistillate, 462 Annals of the Missouri Botanical Garden from which they are easily distinguished by their very long pedicels. These pedicels elevate the sta- minate flowers above the involucre, as is charac- teristic of sect. Thysanocephalus. Also as in sect. Thysanocephalus, the sepals and petals of the staminate flowers are glabrous, concave, and similar to each other in size and texture (Fig. 1A, B). Contrary to the original de- scriptions of C. insolita and P. jenmanii, the sepals are united at the base, and the petals are briefly united or free. Moreover, we did not find elongation of the floral receptacle above the calyx (i.e., an ““androphore””), as is known in some Syngonanthus (not sects. Eulepis or Thysanocephalus), and in most Paepalanthus and Eriocaulon. The descrip- tion of an androphore in this species may have been due to misinterpretation of the fused basal portion of the corolla. The pistillate flowers particularly recall sect. Thysanocephalus. In these the petals are densely pilose in the middle, and far surpass the sepals, with the glabrous apices curved around and en- closing the style branches. The tips of the style branches protrude slightly beyond the petals, and this apparently was misinterpreted by Moldenke as a “caudate appendage” (Fig. 1C, D). The sepals and petals appear to be free, though the hairiness of the medial areas of the petals (where fusion normally occurs in Syngonanthus) makes it difficult to determine whether the petals are merely cohering about the style or truly fused. Shallow or ephemeral corolla fusion in both sta- minate and pistillate flowers is recorded sporadi- cally from sect. Thysanocephalus. Staminate flow- ers with petals nearly free are described for Syngonanthus squarrosus Ruhland (1903) and for S. chrysolepis Silveira (1928). Pistillate flowers with petals free or united medially are described for S. flexuosus Silveira, and petals connate but eventually separating are found in S. glaber Sil- veira (192 Moreover, complete loss of petal fusion in sta- minate and pistillate flower corollas, though rare, is found in Leiothrix fluitans (Mart.) Ruhl. and Rondonanthus (Paepalanthus) capillaceus (Klotzsch ex Koern.) Hensold & Giulietti (1991), two species whose closest relatives have the petals fused in at least the staminate flowers. These two species arian od also share the rheophytic habit with C. jenman Thus, no reliable character remains for isolating this species from Syngonanthus sect. Thysano- cephalus. 'The relationship is further supported by the characteristic campanulate to urceolate invo- lucre found in Carptotepala and sect. Thysano- cephalus. Inclusion of Carptotepala jenmanii within Syngonanthus is therefore proposed as fol- lows. Syngonanthus jenmanii (Gleason) Giulietti 4 ensold, comb. nov. Paepalanthus jenmanit Gleason, Bull. Torrey Bot. Club 56: 14. 1929. Carptotepala jenmanii (Gleason) Mold., Mem. New York Bot. Gard. 9: 278. 1957. TYPE: Guyana. Potaro River, Kaieteur Savanna, Jenman 1032 (holotype, K). ii ee е Mold., Fieldiana, Bot. 28: 114. 1951. : Venezuela. Bolivar: Rio Karuai, SW base " Puri. -tepui, 1,220 m, Steyermark 60703 (holotype, NY; isotype, F, GH). Syngonanthus xeranthemoides var. grahamae Mold., Phytologia 8: 395. 1962. TYPE: Guyana. Potaro River, Mahdia goldfield, V. Graham 154 (holotype, K; isotype, LL). Syn. nov. Additi ronal specimens examined. VENEZUELA. Gran Sabana, Salto Cama, Davidse 4865 (МО); Палерш, Maguire 33516A E bá YANA: Sandwith 1258 (K), Schomburgk s.n. (К); Upper Mazaruni River, Kamarang River, base of Utschi Falls, “Tillett & Tillett 45744 (MO); Upper Mazaruni, Kako River, Tillett & Tillett 45486 MO); Potaro River, Abraham 330 (K); Potaro River, Cobanatuk Falls, Jenman 7486 (K). The closest affinities of Syngonanthus jenmanii appear to lie with the sympatric species S. xeran- themoides (C. Martius) Ruhl., which usually occurs on relatively drier sites. Syngonanthus jenmanii is distinguished within its section by its occurrence along wet riverbanks, in waterfall spray, and even occasionally with the leaves fully submerged. II. COMANTHERA Smith (1937) described the genus Comanthera to accommodate C. linderi Lyman B. Smith. He characterized the new genus by the presence of three distinct types of flowers: (1) the staminate, these pedicellate, with perianth reduced to a ves- tigial tubular sheath at the base of the flower, and with only one stamen, comose at the apex with long hairs; (2) the pistillate, these sessile, with sepals and petals free; and (3) the sterile, these also ped- icellate, with the sepals and petals fre Moldenke (1966) transferred the species Syn- gonanthus kegelianus (Koern.) Ruhl. to Coman- thera, without discussion or further descriptive in- formation. Two years later, he recognized C. linderi Lyman B. Smith as a synonym of C. kegeliana (Koern.) Mold. (1968). He also included in the Volume 78, Number 2 1991 Giulietti & Hensold Synonymization in Syngonanthus 463 synonymy of this species 5. akurimensis Mold. d S. akurimensis var. amazonicus Mo In the same article, he quoted a letter from A. Gorts-van Rijn, who had studied type material of C. kegeliana, and questioned the validity of the genus. She compared the appearance of the sta- minate flowers in Smith's illustration (1937) **with the overripe female flowers, where the fruits have come out and the perianth segments had partly fallen off." Smith's reply in letter was also included, indicating that he was not convinced by her ob- servations. Because the perianth of eriocaulaceous flowers is chaffy and retains its form and color long after anthesis, differences between flowers at different stages may easily be веке аз taxonomic differences. Therefore, flowers rom capitula of C. kegeliana in various stages of mat- uration. Only two types of flowers were observed, staminate and pistillate. At anthesis, these possess the following organization: Staminate flower. Long-pedicellate; sepals three, free, glabrous; petals three, connate briefly at base, glabrous, somewhat surpassing the sepals; stamens three, the filaments filiform, included, anthers di- thecous; pistillodes very small (Fig. 1E, F). Pistillate flower. Short-pedicellate; sepals similar to those of staminate flowers; petals somewhat sur- passing the sepals, connate and long-pilose medi- ally, but separating very easily, free at apex and base, the apical portion somewhat broader than the medial portion and with the lateral margins invo- lute, enclosing the styles; gynoecium with a three- locular ovary, and three elongate simple style branches slightly surpassing the petals (Fig. 1G, H The old flowers persist for some time in the head and present the following organization: Staminate flower. Same as above, but the an- thers lost and the filaments delicate and filiform. Pistillate flower. After fertilization, the pedicels elongate considerably to equal those of the sta- minate flowers. A loculicidal capsule is formed with a membranaceous pericarp. After the liberation of the seed the two valves of each locule recurve toward the floral axis, completely enclosing the narrow bases of the petals. At the same time, the sepals fall, leaving only the pedicel and a smal ring of tissue representing the scar where the sepals were inserted. It is evident that the sterile, long-pedicellate flower portrayed in Smith’s (1937) description and figures (pl. 2, fig. 59) is in reality a staminate flower that has lost its anthers after anthesis. The pistillate flower described and illustrated is in all respects like that of the material examined, except that the petals are depicted as free instead of connate medially. However, in dissection of pis- tillate flowers of this species it is quite easy to separate the pilose membranous petals without awareness of doing so. Smith described the flowers as sessile, but as noted, they are very short-pedi- cellate before fertilization, and may appear sessile in comparison to later stages of development. Moreover, truly sessile flowers are virtually un- known in Eriocaulaceae. The staminate flower described and illustrated as possessing a single stamen comose at the apex is identical in appearance to the female flower after liberation of the seeds, as Gorts-van Rijn suggested. The abscission scar left by the sepals corresponds to the reduced perianth described by Smith; the single dithecous anther corresponds to the two vis- ible ovary locules with the pericarp completely recurved toward the center of the flower (the third locule is of course not evident as the flower is viewed from the side); and the comose apex of the “ап- ther" corresponds to the narrow pilose segments of the petals, which project above the ovary, their bases hidden by the recurved locule walls. The floral structure of Comanthera kegeliana and the variations found in progression from bud to fruiting are characteristic for the majority of species of the genus Syngonanthus, especially those of sect. Eulepis, in which S. kegelianus was first described. It is therefore proposed that the genus Coman- thera be synonymized with the genus Syngonan- thus and its single species be placed in sect. Eu- lepis. Syngonanthus kegelianus (Koern.) Ruhl., Pflanzenr. IV. 30: 273. 1903. Paepalanthus kegelianus Koern. in C. Martius, Fl. Bras. 33(1): 438. 1863. Dupatya kegeliana (Koern.) Kuntze, Revis. gen. pl. 2: 746. 1891. Comanthera kegeliana (Koern.) Mold., Phy- tologia 13: 218. 1966. TYPE: Surinam, Marie- paston, Kegel 1473 (holotype, GOET not seen, photo, K; isotype fragment, B not seen). i Ar B. Smith, Contr. Gray Herb.‏ ا pl. 1937. TYPE: Guyana, Rockstone,‏ ,38 : pu 40 а СН not seen; isotype, NY).‏ dur v pe ку Mold., Phytologia 2: 371.‏ Venezuela. Bo T Cerro urima,‏ : 1947 n 3234 (holotype, US not seen; isotype, NY).‏ v.‏ 464 Annals of the Missouri Botanical Garden ZUELA. BOLÍ. 1705 Additional materias examined. VENE : alo E Uairén, Lasser veo , Atkinson Field, Maas & Westra 3571 (K), Mckee "10681 we Robertson & ‘usin 268 (NY); Rockstone, Gleason 652 (NY). SuRI- M. Jodensavanne, Lanjouw & Lindeman 2984 (К). Syngonanthus kegelianus is readily identified even from descriptions and photos because of the distinct pubescence on the leaves and peduncle sheaths of this species, i.e., stiff, erect hairs on leaf undersurface and appressed white malpighian scales on upper leaf surface and on peduncle sheaths. It is the only member of Syngonanthus sect. Eulepis found in the Guayana Highlands, and is fairly un- usual within its section due to the small, hyaline, relatively unspecialized upper involucral bracts. It has very close affinities with Syngonanthus ulei Ruhl. of Amazonian Brazil, which is only readi- ly distinguished by the long spreading hairs on leaves and sheaths and the lack of white scalelike malpighian hairs. It may represent only a variety of S. kegelianus. Syngonanthus akurimensis var. amazonicus Mold. (Phytologia 3: 42. 1948), which Moldenke considered synonymous with S. kegelianus, is probably identical with S. ulei, as is S. kuhlmannii f. viviparus Mold. (Phytologia 37: 97. 1977). These new additions to the generic synonymy of Syngonanthus are summarized as follows: Syngonanthus Ruhl. in Urban, Symb. Antill. 1: 1900. TYPE: Syngonanthus umbellatus (Lam.) Ruhl. (Eriocaulon umbellatum Lam.). Carptotepala Mold., Fieldiana, Bot. 28: 114. 1951. TYPE: Carptotepala insolita Mold. d rao ан B. Smith, Contr. Gray Herb. 117: . TYPE: Comanthera linderi Lyman Smith, LITERATURE CITED GIULIETTI, A. M. 1984. Estudos Taxonómicos no Gé- nero Leiothrix Ruhland (Eriocaulaceae). Tese de Li- à A. M. А redefinition of the genus Rondonanthus Herzog (Ег- iocaulaceae). Ann. Missouri Bot. Gard. 78: 441 459. Kral, К. 1989 . The genera of Eriocaulaceae in the Southeastern United States. 31-142. J. Arnold Arbor. 70 MoLDENKE, Н. N. 1951 Eriocaulaceae. In: J. Steye ma E pus exploration in Venezuela. Fieldiana, 14-129. Bot. 57. Eriocaulaceae. /n: B. Maguire & J. J. Wurdack, The Botany of the Guayana Highland — Part P nous . New York Bot. Gard. 9: 278-283. Notes on new and noteworthy plants. XLIII. p 13: 218. Additional notes on the Eriocaulaceae. XIV. Bb. 17: 372-395. RuHLAND, W. 1903. в In: А. Епрјег, Das Pflanzenreich, IV. 30. (Heft 13). Wilhem-Engel- . DA. 1928. Floralia Montium, Volume I. Imprensa Oficial, Belo Horizonte (Brazil). A new genus of Eriocaulaceae. 38. Die systematische Stellung der Gat- tung Wardackin (Eriocaulaceae). Flora 177: 335 1987. On — morphological and systematic position of the genus Moldenkeanthus (Eriocaula- ae). Pl. Syst. Evol. 156: 133-141. STUDIES IN NEOTROPICAL PALEOBOTANY. X. THE PLIOCENE COMMUNITIES OF PANAMA —COMPOSITION, NUMERICAL REPRESENTATIONS, AND PALEOCOMMUNITY PALEOENVIRONMENTAL RECONSTRUCTIONS! Alan Graham? ABSTRACT The middle Pliocene Gatun Formation of central Panama consists of carbonaceous sandstones and siltstones the first appearance of substantial quantities of grass pollen (maximum 7.5%), tropical dry forest, (4) increased physiographic diversity, (5) be tter representation of lower montane forests, and (6) possibly, the first indication of developing wetter Atlantic and drier Pacific provinces in southern Central America. An assemblage of 110 fossil pollen and spores, including 13 with no previously known fossil record and 27 unknowns, has been recovered from the middle Pliocene Gatun Formation of central Pan- ama. The systematics of the Gatun flora has been presented earlier (Graham, 1990, 1991а, b), and the geology is summarized in Graham et al. (1985). This paper concludes the study with presentation of the composition, numerical representations, and paleocommunity / paleoenvironmental reconstruc- tions. MATERIALS AND METHODS Extraction, slide preparation, and photographic techniques are discussed in Graham (1985). Ter- minology for vegetation types follows Holdridge (1947; Holdridge et al., 1971), and distributions and community assignments are based on Croat (1978) and D'Arcy (1987). Depths along the core are given in feet, following the original log data. COMPOSITION AND NUMERICAL REPRESENTATIONS The composition, distribution, and percentage for each palynomorph at the nine levels along the Gatun core are given in Table 1. The distribution of the 15 most abundant microfossils (276 or grea er from at least one level) is presented in Figure 1. Sediments in the lower portions of the profile (257.0-253.0') consist of fine- to medium-grained sandstone with interspersed lenses of coalified lig- nite. The sandstone suggests near-shore deposition as rivers and surface waters slow and drop heavier mineral particles first upon entering the ocean. The environmental setting was paralic (marine coastal), tropical to warm-temperate (presence of Rhizoph- ora), estuarine (bay), and low-energy where au- tochthonous organic material accumulates rather than being dispersed by coastal currents. Percent- ages of Rhizophora pollen are relatively low (3.0%, 2.5%, 5.0%; Fig. 1). The lower portion of the core ' Research supported by NSF grants GB-5671, DEB-8007312, DEB-82055926, BSR-8500850, and BSR-8819771. ? Department of Biological Sciences, Kent State University, Kent, Ohio 44242, U.S.A. ANN. MISSOURI Вот. GARD. 78: 465-475. 1991. 466 Annals of the Missouri Botanical Garden TABLE 1. Identification and numerical representation of fossil palynomorphs from the Gatun Formation, Panama. Percentages are based on counts of 200 from slide 1 of each sample. (—) indicates specimen not encountered in tabulation. Depth (feet) Species 257.0 255.5 253.0 250.0 223.5 223.0 222.5 178.0 162.0 Pyrrophyta — = — — 7.5 7.0 7.5 = = Lycopodiaceae Lycopodium Selaginellaceae Selaginella = == == — 1.5 = — — — Cyatheaceae Alsophila 3.5 — 3.5 + 6.0 8.0 8.0 3.0 9.0 Cnemidaria La — 0.5 == — = = 1.5 = Cyathea type 1 — — 1.5 — 1.5 1.5 — 1.5 1.5 type 2 = = = == m — — — — Ophioglossaceae Ophioglossum — — — — — — — — — Polypodiaceae Grammitis — = = — 1.0 = x: = — Pteridaceae Ceratopteris == == — == 1.0 1.5 — 1.5 1.5 Реп Trilete fern spore type 1 0.5 0.5 — — = == — — — type 2 — — — — = — Tet = = Dryopteridaceae Ctenitis type 1 14.0 8.5 8.0 1.5 9.0 14.0 13.0 14.0 26.0 type 2 7.0 13.5 9.5 3.0 11.5 8.5 4.5 24.0 16.0 Podocarpaceae Podocarpus 1.0 1.0 — — = = es _ Gramineae 7.5 6.5 4.5 — — 1.0 — — == Volume 78, Number 2 1991 Graham Pliocene Communities of Panama 467 TABLE 1. Continued. Species Depth (feet) 257.0 255.5 253.0 250.0 223.5 223.0 222.5 178.0 162.0 Palmae type 1 type 2 Аппопасеае Сутђоре ит Aquifoliaceae Ilex Bombacaceae cf. Aguiaria Bernoullia Ceiba Pseudobombax Burseraceae Bursera Cabombaceae Cabomba Combretaceae cf. Bucida Combretum/ Terminalia Compositae Mutisieae type Cucurbitaceae cf. Cionosicys Ericaceae type 1 type 2 Euphorbiaceae Alchornea cf. Glycydendrum cf. Jatropha cf. Stillingia Fagaceae Quercus Juglandaceae Alfaroa/ Oreomunnea Leguminosae Mimosoideae Acacia Caesalpinioideae Crudia Papilionoideae Erythrina Lentibulariaceae Utricularia 4.5 20.0 12.0 6.0 4.5 3.5 1.0 12.0 3.5 18.0 20.0 13.0 1.5 6.0 468 Annals of th Missouri Bond Garden TABLE l. Continued. Species Depth (feet) 251.0 255.5 253.0 250.0 223.5 223.0 222.5 178.0 162.0 Loranthaceae type 1 type 2 Malpighiaceae type 1 type 2 Malvaceae Hampea/ Hibiscus Melastomataceae Meliaceae Cedrela Guarea Myrtaceae Eugenia/ Myrcia Nymphaeaceae (see Cabomba- eae) Onagraceae Hauya Rhizophoraceae Rhizophora Rubiaceae Chomelia type Cosmibuena ype Posoqueria Rutaceae Casimiroa Sapindaceae Allophylus Paullinia Sapotaceae cf. Bumelia Symplocaceae Symplocos Tiliaceae Mortoniodendron Verbenaceae Aegiphila Petrea Моште 78, Митбег 2 Graham 469 1991 Pliocene Communities of Panama TABLE 1. Continued. Depth (feet) Species 257.0 255.5 253.0 250.0 223.5 223.0 222.5 178.0 162.0 Unknown 1 — 0.5 3.5 = — 1.5 — 1.0 — 2 — = 2.5 Е = == — — — 3 == ~ Tee р = = == 4 2.5 5.0 1.5 — — — ы — = 5 ب‎ — — — дні, ЕРЕ аі -— 6 1.5 — — — 2.0 — — — = 7 1.5 — 3.0 — 0.5 — — — 8 — 0.5 — — — — — 9 PEN = -— == = БЕР == == = 10 = - e = = — A: = = 11 == — == = — — — — 12 — 0.5 = = — — — — — 13 — — 1.0 — — — — 14 — — 2.5 — — — — — — 15 — 0.5 — — — — — — 16 — 0.5 -— — — — — 17 — — — 0.5 — — 2.0 — — 18 — — 0.5 — — — — 19 — — — — — — — — — 20 — — E E — Е — — — 2] = = = — — 0.5 — — — 22 — — 0.5 = 1.0 0.5 — — — 23 — — — — — — — 24 — 0.5 2.5 — 5.0 4.5 3.5 — 4.0 25 — — 3.0 == — — — — 26 — — 3.0 — — — — — — 27 2.0 1.5 2.0 0.5 1.5 — — 1.0 — Other unknowns 12.5 9.0 10.0 0.5 12.0 10.5 9.0 8.0 10.5 is characterized by high percentages of monolete fern spores (combined values 21%, 22%, 17.5%) and palm pollen (19.5%, 3276; Table 1), and rel- atively high percentages of grass pollen (7.5%, 6.5%, 4.5%). The figures for grass pollen are sig- nificant when compared to values for older floras from the same region. Maximum percentages for grass pollen in the middle(?) to upper Eocene Ga- tuncillo and the lower Miocene Culebra, Cucara- cha, and La Boca floras of central Panama, and the Uscari sequence flora of Costa Rica (Graham, 1985, 1988a, b, 1989a, 19872) are 0%, 1%, 0%, less than 176, and 0%, respectively. The appear- ance of comparatively large amounts of grass pollen in the middle Pliocene Gatun flora is one of the principal differences between it and older Tertiary floras from the same region. At level 250' a major change in microfossil percentages occurs, with Rhizophora pollen reaching 82% of the total flora. This compares to previous values of 3.0%, 2.5%, and 5% at levels 257.0, 255.5, and 253.0’, respectively (Table 1, Fig. 1). All other major pollen and spore types decrease in value (Fig. 1), or disappear completely. Only 16 palynomorph types (14%) were encoun- tered at level 250’, compared to 48 types (44%) for the preceeding three levels. Monolete fern spores decrease from combined values of 21%, 22%, and 17.5% to 4.5%, and palm pollen decrease from 19.5%, 32%, and 10.5% in the three lowest levels to 4.5% at level 250.0". A change in sediment type, and in the microfossil assemblage, occurs at levels 223.5-222.5'. The sediment particle size changes from a fine- to me- dium-grained sandstone with lenses of lignite, to a finer siltstone with organic matter dispersed through the sediment. Smaller-grained particles remain sus- ended longer upon entering the ocean and arc characteristic of sedimentation further from the shore where currents disperse the microscopic-size organic comp . The principal change in the ed iss at reels 223.5-222.5' is the first 470 Annals of the Missouri Botanical Garden А игорпога Combretum/ Ра!тае Terminalia Ah E Cecil va = Faramea 1 d 4 2104 = >> NY У e 4 NE $—1——9 $47 4 4 2304 + 5 z : М L + = € | | E +++ + L1 LI E + Kam + + bd 1 LAA EE ДЕ! 1 1 1 11101 1 FIGURE 1. Pliocene Gatun Formation, Republic of Panama. SLT = siltstone, 55 = sandstone. Each mark on the bottom scale represents 5%. appearance of the shallow water marine dinoflagel- lates Operculodinium and Spiniferites (Graham, 1989a, p. 51, figs. 2-4). Diversity of palyno- morphs decreases from 44% at levels 257.0- 253.0’, to 32% at levels 223.5-222.5' (Fig. 1), and many are represented by thick-walled fern spores (Alsophila, Pteris type 4, monolete fern spore types 1, 2) and pollen (Лех) that can with- stand long-distance transport, or by wind-dispersed pollen produced in large quantities that can readily be blown into depositional basins (Palmae type 1; Table 1). The upper two levels of the Gatun core (178, 162’) consist of medium-grained sandstone with dispersed organic material. Preservation of the pal- ynomorphs is only fair, diversity is low (21%), and almost all palynomorphs are represented by thick- walled forms (fern spores, Пех, Malpighiaceae) or forms belonging to anemophilous species (Palmae type 1). Collectively, the lithology and composition of the Gatun flora reveals a dynamic landscape with fre- Vertical distribution of the most common palynomorphs at nine levels along a core through the middle CSS = carbonaceous sandstone (dark lines indicate lenses of lignite), quent and significant changes in the physical en- vironment. Sedimentation began in a near-shore setting close to the source area for the palyno- morphs (257.0-250.0'). The site was subsequently inundated by marine waters, reflected first by an increase in Rhizophora pollen (250.0'), and then by the presence of Operculodinium and Spinifer- ites (223.5-222.5'). The sediments and microfossil composition of the uppermost levels (178.0, 162.0") suggest uplift of the adjacent land with steeper gradients carrying coarser-grained sandstone out over the site. Lignitic lenses could not accumulate in this high-energy environment, and plant micro- fossils are represented by thick-walled, resistant spores and pollen (drift specimens), and by wind- dispersed pollen. Similar cycles of deposition are repeated throughout the Gatun core, and are also evident in many other Tertiary formations in this tectonically active region. The genesis of these numerous, short-term cycles was likely relatively rapid uplift and subsidence of the land, superim- posed on a general eustatic rise in sea level, which Volume 78, Number 2 Graham 471 Pliocene Communities of Panama continued until about the mid-Pliocene, when levels dropped, reflecting the early stages of extensive polar glaciation (see illustration in Graham, 1987b, fig. 4). Intense tectonic activity is clearly evident in the structural geology of central Panama (Gra- ham et al., 1 ‚ fig. 2, p. 49, Culebra Formation; Lowrie et al., 1982, fig. 3) The changing position of the source area veg- etation relative to the Gatun core site is reflected in the diversity values noted earlier. These values (viz., percent of the total flora of 1 10 palynomorphs represented at each of the three groups of levels; Fig. 1) show a trend through the profile. The figures are very general, and a number of factors influence the quantitative and qualitative representation of plant microfossils in a depositional basin (e.g., dis- tance of the basin from the source area, sediment type, rate of sedimentation, amount of sediment in each sample, number of samples for each level). Nonetheless, a decrease in diversity and an increase in thick-walled spores and pollen, and anemophilous pollen types, is evident from the bottom to the top of the Gatun core. Levels 257.0, 255.5, and 253.0' (representing 4' of sediment), with the site of de- position closest to the source area vegetation, con- tain 4476 of the Gatun palynomorph types. Levels 223.5, 223.0, and 222.5' (representing 1' of sed- iment), and further removed from the source area, contain 32% of the palynomorph types. Levels 178.0 and 162.0' (16' of sediment) contain 21%, with poorer preservation and evidence of differ- ential preservation (thick-walled pollen/spore types). Levels 257.0-253.0' are most representative of the vegetation growing in the region during Gatun time. The 110 palynomorphs recognized for the Gatun flora are considerably more than has been recorded from the preceeding early Miocene floras in south- ern Central America: Culebra—55 taxa; Cucara- cha—21 taxa; La Boca— 54 taxa; Uscari—44 taxa. One of the principal differences between the Gatun assemblage and older Tertiary floras from the same region is the marked increase in the kinds of palynomorphs present, reflecting greater diver- sity in the developing southern Central American vegetation. PALEOCOMMUNITIES Some of the 110 taxa recognized from the Gatun microfossil flora indicate the presence of a specific docarpus), especially when common associates of that community are also present (e.g., Alfaroa/ Oreomunnea, Quercus), even though these may extend into adjacent communities along transition zones, or occur as edaphically or physiographically controlled local outliers in differe Other taxa range widely through several commu- nity types (e.g., Lycopodium). ost pollen and spore types in the Gatun as- semblage can be identified only to genus, and many of the larger modern genera contain species that grow under widely different ecological conditions. For example, Quercus oleoides Cham. c is a lowland tropical species, while Q. aaata Muller, Q. brenesii Trel., Q. copeyensis Muller, and others grow at higher altitudes in temperate habitats. Oth- er pollen and spores can be identified only to genus or above (e.g., monolete fern spore types 1-5, Gramineae, Palmae, Епсасеае, Malpighiaceae, Melastomataceae). Unknown types 1-27 cou represent taxa from any number of communities, while the pollen of some important tropical families (e.g., Lauraceae) does not preserve in the fossil nt communities. ecord. For these and other reasons, paleocommunity reconstructions based on palynological data alone are tentative, and provide only a general picture of the paleovegetation. The probability of some version of a modern community being represented in a fossil assemblage is enhanced, however, in descending order of importance, by the presence of (1) elements restricted to a particular commu- nity, (2) elements typical of a community, even though they may range into other vegetation types, (3) a combination of several genera that, even though individually widespread, are collectively typical of a given community, and (4) a very large number of genera that occur in a particular com- munity even though none are restricted to, or even typical of, that particular vegetation (e.g., numer- ical abundance). Once a tentative model of paleo- community types has been established, it can be assessed in terms of (1) paleophysiographic and paleoclimatic reconstructions based on other lines of evidence (e.g., paleotemperature curves, fossil vertebrate faunas), (2) plate tectonic data (viz., land-sea relationships), (3) biogeographic data on the timing and direction of migrations, and geo- graphic affinities, derived from studies of the past (fossil occurrences) and present distribution of oth- er taxa, (4) consistency with results from studies of other fossil floras of the same age from the same region, and (5) consistency with trends and patterns E! established from studies of older and younger biotas and from the above collective evidence. Taxa identified from the Gatun flora are ar- 472 Annals of the Missouri Botanical Garden TABLE 2. Distribution of taxa identified from the mid- dle Pliocene Gatun Formation among comparable com- munity types in Panama. Placement is according to prin- cipal or most common occurrence(s), and most range through more than one community. Shallow Water Marine Community Operculodinium, Spiniferites m Floating/Submerged Fresh Water Aquatic Community Ceratopteris, Utricularia, Cabomba Lowland Fresh Water Marsh Community and Fringing Vegetation Monolete fern spore types 1-2 (likely representing Blechnaceae, Polypodiaceae, Pteridaceae), Palmae e 1, Bucida type 1, Tropical Wet Forest (31 taxa) Lycopodium, Selaginella, Cnemidaria, Ctenitis, Pter- is, Aegiphila, cf. Aguiaria, Alchornea, Allophy- lus, Bernoullia, Bucida, Bursera, Chomelia, Com- retum, Crudia, Cupania, Cymbopetalum, Erythrina, Eugenia/ Myrcia, Faramea, сї. Gly- cydendrum, Guarea, Hampea/ Hibiscus, Morto- niodendron, Paullinia, Petrea, Posoqueria, Pseu- dobombax, cf. Stillingia, Symplocos, Terminalia Tropical Moist Forest (38 taxa L ium, Selaginella, Alsophila, Cnemidaria, Bt Cyathea, Ophioglossum, Pteris, Aegiphila, cf. pesca: ; Cionosicys, Combretum/ Terminalia, Cosmibuena, Crudia, Cupania, Cymbopetalum, Erythrina, Eu- enia/ Myrcia, Faram Premontane Wet Forest (27 taxa) Lycopodium, Selaginella, Alsophila, Cnemidaria, Ctenitis, Cyathea, Pteris, Aegiphila, Alchornea, Allophylus, Bernoullia, Cedrela, Combretum/ Ter- m inalia, Cosmibuena, Crudia, Cupania, Cymbo- > Ophioglossum, Alfaroa/Oreomunnea, Aegiphila, Al- ophylus, Bursera, Ceiba, cf. Cionosicys, Com- bretum/Terminalia, Eugenia, Faramea, Hampea, plocos Premontane Rainforest (11 taxa) Alchornea, Bernoullia, Cosmibuena, Cymbopetalum, Erythrina, Hampea, Paullinia, Petrea, Serjania, cf. Stillingia, Symplocos Lower Montane Moist Forest (12 taxa Lycopodium, Selaginella, Alsophila, Cyathea, Gram- mitis, Ophioglossum, Pteris, Podocarpus, Acacia, Alfaroa/ Oreomunnea, Ilex, Quercus TABLE 2. Continued. Tropical Dry Forest (11 taxa) Gramineae, Acacia, Allophylus, Bursera, Cedrela, Ceiba, Combretum, cf. Jatropha, Posoqueria, ania Lower Montane Wet Forest (6 taxa) Lycopodium, Selaginella, Alsophila, Ctenitis, Cy- athea, Pteris Premontane Dry Forest (6 taxa) Gramineae, Acacia, Bursera, Combreatum, Eugenia, cf. Jatropha Montane Moist Forest (4 taxa) Grammitis, Podocarpus, Alfaroa, Quercus Montane Wet Forest (3 taxa Montane Rainforest — none ranged into paleocommunities in Table 2, based on the occurrence of their presumed modern an- alogs. The Gatun Formation is primarily a shallow water marine deposit (Graham, 199 1a, fig. 1), with all terrestrial pollen-bearing material coming from subsurface well cores. A shallow water marine com- munity, and a coastal, warm-temperate to tropical depositional environment, are indicated by the di- noflagellates Operculodinium and Spiniferites. Fringing the coast was a mangrove swamp hi- zophora), with the pollen form Hampea/ Hibiscus likely representing Hibiscus tiliaceous L./H. so- rorius L.f., tropical beach plants. A floating/sub- merged fresh water aquatic community, beyond the influence of marine or brackish waters, 15 SUg- gested by microfossils of Ceratopteris, Cabomba, and Utricularia. Also in the lowlands and not m- fluenced by marine waters were marsh or wet habitat communities reflected by non-tree fern spores (maximum 26%, Table 1) of the Blechna- ceae, Polypodiaceae, Pteridaceae type. Some palm ollen (maximum 20%) probably represents specie? associated with or bordering the marsh community. Bucida is a small tree or shrub occurring in strand thickets and wet forests at elevations from sea level to about 1,000 m. | It is difficult to distinguish between the r lowland forest types on the basis of palynologic evidence because the generic composition of these forests is often similar, and transition zones таў be broad and gradual. The tropical moist forest s represented by 38 taxa that can occur m da formation, the tropical wet forest by 31 taxa 7 premontane wet forest by 27 taxa, and the en montane moist forest by 21 taxa (Table 2). It Volume 78, Number 2 91 Graham Pliocene Communities of Panama 473 likely that some form of all of these communities were represented in the middle Pliocene vegetation of central Panama. The premontane rainforest (11 taxa), lower montane moist forest (12 taxa), and tropical dry forest (11 taxa) are also represented byan r of taxa and/or by elements charac- teristic of these vegetation types. For example, the association of Podocarpus, Alfaroa/ Oreomunea, Пех, and Quercus is most typical of the lower montane moist forest. The presence of pollen of ll taxa that can or typically do occur in the tropical dry forest compares with the following figures from other Tertiary floras in the region: Gatuncillo (middle(?) to lower Eocene, Pana- ma)—4 (+ two widespread) Culebra (lower Miocene, Panama)—7 Cucaracha (lower Miocene, Panama)—0 a Boca (lower Miocene, Panama)—0 Uscari (lower Miocene, Costa Rica)—0 Grass pollen, virtually absent from these older flo- ràs, increases to values of 7.5%, 6.5%, and 4.5% in the Gatun flora (Table 1). Many of the Gramin- tae were probably associated with the tropical dry forest. One of the principal differences between the older Tertiary floras of southern Central America, and the Gatun flora, is the initial appearance of a more fully developed and widespread tropical dry forest in the middle Pliocene. In turn, this suggests that relatively extensive, sustained, and adequate resources for moderate- to larger-sized herds of browsing and grazing mammals, migrating in substantial kinds and numbers between North and South America near the end of the Pliocene (ca. 24 myr), began to appear in the middle Pliocene. Prior to that time, smaller numbers of browsers and grazers were probably supported by shifting, temporal seral зе Ж не: : | гезропзе lo intense tectonic and volcanic activity, docu- mented for southern Central America during the Tertiary (Graham, 1988b, pp. 1476-1478). PALEOPHYSIOGRAPHY Estimates of maximum altitudes in proto-Central merica must be general because of the wide al- tudinal r ange of many genera, and because mon- tane species are underrepresented in sediments ac- Cumulating in lowland he... £2 RM ae g p : Muller (1959) has shown that pollen from upland communities is commonly transported into coastal ments by rivers, and Germeraad et al. (1968) lieve that outwashing by surface runoff is an "прогіапі means of pollen transport in the tropics. accounts, in part, for the better representation of both entomophilous pollen types (e.g., Cymbo- petalum, cf. Cionosicys, Acacia, Erythrina, Hampea/ Hibiscus, Hauya) and pollen from mon- tane species in tropical assemblages than in tem- perate microfossil floras. Thick-walled spores and pollen (e.g., Gramminitis) and wind-dispersed types (Podocarpus) from inland and upland habitats are readily incorporated into coastal deposits. Thus, montane communities may be underrepresented, but not a priori absent from estuarine sediments. The impression from the Gatun assemblage is that low to moderate montane habitats are better represented than in Eocene and lower Miocene floras from the same region. Gramminitis, Podo- carpus, and Alfaroa/Oreomunnea are all low- to mid-elevation montane plants in southern Central America, and they occur together for the first time in the Gatun flora. Also, the number of genera that can occur in the lower montane moist forest is reater in the Gatun flora (12 genera) than in the Gatuncillo (3) or Culebra (6) floras. America (Graham, 1987b, 1988c, d) indicate that all components could be accom- modated within elevations ranging from sea level to about 1,400 m (Graham, 1989b). The Gatun flora suggests slightly higher elevations, to an es- timated 1, m, within pollen rain deposition of the Gatun site. Current maximum elevations are 3,820 m (Cerro Chirripó) and 3,432 m (Volcán Irazü) in Costa Rica, and 3,475 m (Volcán Baru) in Panama, all to the west of the Gatun locality. The middle Pliocene Gatun flora provides the first evidence for low to moderate montane habitats in southern Central America. * ч 1 d increased phys- iographic diversity accounts, in part, for the larger number of genera in the Gatun flora (110) than in other Tertiary floras from southern Central Amer- ica (e.g., 55 taxa in the Culebra flora; 21 in the Cucaracha). These, in turn, sort into a greater montane moist forest; tropical dry forest). In par- ticular, the tropical dry forest may reflect the be- ginning of a differentiation into drier (Pacific) and wetter (Atlantic) zones in southern Central Amer- ica. PALEOCLIMATE Climates of the middle Pliocene in southern Cen- tral America must have been similar to those of the present because all taxa identified from the Gatun Formation, with the exception of Crudia 474 Annals of the Missouri Botanical Garden and possibly some cf. identifications, grow in the area today. An approximation of these climates may be gained by considering those on nearby Barro Colorado Island. There the annual rainfall ranges from 190 to 360 cm (76 to 143 inches), and averages 275 cm (107.3 inches). “Тһе climate is markedly seasonal, with a sharp dry season usu- ally starting in mid-December and continuing until about the beginning of May. During the dry season, only 18-26 cm (7-10 inches) of rain fall" (Croat, 1978, p. 3). The development of a better defined an xtensive tropical dry forest in the middle Pliocene Gatun flora, compared to older Tertiary floras in the region, may mark the beginning of greater seasonality in the region as a consequence of increasing elevations and closure of the marine portal between North and South America. "The atmospheric temperature may vary from as low as 16.5?C (61.7°F) to as high as 35.5°C (95.9°F), with the lowest temperatures being re- corded within the forest during the rainy season and the highest at the Laboratory Clearing in the dry season. With rare exceptions the temperature ranges between 21° and 32°С (70° and 90°F) throughout the year, and the average ambient tem- perature in the Laboratory Clearing is 27°C (77°F). The seasonal variation in monthly averages is just 2.2°С. In the dry season the range of the monthly averages of diurnal temperature is 9-11%C, and during the rainy season it is 8-9?C. Thus the di- urnal temperature variation is greater than the variation between seasons. Temperature on the forest floor is especially constant, the range of the average diurnal temperature being about 6.19C" (Croat, 1978, The Tertiary microfossil floras known for north- ern Latin America have been placed on the global paleotemperature curve of Savin (1977; Savin & Douglas, 1985; Savin et al., 1975) by Graham (1987b, fig. 3). The middle Pliocene Gatun flora (previously considered late Miocene in age, Gra- middle Miocene, Stewart & Stewart, 1980) was deposited near the end of a temperature drop that began in the middle Miocene, and just before a slight rise in the late Pliocene that pre- ceded the final fall toward glacial conditions in the Pleistocene. These temperature fluctuations, af- ; Wolfe, 1978), also affected biotas as far south as 18°N latitude (viz., the Pliocene Paraje Solo flora of Veracruz, Mexico; Graham, 1976). However, at ФМ latitude the Tertiary floras of southern Central America (Uscari of Costa Rica; Gatuncillo, Culebra, Cucaracha, and La Boca of Panama) were deposited under a different set of physiographic conditions. To the north (eastern Mexico) the setting was continental within an area of considerable physiographic relief. To the south the Costa Rican and Panamanian Tertiary floras) the setting was insular with less physiographic relief (Dengo, 1973; Raven & Axelrod, 1974, fig. 4). As noted by Graham (1989b, c), the existence of these two physiographic provinces within northern Latin America influenced the impact of Tertiary paleotemperature changes on the terrestrial biota. In the Paraje Solo flora of Veracruz, Mexico, the impact was augmented by communities already stressed along altitudinal gradients. The Panamani- an Tertiary floras were buffered against the ex- tremes of temperature fluctuations by communities existing under lower-altitude, insular conditions in a region of considerably less physiographic diver- sity. Thus, the dynamics of Tertiary paleocom- munities in southern Central America were driven not so much by temperature fluctuations, as with the tropical flora of southern Mexico, or by changes in sea level (and, in consequence, water tables) as with the biota of the Amazon Basin, but by changes in physiography. These changes, first evident among presently known fossil floras in southern Central America beginning in the middle Pliocene, were relatively rapid uplift and subsidence of land sur- faces, increasing altitudes producing the beginnings of a rain shadow effect, and closure of the isthmian portal which affected oceanic circulation patterns that resulted in an intensification of wet and drier seasons. The biological effects were relatively rapid changes in the number, distribution, and rate of speciation of mangrove and associated lowland veg- etation elements (increasing diversity of the biota), and the development of a more complex and diverse set of communities, including, especially, the lower to middle montane and tropical dry forests. — LITERATURE CITED Скоат, T. B. 1978. Flora of Barro Sonet Island. Stanford Univ. Press, Stanford, Califor D’Arcy, W. С. 1 . Flora of Panama Checklist and Index. Monogr. Syst. Bot. Missouri Bot. Gard., Vol- ume 8 DENGO, С. 1973. Estructura geológica, historia tectóni- ca y morfologia de America Central. 2nd edition. Centro iw de Ayuda Tecnica, e i GERMERAAD, J. H., C. A. HOPPING & J. MULL Palynology of Tertiary sediments from iid areas. Rev. Palaeobot. Palynol. 6: 189-348. GRAHAM, А. 1976. Studies in neotropical paleobotany. . The Miocene communities of Veracruz, Mexico. Ann. Missouri Bot. Gard. 63: 787-842. 1985. Studies in neotropical paleobotany. IV. Volume 78, Number 2 1991 Graham 475 Pliocene Communities of Panama The pue ңе r of Panama. Ann. Missouri Bot. Gard. 72: 504-534 TN Miocene communities and paleoen- vironments vd southern Costa Rica. Amer. J. Bot. 74: и “ч “Tropical American Tertiary floras and a Би exico, Costa Rica, and Pana- 531. J. Bot. 74: 1519-1 1988a. Studies | in neotropical к ү. The lower Miocene communities of Pan —the Culebra Formation. Апп. Missouri Bot. Card. 75: 1440-1466 1988b. Studies in n “о VI. The lower Miocene com ies of Pan the Cucaracha Formation. feti Missouri Bot. Gard. 75: 1467-1479. : . Some aspects of Tertiary vegetational history in the Gulf/Caribbean region. Trans. 11th ا‎ Geol. Conf. (Barbados, 1986): 3:1-3:18. iocene floras and biogeog- raphy s Central America. J. Geol. Soc. Jamaica 25: 15. ERE 989a. ue in neotropical e VIL. The lower Miocene communities of Panam the La Boca Formation. “Ann. Missouri Bot. Card. 76: 50-66 989b. Late таа рашт and уер- etational zonation in M an tral America. Acta. Bot. Neerl. 38: 417-424. 89c. Paleofloristic and paleoclimatic changes in the Tertiary of northern Latin America. . Ра- laeobot. Palynol. 60: 283-293 1990. New angiosperm records from the Ca- ribbean Tertiary. Amer. J. Bot. 77: 897-910. 1991a. Studies in neotropical paleobotany. ҮШ. The Pliocene communities of Panama — intro- А udies in neotropical paleobotany. IX. The Pliocene communities of Panama — angio- sperms (dicots). Ann. Missouri Bot. Gard. 78: 201- 22 ‚ В. Н. STEWART & J. L. STEWART. 1985. Stud- ies in neotropical paleobotany. III. The Tertiary com- munities of Panama — geology of e NE -bearing deposits. Ann. Missouri Bot. Gard. 8 HOLDRIDGE, Г. В. 19 Sale ce M a world plant formations from simple climatic data. Science 105: 68. , W. C. GRENKE, W. Н. HATHEWAY, T. LIANG & J. A. Tost. 1971. Forest Environments in Trop- ical Life Zones: A Pilot Study. Pergamon Press, New York. HUBBARD, В. М. & M. C. BOULTER. 1983. Reconstruc- tion of Palaeogene climate from palynological evi- dence. Nature 301: 147-150. boundary of the Cocos don during the Miocene. Marine Geology 45: 261-279. MULLER, J. 1959. Palynology of recent Orinoco delta and shelf sediments: reports of the Orinoco shelf expedition. Micropaleontology 5: 1-3 Raven, P. Н. & D. I. AXELROD. 1974. Angiosperm biogeography and past continental movements. Ann Missouri Bot. Gard. 61: 539-673 SAVIN, S. M. 77. The dvd of ie earth's surface temperature during the past 100 million years. Ann. Rev. Earth Planet. Sci. 5:319-395. DoucLas. 1985. Sea level, climate, and the Central American land bridge. Pp. 303-32 in F. G. Stehli & S. D. Webb r The Great American Biotic Interchange. Plenum, New York. —— ——— & Е. С. Ѕтени. 1975. Tertiary ay EIS. Geol. Soc. Amer. Bull. 86: 1510 В: 4 H. & J. E STEWART (with the collaboration of W. P. WoopRING). 1980. Geologic Map of a Panama Canal and m Republic of Pana Scale: 1:100,000. U. rv. Misc. Inv Map 1-1232. [Map also едм! in Woodring, 1982. Profess. Pap. U.S. Geol. Surv. 306-F.] у . 1967. Climatic implications Pacific Seience E А paleobotanical interpretation of Tertiary climates in the northern hemisphere. Amer. 703. Sci. 66: 694- LEAF ARCHITECTURE OF THE CHLORANTHACEAE! Carol A. Todzia? and Richard C. Keating? ABSTRACT | | f, (1 ] 1 arinopsis) Leaves of 42 species of Chl genera (Sa arcan ndra, Chloranthus, and Hedyosmum) were cleared and examined to determine msi and interfamilial relationships. The teeth in the Chloranthaceae are quite variable The leaf architecture of the Chloranthaceae is compared with that of onaceae, and, although the Chloranthaceae share a number of features in common with these two families, other о suggest an ancient separation Although leaf architecture of extant angio- sperms has been shown to be useful to identify fossil remains (Upchurch & Wolfe, 1987) and to elucidate intra- and interfamilial relationships (Le- vin, 1986a, b, c; Rury & Dickison, 1977), leaf architecture has been studied in only a relatively minute fraction of angiosperms. The leaf architec- ture of the Chloranthaceae is of particular interest because the position of the family in higher-order angiosperm classification is in dispute, and because some of the oldest known fossil floral parts (Friis et al., 1986) and fossil leaves (Upchurch, 1984a, b) are attributed to the Chloranthaceae. No in- depth studies have been conducted on the leaf architecture of the Chloranthaceae, and few critical studies have been conducted on leaf architecture in general. Especially few studies have been done in the Magnoliidae (Doyle & Hickey, 1976; Hickey & Wolfe, 1975; Klucking, 1986, 1987). The Chloranthaceae are a small pantropical fam- ily with four genera and approximately 70 species. The four genera fall into two distinct groups. Sar- candra and Chloranthus are distinguished by bi- sexual flowers and an herbaceous or suffrutescent habit; they are confined to Indomalasia (Verdcourt, 1986). Ascarina and Hedyosmum, on the other hand, are distinguished by their unisexual flowers and arborescent or shrubby habit. Ascarina ranges from New Zealand to New Caledonia and Malesia with one disjunct species (originally placed in its own genus, Ascarinopsis) in Madagascar (Jeremie, 1980; Swamy, 1953; Verdcourt, 1985, 6 Hedyosmum occurs from Mexico to Brazil and the West Indies with one disjunct species in Southeast Asia (Todzia, 1988). Some authors have suggested that the Chloran- thaceae are among the most primitive families of angiosperms and at the root of the Hamamelid and Magnoliid lines (Burger, 1977; Endress, 1987). The fossil record of the family and closely related forms dates to the Early Cretaceous and includes pollen (see Walker & Walker, 1984, for sum- mary), leaves (Upchurch, 1984a, b), and flowers (Crane et al., 1989; Friis et al., 1986) (see Crane, 1989, for review). The Chloranthaceae also have been postulated to have an important position in the separation of the monocot and dicot phyloge- netic lineages (Burger, 1977). Furthermore, much controversy surrounds the placement of the Chloranthaceae in higher-order angiosperm classification (see Verdcourt, 1985). In recent classifications the family has been placed in the Piperales (Cronquist, 1981), Magnoliales (Dahlgren, 1980), Laurales (Takhtajan, 1980), and the Annonales (Thorne, 1976). Recent studies of floral anatomy strongly suggest placement of the ' This research was supported in part by a National Science Foundation Dissertation Improvement Grant (BSR- 8400923) to C. Todzia and by the No Botanical Garden. We thank G. R. Upchurch, Jr., manuscr ipt. ? Plant Resources Center, Department of Botany, University of Texas, Austin, Texas 7871 oyes Foundation while the senior author was a postdoctoral fellow at the Missouri and an anonymous reviewer for constructive comments on the , U.S.A. * Department of Biological Sciences, Southern Illinois University, Edwardsville, Illinois 62026, U.S.A. ANN. Missouni Вот. GARD. 78: 476-496. 1991. Volume 78, Number 2 1991 Todzia & Keating Leaf Architecture of Chloranthaceae Chloranthaceae in the Laurales near the Trimenia- ceae and Amborellaceae (Endress, 1986, 1987; Endress & Sampson, 1983). Endress (1987) sug- gested that the family may link the Laurales and the Piperales as well as being pivotal to under- standing the separation of the Hamamelidae and the Magnolidae. The present study was undertaken to examine the leaf architecture of the Chloranthaceae in order to provide additional information for generic cir- cumscription and to shed new light on intrafamilial and interfamilial relationships of the family. The specific objectives of this investigation are (1) to provide а comprehensive description of the leaf architecture of the Chloranthaceae; (2) to outline basic trends of specialization in the architecture of chloranthaceous leaves; (3) to look at the relation- ship of leaf architecture and elevation in Hedyos- mum species; and (4) to provide leaf architectural data that will contribute to a better understanding of the interfamilial relationships of the family. MATERIALS AND METHODS Leaves of 45 species of Chloranthaceae repre- senting all four genera were studied (Table 1). The most completely surveyed genus was Hedyosmum in which leaves of 37 of the 40 species were cleared. Leaf clearings were done following Fuchs's Method No. 1 (Fuchs, 1963) with some modification. Leaves were obtained from FAA-preserved material or dried herbarium specimens, placed in 80% ETOH for several days, and then slowly rehydrated. The spec- imens were subsequently placed in a solution of 1% basic fuchsin in 10% sodium hydroxide for 12-36 hours in a 60?C oven and then washed in water for several days. After dehydration, the leaves were immersed in a mixture of 3 parts absolute alcohol and | part concentrated HCl and then washed in absolute alcohol for 48 hours. After two xylene rinses, the specimens were mounted on slides with permount. Leaves with thick waxy cuticles that prevented infiltration of the NaOH-basic fuchsin solution were pretreated with 10046 dichloromethane. After 0.5— 2.0 hours in dichloromethane, the leaves were washed in absolute alcohol before the dehydration series. Leaves with large amounts of tannin deposits that obscured the internal features were pretreated with 1075 ammonium hydroxide for 1-5 minutes and then rinsed well before continuing the basic fuchsin-sodium hydroxide procedure. Many of the principal character states and dis- crete features are summarized in Table 2. The only comprehensive sample here is Hedyosmum. One of the two species of Sarcandra is represented. Only small representations of Ascarina and СМог- anthus are given. Five out of 20 species of Chlor- anthus and three out of 11 species of Ascarina were sampled. CHARACTERS Leaf architectural terminology generally follows Hickey (1973, 1977, 1979). Fifteen characters were examined (Table 2), the following of which require more explanation. (8) Tooth type. The six different variants found in the Chloranthaceae were given the fol- lowing codes: 1A, 1B, 2A, 2B, 2C, and 3. A de- tailed description of each type is provided in the discussion section. In addition to the 1 t where some (9) Secondary vein type. standard | g secondaries are joined in a series of prominent arches, some leaves have the secondaries upturned and gradually diminishing apically inside the mar- gin and thus are intermediate between brochido- dromy and eucamptodromy. The condition where the secondary veins are not looped near the apex is designated EB, and the less common case when the unlooped secondary veins are near the base is called BE. The semicraspedodromous (S) condition may occur when secondary loops branch sending the admedial branch to join the superadjacent sec- ondary vein and the exmedial branch to a tooth. This is seen in Chloranthus (Fig. 5), Hedyosmum (Figs. 11, 14), and Sarcandra (Fig. 2). A less well organized version is like the above but with some branches from secondary arches forming tertiary and higher order loops and other branches entering teeth. Upchurch (pers. comm.) calls this **festooned craspedodromous" (FC), and this can be seen in Ascarina (Figs. 3, 6). (11) Presence of an intramarginal vein. This refers to a collector vein formed from the loops of the secondary veins and is usually present within 2-4 mm of the leaf margin. (12) Presence of an intermarginal vein or fim- This refers to the vein immediately within the leaf margin. Tertiary veins usually unite to form the fimbrial vein. (13) Highest vein order. Leaf veins branch repeatedly, producing higher orders of venation. The primary vein (or midvein) of the leaf is the first rank with the second rank composed of sec- ondaries or the major lateral veins. The next finer brial vein. set arising from both primary and secondary veins is designated the tertiary veins and so on, until the Annals of the 478 Missouri Botanical Garden SN ‘XAL ‘ANDO ‘¥20 ‘AN оо "5 ш 00SZ а цошцо)—елоше7 лорепол ‘OW ‘AVY 9152 sauna?) № тгрој (иоле № zado] zmy) шпзошгора шптшзоХрон sn ш 001° 1-006 те ешеџеј ‘XAL “AN ‘OW ‘HO Я SSZ vizpog ман шпитрирјдиод umusoXpapg XAL ш 009'1-00€'I mbun eureueg ‘OW 9222 1?z]v»unjog X тгрој 1eusar[ № Koy, q urnupi4221nq umnuisoKpa]] ш (OTT э50[ uec тогу 21502) XAL ‘HD 805 vizpo[ Ja3ing ‘М 2su2321107500 шпшѕоХрәр ur L/£'T-0I€£'I puepaog влеше SN ‘НЭ “TI ANNA 2022 иозлгриру MG 5и205210д10 шпшзохраН ш 001'2–-006'1 SPUOZEUIV e[anzaua A XAL ‘ОИ 0029r 11216 № 11222) в12ро |, гриђдги шпшзохраН ш 00v*1-00Z'1 чоә[е 4 e[anzaua д XAL ‘A 96£'£] ирРштоја 1eusar] 9 Áo1y,q Клиа8 umuso&pog] ш 00£'z [елоро 011115 e[anzaua\ Sf] ‘AN НО ‘Я 502'16 41Dwıa21% 'qneT-sugos оргшолрирорпога umuisoKpog] едлеоолопиј “1995 eqpeje] “Sqns шпшѕоќрәң ш OOF пошт] тогу 21507) XAL ‘OW ОССЕ 70 зә штала.) Áa]pueig 152u24Q штшвоХргн ш 6/6 иориәзег) вотешеј S ‘OW Y ‘У 226 70 12 Laow "MG зирти шпшѕоХрәр umulsoÁpaH '1295 umusoÁpay “Sqns шпшвоХрон V/N 9juau() eqn”) SN ‘AN 1908 29/045 иеді wnisoquigns штшвоХргн ш 00g'1-009'I aJUALL) eqn) SN 8226 чоно иеді asuagunuop штшвоХргн V/N susto eqn) SA ‘НЭ #255 P зә иту 'qneT-suog 11420925119 umusoXpag] V/N Зипу#ием y ешт) V 26992 Fuvs] ung"? Y [HW 2707u2140 шпшзохраН э[езиэ40 '12985 umuirsoKpaH :3qns шпшѕоќрәң V/N од “YA ешорајео) мом ON 9879 uosi2yqopy "qneT-sugog s7mpo1qna сү V/N V/N зашто MON SN 2687 әррочә$ чон 'g `2 sisuauiddipyd ү V/N pues? M prep мәм ү 8289 2111427 3 ooH y PULIPOSV V/N V/N V/N у 569 әшаә1 O 3 tad sninoids 7) V/N V/N eun) H9 859 41u2y 19AJQ smjofusngun 7) V/N V/N V/N SN 2959 1102и15 f ипоәрлзәд (ure qongp) sn12212 3 snuju?o[q7) V/N ixguen-) eup ON 0212 әшаәт мпоорлад (aum[g) s€12n07s41204qQ ‘dsqns 219118 `ç V/N uars 'одтеј, SEL ON ‘V 6b8€ 1" 12 poor 19018 *dsqns тезе (“QUAY L) 219518 5 елрпеолес иоцелә[ goutaoid 10 91216 Ánunor) џоцогјјог) sei edg '"Kpnis 91nj99jrqoI1e језј 38398U7UE 10U) ut pasn suaunoads 10} јер uono9[o-) `Т ЯПЧУ[, 479 Todzia & Keating Volume 78, Number 2 1991 Leaf Architecture of Chloranthaceae WSA ш 00$°5 oonuenH nəd ‘XAL ‘OW ‘A EELZ uteig P ZPOL "qneTsuqgog 1427422] штшѕоХрәр WSN ‘SN ‘XAL ‘OW ш 096'1–006'1 unrejy ues nəd ‘XN ‘HO ‘A 6693 #1945 P vizpo qneT-sugos шпирХәфшор штшѕоХрәр Хал ‘054 ‘AN ‘OW ш 008'2-009'c оќештцпд 214010109 ‘WAVE ‘100 ‘NVV ZIP vizpor севезодепо umptonjsup4) wnusokpaH ш 006-082 ә[зог) eureueq XAL ‘ОИ 6252 7129] Áe]pueig шти1л12д025 штш«гоХргн XAL “ANDO “VIO ш 005% әбщэшц”) -е10шв7 Jopenoy “AN ‘OW SISZ $2145) № 112ро] eizpo |, 2719012245 шпшѕоХрәр ш 056'1-002'1 0204) etquio[o) ТОО 5916 ouozoT таро, Huey tunc paj] XAL “ANDO ‘WOO ш 002'2-000'2 euourqotd Jopenoq ‘AN ‘OW ‘A 5272 sauna») 3 vizpo epo] шилорозтир шпшзохраН ANDO 'v90 ш 0eG'z одем дореполј ‘XAL ‘OW 0252 s2una42) № 112ро] eizpo] umsogiujs umusokpag] SN ‘XAL “AN ‘OW ш OOT‘'E ‘еә вэлешеитрий”) еташојо) ‘WAVE ‘YAH ‘Я 100 OF FZ vizpor 11014220 UNu шпшзохраН XAL “AN ш 00S'Z ‘89 ?21eureurpun?) eIquio[o7) ‘OW ‘WOVE ‘A 100 Sere тлгрој зезеоо депо) шпиттдшојоо umusoXpor] sn ш 009'I-00S'I mburq?) eueued ‘XAL AN ‘OW НО 40622 тлгрој Jauser] № Аолу ( umun24102 unusoXpaH sn ш (OGG “82 embonuy eiquojo) ‘XAL ‘AN ‘OW НО 47622 тлгрој ‘qney]-suljog шпирорпой шпшѕоХрән SN ‘XAL ‘S ‘AN ‘OW ‘4 НО “5 ‘4 ‘SVO ш SZ-OI eueleg [22:9 ‘AVY FEES y»vquosmg р тгрој buy хә ‘ı1e 2$и21]15249 шпшзоХраН Хал “ANDO ш 000'1-006 ezejseq Jopenoy *'V9OÓ 0252 sawu11) № тлгрој "'qneT-sugog 122n4ds umuisoXpagy uoneA9[] eoutao1d 10 21015 Аципо) uona вәїзәйс 'penuguo) Т 318V], 480 Annals of the Missouri Botanical Garden WSN ‘XAL “AN ОМА ш 009'z-00T'c oonuenH плод “HO ‘Я SAVY 1226 "1918 № тлгрој PIZpOT wnupianiad wnwsod pap] SN ‘XAL “ANDO ‘¥20 ш OOT‘E-000'E odeN lopenog “АМ ‘OW 0872 sounds) р vizpo epo] uan] umusoXpor] SN ‘XAL ‘ANDO “VIÓ ш 008'2-009'2 Áenzy Jopendy ‘KN ‘OW ‘A SOSZ s2un42) № vizpo[ шогооб) ummupzp224]p0n2 UNUSOL PI sn ш 002'2–005'5 eoe epenzausa ‘AN ‘OW "IOO ‘AVV Lees ulom] 'qneT-sugos шпуојлалра штигоХргн XAL ш 00£'£-00€ € oren Jopenoy ‘VD0 ‘OW &6Ё& $2145) 3 тгрој U9]SIE «M 2SUIJDQUINI wnusod pay шосре еаштедец207) eog Sn ‘AN ‘OW Но 16S yorquiars ииешпцос̧ ^w (oziunw) шпиихош wnUsOA pat] qneT sugog ш QO0T'T-00€'I urunf niod WSN ‘XAL #922 #1215 № ZPOL (иоле № zado] ziny) wnyofijsndup штш«оХргн ANDO ‘VOO qneT-sugos ш 002% eloT Jopenoy “AN ‘OW “A PISS Sale тлгрој (цоле Y zado] ziny) 1un1qDos штш«охргн едлеоолов ји] '1295 eq[eje], "qns шпшѕоХрәң XAL ш 008'I mbunyp BUIBUE f ‘OW 0922 122]0шуэѕЅ P vizpo[ Аошарлој шпирмхош ummusoXpag] зэрюЧлеэозту “joes e[[eje], "dqns umursoÁpoH uoneAo[j] дошлола 10 31818 Алипог) uona вогоо 46 грепициој “| 18V], Volume 78, Number 2 1991 Todzia & Keatin g 481 Leaf Architecture of Chloranthaceae ¡EN ара : ~ e NY 7 X 3 f {д hloranthus spicatus. —6. grum. —9. Неду ultimate vein order present in the leaf is reached (Hickey, 1973). (15) Number of free endings/areole. The ac- tual range of the number of freely ending veinlets within an areole is reported. Although several de- grees of branching are usually present within a single leaf, range of the degree of veinlet branching appears to be consistent within a species. i osmum nutans. — 10. Hedyosmum grisebachii. Scale line = 2 c = DE) 3 Т => ы а => » < as = E SES ср SE Sn ¿e Ss S ом hee Sa » a Lan Ga a `c 1 es NS PS < DATED 2 ^ Ny E 7 WS (16) Leaf rank. According to Hickey (1977, р. 158), leaf venation is organized into four levels or ranks. Leaf rank is based primarily on the high- est vein order in which the veins follow a regular course and delimit areas of relatively consistent size and shape. First rank leaves have secondary: and higher order veins following an irregular course. The intercostal areas are irregular in shape, as are 482 Annals of th Missouri ane Garden Pines 2. БА sect. Microcarpa; 9 s Hedyosmum individual cleared leaves. (4) yu dape el = elliptic y ect. g. Tafalla sect. M UC (2) & osmum; + subg. , la = lanceolate, li = linear, n — Leaf architecture data. (1) In ов the sectional alignments are given as follows: * subg. yosmum sect. Orientale; < subg. sect. Microcarpa; ! subg. (3) Data representative for species, not just narrowly, ob = obovate, ov = ovate, w = widely. (5) мн leaf pleniorem ratio. (6) Leaf apex: acm — acuminate, act — acute, b — broadly, 1 = long, ob = obtuse. (7) Leaf cun — cuneate, n — narrowly, obt — obtuse, rd — nded, trun — truncate. dary vein type: B = brochidodromous, 5. = sen ааа (8) Tooth type: see text for не (9) Secon festooned craspedodromous. (10) Average number of major secondary veins per leaf ce of an dE QE vein. (12) Presence or absence of an intermarginal (or fimbrial) E — eucamptodromous, FC — side. (11) Presence or absen vein. (13) Highest vein order. (14) Areole size. 1 = very mm, 4 — small «0.3 mm. (15) Number of "dl vein endings/ areole. (16) Leaf rank explained in text. large >2 т m, — medium (5) (1) (2) (3) (4) Мм (6) Species Length (cm) Width (cm) (Shape) ratio Apex S. glabra subsp. glabra 2-20 1-8 la to el 2.4:1 l-acm S. glabra subsp. brachystachys 2-20 1-8 la to el l-acm C. erectus 8-29* 3-13* el 5:1 аст A. lucida 2-7 2-4 el, ov 1.5:1 act A. philippinensis 3-14(-18) 2-8(-10) el, ob, ov 1.7:1 ођ, аст A. rubricaulis -12 1-3 a 4.2:1 acm * Н. orientale (9-)15-25 1-3 la to li-ob 10:1 l-acm * H. grisebachii 13 2-3 la to ob-la 4:1 l-acm * Н. domingense 4-6(-8) 1-2 la to n el 3.3:1 acm * H. subintegrum (2-)5-8 1-2 n el, n ob 4.3:1 acm ОН. nutans 5-14 1-2 la 6.3:1 acm ОН. brenesii 6-16 1-4 li-la, n el 4.4:1 l-acm +Н. pseudoandromeda 3-4(-6) 2-3 el, n el 1.4:1 act +H. gentryi 6-16 3-6 ob to el 2.4:1 acm +H. neblinae (4-)6-9 3-6 ov to w el 1.7:1 act, acm +Н. arborescens (3-)5-12(-18) 2-6 el 2.1:1 ас +Н. costaricense (4-)8-15(-20) 2-5(-7) n el, el 3.3:1 l-acm +H. burgerianum 3-6 1-2 ov to ob 3:1 l-acm +Н. bonplandianum 5-13(-17) (2-)4-7 n el, el, ov 1.6:1 acm +Н. racemosum -23 3-8 ob, el, ov 2.6:1 acm +H. sprucei (5-)8-19(-22) (2-)4-8 n el, n ob 2.2:1 acm +H. brasiliense -21 2-6(1) n ov, n ob 3.6:1 acm +Н. goudotianum (3-)6-16(-19) 2-7(-8) п еј, оу, ођ 2.4:1 аст +H. correa (3-)5-9 1-4 ov, ob, el 2.8:1 acm +Н. colombianum 6-15(-18) 2-7 el, ob 2.5:1 acm +Н. crenatum (4-)5-11(-14) 2-4(-5) n el, el 3:1 acm +H. strigosum 8-17(-21) 3-8 el to n el 2.1:1 acm +Н. anisodorum (9-))4-22(-29) 5-8(-10) n el, n ob 2.8:1 acm +Н. чети -12 2-5 n el, el 2.3:1 acm +H. spectabile (6-)10-20 (3-)5-8 el, ov 2.3:1 аст +Н. scaberrimum -27 3-6(-8) el, n el 3.9:1 acm +H. translucidum 5-12(-16) 2-4(-5) n ob, n el 2.4:1 acm +H. dombeyanum -10 2-5 n el, ob 2.1:1 аст +Н. lechleri 4-12 2-4 n el, el 2.7:1 acm ЇН. mexicanum 10-25 2-7 la, ob-la 3.9:1 l-acm eH. scabrum 6-13(-16) 2-7 n el, el, ob 2.1:1 аст eH. angustifolium (5-)12-17 2-5 la, el, ob-la 4.1:1 аст eH. maximum 8-14 4-6 e 2.2:1 acm eH. cumbalense 1-5 0.5-2 n el, el, ob 2.4:1 acm eH. parvifolium (2-)4-9 2-3 n el, el, ob 2.6:1 act, acm eH. cuatrecazanum (5-)8-19 2-8 ob, n ob, n el 2.7:1 acm eH. luteynii 4-8(-13) 3-5(-8) ob, el 1571 аст eH. peruvianum (8-)1 2-22 (3-)5-9 n el, el 2.4:1 acm Volume 78, Number 2 Todzia & Keating 483 1991 Leaf Architecture of Chloranthaceae TABLE 2. Continued. (11) (12) (9) (10) Intr Inter (13) (14) : (16) (7) (8) 2? vein 2? vein mar. ma ighest Areole vein end- Leaf Base Tooth type type no vein vein vein si in rank cun 2B S 4-5 - + 5 2 1-5 2r° cun 2B, 2C EB 8 = — 5 2 0-2 lr! cun 2B EB, S 6-9 — = 6 3 0-2 — cun 3 FC 7 = = 4-5 2 0-5 2r? cun 2A EB 9 = + 5 2-3 0-2 2г! сип 2А ЕС 10 E — 5 2-3 0-4 2r! n cun 2B EB 18 = — 4-5 2 0-1 2r" cun 1, 2A EB 16 + + 5 2 0-3 2r! rd 2A B 7 — + 5 2 2-4 2r? cun 1 ВЕ 8 s + 5 3 2-3 2r! rd, cun 1, 2A S 10 — + 5 2 0-3 2r? cun 2B E 15 — + 5 2-3 0-2 2r! obt, cun 2A B 14 + + 5 3 0-2 2r? b cun 2A EB 12 + + 5 2 1-3 2r? obt, trun 2A EB 11 + + 5 1 0-2 2r? cun, obt cun 2B EB, S 13 + + 5 2 0-2 ar? rd, b cun 2A E 9 + + 5 2 0-2 2r? cun 2A, 2B EB 8 + + 5 2 0-2 2r? cun, obt cun 2B E 8 — — 5 2 0-3 2r! cun, obt cun 1B E 5-12 — + 5 2 0-3 3r! cun, obt cun E 11 - – 5 3 0-3 3r! cun, obt cun 1B, 2A EB 8 — — 5 2 0-3 2r? cun, obt cun 3 E or B 8 — 5 2, 2-3 4r? cun 3 E 7 — — 5 2 0-2 2r! cun 2A E 11 — = 9 2-3 0-2 2r! cun, obt cun 1B, 2A E 6 — — 5 2 0-2 3r? rd, obt rd lA E N/A — — 5 2 0-2 3r? 3 E 9 — + 5 2 0-2 2r! cun, obt cun 3 E 7 + — 5 2 0-2 2r! cun, г 3 E 6 — = 5-7 2 0-6 4r? cun, obt cun 3 E 6-12 — — 5 2 0-2 2r! cun, obt cun 3 E 10-12 = 5 2 0-3 2r* cun 2A E 8-10 T = 5 2 0-2 2r! obt, cun 1B EB 9 T — 5 2 0-2 2r! cun, obt cun 1A, 1B, 2A E 12-14 — 5 2 0-3 2r? cun 1А В 20 + + 5 2 0-1 3r? cun, obt cun 1B EB 15 — - 5 2 0-2 2r? b cun, obt cun 1А ЕВ 10 = 5-6 2-3 0-4 2r? rd, trun 1B EB 13 ae + 5 2 0-2 2r3 cun, obt 2A E 16 = – 5 2 0-2 2r? cun, ob cun 1B EB 14 — + 5 2-3 0-2 3r? n-b-obt cun 1А В 16 — 4 5 2 0-2 Зе cun, obt cun 1B E 18 — + 5 3 0-2 3r! 484 Annals of the Missouri Botanical Garden SG fe PF: IEE АЯ /; УЖЕ Ру EOS >, Y D 2 4 A y 3 if | 8 ЧЕ [А %. || | үр М 7 Y ау, Ga 987, бола СЈАЈА p EAS, rans та 06 Ба e AX zv. 2 Volume 78, Number 2 1991 Todzia & Keating Leaf Architecture of Chloranthaceae 485 the tertiary and quaternary domains. Domains are areas delimited by veins of tertiary and higher rank. Second rank leaves have regularly shaped inter- costal areas, but higher domains are not uniform in size or shape. Third rank leaves have relatively consistent (predictable) tertiary domains. Fourth rank leaves have areoles of predictable or uniform size and shape. Within each of the main ranks, subranks 0-3 are recognized based on the degree of advancement (increasing regularity of pattern) toward the next level. The assignment of subranks would be based on the assessment of the status of the following kinds of morphoclines. Secondary and higher order veins may arise from the next lower level at de- current, acute, orthogonal, or obtuse angles. Vein endings within areoles may be a high number, a low number, rare, or absent. In a brochidodromous leaf there may be several orders of loops, few loops (festooned), no loops, and finally, the straightening of the arches into an intramarginal vein. In each of these sequences, one would assign a low subrank (0-1) for the beginning of the morphocline and high subrank (2-3) at the other end of the mor- phocline. These sequences are read in a direction of increasing specialization as supported by anal- yses of fossil floras. Leaves with low rank appear first in the fossil record, and it seems that leaf venation rank order correlates with the degree of advancement of the whole plant as determined by other evidence (Doyle & Hickey, 1976; Hickey, 1977; Hickey & Wolfe, 1975; Thomas & Spicer, 1987) Leaf length and width were plotted in scatter diagrams using log scales. Samples of up to ten specimens were used for each species, if available. GENERIC DESCRIPTIONS SARCANDRA GARDNER (FIGS. 1, 2, 45, 46) Leaves elliptic, chartaceous, with length/width ratios of 3-4:1. The apex acuminate and the base cuneate to degit Margin usually boldly toothed with deeper sinuses than the other genera. Sinuses rounded. The median vein to the tooth apex arising from secondaries. Petiole very short with a decur- rent wing of laminar tissue on petiole. Secondary venation either semicraspedodromous or brochi- dodromous with basal secondary veins forming loops but tending toward eucamptodromy. Secondary veins alternate, widely spaced with about five per side, arising at an acute angle and curving apically. Intercostal areas large and irregular. Intersecon- daries, about two per intercostal area, mostly sim- ple but sometimes composite. Tertiary veins acute admedially and exmedially. Tertiary and quater- nary domains both random polygonal. Highest vein resolution five. Areole development imperfect, ar- rangement random, and irregularly shaped. Are- oles 1-2 mm with 0-5 veinlets per areole. Druses random or absent. Stone cells present in loose clus- ters near the midvein. CHLORANTHUS SW. (FIG. 5) Leaves 4-30 cm long, elliptic, with length / width ratios of 3-5:1. The apex acuminate and the base cuneate to decurrent. Margin serrate with teeth closely adjacent to the shallow, rounded sinus. Bas- al side of tooth slightly convex. The tooth hyda- thodal with one or more large stomates on the tooth surface distal to the dilated median vein of the tooth. Leaf texture chartaceous and the petiole very short with a decurrent wing of laminar tissue. Secondary venation brochidodromous with a broad row of marginal loops or distally brochidodromous and proximally eucamptodromous with marginal loops that produce tooth veins. Secondary veins acute, usually widely spaced, with 6-9 per side of the midvein, alternate or subopposite. Intercostal areas irregular and gradually apically curving. In- tersecondaries common, 1-2 per intercostal area, and ranging from simple or composite. Tertiary and quaternary domains irregularly polygonal in size and shape. Highest vein resolution five. The imperfectly developed areoles polygonal and ran- domly arranged, and about mm long. Sto- mates common and randomly arranged over the abaxial epidermis. Prismatic crystals occurring sin- gly or in clusters and randomly distributed in the lamina, often in and near teeth but also closer to the midvein. The clusters not like druses but ap- pearing eroded. ASCARINA J. В. & С. FORST. (FIGS. 3, 4, 6) Leaves elliptic, with length / width ratios of 1.5- 4.2:1. The apex acuminate to obtuse and the base cuneate. Margins serrate with the tooth apex close =— FicunES 11-20. linae. —14. H. burgerianum. — Leaf clearings of Hedyosmum. — 11. H. gentryi. —12. H. pup 5. H. arborescens. — 16. H.r . H. neb- Н. costaricense. — 17. racemosum. a brasi- liense.—19. H. mexicanum. — 20. Н. bonplandianum. Scale line = 2 cm 486 Annals of the Missouri Botanical Garden to the sinus. The sinus rounded. Leaf texture char- taceous to coriaceous. Petiole very short, with a decurrent wing of laminar tissue. Secondary ve- nation a combination of festooned brochidodromous and semicraspedodromous, the secondary veins forming a series of loops near the margin and branches of looping secondary veins enter the teeth. Ascarina philippinensis with regular brochidodro- mous arches. Ascarina rubricaulis with strongly decurrent intersecondary veins arising either from the midvein or from a secondary vein. Other As- carina species with + nondecurrent intersecon- dary veins that arise only from the midvein, as in Sarcandra and Chloranthus. 'The 6-8 secondary veins producing somewhat regular to very regular intercostal areas. А few simple or composite inter- secondary veins present. Tertiary and quaternary veins formin ae and polygonal domains. Highest vein order 4-5. The randomly arranged areoles imperfect and irregular to polygonal. Ar- 0.3-1 mm and showing strong tendency to be elongate and parallel to the secondary veins or eoles at right angle to the midvein. Crystals absent. Stone cells free or clustered near the midvein, or in short rows along the midvein. HEDYOSMUM SW. (FIGS. 7-44, 47, 48, 49) Leaves ovate to oblanceolate, elliptic, obovate varying in size from | to 29 cm, with length/ width ratios of 1.5-10:1. The apex mostly acuminate to occasionally obtuse and the base mostly cuneate to truncate or obtuse. Margin serrate with various forms of glandular teeth. Most teeth convex basally with a small, rounded sinus. The tooth apex often close to the sinus. ceous. The petiole very short with a decurrent wing of laminar tissue. Secondary venation brochidodro- mous, eucamptodromous, or transitional, and oc- casionally looped. Venation occasionally semicras- Texture chartaceous to coria- pedodromous. Brochidodromous arches occasionally aligned forming an intramarginal vein. Secondaries often more numerous than 12–15 per side, acute and curving apically. Intercostal areas often very regular. Intersecondaries absent or present as sim- ple or composite. Tertiary and quaternary veins usually irregular and polygonal. Highest vein order present usually 5 (rarely to 7). Areole development imperfect to well developed in about 2 the species. Areole shape polygonal to irregular, usu- ally randomly oriented. Size quite variable within a specimen and varies for the genus between 0.3 and 2.0 mm. Veinlets varying from 0-1 to 2-4 or 0–6 per areole and sometimes branched several times. Crystals absent. Sclerenchyma varied, stone cells, single or clustered, branched spongy meso- phyll cells and branched vein sheath cells. DISCUSSION Interpretation of character state polarities be- gins with the analyses of Hickey & Wolfe (1975), Doyle & Hickey (1976), and Doyle (1978). Hickey and Wolfe derive most of their conclusions from comparisons of leaf clearings based on the ap- proximately 20,000 specimens represented in the two major cleared leaf collections in the United States. Doyle's (1978) and Hickey & Doyle’ s (1977) reviews comment on the one stratigraphic sample that has been analyzed in detail for both pollen and leaf fossils, the Potomac formation of the eastern United States. From those works, it appears that the earliest known angiosperm leaves are simple, entire, pinnately veined, first rank forms, with sev- eral orders of anastomosing veins and freely ending veinlets. The earliest forms illustrated in those stud- ies show a festooned brochidodromous secondary architecture of Hickey's (1979) system. Hickey & Taylor (1989) have refined this model using cla- distic comparisons of primitive genera with sug- gested angiosperm sister groups and with Early Cretaceous angiosperm fossils. The ‘‘ancestral an- giosperm leaf" is conjectured to have a sheathing base with 4—5 separate vascular bundles entering the petiole producing pinnate (brochidodromous) or palmate (palinactindromous) venation in ovate leaves. Chloranthoid teeth are found near the apex. n the chloranthaceous leaves conforming most FIGURES 21-28. Leaf clearings of Hedyosmum. — 21. reanum. — 24. Н. crenatum. — line = 2 cm FIGURES 29-33. Leaf clearings of Hedyosmum. lucidum. —32. H. lechleri. — FIGURES 34-40. ам — 37. Н. angustifolium. — 38. Н. 2c H. goudotianum. — 22. 3. H. с 5. H. мети. — 26. H. spectabile. — 27. H. anisodorum. — 28. H. strigosum. Scale Leaf clearings of Hedyosmum. — 34. H. scabrum. — cuatrecazanum. — 39. H. colombianum. — — 29. H. scaberrimum. — 30. H. dombeyanum. —31. H. trans- 33. H. sprucei. Scale line — 2 cm. H. parvifolium. —36. H. cumba- H. кыы а л Н. luteynii. Scale line = 487 Pa ID. da: Leaf Architecture of Chloranthaceae Todzia & Keating Volume 78, Number 2 1991 Ky | ЕНИ "A ГРБУ |. М MEAS: > уай ыр Za" "КУГ, A) p. $ Missouri Botanical Garden Annals of the 488 DAC TÍA ME x CUTE 102 я а A E 489 Todzia 4 Keating Volume 78, Number 2 Leaf Architecture of Chloranthaceae 1991 2! Ир Ер EC > AS C EC m. e EX AON ‘ c ESAS ORS LEARN <= El 29 22 OD s, aa ome 5 de CA ES e d ! gue 389 | 2 e ср MS SN — BEHETIK d NECA LAS CA ALIS cH oS PPP Se CDA , ox] E ARA Da HERA 5 ОУ ES Lu en E (Mn E Za di АЕ сга € q e» ЖЛЕ 222626725 ЕЕ OMS i da E ES Г | NS FORO \\ ee 5 g e p gi m 490 Annals of the Missouri Botanical Garden e Volume 78, Number 2 1991 Todzia & Keating Leaf Architecture of Chloranthaceae closely to these early leaves, the secondary veins tend to be widely spaced, are never narrowly spaced, and form irregularly shaped intercostal areas. Be- yond this observation, there is no established po- larity regarding the spacing of secondaries. Possible ranalian outgroups give conflicting evidence. Leaves with few, widely spaced secondaries occur in the Amborellaceae and Austrobaileyaceae while leaves with numerous, closely spaced secondaries are found in Trimeniaceae, Winteraceae, Degeneriaceae, and Eupomatiaceae (С. Upchurch, pers. comm.). Transitional secondary venation patterns are common and there is too little comparative infor- mation to polarize the character states with con- fidence. Prominent secondary arches occur and often tertiary and higher order loops form between the secondary arches and the margin (e.g., Sar- candra glabra (Figs. 1, 2), Ascarina lucida (Fig. 3), and A. rubricaulis (Fig. 6)). Higher rank leaves have a higher number of secondaries, and brochi- dodromy tends toward eucamptodromy at the prox- imal end of the leaf (e.g., Hedyosmum grisebachii (Fig. 10), H. costaricense (Fig. 16), H. racemosum (Fig. 17), H. mexicanum (Fig. 19), H. bonplan- dianum (Fig. 20)). The semicraspedodromous pattern is seen where the tooth vein originates prominently from sec- ondary arches of the brochidodromous plan as seen in Sarcandra glabra (Figs. 1, 2), Ascarina phi- lippinensis (Fig. 4), Chloranthus spicatus (Fig. 5), and Hedyosmum nutans (Fig. 9). In Hickey & Wolfe's (1975) taxonomic review, semicraspe- dodromous leaves show up for the first time in the dilleniid genera, a group parallel to and slightly higher than the magnoliids. We have seen no pre- vious discussion relating this patten closely to the brochidodromous leaves and yet it is found among genera in the Chloranthaceae that on other grounds re more primitive than Hedyosmum. Lorence (1985) included a survey of the leaf architecture in his monograph of the relict genera of the Monim- iaceae of the Malagasy region. The patterns illus- trated are mostly brochidodromous but include ad- mixtures of craspedodromous patterns in Ephippiandra madagarcariensis and Tambour- issa ficus, and semicraspedodromous in Ё. tsara- tanensis and T. trichophylla. It appears to us that the developmental and phylogenetic relationships between brochidodromous and semicraspedodro- mous leaves are quite close and need to be ade- quately analyzed. Existing concepts should be re- garded as tentative at this point but we suggest that the semicraspedodromous leaf type is one of, if not the earliest, associated with toothed leaves. The fact that many early ranalian leaves are entire- margined may have helped to obscure this rela- tionship. Fully eucamptodromous leaves can be seen in Н. goudotianum (Fig. 21), Н. steinii (Fig. 25), and H. anisodorum (Fig. 27). Some forms of eu- camptodromous leaves (e.g., H. gentryi (Fig. 11), H. neblinae (Fig. 13), H. arborescens (Fig. 15)) have secondary arches tending to become straight- ened into an intramarginal vein. Without direct evidence, but because of the common association of brochidodromous and eucamptodromous vena- tion patterns, we hypothesize that eucamptodromy is derived from an earlier brochidodromous pro- totype through a shift in ontogeny. The large majority of leaves in this family are elliptic (narrowly elliptic with a length/width ratio of 4.3:1 to widely elliptic with a length /width ratio of 1.4:1). Leaves may be ovate (Fig. 1, Sarcandra glabra) or obovate (Hedyosmum cuatrecazanum (Fig. 38) and H. peruvianum (Fig. 39)). In most cases the tip is acuminate and the base cuneate or decurrent. Leaf size ranges from 10 to 300 mm and length and width seem tightly correlated as seen in Figure 49. In the scatter plots (Fig. 49), the largest leaves . stri- gosum, H. pungens, H. peruvianum) occur in low to mid elevations in the Andes and Central America, whereas species with the smallest leaves (H. domin- gense, H. pseudoandromeda, H. burgerianum, H. cumbalense) are found on exposed ridgetops or high elevations. Species with leaves representing the central values of the scatter plots have a variety of habitat preferences that show no ready corre- lation. Teeth in this family had been referred to by Hickey & Wolfe (1975) as chloranthoid, and Hick- ey & Taylor (1989) regard that type as prototyp- ical for the angiosperms. However, we have noted that the leaf teeth in this family are quite variable and difficult to ascribe to one familial type (see (Hedyosmum sprucei, H. anisodorum, e FIGURES 41-48. 42. H. cumbalense, tyje l d H. gentryi, type subsp. glabra, type ra pu subsp. brachystachys, type 2 оо pls 3. Scale s = 2m Leaf tooth types. Cleared leaves of Hedyosmum and Sarcandra. — 41. H. scabrum, type Та. — 2a.—44. H. iC و‎ A type 2 —47. 2b.— 45. ipea da glabra H. anisodorum, type 3.— H. 492 Annals of the Missouri Botanical Garden m H. subintegrum m Н. arborescens 100 | п Н. orientale о Н. gentryi Ё e Н. pseudoandromeda А 8 e Н. nutans H. pseudoandromed Ё | E B o H. domingense A Н. costaricense En P 50 | : a E rianum na a4 aol Н. brenesii а ид“ E H. grisebachii Е e ean," мр "7€ د‎ A Am a n o. де ФА А ٠° • p А в EJ del о e А x Oo "e ©з 5 10 | O o O 8 - - O er A W Н. bonplandianum = Н. spectabile у . anisodorum O Н. sprucei о Н. anisodo 100 F 5 п ои о T А Н. mexicanum "dg e Н. goudotianum во G a a E e H. brasiliense mu ААА H. correanum Mia sob О Н. racemosum п ново A А Н. сгепаї ogee” d 40 | ОФА А A M. те 2° О мә ^ o0% А 0% A 30 | ene ® А m о 4O^A [e] 0404 * 20 | O А 10 - B L- st С e H. strigosum e Н. peruvianum O H.lechleri О Н. angustifolium Wor А H. translucidum D? e m Н. luteynii ¿e L A Н dombeyanum "Ua п Н. cuatrecazanum ө E = Н. scaberrimum ¿e > А Н. cumbalense бео, ° 8 D ап 50 О Н. pungens до Он^ ^ H. parvifolium و‎ 40 - A A о A Do OO а » = o0 A^ А Аб O 20 | aso А ^ А А А ег A A а - Ё РУМ 6L zx - |__-_---_____| || و‎ 1 i1 i PDE EE 10 20 30 40 50 100 200 300 10 20 30 40 50 100 200 300 Е 49. Hedyosmum leaf length ( aca: were used in each species sample. (x axis) and width ( y axis) in mm. Log scales. Up to ten specimens, when Volume 78, Number 2 1991 Todzia & Keatin g 493 Leaf Architecture of Chloranthaceae -—!——— —— À — на Figs. 41-48). We categorized the leaf teeth of the Chloranthaceae as follows: type la (Fig. 41) is a broad tooth with a median vein usually arising from a secondary vein; it is symmetrically disposed be- tween two smaller lateral and marginal converging veins. The three veins end in a vascular plexus distal to which appears a hydathodal glandular tip. On at least one specimen (Chloranthus sp.) a large stomate was noted on the surface of the tooth apex. Tooth type 1b (Fig. 42) still has the symmetrically converging lateral veins, but the tooth itself is shal- low with only a small glandular пр protruding. Teeth of several pecies (Hed H. subintegrum, H. nutans, and H. sprucei) are not clearly placed into either type la or type lb and are therefore coded simply as type 1 (Table 2). All the remaining teeth tend to have tips strong- ly curved toward the distal end of the leaf. The tooth apex is often oriented very close to the distal margin of the sinus. Type 2a (Fig. 43) has the median vascular bundle fused some distance back from the apex of the vascular plexus to the ex- medial marginal vein. The admedial lateral be- comes a marginal vein following around the sinus. The admedial lateral enters the tooth near the tip of the vascular plexus. In type 2b (Figs. 44, 45) the medial vein and the exmedial vein are fused far back from the tooth apex, and the admedial lateral follows the same path as the previous type. The unique feature is that the vascular plexus extends a greater distance into the attenuate leaf tooth. In type 2c (Fig. 46) the medial vein enters into the tooth with one small lateral vein merging far back from the tooth tip. The pattern is variable. There can be admedial, exmedial, or no lateral veins converging. Type 3 teeth (Figs. 47, 48) have а prominent densely sclerotic tip. They are vas- cularized by a prominent median vein with 0 to 2 Converging lateral veins well inside the leaf margin. © nature of the teeth is currently being examined histologically. It is unclear at this point as to what *Xactly characterizes a chloranthoid tooth, if in fact, such a generalization can be made Since data are available for 37 of the 40 species of Hedyosmum, relationships within that genus can be discussed in greater detail. Specific architectural features Support Todzia’s (1988) classification, Which recognizes two subgenera and five sections (Table 1). Leaves within subg. Hedyosmum all are lanceolate to narrowly elliptic with length/width ratios ranging from 10:1 to 3.3:1. Leaves in the 31 species of subg. Tafalla represented in our Sample are mostly elliptic, with only a few species аута obovate or lanceolate leaves. The length/ Width ratios of subg. Tafalla range mostly from erisebachii > > 1.7 to 3.3:1 with a few exceptions (H. brasiliense, 3.6:1; H. mexicanum, 3.9:1; H. scaberrimum, DE angustifolium, 4.1:1). The median leaf rank differs among sections of the two subgenera presently recognized in Hed- yosmum. Section Orientale and sect. Hedyosmum of subgenus Hedyosmum have median leaf ranks of 2г' and 2r'? respectively, while sects. Micro- carpa, Macrocarpa, and Artocarpoides of subg. Tafalla have higher median leaf rank (2r?, 2r°, 2r respectively). Within sect. Orientale the species with the lowest leaf rank order, H. orientale (де), is clearly the most primitive species in that group. This agrees with Hickey & Taylor's (1989) de- scription of early angiosperm leaves that have very low rank. Within each section of Hedyosmum, however, leaf rank varies widely among species. Although the phylogeny within sect. Microcarpa is still in question (Todzia, 1988), rank may be indicative of species relationships. For example, the following species pairs are more closely related to each other than to other species within sect. Microcarpa, and each pair possesses the same leaf rank: H. gentryi and H. neblinae (2r?); H. cos- taricense and Н. burgerianum (2г?); Н. racemo- sum and Н. sprucei (Зг!); Н. correanum and Н. colombianum (2г'); and Н. goudotianum and Н. spectabile (4r°). Within Hedyosmum all tooth types are present except for type 2C. Section Microcarpa has the most varied array, including 1A, 1B, 2A, 2B, and 3. Species with type 3 teeth (H. goudotianum, H. related by virtue of their widely spaced teeth. Con- sidering its size, sect. Macrocarpa has the most limited range of all the groups with types 1A and 1B predominant and only one instance of type 2A, in H. parvifolium. In subg. Hedyosmum, sect. Hedyosmum and sect. Orientale have 1A, 1B, 2A, and 2B type teeth. An intramarginal vein is present in only a few species, H. pseudoandromeda (Fig. 12), H. gen- tryi (Fig. 11), H. neblinae (Fig. 13), H. arbores- cens (Fig. 15), and H. costaricense (Fig. 16), and appears to be one of the defining characters of this group within the large sect. Microcarpa. Within sect! M. Mos је Veils appears to be Е useful character. Closely spaced secondaries are present in H. parvifolium (Fig. 35), H. cumbalense (Fig. 36), H. cuatrecazanum (Fig. 38), H. peruvianum (Fig. 39), and H. lutey- nii (Fig. 40). These species are more closely related Annals of the Missouri Botanical Garden to each other than to H. angustifolium (Fig. 37) and H. scabrum (Fig. 34) which, in addition to having more widely spaced secondary veins, also have a different pistillate inflorescence morphology. Within Hedyosmum leaf architectural data do not appear to be correlated with elevation. INTERGENERIC RELATIONSHIPS Sarcandra, Ascarina, and Chloranthus all ap- pear to have less well organized leaves than Hed- yosmum. Sarcandra and Chloranthus leaves have widely spaced secondaries with poorly organized loops and unpredictable eucamptodromous pat- terns. Ascarina lucida has a primitive looping pat- tern and semicraspedodromous tooth vasculariza- tion. The teeth of leaves of all three genera are well vascularized. There is some resemblance between these three genera and Hedyosmum subintegrum and Н. bra- siliense. Hed yosmum brasiliense in particular has anastomosing tertiaries. That is, the exmedial ter- tiary of the distal secondary fuses with the admedial tertiary of the next most basal secondary to form a craspedodromous vein that follows the course of other secondaries toward the margin. In general, all Hedyosmum species examined have well-organized secondaries, good areolation, high leaf rank, reduced teeth, and close secondary spacing. It is the only genus that has developed intramarginal veins and specialized crystal pat- terns. АП four genera have teeth within the same size range and vascularization having the same brochi- dodromous/eucamptodromous series. All leaves have generally decurrent basal laminas, acute to acuminate tips, and tertiaries that seldom form predictable patterns. INTERFAMILIAL RELATIONSHIPS Little detailed information on leaf architecture of other ranalian families is available, but we do have a well-illustrated guide to the Lauraceae (Klucking, 1987) and the Annonaceae (Klucking, 1986). Lorence (1985) illustrated leaf clearings of species of Monimiaceae from the Madagascar re- gion. An analysis of these wo of similarities with the Chloranthaceae and a num- ber of features that emphasize their lack of a close relationship. rks shows a number The Lauraceae generally resemble the Chloran- thaceae in having decurrent leaf bases, attenuate leaf tips, ovate leaf shape, and irregularly spaced secondary veins. Other resemblances are found among only some genera, and often only in certain species of these genera. These similarities include irregularly spaced, looped brochidodromous sec- ondaries (Licaria, Litsea, Aiouea, Ата); variable spacing and frequency of secondaries (Nectandra, Ocotea); forked secondaries (Beilschmiedia); eu- camptodromous secondaries (Cryptocarya); and tendencies toward intramarginal veins (Notho- phoebe, Ocotea). Major differences from the Chloranthaceae in- clude entire margins in all lauraceous genera, the presence of lateral primary veins in some species, the presence of often orthogonal tertiaries, a strong tendency toward percurrent tertiaries (Actonor- aphe, Alseodaphne, Beilschmiedia, Cinnamo- mum, Cryptocarya, Litsea, Mezilaurus, and Nec- tandra). Admedial vein endings of several orders are found in Beilschmiedia. Ascending acrodro- mous secondaries are found in Cinnamomum, Cryptocarya, Neocinnamomum, and Neolitsea. The synapomorphies found in the Lau- raceae indicate an ancient separation between the two families. While most phylogeneticists classify the Trime- niaceae with the Lauraceae, away from the Piper- ales (Cronquist, 1981; Takhtajan, 1987), Endress (1987) has called attention to numerous similarities between the Chloranthaceae and their proposed closest relatives, the Trimeniaceae. We compared specimens of 7rimenia from the herbarium of the Missouri Botanical Garden, but specimens of Pip- Trimenia neocale- Lindera, tocalyx were not available. donica Baker f. from New Caledonia has an en- tire-margined elliptic leaf of first leaf rank with a well-developed intramarginal vein. Secondary veins, averaging nine on each side of the midrib, are often branched, giving the secondaries a cladodromous appearance. This leaf shows little resemblance to any we examined in the Chloranthaceae. Trimenia papuana Ridl. from Papua, New Guinea, has a toothed elliptic leaf with about 28, often bifid sec- ondaries per side. They are regular in arrangement, and therefore at least second rank order, with one intersecondary usually found per intercostal area. The secondaries end short of the margin at an undulating intramarginal vein, which follows the outline of the tooth lobes and shallow sinuses. Veins leave the intramarginal vein in a semicraspedodro- mous pattern and appeared to run to the base of the sinuses just distal to the tooth. The tooth has a glandular apiculus not characteristic of any seen in the Chloranthaceae. Trimenia tooth architecture of our specimens does not seem to fit any of the chloranthoid types that we have identified. How- ever, Endress (1987, fig. 191) illustrated a cleared Volume 78, Number 2 1991 Todzia & Keatin 495 9 Leaf Architecture of Chloranthaceae tooth of the 7. papuana, which we would interpret as similar to our type 3 chloranthoid type. The opposite phyllotaxy and leaf base mor- phology of Trimenia are suggestive of the Chloran- thaceae, and in both families tooth vascularization always seems to arise from brochidodromous loops or collectors and never from secondaries directly. In other ways they are quite distinguishable. In the Chloranthaceae, the main tooth vascularization is always direct and never by way of the sinus. No eucamptodromous or brochidodromous patterns were found in the Trimenia specimens examined. In our limited observations of Trimenia, its leaves appear to have specialized independently and to a somewhat higher level than most Chloranthaceae. However, the postulated close relationship is not negated by our evidence. Lorence's (1985) study of the Malagasy Monim- iaceae reveals characters in those Lauralian genera that closely resemble characters found in the leaves of Chloranthaceae. Second rank, festooned bro- chidodromous leaves are found in Monimia and Decaryodendron with eucamptodromy found in M. ovalifolia. Widely spaced secondaries in craspe- dodromous or semicraspedodromous leaves are found in Ephippiandra and Tambourissa. Margins range from entire to dentate in the family. Dentate leaves are noted in Hortenia and Tambourissa but they have monimioid vascularization. That type does not resemble the Chloranthaceae in that the median vein servicing a tooth supports loops on both sides, which decrease in size toward the margin and tooth apex. In Tambourissa, species with en- tire adult leaves have dentate juvenile leaves. How- ever close the two families are later judged to be, it is clear that they share the same grade or level of specialization of leaf architecture. Klucking (1986) studied an extensive sample of the Annonaceae, which allows a detailed compar- ison with the Chloranthaceae. Again, it appears that the differences mitigate against postulating a close relationship. Similarities between the Chloranthaceae and the Annonaceae include an ovate to elliptic shape and usually eucamptodromous venation. Brochidodro- mous leaves are found in a few genera (Mez- zettia, Ártabotrys, and Marsypopetalum) but not mmonly. Decurrent secondaries are found only in Cleistopholis. Attenuate leaf tips are common in most genera. A few species from Klucking's illustrated Anona- ceae sample bear a close resemblance to Sarcan- dra and have a generally low rank. These include Annona reticulata L., Guatteria schomburgkiana C. Martius, Desmopsis mexicana R. E. Fries, Greenwayodendron suaveolens Engl. & Diels, Goniothalamus repevensis Pierre, Porcelia ma- crocarpa (Warm.) Fries, P. nitidifolia Ruiz Lopez & Pavon, Sageraea glabra Merr., Polyalthia sub- cordata Blume, and Oxandra riedeliana R. E. Fries. These specimens do not come from geo- graphically close жыена but rather from the trop- ical range of the Annonaceae. All of the specializations found in the Annona- ceae represent trends not present in the Chloran- thaceae. Among these are strongly percurrent ter- tiaries in most genera, the absence of teeth in all genera, close and regularly spaced secondaries, basal lamina insertion that is emarginate, blunt, incipiently cordate in species of Friesodielsia, Das- ymaschalon, and Ellipeia; strongly ascending sec- ondaries in Platymitra, Mitrephora, and Kings- tonia, obovate shape in Asimina, and emarginate leaf tips in Goniothalamus. Many more studies need to be made of ranalian families, o among the genera having low ran s. Important early dicot trends are abun- dant among ranalian genera, but we have little understanding of the systematic significance of their leaf architecture. LITERATURE CITED BURGER, W. С. 77. The ever and the monocots. Bot. Rev. (Lancaster) 43: -393. CRANE, P. 1989. co Be evidence on the early аве of E план d dicotyledons. Pl. Syst. Evol. 162: s & K. R. ID 1989. Re- produ ctive У си and function in Cretaceous Chloranthaceae. Pl. Syst. Evol. 165: 211-226. — - A. . Ап Integrated System of Classi- on of Flowering Plants. Columbia Univ. Press, New v Yor DAHLGREN, R. M. T. A revised system of pri 2 of the angiosperms. J. Linn. Soc., Bot. -124. Dovie, J. A. 1978. Origin of angiosperms. Ann. Rev. Ecol. Syst. 9: 365-392. & L.J. B. Beck (editor), Origin and Early Evolution of An- giosperms. Columbia Univ. Press, New York. Enpress, P. К. 1986. Reproductive structures and por significance of are primitive angio- sperms. Pl. Syst. Evol. ан -28. 198 . The Chlor "rad reproductive structures and phylogenetic seien Bot. Jahrb. Syst. 109: -226. . B. SaMPsoN. 1983. Floral structure and relationships of the Trimeniaceae (Laurales). J. Ar- nold Arbor. 64: 447-473. Fris, E. M., P. В. CRANE & К. К. PEDERSEN. 1986. Floral evidence for Cretaceous chloranthoid angio- sperms. Nature 320: 163-164. 496 Annals of the Missouri Botanical Garden FucHs, CH. 1963. Fuchsin staining with NaOH clearing for lignified elements of whole plants or plant organs. Stain Technol. 38: 141-144. HickEv, L. J. 1973. Classification of the architecture of dicotyledonous leaves. Amer. J. Bot. 60: 17-33 Stratigraphy and paleobotany of the Golden Valley Formation (Early Tertiary) of western North Dakota. Mem. Geol. Soc. Amer. 150: 1-183. 19 A revised classification of the archi- tecture of dicotyledonous leaves. Pp. 23-39 in C. R. Metcalfe & L. Chalk (editors), Anatomy of the Dicotyledons, Volume 1, 2nd edition. Clarendon Press, Oxford. J. A. DoYLE. 1977. Early Cretaceous fossil evidence _ ее evolution. Bot. Rev. (Lan caster) 43: 3- 9 Reexamination of leaf architectural characters of basal angiosperms and their sister groups: implications for the origin and relationships of angiosperms. r. J. Bot. 76 Ш 245-246. [Abstract.] OLFE. 1975. The bases of angio- sperm ano Ra vegetative morphology. Ann. Mis- souri Bot. Gard. 62: 538-589. . Notes sur le genre Ascarina (Chlor- anthaceae) en Nouvelle-Caledonie et a Madagascar. ansonia, ser. 2, 20: 273-285. KLUCKING, E. P. 1986. € Venation Patterns, Volume 1. Annonaceae. J. Cramer, Berlin. 1987. Leaf venation patterns, Volume 2. incas: J. Cramer, Berlin. Levin, С. A. . Systematic foliar morphology of | iar МЕ ан и Conspectus. Ann. Missouri Bot. Ga 3: 20-8 6b. B densi or morphology of Phyl- lanthoideae (Euphorbiacea etic analysis. nn. Missouri Bot. Gard. 73: 86-98. 1986c. Systematic foliar morphology of Phyl- Бића јеле (Euphorbiaceae) III. Cladistic analysis. Syst. Bot 515-530. LoRENCE, D. H. 1985. A monograph of the Monimi- aceae (Laurales) in the Malagasy region (Southwest Indian Ocean). Ann. Missouri Bot. Gard. 72: 1-165 Кику, P. M. & W. C. DICKISON. eaf venation patterns of the genus aee (Dilleniaceae). J. Arnold Arbor. ы 209-24 Swamy, В. С. L. 1953. А taxonomic revision of the genus Puts Гей Proc. Natl. Inst. Sci. India Outline of the classification of (Lan- 19: . TAKHTAJAN, A. 1980. flowering on (Magnoliophyta). Bot. Rev. caster) 46: 359. : ико Magnoliophytorum. Officina Editoria Nauka. Leningra THOMAS, В. A R. A. SPICER. 1987. The evolution and palaeobiology of land plants. Dioscorides Press, A phylogenetic classification of the angiosperms. Evol. Biol. 9: 35-106. A. 1988 e ga Hedyosmum. Flora Neotropica о 48: 1-138 UPCHURCH, С. R., №. 1984a. Cui evolution in Early Cretaceous angiosperms from otomac group of Virginia and Maryland. Ann. Missouri Bot. Gard. 71: 50. Cuticular anatomy of angiosperm leaves from the Lower Cretaceous Potomac Group. I. Zone I leaves. Amer. J. Bot. 71: 192-202. J. А. Wore. 1987. Mid-Cretaceous to Early Tertiary vegetation and climate: vs from fossil leaves and woods. Pp. 75-105 in E. M. Friis, G. Chaloner & P. R. Crane (editors), The Origin of TS and their Consequences. Cambridge Univ. Press, New VERDCOURT, B. 19 Notes on Malos Chlorantha- ceae. w Bull. 40 - . 1986. Chloranthac n: Steenis (editor), Flora Malesia I. a 12 3-144. WALKER, J. W. & A. G. WAL 984. Ultrastructure of Lower Cretaceous rea uid and the origin and early evolution of flowering plants. Ann. Missouri 64-521. Bot. Gard. 71: 464- NOTES ON THE SYSTEMATICS OF ARISTOLOCHIA SUBSECT. HEXANDRAFE Favio González G? ABSTRACT Using eade d ы ap studies from 123 species of Aristolochia sect. Gymnolobus subsect. Hexandrae, this taxon is reevaluated. Two series (Hexa nov., and are designated as types ndrae and Thyrsica He хае) are recognized. Aristolochia (Thyrsicae) maxima and Aristolochia (Anthocaulicae) iquitensis e, ser. nov.) and two subseries (Anthocaulicae, ser. The Aristolochiaceae are represented in Amer- ica by the genera Aristolochia, Asarum, Eugly- pha, and Holostylis. The noteworthy diversity and wide distribution of Aristolochia (250 species rang- ing from ca. 40°N latitude to ca. 35°S latitude) contrasts with the restricted distributions of Asa- rum (southeastern and northwestern United States) and the monotypic genera Euglypha and Holosty- lis (central South America). Most New World spe- cies of Aristolochia belong to Aristolochia sect. Gymnolobus Duchartre (1854, 1864). While working on a revision of Aristolochia in Colombia, I (Gonzalez, 1989) studied the compar- ative morphology of the 27 species reported from this country, as well as another 96 neotropical species. All species belonged to Aristolochia sect. Gymnolobus Duchartre subsect. Hexandrae Du- chartre (1854, 1864; lectotype here designated, A. ringens Vahl). Based on this study, it was pos- sible to recognize two new series and two new subseries into that subsection. The present treatment appears as an alternative to the classical works of the subsection made by Duchartre (1854, 1864), Masters (1875), Schmidt (1935), and Hoehne (1942). These authors con- sidered the characters of the perigonium to be essential for grouping species; such characters are unstable, however, and suggest that convergence or parallelism have occurred (Gonzalez, 1989). The new series and subseries are recognized b the correlation of characters related to: abscission basal zone in the petiole, whose appearance is here reported for the first time in the Aristolochiaceae; typology of inflorescences; and morphology of cap- sule and seeds. The concepts and terminology used here to describe inflorescences were developed by Mora-Osejo (1987), Troll (1950, 1964, 1969), and Weberling (1981, 1983, 1985). These new concepts made it possible to establish the basic types of inflorescences in Aristolochia and to apply these types to the recognition of natural groups. In order to increase the understanding of the following descriptions, a short explanation of un- familiar concepts related to the inflorescence mor- phology is desirable. Holocaulic plants (sensu Mora- Osejo, 1987) have stem systems (holocaules) with high innovation (branching) capacity, with vege- tative elements (i.e., bracts, buds, and leaves) in- cluded in the floriferous zone (Figs. 1, 2A, 3, 5A) or synflorescence (sensu Troll, 1950, 1964). In Aristolochia, the synflorescence is polytelic (lack- ing a terminal flower on the main axis or the lateral branches of various order); the growth of the shoots is indeterminate. The synflorescence is formed by the florescence (fl, distal floriferous unit of the polytelic main axis; Figs. 1, 3) and the coflores- cences (cfl, florescences of the lateral branches, or paracladia, pc; Figs. 1, 3). The florescence and the coflorescences become equivalent to the an- thotagma (sensu Mora-Osejo, 1987) or floriferous zone, subtended by a vegetative zone called ћу- potagma (sensu Troll, 1964); the hypotagma (hptg, s paper contains results from a thesis submitted as a partial requirement for the M.Sc. degree in Systematic ' This Botany, Institut e Ciencias Naturales, Universidad Nacional de Colo mbia. I am deeply grateful for the critical revision and ба! wise direction of Luis Eduardo Mora-Osejo (Universidad Nacional de d and Enrique Forero Missouri Botanical Garden). Mora Ose o and B. ? [nstituto de Ciencias Naturales; Universidad Nacional de Colombia, Ap. Ae. nara helped with the Latin descriptio 7495, posui: Colombia. ANN. MISSOURI Вот. GARD. 78: 497-503. 1991. 498 Annals of the Missouri Botanical Garden URE l. Diagram of a frondose, polytelic, thyrsoid holocaule, from Aristolochia ser. Thyrsicae; antg, we tagma; cfl, coflorescence; fl, florescence; fp, partial florescence; frd, frondulose pherophyll; frf, frondose gory js hptg, hypotagma; pc, paracladia of first (pe,), second (рс,) and third (pe,) order; yiv, innovation bud; тар, anthop acladial zone; zi, inhibition zone; ziv, enrichment or innovation zone. Figs. 1, 3) includes the inhibition zone (zi) and the innovation zone (ziv). In the species of Aristolochia ser. Thyrsicae, the partial florescences (ultimate elements of the florescence and the coflorescences) are rhipidia or helicoid cymes (Figs. 1, 2A, B); this implies that the synflorescence is thyrsoid. On the other hand, the synflorescence in the species of Aristolochia ser. Hexandrae subser. Hexandrae is racemose (botrytic) because each partial florescence is re- duced to a terminal flower (Figs. 3, 5A, B). In both taxa the pherophylls (vegetative elements that sub- tend each partial florescence) are leaves, and 7 synflorescence is therefore frondose (frf, Figs. , 3); however, in some higher-order paracladia z pherophylls become frondulose (frd, Figs. 1, ^: sensu Weberling, 1981). In contrast, the inflorescences in the ant | species of Aristolochia subser. Anthocaulicae restricted to lateral sh a les (sensu Mora-Osejo, 1987); they tend to be strongly ii mogeneous and specialized for the pe function. Each shoot is a short paracladium; T implies that it is truncate and has indeterminá hocaulic are called Volume 78, Number 2 1 fms ; МЕ 113 fi. Е Е i НЕ } ij } González G. Aristolochia subsect. Нехапагае 499 FIGURE 2. Вогеѕсепсе; br, bracteol latticelike dissepiment. D, E. Seeds. — D. Adaxial view. — growth. [ts vegetative impulse is not strong; its vegetative elements (buds, internodia, and leaves) ul extremely reduced; the shoot is unbranched T IS occupied totally by the coflorescence (the oo disappears); these shoots correspond to „16 paracladial anthoblast model (sensu Mora-Ose- Jo, 1987). These paracladia are restricted to a basal “one of the plant, here called the anthoparacladial Aristolochia ser. Thyrsicae. Diagram of diagnostic characters.—A. Synflorescence.—B. Partial е; zb, abscission basal zone of the petiole (zbp) and the peduncle (фра). — C. Capsule; dc, eds. — 1 1 . Abaxial view. zone (zap, Fig. 4). As in the species of Aristolochia subser. Hexandrae, these floriferous shoots are racemose, but each flower is subtended by a brac- teal pherophyll (fb, Figs. 4, 5F, С). The outline of the new taxonomic treatment here proposed is the following (the order of the taxa belongs with the typological derivation, discussed in González, 1989): 500 Annals of the Missouri Botanical Garden FIGURE 3. Diagram of a frondose, polytelic Hexandrae. (Abbreviations as in Fig. 1.) Aristolochia subsect. Hexandrae ser Thyrsicae ser. nov. ser. Hexandrae subser. Hexandrae subser. Anthocaulicae subser. nov. Aristolochia ser. Thyrsicae F. González, ser. nov. Hor Figures 1, 2. Plantae ове synflorescentia frondosa, thyrsi- a; paracladia interdum pseudocauliflora vel cauliflora, ad Cen даап reducta; florescentiae partiales ripidiis vel cymis helicoidalibus contormatae: Flores brac- OTYPE: Aristolochia maxima Jacq. I <. -antg=c #1 . racemose holocaule, from Aristolochia ser. Hexandrae subser. teolati. Zona dcn cdm d petioli et pedunculi. Dis- sepimenta capsulae cancellata. Semina transversim ob- longa, plana, bialata, s раж quam adaxialis brevior. Holocaulic plants, synflorescence frondose, thyrsoid, paracladia sometimes pseudocauliflorous or cauliflorous and then reduced to a partial flo- rescence; partial florescence a rhypidium or heli- coid cyme; flowers bracteolate. Petiole and pedun- cule with apsule with dissepiments latticelike. Seeds transverse-oblong, two-winged, abaxial wing shorter than adaxial. asal abscission zo Distribution: Argentina). Neotropics (southern Mexico to Volume 78, Number 2 González G. 501 1991 Aristolochia subsect. Hexandrae sr ST 347 DN Y — nom NEU ant Me B x E г gf pt fb— dE an E 4. Diagram of an anthocaule from Aristolochia ser. Hexandrae subser. Anthocaulicae; ap, paracladial anthoblast; fb, bracteal pherophyll; fp, partial florescence; nom, nomophyll; yiv, vegetative innovation bud; zap, anthoparacladial zone. The following species belong in Aristolochia ser. Thyrsicae: A. acutifolia Duchartre, А. colossifo- lia Hoehne, А. mathewsii Duchartre, А. maxima Jacq., А. melastoma Manso, А. ovalifolia Du- chartre, А. pannosoides Hoehne, А. pfeiferi Ваг- ringer, А. silvatica Barb. Кодг., А. sprucei Mast., A. tonduzii Schmidt, А. translucida Pfeifer, and A. trianae Duchartre. Aristolochia ser. Hexandrae. Figures 3-5. Holocaulic or anthocaulic plants, synflorescence racemose; partial florescence reduced to the ter- minal flower. Flowers ebracteolate (except in 4. grandiflora). Petiole and peduncle without basal abscission zone. Capsule dissepiments entire. Seeds rhomboid, ellipsoid, or from wide- to narrowly ovoid, unwinged or having one wing. Into this series are recognized two subseries, described below. Aristolochia ser. Hexandrae subser. Hexan- drae. Figures 3, 5A-E. Holocaulic plants; synflorescence with distal, frondose florescence and coflorescences; flores- cence and coflorescences sparsely flowered with elongate internodes, more than 1 cm long; pher- ophylls frondose, sometimes frondulose, petiolate, 10 mm long and Z6 mm wi Distribution: Neotropics and adjacent sub- tropical, temperate areas; the highest density of species occurs in the Neotropics. he species now recognized in Aristolochia ser. exandrae subser. Hexandrae are: A. albertiana 502 Annals of the Missouri Botanical Garden Th "ik A ү ч * dh ~ an A ири m 35 LI L4 5 : IDR Syn flores scence. -B Capsule; de, entire dissepiment. — D. E. Seeds. F-I. Aristolochia ser. Hexa Two partial florescences at the axil of one each bracteal meu (fb). — Н, I. Seeds (not С). to scale, except B an Ahumada, А. andahuaylensis Ahumada, А. an- guicida Jacq., А. angustifolia Cham., А. arcuata Mast., А. argentina Griseb., А. barbouri Barrin- ger, A. bilobata L., A. brasiliensis Mart. & Zucc., A. brevifolia (Cham.) Haum., 4. burchellii Mast., А. burkartii Ahumada, А. cabrerae Ahumada, А. castellanosii Ahumada, А. caudata Jacq., А. cere- sensis Kuntze, А. claussenii Duchartre, А. clavi- == LE IO TU TPTRPTVITIIIT EA EL ES or Ит, 54 e axil of a frondose (frf). — С. ae subser. Anthocaulicae. —F. denia Griseb., А. compta Hoehne, А. cordigera Willd. ex Duchartre, 4. curviflora Malme, A. cym- bifera Mart. & Zucc., А. chachapoyensis Ahu- mada, 4. chilensis Miers, А. deltoidea НВК, did yma S. Moore, А. ehrenbergiana Cham., А. ekmanii Schmidt, А. elegans Mast., А. eriantha Mart. & Zucc., А. esperanzae Kuntze, А. fim- briata Cham., A. fuertesii Urb., А. gabrielis Du- Volume 78, Number 2 1991 González G. 503 Aristolochia subsect. Hexandrae chartre ex Briq., 4. galeata Mart. & Zucc., А. gehrtii Hoehne, A. gibertii J.D. Hook., A. gigan- tea Mart. & Zucc., A. ginzbergeri Ahumada, А. glandulosa J. Kickz fil., A. glossa Pfeifer, A. grandiflora Sw., A. guianensis Poncy, A. haiten- sis Ekman & Schmidt, 4. hatschbachii Ahumada, Schmidt, 4. labiata Willd., А. leptosticta Urb., A. lindeniana Duchartre, A. lingua Malme, А. lingulata Ule, A. loefgrenii Hoehne, А. macror- rhyncha Hoehne, A. macrota Duchartre, A. mal- meana Hoehne, А. manaosensis Ahumada, А. marianensis Ahumada, А. melanoglossa Speg., А. mishuyacensis Schmidt, А. mossii S. Moore, A. nummularifolia HBK, A. odoratissima L., A. orbicularis Duchartre, А. papillaris Mast., А. passifloraefolia A. Rich., A. paulistana Hoehne, A. peltata L., A. peltato-deltoidea Hoehne, A. physodes Ule, A. pilosa HBK, A. pohliana Du- chartre, Æ. prostrata Duchartre, А. pubescens illd., А. pucarensis Ahumada, А. raja Mart. & Zucc., А. ringens Vahl, А. rugosa Lam., А. rumi- cifolia Mart. & Zucc., А. schreiteri Ahumada, А. schultzeana Schmidt, А. schulzii Ahumada, А. smilacina (Klotzsch) Duchartre, Æ. stomachoides Hoehne, А. taliscana Hook. & Arn., А. tentacula- ta Schmidt, А. tigrina A. Rich., A. triangularis Cham., А. trilobata L., А. truncata Fielding & Gardner, А. urupaensis Hoehne, А. хегорћупса R. E. Schultes, 4. warmingii Mast., and А. wed- dellii Duchartre. Aristolochia ser. Hexandrae subser. Antho- caulicae F. González, subser. nov. HOLOTYPE: Aristolochia iquitensis Schmidt. Figures 4, 5F-I Plantae anthocaulicae; synflorescentia truncata, ad an- thoblastos paracladiales, pseudocaulifloros vel caulifloros bracteosos reducta; anthoblasti paracladiales ad coflores- centiam saepe confertifloram reducti; internodia coflores centiae constricta, ad 1 cm longa. Phorophylla bracteosa sessilia vel subsessilia, usque ad 7 mm longa et 4 mm lata Anthocaulic plants; synflorescence reduced to paracladial anthoblasts, pseudocauliflorous or cau- liflorous, bracteate; paracladial anthoblasts reduced to the coflorescence (hypotagma absent), usually densely flowered, the internodes constricted, to 1 cm long; pherophylls bracteal, sessile or subses- sile, to 7 mm long and 4 mm wide. Distribution: Primary, wet tropical forests, es- pecially the Amazon basin. The following species are recognized in this sub- series: Aristolochia bukuti Poncy, А. cauliflora e, А. cordiflora Mutis ex HBK, А. cornuta ast., А. cremersii Poncy, А. flava Poncy, А. goudotii Duchartre, А. guentheri Schmidt, A. iquitensis Schmidt, А. Кири Schmidt, A. lage- siana Ule, A. leuconeura Linden, and A. ruiziana Duchartre. LITERATURE CITED DUCHARTRE, P. 54. Tentamen methodicae divisionis generis Aristolochia. Ann. Sci. Nat. Bot., ser. 4 - ^ —. у е In: А. De Candolle, ла ин к GONZALEZ, К. evisión taxonómica del género d n O en Colombia. Tesis ncias Naturales, Universi- dad Nacional de Colombia, Bo ota. 1875. ESE E EA In: C. F. P. von Marius: Flora Brasiliensis 4: 77-114, t. 17- 26. Мока-Озејо, L. E. 1987. Estudios morfológicos, au- toecológicos y sistemáticos en Angio spermas. a demia Colombiana de Ciencias Exactas, Fisicas y Naturales, Bogotá. Ѕснмірт, О. С. 1935. Aristolochiaceae. In: A. En & K. Prantl, = Natürlichen Pflanzenfamilien, ed. 2, 16b: 204- TROLL, W. 1950. bor den Infloreszenzbegriff und seine Anwendung auf die blühende Region Krautiger Pflan- zen. Akad. Wiss. Lit., Abh. Math. Naturw. Kl. Jg. 1950 (15): 377-415. . 64, 1969. Die Infloreszenzen spre. = fap ee in Aufbau а 6 2. Gustav Fisher Verlag, Jena & Stuttgart. WEBERLING, F. 1981. i u Blütenstände. Verlag Eugen U . 1983. Fundamental featur rescence к: Bothalia 14: 917 1985. Aspectos modernos de las idu de las inflorescencias. Bol. Soc. jdm Bot. 24(1- 2): 1-28. of 2 inflo- Peter Goldblatt? and James E. Henrich? CALYDOREA HERBERT (IRIDACEAE-TIGRIDIEAE): NOTES ON THIS NEW WORLD GENUS AND REDUCTION TO SYNONYMY OF SALPINGOSTYLIS, CARDIOSTIGMA, ITYSA, AND CATILA' ABSTRACT Described in 1843 by the British botanist Wil- liam Herbert, Calydorea was erected to include a species of bulbous and plicate-leafed Iridaceae from Chile, Sisyrinchium speciosum, which was mis- placed in Sisyrinchium. Two features distinguished Calydorea from other genera of lridaceae then | h || had a hh L and plicate leaves, and the same basic vegetative morphology: free stamens; and a style with long, slender, un- divided branches loosely alternating with the sta- Ву 1843, 11 genera with similar vegetative morphology had already been described, and more than 15 additional ones have since been added. Except for some obvious synonyms, each differed largely in its floral morphology and was charac- terized by unusual stamen and/or stylar special- izations, and sometimes by differences in the dc ative size and shape of the inner and outer tep whorls. The bulbous and plicate-leafed genera of Iridaceae are now treated as members of Tigri- dieae, one of four tribes (Goldblatt, 1990) of Iri doideae, the others being Sisyrinchieae, Irideae, and Mariceae. Iridoideae is the only one of the „оо Иа t BE И ' Support for this study by grants DEB 81-19292 and BSR 85-00148 from the U.S. National Science Foundation is gratefully acknowledged. W our cladogram using the HENNIG86 program. 2 е also thank C. J. Humphries for his critical review of the manuscript and for confirming B. A. Krukoff Curator of African Botany, Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166- 0299, U.S.A * Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166-0299, U.S.A. ANN. Missouri Bor. GARD. 78: 504-511. 1991. Volume 78, Number 2 1991 Goldblatt & Henrich Notes on Calydorea 505 four subfamilies of Iridaceae that occurs in the New World (Goldblatt, 1982, 1990). A notable feature of Tigridieae (Figs. 1, 2) is the unusual degree of variability in the style branches, which frequently form a highly specialized and complex structure, intimately associated with the stamens, which are also sometimes specialized. This con- trasts markedly with the vegetative uniformity in the tribe. More than 25 genera of plicate-leafed Iridaceae with bulbs have been described, and all but a few of these appear to be very weakly founded. They are largely based on trivial differences in style or stamen structure. Compared with Iridaceae—Iri- doideae in Africa and Eurasia, the generic limits of the Tigridieae seem too finely drawn, and it seems timely to attempt to reduce some of the genera that differ so little from each other. Not only is some degree of consistency in the criteria for recognition of genera desirable at least within the family, but there is a need to define the New World genera of Iridaceae so that usable keys and descriptions can be written. Cardiostigma Baker (1876), Salpingostylis Small (1931), Catila Ra- venna (1983), and /tysa Ravenna (1986) are four such examples of weakly defined genera, and they are reduced to synonymy in Calydorea here. Also critically considered in relation to Calydorea are Gelasine, Onira, and Eleutherine, which are at least partly defined by, or contain species with, slender, undivided style branches. CARDIOSTIGMA Based on the southern Mexican C. longispa- thum (Klatt) Baker, Cardiostigma (Baker, 1876: 102) was reduced to synonymy in Sphenostigma Baker (1876: 124) by Bentham & Hooker (1883) soon after it was described. The latter is founded on the Brazilian S. sellowianum (Klatt) Baker. Baker (1892) expanded Sphenostigma to include 11 species, and Foster (1945) added three more. Ravenna (1977) has, however, made a convincin argument for transferring the type of Sphenostig- та, but not other members of the genus, to Ge- lasine, in which it is now known by an earlier Synonym, G. coerulea (Vellozo) Ravenna. With Sphenostigma thus included in Gelasine, Ravenna (1977, 1979, 1986) transferred most of the erst- While species of that genus to Ennealophus N.E. Br., Phalocallis, Cardiostigma, and the new gen- ега Ainea and Lethia. One species that Ravenna has not dealt with to date is Sphenostigma coe- lestinum (= Salpingostylis coelestinum), and this is discussed in detail below. Cardiostigma sensu Ravenna comprises three species, C. longispathum (the type), C. mexican- um (К. Foster) Ravenna, and C. hintonii (В. Fos- ter) Ravenna. Ravenna never explained his concept of Cardiostigma, but its only notable feature seems to be its flower. This is secund and consists of a blue perianth of subequal spreading tepals, not clearly divided into a limb and claw and without perigonal nectaries; free stamens; and a slender, eccentric style with three short, simple branches. This at least applies to C. longispathum and C. mexicanum. In C. hintonii the three well-defined style branches are each deeply forked for about two-thirds of their length. This style corresponds exactly to that of Ainea (Ravenna, 1979) and Alophia (Herbert, 1840), and we assume that it is related to these genera rather than to the other species of Cardiostigma. We will not deal further with C. hintonii here. SALPINGOSTYLIS This genus is monotypic, containing only the Florida species Salpingostylis coelestinum (basio- nym /xia coelestina Bart.), which was, as recently as 1974, regarded as a not particularly unusual member of Sphenostigma (Goldblatt, 1974). Small (1931) described Salpingostylis and regarded it as monotypic, but his opinion was largely ignored (Foster, 1945; Goldblatt, 1974). Only after 1977, when Ravenna reduced Sphenostigma to synon- ymy in Gelasine, has Salpingostylis, more or less by default, been recognized. A careful comparison of the available collections of Salpingostylis апа Cardiostigma sensu Ra- venna leaves us in no doubt that, with the exception of C. hintonii (mentioned above), these are con- generic. Salpingostylis coelestinum has the iden- tical flower described above for C. longispathum and C. mexicanum, secund in orientation with sub- equal tepals, free stamens, and a long style. Ex- amination of ample herbarium material and some accompanying illustrations made from live plants also makes it clear that the style is divided into distinct branches, not obscurely lobed as is some- times thought (Fig. 1A). A reading of the literature and examination of living and preserved specimens have not indicated any reason to separate Salpin- gostylis and Cardiostigma from one another, and little to distinguish them from Calydorea (Fig. 1 B). ITYSA Described by Ravenna in 1986, /ty pri two species of northern South America. Both are broadly similar to Calydorea and Cardiostigma in 506 Annals of the Missouri Botanical Garden FIGURE 1. Flowers coelestina (Salpingostylis) (with habit). — B. Calydorea sp. — , style, and stamen detail and enlargement of style apex in species of Calydorea. — A. C. 7. gardneri (Itysa). — D. C. pallens. —E. C. amabilis (Catila). Habit and flowers full size; stamens and style x 3; style branch apices much enlarged. vegetative morphology, and they have secund flow- ers with subequal inner and outer tepals lacking differentiation into a limb and claw. The filaments are free, and the long style divides above the level of the anthers into three short, slender branches (Fig. 1C), the structure of which closely resembles that in Calydorea. One peculiarity of the genus is that the bases of the anthers are coherent around the style. Except for this unusual feature, both species of /tysa seem to correspond to the species of Cardiostigma in particular. CALYDOREA Flowers of Cardiostigma, Salpingostylis, and Itysa correspond closely to the type and most other Моште 78, Митбег 2 1991 Goldblatt 4 Henrich Notes on Calydorea 507 species of Calydorea (Fig. 1B) with the following differences. The flower is upright not secund, and the style is shorter, dividing opposite the base of the anthers. The style branches are thus compar- atively longer and more slender, and by a weak twisting lie loosely between the anthers. (This is not a fundamental difference in the position of the ovary or style branches in relation to the stamens.) In Salpingostylis and Cardiostigma, the style di- vides just above the apex of the anthers and the exact position of the style branches relative to the anthers is obscured because the style is weakly eccentric (whether by design or the effect of gravity in the secund flower is not known) (Fig. 1A). How- ever, the structure of the style and style branches of Salpingostylis, Cardiostigma sensu Ravenna, Itysa, and Calydorea is very similar (compare details of style arms in Fig. 1A-D). As far as we can tell from live specimens of a Calydorea sp. from Argentina, and an illustration of the closely related C. xiphioides (Poeppig) Es- pin. (synonym C. speciosa), there are no other taxonomically significant differences between Caly- dorea, Cardiostigma, Salpingostylis, and Itysa. The secund flower, longer style, short style branch- es, and basally connate anthers are inadequate criteria on which to base genera in Iridoideae. Thus, as presently understood, Cardiostigma, Salpin- gostylis, and Пуза cannot be upheld and are here reduced to synonymy. The minor differences that we have outlined, in conjunction with the correlated geographic disjunction, appear to merit sectional separation within one genus at best. The necessary nomenclatural changes are presented below, to- gether with short notes about each species. CALYDOREA PALLENS AND САТА RAVENNA When other species currently assigned to Caly- dorea are taken into account, the circumscription of the genus is somewhat expanded. In C. pallens Griseb., the style is shorter than in C. xiphioides and Calydorea sp., and the branches lie opposite the anthers, though not appressed to them (Fig. 1D); the tepals are weakly unguiculate; and the inner tepals are smaller than the outer. Calydorea pallens bears a remarkable resemblance to Catila amabilis Ravenna (Fig. 1E), the type and only species of Catila (Ravenna, 1983), which differs significantly from it only in having a larger flower with more strongly unguiculate tepals of a darker purple color and in the anthers being closely applied to the style branches. The inner tepals of Саша resemble many other Tigridieae in having the base of the limb folded back on itself and bearing a median zone of nectariferous tissue. There is no basis for the generic separation of Calydorea pal. lens and Catila. PHYLOGENETIC CONSIDERATIONS The basic floral morphology in Tigridieae ap- pears to be similar to the complex structure en- countered in Cypella (e.g., Ravenna, 1981a, b), which is very similar to the basic flower of Mariceae (Fig. 2A), sister taxon of Tigridieae (Goldblatt, 1990). Here the tepals are clawed with the claws forming an open cup, and the inner tepals are smaller than the outer, each folded back on itself at the base of the limb, in which area there is a zone of nectaries surrounded by an area of con- trasting coloration. The style branches are thick- ened and compressed, and terminate in a pair of flat appendages, called crests because of their re- semblance to similar structures in /ris. The fila- ments are free, usually threadlike, and too weak to support the anthers, which are closely applied (not united) to the abaxial side of the style branch. The anthers usually reach to just below the stigma, a transverse band of tissue (Fig. 2A) at the base of the crests on the abaxial surface of the style branch. (The stigma may bear a pair of small appendages similar to the crests.) This complex structure is thought to be basal (plesiomorphic) for a tribe in which much simpler stylar structures also occur (as in Calydorea) be- cause the flowers of the central genera of Mariceae, Neomarica and Trimezia, have exactly the same structure. The genera of Mariceae, however, have a rhizomatous rootstock and plane leaves, and these features are less specialized than the bulb and pli- cate leaves of Tigridieae. In the other New World tribe of Iridoideae, Sisyrinchieae, which also has a rhizomatous rootstock and plane leaves, the inner and outer tepals are similar, not normally unguic- ulate, lack a zone of nectaries, and have style branches that are always simple, filiform structures sometimes reduced to short lobes. Thus, the pattern of floral evolution in Tigri- dieae, as presented in the cladogram (Fig. 3), is best seen as a progressive reduction from the com- plex to the simple (Figs. 1, 2). Goldblatt (1990) presented this argument in detail. To consider the phylogenetic relationships as having a different pat- tern, from the simple to the complex, is tempting. It is, however, less parsimonious as this means that either the exact type of flower structure evolved independently in Mariceae and Tigridieae, or that Mariceae are derived from supposedly specialized 508 Annals of the Missouri Botanical Garden FIGURE 2. denen Hu ie stamen vin and enlargement of style pe apex in selected Iridaceae.— A. ure р. Trimezia steyermar ra unguiculat Tigridieae and acquired a rhizome and plane leaves secondarily. Both scenarios seem unlikely. In Iri- deae, a similar progression from the complex type of flower to the simple seems likely (Goldblatt, 1981, 1986, 1990) and the same pattern seems to have occurred in Cipura, a small genus of Ti- gridieae (Goldblatt & Henrich, 1987). In the scenario outlined above, Calydorea ap- pears to be a moderately specialized member of Tigridieae. It most likely evolved from a Cypella- like ancestor in which the style branches lost their thickening and apical appendages, and differences between the inner and outer tepals were reduced (Fig. 3). The flower of Catila amabilis is inter- preted here as being close to this ancestral type, and Calydorea pallens represents another step in . Gelasine elongata (= G.a Habit and flowers full s size; ise and s tyle > x (3 style branch apices much eee. Eleutherine bulbosa. this process, leading to species with flowers like С. speciosa, and ultimately to such Northern Hemi- sphere species as C. coelestina and C. longispa- tha, with their secund flowers and elongate, ec- centric style (Fig. 3). Other genera of Tigridieae with simple style branches resembling those of Calydorea are Eleutherine Herbert (1843b), which comprises two species; one species of Gelasine Herbert (1840), С. elongata (= G. azurea); and Onira Ravenna (1983). Regarding Eleutherine (Fig. 2D), we feel that it is unrelated to Calydorea. Its unusual red- dish bulb, large subterminal cauline leaf, and short inflorescence spathes seem to isolate it from pos- sible relatives. The reduction of this genus should be considered when its immediate ancestry can be Volume 78, Number 2 Goldblatt & Henrich 509 1991 Notes on Calydorea © © e o? о < + 11 anthers basally 14 style eccentric c oherent MARICEAE ES 2 —T-12 anthers alt. 6 flower secund style arms from anthers 10 stamens united 7 tepals subequal unresolved branching involving Cypella, Cipura d perhap h 5 flowers blue TIGRIDIEAE 9 lacking nectaries 1 bulb 2 leaf with pseudomidrib 3 leaf plicate 8 claws laxly spread filaments weak style branches thickened inner tepals with nectaries oil glands FIGURE 3. Phylogeny on ~ genera and species closely allied to Calydorea suggested by the cladogram based the e list and x presented in Table 1. The major branching patterns in Tigridieae and sister ibe Mariceae (from Goldblatt, 1990) are indicated to give some perspective to ар Onira—Calydorea clade. Polarity of unnumbered characters from Goldblatt (1990). E 1. Data matrix and character list for the cladogram (Fig. 3). Presence of the specialized condition is Mene by 1; absence by 0; state inapplicable by 2 Character number Genus 1 2 3 4 5 6 8 9 10 11 12 13 14 15 Мапсеае 0 | 0 0 0 0 0 0 0 0 0 0 0 0 0 Calydorea | 0 1 1 1 0 1 ? 1 0 0 1 1 0 0 С 1 0 1 1 1 0 0 1 1 0 0 0 1 0 0 Salpingostylis 1 0 1 1 1 1 1 ? 1 0 0 0 1 1 1 Cardiostigma | 0 1 1 1 1 1 ? 1 0 0 0 1 1 1 Itysa 1 0 1 1 1 1 1 2 1 0 1 0 1 0 1 Саша 1 0 1 1 1 0 0 1 0 0 0 0 1 0 0 Onira 1 0 1 1 1 0 0 0 0 0 0 0 1 0 0 Cypella 1 0 1 1 0 0 0 0 0 0 0 0 0 0 0 other genera, e.g., Gelasine, Herbertia 1 0 1 | 0 0 0 0 0 1 0 0 0 0 0 Characters: the еа (apomorphic) states listed first followed by the presumed ancestral (plesiomorphic) condition. l. rootstock а bulb—rootstock a rhizome; 2. leaf with pseudomidrib— leaf without pseudomidrib; 3. leaf plicate— leaf plane; 4. basic ке number x = 7 —base number not x = 7 (but unknown); 5. flowers kgs of blue— flowers dn of yellow; 6. flower зесипд — flower upright; 7. tepals subequal and not clawed— tepals unequa al and clawed; 8. tepal же اا‎ pae ене sear an open cup; 9. tepals lacking nectaries—zone of nectaries present on inner tepals; united — ns free; 11. anthers coherent at base —anthers separate from one another; 12. anthers к ен аңа ones the style шы tia opposite the style arms; 13. style branches slender and not forked —style arms thickened and apically forked; 14. style eccentric —style central; 15. style branches distant from the anthers— style branches at same level as anthers. 510 Annals of the Missouri Botanical Garden identified. The subterminal cauline leaf, rather fleshy bulb scales, and basic chromosome number, x = 6, suggest that it is possibly a derivative of Ge- lasine. Base number in Tigridieae is x — 7 (Gold- blatt, 1982), and apart from Eleutherine, x = 6 is known in the tribe only in G. elongata (reported as G. azurea) (Kenton & Rudall, 1987). The sim- ilarity in the karyotypes in these two is striking and may indicate a relationship. Gelasine elongata (Fig. 2C) has united filaments and is a large plant, resembling С. coerulea and G. uruguaiensis vegetatively. These two species also have united filaments but more elaborate style branches that have paired crests and a broad trans- verse, abaxial stigma (Ravenna, 1984). Gelasine elongata is an autogamous, complex heterozygote (Kenton & Rudall, 1987) with a reduced basic chromosome number, n — 6, unlike most Tigri- dieae, which have x — 7. It seems likely that the apparently simple flower and style of G. elongata are the result of reduction from the more complex structures present in other species of Gelasine. Chromosome number is known in only one other Gelasine, С. uruguaiensis, n = 7 (Ravenna, 1984). Presumably, Gelasine evolved from a Cypella-like ancestor, and the major distinction between the two genera is in the filaments, free and long in Cypella, short and united in Gelasine. We assume that similar patterns of floral evolution occurred in Gelasine and Calydorea. In both genera species have evolved with flowers with subequal unclawed tepals and simple cylindric style branches lacking crests. ONIRA Described in 1983 by Ravenna, Onira includes only O. unguiculata (Baker) Ravenna, based on Herbertia unguiculata. It is a small plant with large, pale blue-green flowers (Fig. 2B). Apart from the color and size of the tepals, and the well- developed tepal claws, the flower is unusual only in the short style, which divides below the middle of the filaments, and the comparatively long, slen- der style branches (Fig. 2B detail). The filaments are free and the anthers are appressed to the style branches and are closely held to them by the re- curving of the anther sacs after anthesis. The style branch apices resemble fairly closely those of Caly- dorea pallens and Catila except for the small, irregular outgrowths of nonstigmatic tissue on the stigma lobes (Fig. 2B detail). These may represent the vestigial remains of the style crests that are so prominent in, for example, Cypella, and hypoth- esized as plesiomorphic for the tribe. The tepal whorls are well differentiated, the inner being much smaller than the outer, and strongly folded back at the juncture of the limb and claw. This differ- entiation is much more pronounced than in Саша. We feel that Onira and related genera should be studied further before a decision is made regarding their status and disposition. SYSTEMATIC TREATMENT Кез. те velud. Bot. Reg. 29, Misc. Matter.: .18 YPE: C. speciosa (Hook.) Herbert n C. dine (Poeppig) Espin.). ул Вакег, Ј. Вог. Гопдоп 14: 188. 1876. : C. longispathum (Herbert) Baker (= Caly- P a longispathum (Herbert) Baker). Satpingosiyiis Small, J. New York Bot. Gard. 32: 161. E: S. coelestinum TA ) Small (= Caly- i. rea iind. (Bartr.) G Catila Ravenna, Nordic J. Bot. ҮП ТҮРЕ: m amabilis avenna (= С. amabilis Ravenna) Goldb Itysa Ravenna, Nord. : 582. en TYPE: /. gardneri (Baker) Ravenna (= Calydorea gardneri aker). NOMENCLATURE OF SPECIES OF CALYDOREA DISCUSSED IN THIS PAPER, INCLUDING NEW COMBINATIONS 1. Calydorea coelestina Е ) Goldbl. & Hen- rich, comb. nov. BASIONYM: /xia. coelestina Bartr., Trav. ed. 1, Philadelphia. 155 (1791) [ed. 1, London, 153 & pl. 3 (1792)]; Sal- pingostylis coelestina (Bartr.) Small, J. New York Bot. Gard. 32: 161 (1931). [See Gold- blatt (1974) for additional homotypic syn- onyms. | TYPE: Illustration in Bartram, Travels, pl. 3. 2. Calydorea longispatha (Herbert) Baker, J. Bot. London 14: 189 (1876). BASIONYM: Ge- lasine longispatha Herbert іп Benth., Pl. Hartweg. 53 (1840). Cardiostigma longispa- tha (Herbert) Baker, J. Linn. Soc. Bot. 16: 102 (1878); Sphenostigma longispathum (Herbert) Benth. & Hook., Gen. Pl. 3(2): 695 (1883). TYPE: Mexico. Hartweg 403 (BM, C) [See Foster (1945: 12) for additional homo- typic synonyms. ] 3. Calydorea mexicana (R. Foster) Goldbl. & Henrich, comb. nov. BASIONYM: Sphenostig- a mexicanum R. Foster, Contr. Gray Herb. 155: 13(1945); Cardiostigma mexicana (R. Foster) Ravenna, Notic. Mens. etc. (1978). [Additional homotypic synonyms in Foster (1945).] ТҮРЕ: Mexico. Mexico: Temascalte- pec, Carboneras, Hinton et al. 8010 (holo- type, GH; isotypes, Е, MO, NY, US). Volume 78, Number 2 1991 Goldblatt & Henrich 511 Notes on Calydorea A у pedes gardneri Baker, J. Bot. London 8. 1876. я gardneri m Ra- venna, Nord. J. . 6: 582. 1986. TYPE: Brazil. Раш: Boa nt uiis 2322 (holotype, BM not seen). [Known to us from illustrations in Ravenna (1986).] . Calydorea venezolensis (Ravenna) Goldbl. & Henrich, comb. nov. BASIONYM: /tysa ve- nezolensis Ravenna, Nord. t. 6: 584 1986. TYPE: Venezuela. Guarico: Ravenna 2178 (holotype, Herb. Ravenna not seen). Known to us only from the illustrations in the protologue. ] Calydorea amabilis (Ravenna) Goldbl. & Henrich, comb. nov. BASIONYM: Catilla ama- bilis Ravenna, Nord. J. Bot. 3: 197. 1983. TYPE: Argentina, Entre-Rios, Ravenna 3 (ho- lotype, Herb. Ravenna not seen). [Known to us from the illustrations in the protologue and from living material. ] л ^ LITERATURE CITED T ч i 1876. се Iridacearum. J. Linn. Soc. : 61-18 p Handbook of the Irideae. George Bell ve Sons == С. “р. Hooker. 1883. Genera Plan- m 3 e ы 1945. = in Iridaceae, — III. Contr. : 3- GOLDBLATT, P. 1974. тш of the bulbous Iridaceae of North America. Brittonia 27: 373-385. 1977. Herbertia (Iridaceae) reinstated as a valid generic name. Ann. Missouri Bot. Gard. 64: 378-319. 981. Systematics and biology of Homeria (Iridaceae), Ann. Missouri Bot. Gard. 68: 413-503. 19 Chromosome cytology in relation to н е. aroma of neotropical Iridaceae. 8. Syst. Вог. 7: 186- г. Convergent evolution of the Homeria flower type in six new species of Moraea (Iridaceae- Irideae) i in southern Africa. Ann. Missouri Bot. Gard. 73: 102-116. 1990. Phylogeny and classification of Irida- ceae. Ann. Missouri Bot. Gard. 77: 627. & J. E. HENRICH. 1987. Notes on Cipura (Iridaceae) in South and Central America, and a new species from Venezuela. Ann. Missouri Bot. Gard. 74: 333-340. HERBERT, W. 1840. Gelasine azurea; Calydorea. Bot. Mag. 66: t. 3779. 1843a. жыз" не Edwards's Bot. Reg., Misc. Matter 29: 1843b. E ho HE anomala. Edwards's Bot. Reg. 29: t. 57. KENTON, А. & P. RUDALL. 1987. An unusual case of complex heterozygosity in Gelasine azurea (lrida- ceae), and its implications for reproductive biology. Evol. Trends Pl. 1: 95-103. MoLsEED, E. 1971. The genus Tigridia (Iridaceae) of Mexico and Central America. Univ. California Publ. Bot. 54: 1-113. RAVENNA, Р. 1977. Notas sobre Iridaceae V. Not. Mens Mus. Nac. Hist. Nat. (Santiago) 21(249): 7-9. 979. Ainea, а new genus of Iridaceae from Mexico. Bot. Not. 132: 467-469. 1981a. A submerged new species of Cypella (Iridaceae), and a new section for the genus (s. str.). N . Bot. 1: 489-492. 981b. Eight new species and two new sub- species of Cypella (Iridaceae). Wrightia 7: 13-22. ——— 1983. Саша and Onira, two new genera of South American Iridaceae. Nord. J. Bot. 3: 197- 205 1984. The delimitation of Gelasine (Irida- ceae), and G. uruguaiensis sp. nov. from Uruguay Nord. J. Bot. 4: 347-350. 1986. Пуза and Гега, two new genera of neotropical Iridaceae. Nord. J. Bot. 6: 581-588 SMALL, J. Е. 1 Bartram’s Ixia Appr redis- covered. J. New York Bot. Gard. 32: 155- CHROMOSOME NUMBERS IN THE GENUS NASSAUVIA (COMPOSITAE-MUTISIEAE)' . Hunziker,’ А. Е. Wulff? о D. Ј. Crawford,* and . Cris ABSTRACT Meioti c chromosome numbers were determined (n been investigated cytologically: N. axillaris, №. chubutensis, М. glome 11) and М. aculeata (with n = ca. 44) ) and n — ca. 44 (8 uniflora (all with п = or the genus: n = 11 (2x), n = 22 (4x cytological data are yet available for t hem. The most closely related genus for which chromosomal data = 11 and ca. 44; x = 11) and observations made for eight . The ии reveal three known chromosomal levels are available is Moscharia with n = 20, which suggests that it and Nassauvia could be based on x = 11 (or perhaps x = The genus /Vassauvia Comm. ex A. L. Juss. of the Compositae belongs to the tribe Mutisieae, sub- tribe Nassauviinae. It comprises 38 species (Ca- brera, 1982; Arroyo & Marticorena, 1988) dis tributed from southern Bolivia along the Andes to southern Patagonia, Tierra del Fuego, and the Mal- vinas (Falkland) Islands. The genus demonstrates a trend toward reduction of numbers of flowers in each capitulum and aggregation of capitula into secondary inflorescences, or pseudocephalia. The cytology of this genus has been poorly stud- ied. Chromosome numbers are known only from four (10%) of the species. The following numbers have already been documented: n — / darwinit, У. magellanica, and N. serpens; and n = 22 for N. gaudichaudii and N. magellanica (Moore, 1967, 1981; Crisci, 1974b; Cabrera, 1982). The present study presents chromosomal data for additional species of the genus and com- ments on their systematic and evolutionary signif- icance. MATERIALS AND METHODS Meiotic investigations were made from young inflorescences fixed in modified Carnoy's solution (4: 3 : 1, chloroform : ethanol : acetic acid) or mod- ified Newcomer's solution (Hunziker, 1966). The latter proved superior. Root tips were squashed in 45% propionic acid haematoxylin (2%) with ferric citrate (1%) used as mordant (Saez, 1960; Nunez, 1968). Preparations were mounted in Euparal after removal of the coverslip by freezing with carbon dioxide. Voucher specimens are deposited at LP. The sources of the materials collected in Argentina are presented in Table 1 RESULTS AND DISCUSSION In the diploid species (V. axillaris, N. chubu- tensis, N. darwinii, N. glomerulosa, N. lagascae, М. revoluta, and N. uniflora) meiosis was regular with 2n = ИП (Fig. 1A, C-H). In contrast, N. aculeata (Fig. 18) was found to be octoploid with n = ca. 44 bivalents. Except for V. darwinii, all of these species are reported here for the first time. With these new counts, 11 species (29%) of Nas- sauvia, including species from each section, have been analyzed cytologically. The basic number is clearly x — 11, with three ploidy levels: 2x, 4x, and 8x. This basic number is also known in other genera of the Nassauviinae, e.g., Holocheilus Cass. ' Appreciation is expressed to: recep for financial support to JHH and JVC; the National Science p Dune P (Grant INT-8402888) to TFS and DJC; — Científico, and AFW to the Persona ] de . Gentile for help i Moreno, and J. Zech for additional assistance in field studies. JHH and JVC belong to the Carrera del се А о a la Investigacion Cientifica (CONICET). ‚ Зап Isidro 1642 an Universidad de in field ПАША and M. M Cigilano, P. Gent Argentina. Buenos Aires, 1428 Buenos Aires, Argentina. „А. ' Department of Plant Biology, The Ohio State University, Columbus, Ohio 43210, U.S * División Plantas Vasculares, Museo de La Plata, 1900 La Plata, Argentina. ANN. MISSOURI Bor. Garp. 78: 912-515. 1991. Volume 78, Number 2 Hunziker et al. 513 1991 Chromosome Numbers in Nassauvia TABLE 1. New chromosome counts of species of Маззашта. Vouchers on deposit at LP, OS, and SI. Num- ber of Chromosome cells Taxon count studied Location and voucher! N. aculeata (Less.) Poeppig n — ca. 44 3 Prov. Chubut, Dep. Futaleufü, 45 km SE de Esquel, var. robusta (Cabrera) Ca- sobre ruta 40, hacia Tecka, 4 Dec. 1984, TFS, DJC, brera JVC, MMC & MG 6842 N. axillaris (Lagasca) D. Don п = 11 3 Prov. Chubut, Dep. Гапршпео, 37 km al E де Tecka, ruta prov. 63 hacia Quichaura, 6 Dec. 1984, TFS, JC, JVC, MMC & MG 6868 N. chubutensis Speg. п = 11 6 Prov. Chubut, Dep. Mártires, 9 km al S de Las Сћараз (ruta 31), 8 Dec. 1984, TFS, DJC, JVC, MMC & 6946 М. darwinii (J. D. Hook. & Arn.) п= 11 24 Prov. Chubut, Dep. Languineo, ca. 6.5 km E de Lago O. Hoffm. & Dusen Vinter, 7 Dec. 1984, TFS, DJC, JVC, MMC & MG 4 N. glomerulosa (Lagasca) D. п= 11 18 Prov. Chubut, Dep. Languineo, 10 km E de Tecka, ruta Don hacia Quichaura, 6 Dec. 1984, TRS, DJC, JVC, MMC & MG 6863 N. lagascae (D. Don) F. Mei- n = 11 12 Ргоу. € za, Ll San Carlos, Borde SW de Laguna Diam , 3,370 m, 27 Jan. 1985, JHH, AFW & NGG pres N. revoluta D. Don n=11 {4 Prov. Chubut, Dep. Futaleufú, Cerro La Hoya, 15 km NE de Esquel, 5 Dec. 1984, TFS, DJC, JVC, MMC & MG 6857 N. uniflora (D. Don) Hauman п = 11 15 Prov. Mendoza, Dep. Las Heras, alrededores Laguna Los Horcones, 3,180 m, 16 Jan. 1985, JHH, AFW & NGG 11163 N. uniflora (D. Don) Hauman n=11 24 Prov. Mendoza, Dep. Las Heras, Laguna Los Horcones, 3,050 m, 17 Jan. 1985, JHH, AFW & NGG 11183 [Fig. 1А] N. uniflora (D. Don) Hauman п=11 4 Prov. Mendoza, Dep. Las Heras, Playa Ancha, Feci del Tolosa, 3,660 m, 18 Jan. 1985, JHH, AFW & NGG 11195 N. uniflora (D. Don) Hauman n = 11 2 Prov. Mendoza, Dep. Luján, Las Cuevas, 3,160 m, 19 Jan. 1985, JHH, AFW & NGG 11205 ! JHH, AFW, and NGG = J. H. Hunziker, А. Е. Wulff, and N. hs oo weis TFS, DJC, JVC, MMC, and МС = T. F. Stuessy, D. J. Crawford, J. V. Crisci, M. M. Cigliano, and M. (п = 11; Coleman, 1968; also reported with n = the ploidy jump occurred (e.g., Poggio & Hunziker, 18, Crisci, 1976) and Marticorenia Crisci (n = 1986). Of all the species of Nassauvia, М. acu- 22; Crisci, 1974a). However, these two genera are leata is particularly variable morphologically and not included in the same evolutionary cluster in has a wide geographic distribution. Taxonomically the phylogenetic reconstructions of the subtribe by — this has resulted in the recognition of three varieties Crisci (1974b, 1980). (Cabrera, 1982). The wide range of №. aculeata Although most of the species of Nassauvia are тау be due at least in part to its polyploid condition diploids, three E are known. Nassauvia and genetic diversity. More populational sampling aculeata is n — 4 (octo plaid]. The other is needed to determine if infraspecific chromosomal counted species E sect. aa are diploid races are present, such as documented in N. ma- (n = 11; N. darwinii and N. uniflora). The chro- gellanica J. Gmelin (see below) and as known in mosomes of №. aculeata are smaller than those of other members of the family (e.g., in Melampo- the diploids, a relationship also observed in other dium L.; Stuessy, 1971). groups (Darlington, 1963; Grant, 1976). This could The second polyploid in Nassauvia is №. gau- be due to (1) smaller chromosome size in diploid — dichaudii, a tetraploid with n = 22. The five other progenitors, or (2) chromosome size reduction after species of this section, №. ameghinoi Speg., N. Annals of the ri Botanical Garden Missou 514 Volume 78, Number 2 1991 Hunziker et al. 515 Chromosome Numbers in Nassauvia hillii, N. juniperina Skottsb., N. pentacaenoides, and N. pygmaea (Cass.) J. D. Hook, are unknown chromosomally. Material of №. hillii and N. junip- erina was attempted without success. The third case of polyploidy in Nassauvia is N. magellanica in which both n = 11 (Moore in Crisci, 1974b) and n = 22 (Moore, 1981) have been recorded. This infraspecific polyploidy occurs in a taxon of the relatively advanced sect. Nassauvia, which contains species with high degrees of head condensation. Within the subtribe Nassauviinae, Nassauvia is closest both phenetically and cladistically to Trip- tilion Ruiz Lopez & Pavon and Calopappus Mey- en (Crisci, 1974b, 1980; Crisci & Freire, 1986). As chromosome counts are not yet available for these two genera, however, the generic relation- ships cannot be assessed cytologically. The next closest relatives of Nassauvia are probably Poly- achyrus Lagasca and Moscharia Ruiz Lopez & Pavon (Crisci, 1974b, 1980), and the latter has n = 20 (Crisci, 1974b). This suggests a base number of x — 10 or 11 for the ancestral stock of these genera and that Moscharia is probably polyploid. Other more nee relatives might include Pan- phalea (n = cheilus (n = 11 and 18), and Perezia ^ = 4, 8. 12), but the relationships of these genera are uncertain and differ depending upon the phenetic or cladistic algorithm used. Oth- er recorded chromosome numbers in the subtribe (summarized in Crisci, 197 4b) are n = 27 in Acour- па D. Don, n = 18 in Jungia L. f., n = 22 for Marticorenia, and n — 27 for Proustia Lagasca and Trixis P. Browne. For insights on evolution within Nassauvia, the most likely outgroup of Nassauvia seems to be Calopappus, which possesses large single capitula (Crisci & Freire, 1986). The other closely related genus, Triptilion, seems best regarded as having evolved from Маззашта. Triptilion has different aggregations of heads, but most are moderately clustered in the axils of the leaves. We regard this as perhaps having come from sect. Panargyrum of Nassauvia, which has similar levels of inter- mediate head aggregation. The most primitive sec- tion of Nassauvia is probably Mastigophora, par- ticularly №. pentacaenoides Speg., which resembles Calopappus in head aggregation, leaf arrange- ment, and rhizomatous habit. LITERATURE CITED ARROYO, M. T. K. & C. MARTICORENA. 1988. A new species of the South American genus Nassauvia (Compositae: Mutisieae) from Chilean Patagonia. Brittonia 40: 332-334. CABRERA, А. L. 1982. Revisión del género Nassauvia (Compositae). Darwiniana 24: 283-379. COLEMAN, J. R. 1968. Chromosome numbers in some Brazilian Compositae. Rhodora 70: — 240 Crisci, J. У. 19 articorenia: a w genus of аы (Compositae). J. Arnold res 55: 38- A numerical-taxonomic study of the subtribe Nassauviinae она Mutisieae). Ј. Аг- nold Arbor. 55: 568-610 Nümeros cromosómicos en los géneros Panphalea y Holocheilus (Compositae, Mutisieae). Hickenia 1: 17-20. Evolution in the subtribe Nassauviinae (Compositae, Pru a phylogenetic reconstruc- tion. Taxon 29: 213-224. & S. Е. 30 EIRE. 1986. El género MAR (Compositae, Mutisieae). Caldasia 15: Бетон. C. D. 1963. Chromosome ча and the Origin of Cultivated Plants. С. Allen & Unwin Ltd., London GRANT, W. F. 1976. The evolution x "rimi and pud in arboreal plants. Taxo 75-84. HUNZIKER, J. H 66. _Numeros cromosómicos y cario- tipos de gropyron 6. y Elymus (Gramineae). Kurtziana 3: 151-15 MOORE, p. M. 1967. Chromosome numbers of Falkland Islands angiosperms. Brit. Antarct. Surv. Bull. 14: 69-82. 1981. Chromosome numbers of fuegian an- giosperms. Bol. Soc. Brot., ser. 2, 53: 995-1012. NuREZ, O. 1968. An acetic-haematoxylin squash meth- od for small chromosomes. Caryologia 21: 115-119. Росс, L. € HUNZIKER. Nuclear DNA ontent variation in Bulnesia. Heredity 77: 43-48. T F. A. 60. El empleo | de la Beinen acética oc. Biol. ed. Systematic relationships in the white-rayed ipea of Melampodium (Compositae). Briten 23: 190. FiGURE 1. D. №. axillaris. — E. N. darwinii. — G — diakinesis, H. — prometaphase 1. Meiotic chromosomes of species of Nassauvia. — А. М. uniflora. — B. №. асшеала. м ЛЕ revoluta. — chubutensis. — С. N. lagascae. —H. N. glomerulosa. All same scale. A- DASYLIRION WHEELERI VAR. DURANGENSE: А NEW COMBINATION IN THE NOLINACEAE' Joseph E. Laferriere? ABSTRACT Fruits from 34 specimens of a single population of iori from the Sierra Madre Occidental of Chihuahua, ose of herbarium specim durangense are conspe durangense is also reported for the first time fron ph ns of | wheeleri and D. Miei from outside the study ge between the tw t was concluded that D. The genus Dasylirion Zucc. is a member of the Nolinaceae with approximately 15 species native to Mexico and the southwestern United States. The plants are large, perennial, rosette herbs with strap- like leaves and flowering stalks up to 6 m tall. Various authors have merged this family with the Liliaceae or the Agavaceae (Cronquist, 1981), but more recent workers have supported separating the group as a distinct family (Dahlgren et al., 1985). Trelease (1911) described several new species and varieties of Dasylirion based largely on dif- ferences in the morphology of the three-winged, one-seeded samara. In particular, D. durangense Trel., a species described from a single specimen collected from Durango, Mexico, was said to have fruits “broadly elliptical-cordate, 7-8 х 9 mm," and the style “scarcely half as long as the rather open deep notch." This was contrasted with the more widespread D. wheeleri S. Wats. in Rothr., which was characterized by fruits ““round-obovate, 6-7 x 1-9 mm, the style normally about equalling the open moderately deep notch." To date, only two publications have compared these two species, both of which have relied almost exclusively on Trelease's descriptions (Standley, 0; Conzatti, 1947). The present analysis of a population in which both taxa occur sympatrically was conducted to ascertain the utility of the length of the style and the shape of the capsule wings in distinguishing D. wheeleri from D. durangense and to determine whether the distinction between the two taxa should continue to be recognized at the species level. MATERIALS AND METHODS The site chosen for study on sympatric popu- lations was the village of Nabogame, Chihuahua, Mexico. The town lies at 28°30'N, 108°30'W, approximately 1,800 m in the Sierra Madre Oc- cidental, approximately 10 km east of the frontier with the state of Sonora. This site is ideal for such a study since plants referable to both D. wheeleri and D. durangense may be found growing within a few meters of one another. This is the most northerly location from which D. durangense has been reported in the literature (Fig. 1). Fruits were collected from 34 individuals within 2 km of Nabogame. This sample represented the maximum number of individuals that could be col- lected within a two-week period. Most of the plants were located on extremely steep granitic slopes subject to less disturbance than other habitats in the area. Since the shape of the wings changes as the capsule matures, fruits were collected in De- cember 1987, after they had fully dried. From each individual, ten fruits were selected for mea- surement on the basis of absence of insect damage and teratological deformities, and the presence of an unbroken style. Measurements were made as shown in Figure 2 For comparison, herbarium specimens of D. ! This work was supported in part by NSF doctoral dissertation grant number BNS-8601339 to the author and Willard Van Asdall. I thank Deborah Ellstrom, Charles Mason, Richard Felger, Geoff Levin, Don Kyhos, Billie Turner, William Weber, Howard Scott Gentry, and п Dice for ашан чыш advice and assistance, and the curators of COLO, LL, MO, MEXU, and US for loan of and access to specirr ? Department of Ecology and Evolutionary Biology, University of Arizona, Tucson, Arizona 85721, U.S.A. ANN. Missouni Bor. GARD. 78: 516-520. 1991. Volume 78, Number 2 Laferriére 517 1991 Dasylirion wheeleri var. durangense Y т Апгопа Y Y v New Mexico v У + Vy У г ҮҮ + Техаз Sonora Y A R У МЕХА AAA AAS ААА А RRA IRS МАЧА PAAPAAAA УМ УМА УУ Oe AAA ААА ААА AAAI DES => AAA А AAA VASA «ААРАААА SS I-II TES СУМУ МДУДУДУЛУЛУДУД И УМУ УУ УМУ LTISIIIIAA А ^v AAA S 772227222227 A ЗА Durango VAAN УКААААААААА IIA М 223 a - СИ МИ ы ылы МЕ 2 t) Mw wheslen ЕД durangense Zacatecas a^ WA B^ МА _ КОЛОСУ А e Nabogame population 22733 A inwh FARR У in which both occur AAAS ВУ А 1. Map showing Nabogame and the distribution of Dasylirion wheeleri and D. durangense. Diamond represents Nabogame population, where both taxa are presen wheeleri and D. durangense from outside the Na- bogame area were also measured. Fourteen spec- imens of D. wheeleri and six of D. durangense with mature fruits were examined, including both the type specimen of D. durangense and the one syntype of D. wheeleri with mature fruits. The same measurements were made on these specimens as on the fruits collected from Nabogame. In addition, leaf and flower characters were ex- amined on a much larger number of specimens, but no consistent differences between the two taxa were noted. In particular, the size of the teeth and of the denticules between them, characters used by Trelease in his taxonomic discussions, varied widely in northern and in southern collections. RESULTS Averages and standard deviations of the three samples are given in Table 1. An analysis of vari- 518 Annals of the Missouri Botanical Garden E pw | Drawing of capsule from type specimen of Dasylirion durangense showing measurements made for FIGURE 2. Peur analysis. A represents the length of bu style, B the length of the yle, D the width of the notch, i.e., the from the base of the fruit to the base of the their point of maximum distance from the hu о measurements were calculated F, the notch depth, i.e., the hei entire fruit, C the length ud the main axis distance betwee ght of the wings above the base of the style (B — C), F). and C, the ratio of the length of the style to the depth of the b (A/ ance (ANOVA) test was performed on each of the seven characters; highly significant differences were found between the three samples at the 1% level for all seven characters. The data were subjected to а Newman-Keuls’ specific comparison test to ascertain where the differences lie (Linton & Gallo, 1975); a Student's r-test yielded results identical to those produced by the Newman-Keuls' test. Re- sults are indicated in Table 1 The data gathered from herbarium specimens showed statistically highly significant differences ee of the eight char- acters measured: total length, notch depth, and between the two taxa in thre style/notch ratio. In addition, there was a signifi- Volume 78, Number 2 1991 Laferriére 519 Dasylirion wheeleri var. durangense TABLE 1. Means and standard deviations of Dasylirion fruit measurements. Data for D. wheeleri and D. durangense represent herbarium specimens collected outside the Nabogame area D. wheeleri D. durangense Nabogame population N 14 6 34 A (style length, mm) 0.80 + 0.09 0.90 + 0.10 0.95 + 0.11** B (total length, mm) 6.48 + 0.7888 8.36 + 0.78** 8.29 + 0.80** C (axis length, mm 9.54 + 0.63 6.12 + 0.79 6.25 + 0.70** D (notch width, mm) 2.39 + 0.76 2.41 + 0.44 3.94 + 0.74**§§ E (total width, mm) 5.02 + 0.95§ 5.97 + 0.66* 6.77 + 0.79**§ F (notch depth, mm) 1.21 + 0.38§§ 2.24 + 0.62** 2.04 + 0.51** G (style/notch ratio) 0.70 + 0.1788 0.26 + 0.11** 0.53 + 0.17**$8 — Significantly different from D. wheeleri (5% level) pnis a Newman-Keuls' or a t-test. — Highly significantly different from D. wheeler үз Significantly different from D. durangense (5% le же §§— Highly significantly different from D. durangense 1% level). cant difference in total width at the 5% level. The Nabogame data differ statistically from the D. wheeleri sample in all characters measured, even showed highly significant differences between Na- bogame and urangense, while total width showed a difference significant at the 5% level. Despite these statistical differences, for each char- acter individuals in the Nabogame population exist which exhibit the characteristics of each of the two recognized taxa. Indeed, in every case the variation in the Nabogame population exceeds the entire range between the mean values for the two taxa. n addition to the characters included in the statistical analyses, there is considerable variation in general wing shape, a character less readily quantifiable than those chosen for measurement. Similar variation is also apparent among the her- barium specimens. Most of this is between individ- uals, but some is between different fruits from the same plant. DISCUSSION The length of the style in fruit is useless as a diagnostic character. The shape of the wings may have some taxonomic value, but there is a great deal of phenotypic variation even within a single population. It thus appears that D. wheeleri and D. idc i should not be maintained as sep- arate specie t does seem desirable, however, to recognize D. durangense as a variety of D. wheeleri, in view of the consistent difference in notch depth between specimens from the southern part of the range and those from further north. Yet there is considerable variation and intergradation between the two taxa. It is not known whether there exists a north-south cline connecting the two extremes or a narrow zone of intermediates in the area around Nabogame. In terms of notch depth, the Nabogame population appears more closely related to var. durangense, but it spans the range between the two extremes. Even in terms of notch width, the one character statistically separating the population from both of the recognized taxa, the variation is so great that if this one trait was used as a diagnostic character to describe a third variety, different individuals in the Nabogame population would be assignable to each of the three taxa. The notch-depth character singled out by Trelease is only part of a much broader variability than implied by his recognition of D. durangense as a distinct species. TAXONOMY In view of the preceeding, the following change is indicated: Dasylirion wheeleri S. Wats. in. Rothr. var. urangense (Trel.) Laferriere, comb. et stat. nov. Proc. TYPE: Dasylirion durangense Trel., Amer. Philos. Soc. 50: 438. 1911. Palmer 557. The name **wheeleri" has precedence because it was first published in 1878 (Rothrock, 1878). The present paper is the first to report D. wheeleri var. durangense from Sonora (Gentry, 1972). The two varieties may be distinguished by the following ey: la. Distal wing tips € s extending more than bove bas е. var. durangense lb. Distal wing tips и less than 1.5 mm above base of style var. wheeleri 520 Annals of the Missouri Botanical Garden Selected specimens examined. Dasylirion whee- - MEXICO. CHIHUAHUA: Nab 1 1099W, 11 Aug. 1940, White 3234 (LL). UNITED STATES. ARIZONA: Cochise Co., 11 Nov. 1967, Barr 67-437 (ARIZ); 2 Jan. 1983, Yatskievych & Windham 83-03 (ARIZ); Gila Co., 16 Sep. 1973, Engard 105 (ARIZ); Pinal Mts., Gila Co., 30 July 1892, Toumey 449 (MO); Graham Co. Wolf & Everett 11404 (ARIZ); Pima Co., 8 Ый E of Vail, 7 Sep. 1938, Crooks, Thornber & Benson s.n. (ARIZ); Pinal Co., Oracle, 9-13 1905, Thornber s.n. (ARIZ); Mazatzal Mts., Sep. 1911, Holzinger s.n. (MO); TEXAS: N of El Paso, 29 Sep жы. Warnock 14255 (LL). Dasylirion wheeler var. durangense. MEXICO. CHIHUAHUA: Nabogame, 2 28930'N, 108°30'W, 15 Nov. 1988, Laferriére 2268 (ANSM, ARIZ, MEXU, МО); Quirire, Mpio. Batopilas, 3 June 1984, Bye, Davis, Ran- dolph & Gerson 12854 (COLO). DURANGO: Durango City, Palmer 557 (type) (MO, MEXU); km 249 on Highway 45 between Durango and Zacatecas, 24 July 1985, Starr 903 (ARIZ); 9 km W of Cd. Durango on Highway 40, 6 Aug. 1975, Gentry & Engard 23613 (ARIZ); 24 km c rca, on Highwy 30, 26 July 1958, Correll & Johnston 20223 (LL); km 40 Durango-Mezquital High- way, 22 Apr. 1982, Gonzalez s.n. (LL); Durango, July 1911, Ochoterena s.n. (МО); ой Maicoba, June T 1968, Pennington 91 (LL). ZACATECAS: Zacatecas, Mar- roquin s.n. (MEXU). LITERATURE CITED СОМ2АТТІ, С. 1947. Flora Taxonómica Mexicana, Tomo II: Monocotiledóneas Diperiantadas- Е е Inferováricas. I | d Mexicana de a Natu- ral, México, D.F., Mexico. а of a 1939. edition published үк м їп o а.] CRONQUIST, А. Int ы System of Classi- fication of Flowering Plants. Columbia Univ. Press, New York. DAHLGREN, В. M. T., Н. T. CLIFFORD & P. Е. YEO. 1985. The Families of the Monocotyledons: Structure, Evo- lution and Taxonomy. Springer-Verlag, Berlin. LINTON . GALLO, JR. 1975. e Practical Statistician. Brooks Cals Publishing, М Cal. ifornia. Gentry, Н. 5. 1972. The Agave family in Sonora. U.S.D.A. P deve Handbook 527. ROTHROCK, J. . Report upon United States боо Surveys West of the One Hundredth . Washington: U.S. O © < c іа! e] 5 oO к= | En = 3 = ga о o : o and shrubs of Mexico (Ы асова Bétulacsáe Contr. U.S. Natl. Herb. : 1-169. TRELEASE, W. Amer. Philos. 11. The desert pt group Nolineae. Proc. Soc. 50: 404-443. NOTES TWO NEW SPECIES OF HILLIA SUBG. RAVNIA (RUBIACEAE) FROM VENEZUELA AND MEXICO Hillia Jacq. is a genus of about 24 species of succulent, usually epiphytic shrubs found through- out the moist and wet Neotropics. Subgenus Кат- nia (Oersted) C. M. Taylor includes three species found in Costa Rica and Panama, plus the two species described below. The characteristics of this subgenus have been presented previously (Taylor, 1989). Recent collections have brought to scientific light the following two undescribed species, which are placed in subg. Ravnia based on the form and color of their corollas, which are tubular to nar- rowly funnelform, generally red in color, and pre- sumably bird-pollinated. Hillia rivalis C. M. Taylor, sp. nov. TYPE: Ven- ezuela. Territorio Federal Amazonas: Depto. Atabapo, Cerro Marahuaca, 3?43'N, 65°30'W, 1,200 m, 16 Oct. 1988 (fl, fr), R. Liesner 24900 (holotype, MO; isotype, VEN). Fig- ure 1. Frutex subsucculentus, saxicola, glaber, ca. 1 m altus. Folia peranguste ap 38-100 x 5-10 mm, apice basique acuta; petiolo 6-12 mm longo; stipulis 10-13 x 3-4 mm. Flores solitarii, 6-meri; pedunculo 1-2 mm longo; lohulis calycinis peranguste triangularibus, 18-36 -2 mm; corolla а externe sordide bes. interne luteoviridi, tubo 3 8-10 x 6-10 mm; cylindrica, laevis, 50-65 x 8-10 mm, stipite 10-15 mm Glabrous, somewhat succulent, lithophytic shrubs to 1 m tall; bark smooth, red-brown to gray-brown. Leaves isophyllous; blades narrowly elliptic, 3.8— 10 x 0.5-1 cm, narrowly acute at apex and base, coriaceous and hard upon drying, glabrous; sec- ondary veins 4—5 pairs, obscure, pinnate, ascend- ing, the distal 1-2 veins looping to interconnect, the midrib prominulous abaxially, the margins slightly involute; petioles 6-12 mm long, glabrous; stipules narrowly elliptic or lanceolate, 10— 3-4 mm, acutely to broadly angled, membrana- caducous. Flowers solitary, ceous, glabrous з ANN. 6-merous; peduncles 1-2 mm long; bracts linear, 12 x ca. 0.5 mm, acute; ovary cylindrical, ca. 1 cm long, glabrous; calyx limb green, coriaceous, glabrous, divided nearly to base, tube 0-1 mm long, the lobes very narrowly triangular to linear, 18-36 mm long, 1.5-2 mm wide at base, acute; corolla funnelform, membranaceous, dull red ex- ternally, yellow-green internally, glabrous through- out, tube 3.2-5.9 cm long, 3-3.5 mm wide at base, the lobes triangular, 0.8-1 cm long, 6-10 mm wide at base, rounded to acutely angled, re- flexed to revolute; stamens partially exserted, fil- aments 3—4 mm long, inserted ca. 4-6 mm below mouth of tube, anthers 8-9 mm long; style 3.4- 6.5 cm long, stigma exserted, capitate, ca. 1-2 mm diam. Capsules cylindrical, woody to rather papery, brown, smooth, 5-6.5 cm long, 0.8-1 cm diam., borne on a stipe 1-1.5 cm long; seeds 1.5 X ca. 0.3-0.5 mm, with trichomes 1.2-1.5 cm long. This new species was found growing on rocks in and along rivers, a habitat that has not been previously reported in Hillia, and which is em- phasized in the specific epithet. One site was in- undated a few days after the specimen was collected (R. Liesner, pers. comm.); presumably these pe- rennial plants are adapted to such disturbance. In habit, leaf morphology, and capsule texture, H. rivalis strongly resembles the Mesoamerican shrub Lindenia rivalis Bentham, which grows in similar sites. These genera are apparently not closely re- lated, and the resemblance is apparently one of convergence. illia rivalis can be distinguished by its re- markably long, narrow calyx lobes, funnelform co- rollas that are red externally but yellow-green in- ternally, relatively narrow leaves, and riverine habitat. The floral morphology of this species strongly resembles that of several other species of Hillia with funnelform green corollas and pinnate leaf Missour! Bor. GARD. 78: 521-523. 1991. 522 Annals of the Missouri Botanical Garden FIGURE 1. Hillia rivalis (Liesner 24900 MO), habit. venation, notably Н. psammophila Steyermark, tionship between the former genus Ravnia and Н. illustris (Vellozo) К. Schumann, Н. foldatsii Hillia. Steyermark, and H. saldanhaei K. Schumann This is the first report of а member of this (Taylor, in prep.). The characteristics found in this subgenus from South America. The rather wide species provide further evidence of a close rela- ranges of calyx lobe, corolla, and style dimensions Volume 78, Number 2 1991 Notes 523 FIGURE 2. Hillia oaxacana (Garcia M. & Torres C. 1523 МО), habit. are derived from measurements of apparently ma- ture structures all found in the same collection. Comparable variation is found in other species of Hillia (Taylor, in prep.). Paratype. VENEZUELA. TERRITORIO FEDERAL AMAZONAS: Depto. Atabapo, Cerro Marahuaca, 3?38'N, 65?28'W, 1,225 m, 19 Feb. 1985 (fr), В. Liesner 17712 (MO). Hillia oaxacana C. M. Taylor, sp. nov. TYPE: Mexico. Oaxaca: distrito de Putla, Arroyo de San Isidro, 3 km al este de Concepción, Bue- navista, 10 June 1985 (fl), А. Garcia M. & R. Torres C. 1523 (holotype, MO; isotype, MEXU). Figure 2. Frutex succulentus ای‎ a Cra Folia el- liptica, 4.2-6 x 2-2.5 cm, apice acuminata, basi cu- neata; petiolo 5-8 mm longo. Flores s solitarii, сше, pedunculo 2-3 mm longo; lobulis calycinis Scd 6 x 1.5-2 mm; corolla tubiformi vel anguste infu ndi. buliforme, tubo rubro ca. 3.8 cm longo, lobulis viridibus 9 X 2.5 mm; antheris inclusis. Capsula ignot Glabrous, succulent, hemiepiphytic shrubs; bark smooth, gray. Leaves isophyllous; blades elliptic, 4.2-6 x 2-2.5 cm, acuminate at apex with the tip ca. 5 mm long, cuneate at base, subcoriaceous, glabrous; secondary veins 3—4 pairs, rather ob- scure, subpalmate, ascending, not looping to in- terconnect, not prominulous, margins straight; pet- ioles 5-8 mm long, glabrous; stipules not seen. Flowers solitary, 6-merous; peduncles 2-3 mm long; bracts absent; ovary cylindrical, ca. 3 mm long, glabrous; calyx limb green, membranaceous, glabrous, divided to base, the lobes narrowly tri- angular, 6 mm long, 1.5-2 mm wide at base, rounded; corolla tubular to narrowly funnelform, carnose, red in tube, green on lobes, glabrous throughout, tube 3.8 cm long, 2 mm wide at base, the lobes triangular, 2.5 x 2.5 mm, broadly round- ed, spreading to reflexed; stamens included, in- serted above middle of tube, anthers 4 mm long, h mm below mouth of tube; style 4 cm long, stigma exserted, with two clavate branches ca. 3 mm long. Capsules not known. Hillia oaxacana can be distinguished by its tubular red corollas with green lobes and its leaves with the secondary veins all arising from below the middle of the blade (““subpalmate”). In its corollas, H. oaxacana is similar to H. longifilamentosa (Steyerm.) C. M. Taylor of Costa Rica and Panama, which differs in its completely red-orange corollas and long-exserted stamens. This new species is also similar to H. saldanhaei K. Schumann of Brazil, a poorly known species with similarly shaped but green corollas with lobes 6-7 mm long. Similar subpalmate leaf venation is found only in H. ulei K. Krause of Amazonian South America, which differs in its broadly funnelform green corollas. (Hillia ulei is here treated to include H. viridiflora Kuhlmann & Silveira, H. irwinii Steyermark, and H. schultesii Steyermark.) Reported here is the northernmost record of a member of subg. Ravnia. Specimens of the type collection were originally annotated and distributed as H. macrocarpa Stan- dley & Steyermark, which H. oaxacana keys out to in the Rubiaceae treatment for the Flora of Guatemala (Standley & Williams, 1975). I thank R. Liesner for field data, B. Manara for the illustration of H. rivalis, R. E. Gereau and B. Manara for assistance with Latin composition, and P. Berry, B. Holst, and R. E. Gereau for critical comments. LITERATURE CITED STANDLEY, P. C. € Г. O. WILLIAMS. 1975. Flora of бүг и IX (Rubiaceae). Fieldiana, Вог. ae Revision al ae subg. Ravnia (Rubiaceae: Cinchonoideae). Selbyana 11: 26-34 rep. Revision of Hillia (Rubiaceae: Cin- а — Сћапоне M. Taylor, Missouri Botanical Gar- den, P.O. Box 299, St. Louis, Missouri 63166- 0299, U.S.A. ARDISIA LUNDELLIANA, A NEW SPECIES OF MYRSINACEAE FROM PANAMA The efforts of the Missouri Botanical Garden’s plant exploration program in Panama continue to yield new and phytogeographically interesting taxa, even in areas as well known as western Panama. During routine identification of plants collected by Gordon McPherson, I discovered a hitherto un- described species of Ardisia, which is described Ardisia lundelliana Pipoly sp. nov. TYPE: P ama. Chiriqui: vic. Fortuna Dam, forested slopes along ridge at S boundary of watershed, 8°45'М, 8°15'W, 1,250 m. Tree 4 m; calyx brown; corolla pale pink; fr. green, 28 Apr. 1986, G. McPherson 9107 (holotype, МО, photocopy, US). Figure 1. Ob ramulos foliorum costas abaxiales necnon inflores- rfu s cernuas, flores icellos recurvatos, antheras lineo-lan- ceolatas et longitudinaliter dehiscentes А. brenesii arcte affinis, sed 2 ea foliis praeclare petiolatis (non sessilibus), 5 (non dimorphis), ellipticis (non oblanceolatis vel ta nec oblanceolatis vel oblongis), basibus cu- neatis (non e rar qe apicibus acutis (non acu- minatis), sepalis 0 (non 1.4-1.6) mm longis mar- ginalibusque integerrimis (non piam petalisque 12.0- 13.9 (non 3.9-4.5) mm longis, intus prope basin verru- cosis (non laevis) statim cognoscitur Tree to 4 m, the branchlets, petioles, abaxial leaf costae, inflorescence rachis, pedicels and calyx rufous-stellate tomentose and furfuraceous stipi- tate-lepidote. Leaves petiolate, the petioles mar- ginate, (1.1-)1.5-2.0 cm long, the blades coria- ceous, elliptic, (7.0-)11.0-16.7(-18.0) cm long, (4.0-)-5.2-7.7 cm wide, apically acute, basally cuneate, glabrous above, densely furfuraceous stip- itate-lepidote below, the margins entire, revolute. Inflorescence terminal, pendent, paniculate, 8.0— 15.0 cm long, the peduncle 2.3-2.7 cm long, the rachis linear in flower, becoming flexuous in fruit. Flowers corymbose; pedicels terete, 0.6-1.4 cm long; sepals erect, coriaceous, dextrorsely con- torted, widely ovate, 5.2-6.0 mm long, 3.6-4.0 mm wide, apically acute, asymmetric, basally au- riculate on one side, the auricle and margin scar- ious, glandular-ciliate, densely rufous stellate and furfuraceous lepidote without, sparsely stipitate- lepidote basally within, black lineate-punctate; pet- als reflexed at anthesis, chartaceous, lanceolate, 12.0-13.9 mm long, 3.9-4.5 mm wide, apically attenuate, slightly asymmetric, glabrous, densely black punctate and lineate-punctate, pink, verru- culose basally within, the margin white, subentire, irregular; stamens 10-12.5 mm long, the filaments hyaline, 6.0-6.3(-6.9) mm long, glabrous, the an- thers linear-lanceolate, concolorous, 4.0-5.9 mm long, apically apiculate, basally sagittate, dorsi- xed, the connective inconspicuously black-punc- tate dorsally, longitudinally dehiscent; ovary glo- 5.3-5.7 mm tall, 3.0-3.1 mm diam., densely orange-punctate; style 9.1-9.4 mm, exserted at anthesis, eccentric or not, densely and prominently orange-punctate; stigma uius placenta glo- bose, obtuse apically, m long and in diameter, the ovules 19-24, п. Fruit glo- bose, 5-7 mm diam. (immature), prominently ver- rucose, inconspicuously punctate. Ardisia lundelliana belongs to a group of spe- cies often referred to the genus Auriculardisia undell. Because the salient features defining this group (Lundell, 1981) have been encountered in a number of paleotropical members of the more broadly defined Ardisia Swartz, this species is as- signed to Ardisia until a more thorough, pantrop- ical survey is completed (Pipoly & Stone, in prep.). With rufo stellate-tomentose abaxial leaf midribs, inflorescence rachises and branchlets, pendent, subcorymbose flowers on panicles, and linear-lan- ceolate and longitudinally dehiscent anthers, 4. lundelliana is clearly most closely related to 4. brenesii (including А. limonensis Lundell). How- ever, lundelliana is readily separable by its petiolate (not sessile), monomorphic (not dimor- phic), and elliptic (not oblanceolate or obovate) leaves with cuneate (not auriculate) bases and acute (not acuminate) apices. Floral characters allowing easy recognition include the longer, entire sepals ANN. Missouni Bor. Ganp. 78: 524-526. 1991. Volume 78, Number 2 1991 Notes 525 piel и 4 Calyx, sh and је мера corolla surface basally verrucose. (A-D drawn from holotype, McPherson. 9107, line, entire margin. — C. F and longer, verrucose petals. Among these char- acters, a preliminary survey revealed that the ver- rucose petals represent an autapomorphic char- acter state. The dimorphic leaf character present in A. bre- nesii refers to the production of modified leaves en P Duke Ardisia — Pipoly.—A. Branchlet, showing pendent inflorescence, tomentose costae. — lower, at anthesis. — D. Dissec B. ted flower, showing another dehiscence MO.) axillant to the inflorescence of that species, a fea- ture thought to autapomorphic to it (Pipoly, sub- mitted). Ecologically, A. brenesii is restricted to riparian areas in tall wet forest or moist premontane forest, whereas A. lundelliana is a montane forest, ridge-top species. Clearly, much more exploration 526 Annals of the Missouri Botanical Garden of western Panama (especially the Atlantic slope) is needed, because this and some 50 other species of Ardisia collected from there are known only from their type specimens. It is with great pleasure that I dedicate this species to Cyrus Longworth Lundell, who has just completed his sixtieth year of research in Meso- american Myrsinaceae. My studies in Myrsinaceae were supported by a Smithsonian Postdoctoral Fellowship. Support for illustration is provided by a Smithsonian Research Opportunity Fund Grant, no. 1233F809. I thank Peggy-Ann Duke for her skillful illustration. LITERATURE CITED LuNDELL, C. L. 1981. о Myrsinaceae — VI. Phytologia 49: 341-354 PriPoLY, J. (submitted). Myrsinaceae. In: W. D. Stevens, editor, Flora de Nicaragua. Missouri Botanical Gar- den, St. Louis. — John J. Pipoly III, Department of Botany NHB 166, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560, U.S.A. Present address: Missouri Botan- ical Garden, P.O. Box 299, St. Louis, Missouri 63166-0299, U.S.A. PELTOSTIGMA PARVIFLORUM (RUTACEAE), NUEVA ESPECIE DE COSTA RICA Y COLOMBIA La tribu Zanthoxyleae (Rutaceae: Rutoideae) (Waterman, 1975). Esta tribu incluye la subtribu Choisyinae, compuesta por cinco pequenos géneros con poca cohesión geográfica, de los cuales dos, Choisya Kunth y Peltostigma Walp., figuran en la flora neotropical (Engler, 1931). Choisya tiene siete especies desde el suroeste de los Estados Uni- dos hasta el sur de México (Muller, 1940), mientras que Peltostigma contaba hasta ahora con dos es- pecies de Jamaica y desde el noroeste de México (estados de Durango y Sinaloa) hasta Costa Rica (Morton, 1933; Standley & Steyermark, 1946; F Chiang, com. pers.) que son notables por sus flores grandes y muy fragantes. Los programas de co- lección botánica en Costa Rica y Colombia y el estudio de ejemplares en los herbarios CR, HUA, y MO han dado como resultado el descubrimiento de una nueva especie de Peltostigma, muy se- mejante a las precedentes en su arquitectura floral pero diferenciada de éstas por sus hojas unifolio- ladas y el tamano muy reducido de sus flores. Peltostigma parviflorum Q. Jiménez & Ge- reau, sp. nov. TIPO: Costa Rica, Prov. San Jose, 943'10"N, 84?23'10"W, Zona Protec- tora La Cangreja, Santa Rosa de Puriscal, 400-500 m, 4 septiembre 1987 (fl), О. Л- ménez M. & L. J. Poveda A. 494 (holotipo, CR; isotipos, Е, K, MO). Figura 1. Frutices arbores i aoe m alti, plerumque ramosis- simi, cortice obscuro, par vegetativis omnibus glabri. Folia паню o pre ulvinatum аи denti) 0.2-5.6 cm longo, adaxialiter canaliculato, inter- t e utrinque glanduloso- -punctata; nervis secun- ariis 6-8 ante marginem anastomosantibus; nervis in- ES ex tertiariis subparallelis laxe reticulatisve vix diversis. Inflorescentiae axillares solitariae vel hae 2- 5 fasciculos formantes; pedunculis 1- usque 3(-5)-floris, unifloris interdum pari bractearum minutarum munitis; ANN. Missouni Bor. GARD. 78: pedunculo (1-)3-38 mm longo, glabro; pedicello 0-21 m longo, glabro usque minute puberulo. Flos sepalis 3 pallide viridibus, late ovatis usque suborbicularibus, subae- qualibus, 2.4-3.0 x 2.3-2.9 mm, abaxialiter glabris us- que puberulis, ciliolatis; petalis 3 albis cremeisve, in vivo concavis conchiformibusque, in sicco late ovatis usque suborbicularibus, 8.2-10.1 x 8.0-9.7 mm, ciliolatis; dis- co nectarifero carnoso, lobulato, coacto, aurantiaco; stam- inibus 10-12 inaequalibus filamentis arcuatis antherisque super stigmata contiguis; filamentis albis, linearibus, alatis; antheris luteis, ovoideis, thecis basi parum divergentibus; ovario pubescenti ex segmentis 6 rhomboideis basi tantum connatis constanti; stylis subterminalibus 0.5-0.7 mm longis; stigmatibus luteis in massam unam obloideam coal- in quoque segmento ovarii solitario. Fructus ex turitate distinctis glabrisque constans; mericarpiis oblique ellipsoideis, 1.5-2.1 1 cm; endocarpio luteo cum semine caduco; semine in mericarpio solitario, fusiforme-ellipsoideo, papilloso, arillodio membranaceo la- tere hilari munito. Arbustos y árboles 1-11 m de alto, hasta 12 cm de DAP, usualmente muy ramificados, la cor- teza oscura, las partes vegetativas glabras. Hojas .6 cm incluyendo el ресто о pulvinado), adaxialmente acanalado, oca- unifolioladas; pecíolo 0.2-5 sionalmente angostamente alado; lámina ovado- iig a ovado-oblonga, 3.2-20.6 x 1.2-10.2 m (L/A = 2.0-2.9), el apice ica hasta acu- minado y agudo, la base asta cuneada, los márgenes enteros hasta levemente sinuados, submembranácea a cartácea, glanduloso-punteada por ambas caras; nervios secundarios 6-8, anas- tomosándose sin alcanzar el margen; nervios in- tersecundarios poco diferenciados de los terciarios, estos subparalelos hasta laxamente reticulados. In- florescencias axilares, solitarias o en fasciculos de 2—5; pedúnculos 1- a 3(-5)-floros; pedúnculos uni- floros con o sin un par de brácteas menudas de- marcando la base del pedicelo; pedünculo (1-)3- 38 mm, glabro; pedicelo 0-21 mm, glabro hasta menudamente puberulento. Sépalos 3, verde pali- dos, anchamente ovados a suborbiculares, subigua- les, 2.4-3.0 х a puberulentos, ciliolados; pétalos 3, blancos a blan- co-cremosos, cóncavos y conquiformes en vida, 927-530. 1991. .3-2.9 mm, abaxialmente glabros 528 Annals of the Missouri Botanical Garden FIGURA 1. Peltostigma paro iflorum. А. Plan anchamente ovados a suborbiculares cuando pren- sados, 0.1 x mm, ciliolados; disco nectarifero carnoso, lobulado entre los segmentos del ovario, aterciopelado, anaranjado; estambres 10-12, algo desiguales, los filamentos arqueados y las anteras contiguas encima de los estigmas; filamentos 2.1-4.4 mm, blancos, lineares, alados y frecuentemente con el ápice de ancho subigual al de la antera; anteras ovoides con las tecas algo divergentes en la base, amarillas; tecas 1.5-2.1 х 5-0.9 mm; ovario pubescente, formado por 6 nta en fruto. — B. Flor. — C. Botón floral. flor con una porción de dos pétalos y seis estambres eliminados. — E. Estambres en vista lateral, ventral y dorsal.— F. Hoja. (А de Grayum et al. 4955; B-E de Jiménez et al. 710; F del isotipo en MO). — D. Vista interior de una segmentos romboides unidos solamente en la base, los segmentos 1.4-1.7 x 0.7-0.9 mm; estilos sub- terminales, 0.5-0.7 mm; estigmas amarillos, uni- dos en una masa obloide, ésta 0.5-1.3 mm de alto, 1.0-1.5 mm de ancho; óvulo 1 en cada segmento. Fruto formado por 2-6 mericarpios unidos en la base, puberulentos cuando jóvenes, separados y glabros en la madurez; mericarpios oblicuamente elipsoides, reticuladamente rugosos en ejemplares en herbario, 1.5-2.1 amarillo, caedizo con la semilla; semilla 1 en cada 6-1.1 cm; endocarpio Volume 78, Number 2 1991 Notes 529 mericarpio, fusiforme-elipsoide, papilosa, 7.9–10.1 1 mm, con un arilodio (porción no lig- nificada del endocarpio) membranaceo а lo largo de su lado hilar. Ejemplares adicionales examinados. COSTA RICA. UANACASTE: Rancho Harold, región del Мојс 30 julio 1986 (fl, fr), 1 А. & A. M. MO); 10?47'N, 85?18'W, Parque Nacional Rincón de la Vieja, above Hacienda Santa Maria, -1,200 m, 28 enero 1983 (8), Davidse et al. 23449 (CR, MEXU, MO); (BM, CR, МО); 10?40'N, 85?03'W, Hacienda Monte- zuma, Volcán Tenorio, 600-700 m, 25 enero 1985 (£l), Grayum et al. 4955 (CR, MEXU, MO); 10?48'N, 85?19'W, Parque Nacional Rincón de la Vieja, Hacienda -850 m, 16 agosto 1987 (fr), G. He- acienda Santa Maria, 17 octubre 83°28'W, Santa Cecilia, La Cruz, Estación Pitilla, faldas Cerro Orosilito, 700 m, 14 junio 1989 (fl), Q. Jiménez M. et al. 710 (CR, MO). PuNTARENAs: 300 m N of Monteverde main road on Quebrada Máquina, ca. 1,400 m, 14 julio 1990 (fr), Woodruff et al. 363 (CR, TEX). COLOMBIA. ANTIOQUIA: 5°53’М, 74°39'W, Mpio. San Luis, Сапоп del Rio Claro: sector occidental, margen izquierda, 325-400 m, 2 diciembre 1983 (fr), 4. Cogollo 1003 (HUA, МО); sector norte, margen izquierda, 340-500 m, 24 diciembre 1983 (fr), А. Cogollo 1069 (HUA, МО); sector oriental, margen derecha, 325 m, 2 junio 1984 (8), А. Cogollo 1766 (HUA, MO); camino al Refugio, 600 m, 11 octubre 1982 (fl, fr), E. Rentería et al. 2845 (HUA, MO). Las poblaciones conocidas de Peltostigma par- viflorum forman dos morfotipos segün su posición geográfica: la “raza norteña” consiste en las co- lecciones citadas de Guanacaste y Puntarenas, Costa Rica; y la “raza sureña” consiste en el tipo y las colecciones citadas de Antioquia, Colombia. Las diferencias entre estas dos razas son las siguien- "raza norteña”: hojas con peciolos de 0.2-2.1 cm de largo y poco variables, el peciolo mayor en un individuo dado sobrepasando al menor por 0.5-1.5 cm de largo; foliolo 3.2-16.0 x 1.2- 7.3 cm, el ápice obtuso u obtusamente acu- minado, la base aguda hasta redondeada; pé- talos 8.2-9.3 x 8.0-8.5 mm "raza sureña”: hojas con peciolos de 0.3-5.6 ст de largo y muy variables, el peciolo mayor en un individuo dado sobrepasando al menor por 2.0—4.5 cm de largo; foliolo 4.2-20.6 x 2.2- 10.2 cm, el ápice agudamente acuminado y la base usualmente atenuada; pétalos 9.8- 10.1 x 8.5-9.7 mm Aunque consistentes dentro de sus áreas geográfi- cas, tales diferencias vegetativas podrian reflejar las condiciones ecológicas de las localidades, y la carencia de material con flores nos impide esta- blecer una correlación significativa entre el tamano de los pétalos y las caracteristicas foliares. La “raza sureña” se encuentra en alturas de 325-600 m.s.n.m., y en Costa Rica habita el bosque tropical muy hümedo (Tosi, 1969) con clima muy ћитедо caliente con una corta estación seca (Herrera, 1985), asociada con Bauhinia тапса Standley (Caesalpiniaceae), Coccoloba tuerckheimii J. D. Smith (Polygonaceae), Croton killipianus Croizat (Euphorbiaceae), C. schiedeanus Schldl., Hampea platanifolia Standley M eem Hyperbaena Smith) Standley (Menisper- maceae), Olmedia aspera Ruiz López & Pavon (Moraceae), Otoba novogranatensis Mold. (My- risticaceae), Paullinia rugosa Benth. ex Radlk. (Sapindaceae), Schizolobium parahybum (Vell. Conc.) S. F. Blake (Caesalpiniaceae), Turpinia oc- cidentalis Don (Staphyleaceae), Unonopsis theo- bromifolia Zamora & Poveda (Annonaceae) y Vi- tex cooperi Standle leptobotryosa (J. D. 135 у (Verbenaceae). La “raza norteña” habita en alturas de —1,400 m.s.n.m. en bosque tropical muy hümedo en transición a premontano y bosque pluvial premontano (Tosi, 1969) con clima hümedo caliente con una modera- da estación seca (Herrera, 1985), frecuentemente asociada con otras Rutaceae (Angostura nicara- guensis (Standley & L. O. Williams) T. Elias, Ery- throchiton lindenii Planchon, Galipea granulosa Kallunki, y Stauranthus perforatus Liebm.). En espera de mas datos, especialmente medidas de partes florales, dejamos los dos “‘razas”’ sin recono- cimiento nomenclatural. Stauranthus Паша, ез ип género de Rutaceae conocidas са amente con dos especies de México (Standley, 1923). Stauranthus perfora- tus esta actualmente conocida fuera de la Repub- lica Mexicana por las colecciones siguientes: Costa RICA. CARTAGO: along Camino Raiz de Hule, SE of Platanillo (Tsipiri), 1,200-1,400 m, 1 julio 1976, Croat 36799 (MO). PUNTARENAS: Reserva Biologica Monteverde, 8 agosto 1975, L. J. Poveda 1112 (CR, MO). PANAMA. COCLÉ: hills above El Valle, ca. 1,000 m, 24 diciembre 1972, Gentry 6874 (MO) LITERATURA CITADA ENGLER, А. 1931. Rutaceae. /n A. Engler & K. Prantl, Die Natürlichen Pflanzenfamilien, ond edition, 19a: 187-359 HERRERA, W. 1985. Clima de Costa Rica. Editorial Universidad Estatal a Distancia, San José, Costa Rica. Morton, C. V. 193 A new species of Peltostigma from Mexico. Proc. Biol. Soc. Wash. 46: 83-84. Annals of the Missouri Botanical Garden MULLER, C. H. . A revision of Choisya. Amer. idl. Pago 24: 729-742. STANDLEY, P. C. 23. Rutaceae. /n: P. C. Standley, rees — sina of Mexico. Contr. U.S. Natl. Herb. 23(3): 5 53 VE STEYERMARK. 1946. Rutaceae. т: P. C. Sta idis & J.A a Flora of Guatemala. Fieldiana, Bor. 24(5): 389- Tosi, is 19 Mapa о M la clasificación onas de vida del mundo de L. R. Holdridge. iN Cientifico Tropical, San Jose, Costa Rica. WATERMAN, P. С. 1975. Alkaloids of the Rutaceae: their distribution and systematic significance. Bio- chem. Syst. Ecol. 3: 149-180. —Quirico Jimenez M., Instituto Nacional de Bio- diversidad, Apdo. 3100, Santo Domingo de He- redia, San José, Costa Rica y Roy E. Gereau, Missouri Botanical Garden Herbarium, P.O. Box 299, St. Louis, Missouri 63166-0299, U.S.A. TWO NEW SPECIES OF TRIXIS (ASTERACEAE-MUTISIEAE) FROM PARAGUAY Two new species of Trixis are described from Paraguay. Trixis cerroleonensis Soria & Zardini from Occidental Paraguay is most similar to T. ophiorrhiza Gardn. but has much smaller phyl- laries. Trixis ravenii Zardini & Soria from Oriental Paraguay is most similar to T. spicata Gardn. but has deltoid leaves decurrent at the base. The genus Trixis (Asteraceae- Mutisieae) com- prises 60 neotropical species ranging from the southern United States and Mexico to northern Chile, northern Argentina, and Uruguay. The ge- nus is in need of revision; only local treatments are available (Anderson, 1972; Baker, 1884; Cabrera, 1936; Loja, 1969). During our exploration work in Paraguay, we have found two new species, which we describe here. Trixis ravenii Zardini & Soria, sp. nov. TYPE: Paraguay. Amambay: Arroyo Estrella, 22°20'5, 56?20'W, 7 May 1989, Soria, Ba- cigalupo & Fortunato 3825 (holotype, ЕСО). Figure Herba 0.50-1 m. Caule partim alato inferne simplici, u bracteis lanceolatis dense se pilosis, extimis paucis. Flores 5. Achenia cylindrica pilosa rostrata. Pappus albus. Herbs 0.50- 1 m tall, erect, the individual stems with few branches. Stems partially winged, 3-5 mm wide, pilose. Leaves sessile; blades 3-6 cm long, 2-6 cm wide, rhomboid or deltoid, acute, decurrent at the base, pilose below, sparsely pilose to hispidulous above; margins dent Mid midrib and nerves prominent below. Inflorescence spi like. Accessory bracts 2-3, 5-8 m cas 3 mm wide, lanceolate or ovate, acuminate, боз below, convex, margin entire; midrib and nerves prominent above. Phyllaries in two series, the out- side 2-3, 6-8 mm long, 2-3 mm wide, margin entire, acuminate, pilose below, midrib and nerves prominent below; the inside 5-6, 10-11 mm long, 2 mm wide, lanceolate to ovate, pilose below, mar- gin entire, apex acuminate, midrib and nerves ANN. MISSOURI Bor. GARD. 78: prominent above. Receptacle 4-5 mm wide. Flo- rets 5. Tube of the corolla 5-6 mm long, 0.1-0.2 mm wide at the base, ] mm wide at the mouth, glabrous or with few trichomes or glandular hairs; outer lip 1.5-2 mm long, 0.1 mm wide, glabrous or with few hairs, flat; lobes of the inner lip 1.5— 2 mm long, 1-1.5 mm wide at the base, glabrous or with few trichomes at the apex. Filament 1.5- 2 mm long. Anthers 4.5-5 mm long, appendages 0.5-1 mm long; thecae 2.5-3 mm long; tails 1- 1.5 mm long. Style 5.5-6 mm long below the bifurcation, the branches 1-1.5 mm long. Achenes rostrate, 3—7 mm long, 0.3–0.5 mm wide, covered with glandular and nonglandular hairs. Pappus 7- 9 mm long. Trixis ravenii is a distinctive species, most sim- ilar to Trixis spicata Gardn. from Central Brazil, which has winged stems and oblong leaves. We dedicate this species to Peter H. Raven for his efforts in institutionally establishing Paraguayan botany 10 years ago. Trixis cerroleonensis Soria & Zardini, sp. nov. TYPE: Paraguay. Chaco: Cerro Leon, 8 Aug. 1983, Hahn 1555 (holotype, PY; isotype, MO). Figure 2. Frutex dense piloso et glanduloso. Caule haud alato. Folia sessilia ellipticis vel lanceolatis denticulata. Capitula corymbosa. Involucrum campanulatum 1.3-1.5 mm lon- gum, bracteis linear-lanceolatis, extemis paucis. Flores 20-30. Achenia cylindrica pilosa rostrata. Pappus albus. Shrubs 1-1.5 m tall, erect. Vegetative parts, accessory bracts and phyllaries with glandular and nonglandular hairs. Stems not winged, glandular and pilose; internodes (0.1-)0.5-1.5 cm long. Leaves subsessile; blades 2.5-10 cm long, 1.5-2 cm wide, usually elliptical or lanceolate to oblan- ceolate, acuminate, attenuate at the base, glandular and pilose above, glandular and sericeous below; margin denticulate; midrib and nerves prominent below. Inflorescence a corymb; peduncles 1-2.5 cm long, glandular and sericeous. Accessory bracts 5-6; 10-20 mm long, 2-6 mm wide, lanceolate 531-534. 1991. 532 Annals of the Missouri Botanical Garden N 7. [= NY AM Ñ Q ) 777 АК N Y) ИД Wy) 4 SE уу SS í y Ug у K \// ^w Q Ny Y, Y FIGURE l. Trixis ravenii Zardini & x10.—E. Style, x10.—F. Achene x to oblanceolate or elliptic, acute, sessile or subses- sile, glandular and pilose or sericeous on both sur- faces, margin entire, flat. Phyllaries 10-12, 13- 15 mm long, 1.5-2 mm wide, linear-lanceolate, acute, convex, with glandular and sericeous hairs. A Soria. — A. Habit, x 0.5. — B. Capitulum, x8.— C. Flower, x 10. — D. Anther, 10. Receptacle with nonglandular hairs. Florets 20- 30; tube of the corolla 6-7 mm long, glabrous or with glandular and sericeous hairs; outer lip 2.5 mm long, glabrous; lobes of the inner lip 2 mm long. Anthers 5-6 mm long, tails 2 mm long. Style Volume 78, Number 2 Notes 533 1991 FIGURE 2. Trixis cerroleonensis Soria & Zardini.— A. Habit, x0.5.— B. Capitulum, x4.—C. Flower, x5.— D. Anther, x 10. — E. Style, x 10. — F. Achene, x10. 5 mm below the bifurcation, branches 1 mm long. is most similar to Trixis ophiorrhiza Gardn. from Achenes 7 mm long, 0.5 mm wide, papillose, ros- Central Brazil, which has larger phyllaries. trate. Pappus white. We thank C. Jeffrey (K) for allowing access to photographs of the types, and Gloria Delmas and Trixis cerroleonensis is a distinctive species. It Adrian Eliceche for help with the illustrations. 534 Annals of Missouri Bruns Garden LITERATURE CITED ANDERSON, C. 1972. A monograph of the Mexican and Central American species of Trixis (Compositae) Mem. New York Bot. Gard. 22: 1-68. BAKER, J. G Compositae- Mutisieae. /n: Martius, Flora Ва епа 6(3): 382-392. CABRERA, А. L. 1 Las especies argentinas y uru- га del genero “Trixis. Rev. Mus. La Plata (n.s.) : 31-86. ГОЈА, B. H. A. 1969. Revisión de las especies peruanas del género Trixis бш Publ. Mus. Hist. Nat Javier Prado, Bot. 25: 1- —Nelida Soria, Facultad de Ciencias Químicas, C.C 1055, Asunción, Paraguay; and Elsa Zardini, Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166-0299, U.S.A INTRASPECIFIC VARIATION IN PAVETTA RIGIDA (RUBIACEAE): ESTIMATES OF RELIABILITY OF TAXONOMIC INFORMATION Variations in number of flower buds per inflo- rescence, number of inflorescences per plant, ap- proximate plant height, leaf blade length and width ranges, number of secondary lateral leaf veins, and leaf length/ width ratio are analyzed for a sample of 11 plants of Pavetta rigida near Lake Barumbi Mbo, Southwest Province, Cameroon. Lowest prev- alence of bacterial leaf nodules is noted to coincide with lowest flower number among plants sampled. Pavetta rigida provides quantified evidence that, for purposes of plant identification, ranges of vari- ation in individuals being identified and ranges of variation in keys and descriptions based on one or few collections can often be doubled in attempts to fit specimens being identified into existing tax- onomic groups. Pavetta L. (Rubiaceae subfamily Ixoroideae) is a paleotropical genus of ca. 400 species, mostly forest shrubs, having four-merous flowers, nonre- curved stigmas, and the “Ixoroid” pollination mechanism sensu Bremekamp (1934) and Rob- brecht (1988). Eleven individuals of P. rigida Hiern subg. Pavetta were found in flower bud shortly before anthesis near a path around the perimeter of Lake Barumbi Mbo near Kumba, Southwest Province, Cameroon, inside the crater containing the lake. The distance between the farthest apart of the 11 individuals was estimated to be ca. 1 km straight line distance or nearly М of the circum- ference of the lake. The nearest two individuals were separated by ca. METHODS Field recordings of the ranges of variation in the following characters were made from April 18 through April 23: numbers of flower buds per in- florescence in inflorescences nearing anthesis, numbers of inflorescences per plant, approximate plant heights, leaf blade lengths and widths, and numbers of lateral veins on one side of leaf midribs. Leaf length/width ratios were calculated later. Ranges and midranges of character states for in- dividual plants, and the resulting total range in the plants studied, were determined from the data in Figure 1. The mean per-plant range was deter- mined for each character and compared with the total range for the 11 plants in the study. Com- parisons of the ranges of variation observed in this study were then made with earlier species descrip- tions of P. rigida (Hiern, 1877; Bremekamp, 1934). RESULTS FOR QUANTITATIVE CHARACTERS Raw data are shown in Figure 1. Overall ranges, as well as means and standard deviations of the individual plant midranges, are presented in Table 1. Mean per-plant ranges are expressed as per- centages of the total range of the 11 plants in able 2. The comparisons of ranges of variation observed in this study with earlier species descrip- tions are presented in Table 3. Results presented in the figure and tables not thought to be self- explanatory are discussed below. NUMBER OF FLOWER BUDS PER INFLORESCENCE Flower buds were counted on each inflorescence nearing anthesis on each of the 11 plants. Three plants had one such inflorescence, seven had two such inflorescences, and one had four such inflo- rescences. The mean of the midranges in number of flowers per inflorescence on each individual of 82.4 buds and standard deviation of 32.6 buds (Table 1) do not fully express the pattern of vari- ation in this character. Of the 11 plants, five had all inflorescences nearing anthesis with 86-124 flower buds, five had all with 50-72 buds, and one had all with 19-41 buds. The 11 plants can thus be divided into three nonoverlapping groups with regard to number of flowers per inflorescence. If the original species description had included num- ber of flowers per inflorescence and had been based on any one of these plants, the probability that the ANN. Missouni Вот. GARD. 78: 535-538. 1991. 536 Annals of the Missouri Botanical Garden TABLE 1. standard deviations in 11 individuals of Pavetta nu Total character state ranges, mean Stand- ard devia Mean tion of mid- of mid ranges ranges or or indi- indi- vidual vidual Character Range plants plants Number of buds per inflorescen 19-124 82.4 32.6 cvs of i E per plan 1-5 med of lateral veins! 6-15 9.6 1.0 Leaf blade length, c 14-48 29.0 3.2 Leaf blade width, 5-17 11.5 1.4 ж blade length/ width 2.1-3.2 2.6 0.4 Plant pem m 1.0-2.5 = — ' Only 10 plants scored for this character. second plant found, even if in the same area, would have matched the description would have been less than 0.5. The original description (Hiern, 1877) did not mention this character and Bremekamp (1934), in his monograph of Pavetta, indicated 100." flowers to be “‘circ. NUMBER OF INFLORESCENCES PER PLANT As seen in Figure 1, almost two-thirds of the plants bore two inflorescences. Nearly, but not quite all, of the inflorescences were nearing an- thesis. NUMBER OF LATERAL VEINS ON ONE SIDE OF MIDRIB Lateral veins included in these counts are only those veins originating at the midrib and extending all the way to a submarginal position on the leaf blades and larger in diameter than other veins originating from the midrib—1.e., secondary veins sensu Hickey (Hickey, 1973) originating from the midrib. This character was determined for each of 10 individuals. LEAF BLADE LENGTH / WIDTH RATIO This ratio was determined for each of the 11 plants as the midrange of leaf blade lengths divided by the midrange of leaf blade widths. The within- plant ratios so determined were scored to find the overall range in leaf length /width ratios. PLANT HEIGHT Plant height was estimated to the nearest 0.25 m. COMPARISON WITH EARLIER SPECIES DESCRIPTIONS As shown in Table 3, earlier species descriptions encompass considerably less than the complete range of variation revealed by measurements of 10 or 11 plants in all characters for which there are comparable data; the present study of 11 in- dividuals failed to detect the complete range of variation known in plant height and leaf blade length. OBSERVATION OF FEWER FLOWERS ON PLANT WITH PAUCITY OF BACTERIAL NODULES The one plant (voucher: Manning 1777 MO) whose two inflorescences had the least flowers (41 and 19) was also the plant with least extensive leaf growths, termed bacterial nodules (Bremekamp, 1934; Lersten & Horner, 1976) or leaf galls (Rob- brecht, 1988), characteristic of most Pavetta spe- cies. These growths were not prevalent on the plant and individual leaves sometimes lacked them, un- like all the other individuals in this survey. This TABLE 2. Ranges of variation in individual plants compared to total range of variation in 11 plants. Mean % of Range of ranges Mean range Total range Character in individuals рег individual of 11 plants рег individual Number of buds per inflorescence! 1-38 13.5 105 13 Number of lateral veins 3-7 5.3 10 53 Leaf blade length, c 10-29 17.9 34 53 Leaf blade width, cm 4.5-11 6.6 12 55 ' Only eight plants were included for this character because the other three had only one near-anthesis inflorescence ach. Volume 78, Number 2 Notes 537 1991 A. Numbers of flower buds per inflorescence. Inflorescences on the same plant are connected by horizontal lines. oe e O) е——5 Фо 4 9 =P toe фә Г. = — > 0 20 40 60 80 100 120 B. Plant heights, m. C. Leaf length/width ratios Н a E 1 1.5 2 2.5 31 2.4 2.7 3.0 3.2 D. Leaf lengths (@—®) and widths ( e———»). Each horizontal line represents the range in length or width for one individual in cm. Length and width ranges are shown side by side for each individual. ~ d Г ч | Y да = с - а. 4 Сон d Y Y oe В سس‎ @ ————& ۾ ټيټ م ت‎ yg ans a "= =D в $ —— — о nn 5 10 15 20 25 30 35 40 45 48 E. Numbers of inflorescences F. Ranges in numbers of per plant lateral veins on one side of midrib. Each horizontal e i line represents one ө ө | — И 2 ? individual. 6 8 10 12 1415 FIGURE l. Intraspecific variation in selected characters of 11 individuals of Pavetta rigida. plant appeared otherwise normal. The possible function of these nodules has been a topic of con- siderable interest (Lersten & Horner, 1976). To Only 13.5/105, ог ca. Y, of the total range in the possible functions may now be added promotion numbers of flower buds per inflorescence in this of initiation and development of increased numbers geographically limited sample of Pavetta rigida of flowers. was typically expressed in single individuals with DISCUSSION AND CONCLUSIONS 538 Annals of the Missouri Botanical Garden TABLE 3. Comparison of ranges of variation stated for Pavetta rigida in its original and amended descriptions and from recent field measurements on 11 individuals with total reported ranges from all three sources. Blanks mean no data available. Hiern data converted to metric units from original. Range from Range from Hiern Bremekamp Range from (1877) 93 aut % of % of % of Total range Character Raw total Raw total Raw total Number of lateral veins 10-12 30 -11 40 6-15 100 6-15 Leaf blade ка cm! 12.7-21.6 25 15-24 25 14-48 96 12.7-48 Leaf blade ; 6.75-8.5 15 7.5-11 29 5-17 100 5-17 Leaf blade eeth widik ratio 2.5 2.1-3.2 100 2.1-3.2 Plant height, m 3.0-3.7 26 1.0-2.5 56 1.0-3.7 ' Hiern leaf blade lengths obtained by subtracting petiole lengths from leaf lengths. more than one inflorescence. This cautions against using flower number as a defining character when delimiting taxa or creating keys based on few spec- imens. For vegetative characters, ca. % of the total range of variation of this sample of P. rigida was typically expressed in individuals (5376 for number of lateral veins, 53% for leaf blade length, and 5596 for leaf blade width). Examination of her- barium specimens may thus actually reveal less than Yê of the total range of a similar sample for these characters, and ranges of variation in leaf characters stated in descriptions based on single or few collections should perhaps be at least doubled by users of keys and descriptions to best estimate true ranges of variation. There is also significant variation among individuals in each of the other characters surveyed. e plants examined for this study, taken from a limited geographical area, may be reproductively within the same population. They cover a minute fraction of the geographical range of the species, which has been reported from Bioko Island, Equa- torial Guinea, and a fairly large portion of the forested area of Cameroon. Thus, the results of this study may still understate the actual range of variation in characters surveyed. Table 3, com- paring results of this study with observations of earlier e indicates this to be the case for plant hei ften a and descriptions of tropical plants are difficult to use at species and infraspecies levels because their character state ranges do not exactly fit specimens under study. It then becomes nec- essary to judge whether (1) the plant at hand rep- resents an extension in the range of a particular character state, in which case the specimen may be referred to a previously described taxon; (2) the lant represents a new taxon; or (3) the user has made an earlier error in identifying the plant. The results reported here emphasize that at least for certain quantitative characters, the first of the three p is to be expecte I thank Rick Clinebell, Peter Imrey, and George Rogers for helpful suggestions. This work was sup- ported in part by a grant from Sigma Xi, the Scientific Research Society, by the Alumni Fund of the Missouri Botanical Garden, and by the Bi- ology Department of Saint Louis University. All of this assistance is gratefully acknowledged. LITERATURE CITED кап C. E. В. 1934. А monograph of the genus Pavetta L. Repert. Spec. Nov. Regni Veg. 37: 1- HICKEY, 1. Ј. 1973. e BA of the architecture of dicotyledonous leaves. ot. -33. .P. 1877. Rubiaceae. e D. F. Oliver (editor), Flora of Tropical Africa 3: 178- LERSTEN, N. К. & Н. T. HORNER, ЈЕ. 1976. Bacterial leaf nodule symbiosis in angiosperms with emphasis on Rubiaceae and Myrsinaceae. Bot. Rev. (Lancas- ter) 42: 145-214 ROBBRECHT, E. deus Tropica woody Rubiaceae. Opera Bot. Belg. 1: 1- —Stephen D. Manning, Department of Biology, Saint Louis University, St. Louis, Missouri 63103, U.S.A., and Missouri Botanical Garden, Р.О. Box 299, St. Louis, Missouri 63166, U.S.A. Pres- ent address: Biological Survey, New York State Museum, Albany, New York 12230, U.S.A. Volume 78, Number 1, pp. 1-272, of the ANNALS ОЕ THE MissoURI BOTANICAL GARDEN was published on February 26, 1991. Monograph in Systematic Botany from the Missouri Botanical Garden Number 37 Catalogue of the Flora of Missouri George Yatskievych and Joanna Turner хй, 345 pp., 5% x 8% їпсһез, phficover $9.00, s tia The Flora of Missouri Project is a ios effort таван ће Missouri Department of Con- servation and the Missouri Botanical Garden to produce a fully revised edition of Julian Steyermark's well-known Flora of Missouri, first published in 1963. The Catalogue. represents a first step. e this goal and contains an annotated checklist of the 2,622 sp d hybrids, as well as z) numerous infraspecific taxa, that are eee мара to grow outside of cultiyátion в in the state. : The ed features: e Updated nomenclature 2127. | € Common names : e. . Cross-referencing | to обата s Tiele s page r number * Additions to the flora indicatd with asterisks ri, » Introduced. taxa Лот): with diamonds | hae То pace an order, dae vs der is OS. Pen $1.50 for as Leak uel 9:90 € oe book; non E 2 each additional book. ‘Orders Mori to repaid, а $1.00 fee vil be added to wens раме e are made until parents is sedie, Mert forin with m check or дека veri Зана to Missouri vi Botanical Garden Y Ce т m J the Flora Џ Missouri to CONTENTS Biological Extinction in Western Ecuador C. H. Dodson & А. OCA E ARS | D. of зае ШОКАН? Using Protein Sequences Р. 6. Martin & J. M. Dowd DI 4 Be | Relationshi fa the Genera of Angiosperms in the Fanjing Shan Mountain Range, ES Guizhou, China Tsun-shen Ying, David E. Boufford & Yuling Tu ... | | ; С ; Revisión del Genero Sireptostachys ола РАЗЫН АДА su dM Sistematica dentro ERIS Tribu Paniceae Osvaldo Morrone & Fernando Zuloaga... ^ | _ А Key to the Genera of Lauraceae in the New World Henk van der Werf on. | Two. New Genera of. Neotropical Lauraceae and Critical Remarks on the Generic УМ ` Delimitation _ Jens б. Rohwer, Hans Georg Richter & Henk van der Werff- € 8 rS eth a New Genus from Costa Rica Henk van der Жа] & Peter : Ke lress .. | = New Species of Lauraceae from Ecuador: dnd Pus | Henk. van Der Werff С of Venezuela Nancy Herse a на J Revision es Я of ES Genus, Rondonanthus ХУМА (Eriocaulaceae) Nancy UNS _ Hensold & Ana Maria Giulietti .. XM > praia of the Genera Comanthera. and: SS with Syngonanthus Eso Te Studies i in ‘Neotropical Paleobotany. ` Ж; The Pliocene Cobian of Panama— Composition, TM Numerical. Representations, and Ра > сащасвае) — Ana Maria Giulietti & Nancy Hensold === = Mbit они Берези > ice Ead И: Todsia & Richard с Kenta: pe 4 6 0 ا‎ matics шайр: Sivek: Hexandrae | ` Favio González G. pud To : ET Сабе еа Herbert (Iridaceae- Tigridieac]: Notes « on This New World Genus and Reduction — узора СГ Зор) ingosty lis, Car ке Бую, and Catila | Peter Sola v c James E. Henrich .. ФЕ; Wulf. IE Stuessy, D. J. Crawford & EIC E SE TÁ : ~ i Chromosome Numbersi in de Chin Nassauvia: (Coupes Майин) | S. H Hunziker, AS ee ue wheeleri v var. ене А New Combination i in the Nolinaceae > p Y NOTES D Charlotte М. Taylor ... Jiménez M. & Roy E. Gereau Ela Zardini .. Се ыы NS те New. Bid >" Hillia E Каша (ulace) from. Venezuela E Mexico 4 E RR è А lundelliana, a New swa of ace fon SESS S 35 Pipoly ш E: x | : уе Peltostigma parviflorum (Rutaceae), Nueva Especie. de Costa Rica у. Cor. ba» Quirico ++ | à oa Two New Species of. Trixis Users Mai f from. Br. К Soria 4 & je Annals of the Missou П Volume 78 \ 7 Number 3 ¡22 | VCS Nos ж Volume 78, Number 3. Fall 4994 Annals of the Missouri Botanical Garden The Annals, published quarterly, contains papers, primarily in systematic botany, con- tributed from the Missouri Botanical Garden, St. Louis. Papers originating outside the Garden will also be accepted. Authors should write the Editor for information concerning arrangements for publishing in the ANNALS. Instructions to Authors are printed in the back of the last issue of each volume. Editorial Committee Marshall R. Crosby Editor, Missouri Botanical Garden А Amy Scheuler . Managing Editor, Missouri Botanical Garden Diana Gunter Editorial Assistant, T Missouri Botanical Cordis = Magdalen Lampe’ | ынс ко Peter Goldblatt Gerrit Davidse Missouri Botanical Garden John D. Dwyer Missouri Botanical Garden & Saint Louis University Missouri Botanical Garden | J Dale E. Johnson Missouri Botanical Garden Henk van der Werff Missouri Botanical Garden For резал то Eleven, Р.О. Box 299, St. Louis, MO 63166-0299. - Subscription price is $75 per volume U.S., $80 — _ Canada and Mexico, $100 all other countries. Four _ issues per volume. Beginning in 1991, recipients of the ANNALS will automatically receive Novon, the - _ Garden's newly ванне scientific quero. © OS = 1991 -. _ POSTMASTER: Send address changes to ANNALS OF THE MISSOURI BOTANICAL GARDEN, Department Eleven, P.O. Box 299, St. Louis, MO 63166-0297 The ANNALS OF THE MISSOURI BOTANICAL pn (ISSN 0026-6493) is published quarterly by the ; Missouri Botanical Garden, 2345 Tower Grove Ar - enue, St. Louis, MO 63110. Second class postage paid at St. Louis, MO and additional mailing offices. Volume 78 Number 3 1991 Annals of the Missouri Botanical Garden WZ A REVISION OF ANTHURIUM SECTION PACHYNEURIUM (ARACEAE)! Thomas B. Croat? ABSTRACT s is the first published revision of Anthurium sect. Pachyneurium since that of Engler in 1905. Section ler Pachyneurium is one of 19 sections (all but one of ned were treated ЗА Еп Mer The section consists of 114 species е ing 126 t series Multinervia Croat, with 110 and are described as new D^ in his revision of Anthurium). two se Pachyneurium Schott, and the new 16 taxa ا‎ а Forty- Чан ud ү а nine subspecies or varieties) Section Pachyneurium is one of 19 recognized for Anthurium and is perhaps the most easily rec- ognizable and well-defined group in the genus. It was for this reason, despite its size and taxonomic difficulty, that this section was chosen as a starting point for a revision of the genus. Section Pachy- neurium is generally defined by its frequently rosu- late or **bird's-nest'' habit, its short, densely rooted caudex, the commonly short-petiolate, oblanceolate to obovate, mostly coriaceous leaf blades with usu- ally free-ending primary lateral veins and, most importantly, the involute vernation (rolled inward from both margins) of the developing leaves (Fig. 1). All other sections of Anthurium, and indeed all other genera of Araceae (except Lagenandra), have convolute vernation (Fig. 2), with one margin rolled inward toward the midrib and the other mar- gin rolled around the midrib and the rolled up opposite margin (like a continuous coil in cross section). The section consists of two series with most species in series Pachyneurium. The other series, Multinervia, is restricted to the Andes of South America, especially in Ecuador. Series Pachyneu- rium contains all the species originally included in grex Pachyneurium by Schott (1860), and most of those included by Engler (1905) in sect. Pachy- neurium. Series Multinervia contains species un- known to Schott and mostly unknown to Engler. The few species of series Multinervia treated by Engler were incorrectly placed in sect. Polyneu- rium or in sect. Urospadix, the latter having been employed by Engler as a rather broad “dumpin ground." See Appendixes 1 and 2, respectively, ' This 8306297. poe for publication was provided in "E. study was completed with support from National Science Foundation grants DEB 80-11649 and BSR art by National Science Foundation grant BSR-8 8914018. m pa Schulze, Curator of Botany, Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166-0299, U.S.A. ANN. MISSOURI Вот. Garp. 78: 539-855. 1991. 540 Annals of the Missouri Botanical Garden for the ultimate disposition of species treated in sect. Pachyneurium by Schott and by Engler. See Appendix 3 for the disposition of all currently rec- ognized Pachyneurium species and their sectional placement in the revisions of both Schott and En- gler. METHODS AND MATERIALS This revision was based on field studies in Central America between 1967 and 1979 and in South America in 1969 (Colombia), 1972 (Peru), 1976 (Colombia and Ecuador), 1980, 1983, 1984 (west- ern South America), 1984 (Venezuela), and 1986 (western South America and Brazil). All but 28 of the 125 taxa were studied live or are under cul- tivation at the Missouri Botanical Garden. Except for those so indicated, all descriptions can be as- sumed to have been prepared from both living and dried material. Morphological characters were cod- ed directly into a computerized database to ensure parallel and sortable descriptions. The aroid de- scriptions database contains 228 characters that are used to describe the morphological diversity expressed by Anthurium. The database also pro- vides a means of sorting species by characters for writing keys, as well as for compiling lists of various character states. [n addition, the database can be put to future use for identifications and for additions of new species for other treatments. The description file is directly tied to a nomenclatural database containing all species names and publication data for all Araceae species stored in Tropicos (Crosby, 1986; Crosby & Magill, 1986). Each species is represented by a unique number, which ties it to data in other files associated with the species. Dis- cussions and exsiccatae are also stored in separate files, as are synonyms, but all parts are automat- ically reassembled for the final treatment. Herbar- ium specimens can be added to the exsiccatae database at any time and sorted in a standardized manner before being printed out. Thus, exsiccatae records can be added up to the last moment, even when the manuscript has been completed. Species descriptions are decoded into narrative text auto- matically from the descriptor base and require only minor editing to put them in a publishable form. Final treatments may be saved on hard disk or stored on tape, but the coded description data (which use very little core storage space) are a part of the permanent Tropicos database and can be used for other projects such as floras or ecological studies. The terminology and usage in the descriptions in this paper are defined by Croat & Bunting (1979). The color of the spadix, unless otherwise indicated, refers to the color at anthesis, i.e., when the spadix is producing stigmatic droplets or when fresh sta- mens are emerging. Colors referenced in the de- scriptions that follow are taken from the color chart by Berlin & Kay (1969). This is hereafter cited as his color chart, available from the University of California Press, is a reproduction of the Munsell Color Array of 40 hues, at maximum saturation, with nine degrees of brightness. The B & K color chart represents 40 hues in the vertical columns and nine degrees of brightness in the hor- izontal rows. Colors are arranged in 10 basic clus- ters with four different hues per cluster, ranging from red through yellow, green, blue, purple, and finally red-purple. The four columns for each color cluster are numbered 2.5, 5, 7.5, and 10. These numbers are repeated for each basic color type. The colors from the B & K color chart are read by first reporting the color, then the row followed by the column. For example, the third color in the fifth row in the red area would be called Red 5/7.5. The second color in the eighth row would be called Red 8/5 All estimates of ecological zones given in this paper are based on either Holdridge Life Zone maps for most Central American countries and for Pan- ama or on the “Мара de tipos de vegetación de la Repüblica Mexicana" (Flores et al., 1971). As yet no study has been made to correlate the veg- etation types represented on the Mexican map with those of the Holdridge Life Zone system. For an understanding of the latter system, see Holdridge et al. (1971). For South America, maps based on the Holdridge System were used for Bolivia, Co- lombia, Ecuador, Peru, and Venezuela. Mention should be made of the distribution of herbarium material of Araceae under cultivation. Herbarium material may consist of one of three kinds: (1) fertile original (wild) collections; (2) sterile original collections with an inflorescence added from the cultivated plant of the same number (generally the same individual from which the field specimens were made); and (3) material collected entirely from the cultivated plant. Specimens based entirely or in part on cultivated material are clearly indi- cated as such in herbaria. ACKNOWLEDGMENTS I acknowledge the assistance of numerous hor- ticulturists who provided living material, especially M. R. Birdsey, C. McDaniel, J. Williford, R. Burle- Marx, S. Mayo, M. Carmichael, J. Brenner, S. Thompson, F. Fuchs, and T. Fennell, Jr., and to others who provided observations on leaf vernation Volume 78, Number 3 1991 541 Anthurium sect. Pachyneurium and fruit color, namely J. Banta, C. Fleming, M. H. Grayum, and R. Sheffer. R. Sheffer also has assisted greatly by studying the cytology and breed- ing behavior of Pachyneurium at the University of Indiana Northwest. I thank the following insti- tutions, which loaned or allowed me to study their specimens: А, AAU, B, BH, BM, BR, C, CAS, CAY, CHAPA, CM, COL, CR, DUKE, ECON, ENCB, F, ЕТС, С, GB, GH, GUA, HUA, IAN, IBE, INPA, ITIC, K, LAM, LL, M, MEDEL, MEXU, MG, MICH, MY, NY, P, PMA, QCA, QCNE, R, RB, S, SEL, SI, TEX, U, UC, UCLA, UFMG, UB, US, USM, VEN, W, WIS, WAG, XAL. Also acknowledged are reviewers D. H. Nic- olson and M. H. Grayum, especially the extensive work with the manuscript by Grayum. I also thank Nicolson for his considerable help with thorny no- menclatural problems with Pachyneurium. Special credit goes to volunteer workers, in- cluding A. E. Westhoff and Patricia Croat, for preparation of Latin descriptions, and to the latter for final editing; to Kay Rossmann for typing and editing all parts except the descriptions and exsic- catae; to Edwina Medlock for a two-year pheno- logical study of leaf, flower, and fruit production and for her cross-pollination program; to my tech- nical staff for assisting with the project, including past research assistants Frances Mazanec (1980- 1985) and Honora Murphy (1986); to the green- house staff, Petra Malesevich and Mary Nyswon- ger, for answering repeated queries, transporting a myriad of plants to the office, pollinating plants, harvesting fruits, and keeping records on pheno- logical changes; and to Rob Wilds and Dan Mount for participating in the pollination program. Most special thanks to my current research as- sistants, who helped me bring the project to com- pletion: to Anna Brzyski for describing plants, en- tering descriptions, and assembling the final manuscript for descriptions, exsiccatae and species excluded; to Dylan Hannon for assisting in all phases of the project, especially in generating the South American species key, assembling illustrations and final editing, and to Petra Malesevich, who assisted in the final editing process. The revision could not have been finished without the able assistance of these dedicated staff and volunteers. HISTORY OF SECTION PACHYNEURIUM The first species of sect. Pachyneurium as it is recognized here was described by Linnaeus (1763) in Species Plantarum as Pothos crenata, and was transferred to Anthurium by Kunth in 1841. Three other species clearly belonging to sect. Pachyneu- rium were described in the genus Pothos: Pothos crassinervia was described by Jacquin in 1793 and transferred to Anthurium by Schott in 1832. The other two were Pothos solitarius Vell. Conc. described in 1829, and P. maxima Desf., described in 1832. The former was transferred to Anthurium by Schott in his Prodromus (1860), while the latter was transferred to Anthurium by Engler in 1905 and is here treated as a nomen dubium owing to the inability to place it with certainty in any cur- rently recognized species. The genus Pothos as now circumscribed is restricted to Africa and Asia, especially Southeast Asia. Modern suprageneric classifications do not place Pothos close to An- thurium. Grayum (1984) places it in a separate tribe, Potheae in the subfamily Pothoideae. Gray- um considers Anthurium isolated but closer to Po- thos than to any other genus. He places Anthurium in a separate tribe Anthurieae in the same subfam- ily. Bogner & Nicolson (in press) remove Anthur- ium from the Pothoideae altogether and place it in the subfamily Lasiodeae, leaving Pothos in the Pothoideae with only two other closely related gen- era, Pedicellarum and Pothoidium. The epithet Pachyneurium was first used in 1860 by H. W. Schott in his first revision of the Araceae in Prodromus Systematis Aroidearum. Although no Anthurium species were treated in his earlier Synopsis Aroidearum in 1856, 20 spe- cies were treated in the Prodromus. Half of these were described by Schott himself in that work or in earlier works during the previous decade. Others were described by Karl Koch or Frederick Lieb- mann during the same decade. In addition to the four species described as Pothos mentioned above, only one species of sect. Pachyneurium, A. cras- sinervium (Jacq.) Schott (described as Pothos cras- sinervia in 1793), was described before the early 1850s. Three of the 20 names included in Schott’s 1860 revision, namely A. acaule (Jacq.) Schott, A. hookeri Kunth and A. aduncum (Vell. Conc.) Schott, were published before 1850 but these have been excluded from Pachyneurium and placed in other sections (see Appendix 1). Of the remaining Pachyneurium names in Schott’s 1860 revision, only nine species remain. These include A. wagenerianum K. Koch & Bouche, A. fendleri Schott, A. spectabile Schott, A. crassinervium (Jacq.) Schott, A. affine Schott, A. schlechtendalii Kunth, А. crenatum (L.) Kunth (erroneously treated by both Schott and Engler as A. acaule (Jacq.) Schott (Mayo, 1982)), A. dom- beyanum Brongn. ex Schott, and A. solitarium (Vell. Conc.) Schott. Still other species of sect. Pachyneurium were described in Schott’s revision, 542 Annals of the Missouri Botanical Garden but these he included in other groups (termed "greges" by Schott). These included А. protensum Schott (included in grex Erythropodium), А. oxy- pum Poeppig and А. consobrinum Schott (grex Oxycarpium), and А. oerstedianum Schott (in- cluding its synonym, А. cuspidifolium Schott) in grex Xialophyllium. Thus, а total of 13 species of Pachyneurium were included in Schott's revi- sion, even though some were not recognized as such. In Engler's 1905 revision, 62 names were in- cluded in sect. Pachyneurium (Appendix 2). Of this total, only 18 species were actual Pachyneu- rium species not already included in Schott's re- vision. Fourteen of the names included by Engler (in reality only 12 species, since two were syn- onyms), including A. hookeri Kunth, А. hacu- mense Engl., A. weberbaueri Engl., and all species numbered 52 through 62, have subsequently been proven to be non-Pachyneurium (see Appendix 2). Ten names included by Engler were synonyms of earlier names in Schott’s revision, three names were synonyms of older names in Engler’s own revision and four names have been excluded be- cause of confusion in the nomenclature (see Species Excluded section). These are A. maximum (Desf.) Engl., A. cymatophyllum K. Koch & Sellow, A. agnatum Schott, and A. tricarinatum Sodiro. In addition to the 18 species that Engler added to those Schott had included in his grex Pachy- neurium, Engler included the following elsewhere: A. oxycarpum (sect. Oxycarpium), A. pallatan- gense Engl. (sect. Polyneurium), A. oerstedianum Schott (including A. cuspidifolium Schott, in sect. Urospadix, series Obscureviridia), A. linguifolium Engl. (sect. Urospadix, series Flavescentiviridia), and A. spathiphyllum N. E. Br. (sect. Episeio- stenium). Of these, only А. pallatangense and А. linguifolium had not been treated by Schott. hus, in the 45 years between the publication of Schott’s revision in 1860 and Engler’s treatment for Das Pflanzenreich in 1905, an additional 22 valid species were described, bringing the total number of Pachyneurium species to 33. The spe- cies included by Engler, but not by Schott, in the present concept of sect. Pachyneurium were as follows (in the order presented in the revision): pendulifolium N. E. Br. paraguayense Engl. Joseanum Engl. (= A. protensum Schott) jenmanii Engl. uleanum Engl. tarapotense Engl. . cubense Engl. >a RR А А А А. leonianum Sodiro A. acutifolium Engl. A. latissimum Engl. A. barclayanum Engl. A. ernestii Engl. A. salviniae Hemsley A. lindmanianum Engl. A. martianum K. Koch & Kolb A. selloum K. Koch A. pallatangense Engl. A. linguifolium Engl. Until recent times, relatively few taxa of Расћу- neurium had been described subsequent to Engler's revision. À list of the names published subsequent to Engler's time and prior to the beginning of this work is presented here: А. agoyanense Sodiro = А. dombeyanum Brongn. ex Schott var. agoyanense (1905) — Brongn var. eleutheroneuron Sodiro (1905) = beyanum Brongn. A. angustilaminatum Engl. var. albidum Sodiro (1906) = inatum Engl. var. brevipes Sodiro (1906) — inatum Engl. var. crassum Sodiro (1906) — inatum Engl. var. gladiatum Sodiro (1906) = laminatum Engl. A. atropurpureum R. Schultes & Maguire var. atropurpureum (1953) var. apertum R. Schultes (1954) = A. bon- plandii R. Schultes & Maguire subsp. bon- plandii A. bonplandii Bunting (1975) A. concolor K. Krause (1932) А. fasciale Sodiro (1905) A. giganteum Matuda (1950) = Hemsle A. guayanum Bunting (1975) = A. bonplandii Bunting subsp. guayanum (Bunting) Croat lanjouwii Jonker & Jonker (1966) . maguirei A. Hawkes (1948) ottonis K. Krause (1932) rodrigoi A. Hawkes (1948) = A. paraguayense Engl. superbum Madison (1978) tessmannii K. Krause (1932) = A. ngl. A. wurdackii Bunting (1975) А. dombeyanum A. dom- А. angustilam- A. angustilam- А. angustilam- A. angusti- А. salviniae aana EN uleanum Of the 21 names listed above, 10 still represent valid taxa, while the remainder have here been Volume 78, Number 3 1991 Croat 543 Anthurium sect. Pachyneurium synonymized. This raises to 46 the number of taxa of sect. Pachyneurium at the time this researcher began working with the Araceae (in approximately 1976). Subsequent to 1976, an additional 16 spe- cies of sect. Pachyneurium have been described. In “The Genus Anthurium (Araceae) in Costa Rica," Croat & R. A. Baker (1979) described six additional species: 4. brenesii, A. prolatum, schottianum, A. seibertii, A. standleyi, and A. upalaense. Two more species were described in preparation for a revision of Anthurium for Pan- ama, namely A. luteynii and А. purpureospathum (Croat, 1981); two species, 4. halmoorei and A. salvadorense, were described in “А Revision of the genus Anthurium for Mexico and Middle Amer- ica" (Croat, 1983) and two more, А. nervatum and А. pseudospectabile, were described in “А Revision of the Anthurium of Panama" (Croat, 1986). In addition, a number of Venezuelan species have been described by Bunting (1986, 1988, 1989) since this project began. These include А. xanthoneurum and А. quanchezii (1986), А. vi- nillense (1988, along with three others introduced into synonymy), and А. iramirezae (1989). Fi- nally, one species, А. plowmanii, was described during the preparation of the treatment of the Araceae for the “Flora de Paraguay" (Croat, 1987). The current work includes a total of 114 species, representing 126 taxa, 48 of those being described for the first time. This is 14776 more species than were recognized at the outset of the work. SECTIONAL RELATIONSHIPS WITHIN ANTHURIUM Section Pachyneurium is one of 18 sections recognized by Engler (1905). While Schott (1860) recognized 28 groups (and this probably will prove to be closer than Engler to the actual number of needed sections in Anthurium), he was dealing with only a small percentage of the species known to Engler. Engler no doubt found it necessary to broaden the concepts of the subgeneric groups in order to include all the new species. In doing so, some groups lost definition. This is particularly true in the case of the species he placed in sections Belolonchium, Urospadix, Polyphyllium, and Xialophyllium, all of which (and especially the first) contain what appears to be a wide variety of seemingly unrelated species. A discussion of these groups, their typification, and proposals for some realignment will be the topic of a future paper. Section Pachyneurium is distinct in its ecolog- ical requirements and is the most edaphically adapt- ed of all the sections to xeric growing conditions. Because of this, the taxa in the section are most prevalent in areas where there is a pronounced dry season, especially in areas of tropical moist forest. While species in sect. Pachyneurium occur in most life zones, most occur in tropical dry, premontane dry, and tropical moist forest. Relatively few spe- cies occur in tropical wet, premontane wet, or montane wet forest life zones. No Pachyneurium species is known from pluvial tropical forest. In habit, especially in terms of their short stems, short, densely rooted internodes and mostly erect leaves held in a tight rosette, Pachyneurium spe- cies resemble most closely species of sect. Por- phyrochitonium. Indeed, some members of that section, for example Anthurium hacumense Engl. placed in sect. Pachyneurium by Engler), are often confused with Pachyneurium (Fig. 5). Still, species of Porphyrochitonium differ by having dark glandular punctations on one or both leaf surfaces, and in having convolute, not involute, vernation and berries which are often depressed apically with two or more seeds per locule. In addition, sect. Porphyrochitonium is ecologically very different, inhabiting some of the wettest life zones such as pluvial rainforest. The Choco region of Colombia, for example, which is one of the wettest places on earth with more than 500 inches (2,272 mm) of rainfall per year, is the center of diversity for sect. Porphyrochitonium. Possibly also a member of sect. Porphyrochitonium is An- thurium hookeri Kunth, which resembles sect. Pachyneurium more than any other non-Pachy- neurium (Fig. 6). It differs from typical Porphy- rochitonium in being much larger and in having scalariform secondary lateral veins and perhaps belongs in a section of its own. Much more easily confused with sect. Pachy- neurium than sect. Porphyrochitonium is a rel- atively small group of species which also has the "bird's-nest" habit and shares similar oblong to oblanceolate or obovate leaf blades, but without involute vernation (a feature considered an absolute requirement regardless of habit or blade shape). Typical of this group is A. michelii Guill. (Fig. 3), once considered to be a member of Pachyneurium (Croat, 1983). These species, with epunctate, ob- lanceolate to oblong-elliptic blades and generally short stems with short internodes are mostly small plants not easily confused with typical Pachyneu- rium species, but some of the larger ones, such as А. michelii, are in fact easily confused with Pachy- neurium. This group of plants has not been placed with certainty in any group, though it is tentatively placed in a new section including A. decurrens Poeppig, placed by Schott (1860) in his grex Ox- ycarpium. Since only three species, A. decurrens — 544 Annals of the Missouri Botanical Garden Poeppig, А. oxycarpum Poeppig, and А. conso- brinum Schott were included in Schott's grex Ox- ycarpium, and since the two latter species are now laced in sect. Pachyneurium (along with the ep- ithet sect. Oxycarpium, which is now synonymous with sect. Pachyneurium) only А. decurrens re- mains. The name for this new section will be p posed in a future paper. This new section is p only section (other than sect. Urospadix) with short internodes and elongate, epunctate leaf blades with convolute vernation. Section Urospadix is, 1 ђе- lieve, distinct from the latter section based on gen- erally closer, more uniform primary lateral and its more restricted geographical distribution, mostly in southeastern Brazil. Section Urospadix (Fig. 4) is the next most likely group to be confused with Pachyneurium. These two are probably not closely related owing to their lack of involute vernation (although one probable species of sect. Urospadix has hybridized veins with а Pachyneurium— see section on breeding behavior). Given that this character state, a con- servative ontogenetic feature, is present only in sect. Pachyneurium and in only one other genus in the family, Lagenandra, it seems apparent that sect. Pachyneurium has no close relatives. Nev- ertheless, sect. Urospadix has many features in common with Pachyneurium, especially the short stems, close internodes, frequently rosulate habit and the presence of elongate, often short-petiolate leaf blades. In gross morphology they differ from members of series Pachyneurium in their generally closely spaced and moderately obscure primary lateral veins. Some members of sect. Urospadix are very similar to members of series Multinervia, a group mostly restricted to the Ecuadorian Andes. From species of series Multinervia they differ prin- cipally in having supervolute vernation, and in be- ing geographically isolated in the geologically more ancient parts of eastern South America, namel Brazil. Section Urospadix ranges principally from the state of Bahia to Sào Paulo with at least two outlying species (tentatively placed in this section), one in the Venezuelan highlands (4. yutajense Bunting) and one from the Cordillera de la Costa (A. lilacinum Bunting). Some other groups of Anthurium, especially species currently placed in sect. Belolonchium in Central America and in the West Indies, have species that are similar to those few species of сота with cordate or subcordate blades, e.g., A. venosum, А. ranchoanum, A. schottian- um, ‚А. standleyi, A. colonicum, and А. cotobrusii. It is believed that this latter group of Pachyneu- rium has independently evolved the cordate con- dition rather than having independently acquired involute vernation. The reasoning here is that leaf vernation is a conservative ontogenetic character whereas leaf blade shape in the genus is a highly plastic one. Indeed, leaf shape may be highly vari- able even on individual plants (see section on leaf blades under Morphology). The Centr now well known, and all suspected species have been screened for the presence or absence of in- volute vernation. Despite the fact that the endemic South American Pachyneurium species are less well known, many cordate South American species ave been studied in cultivation, and no Pachy- neurium species have been found among them. Thus the phenomenon of subcordate- or cordate- leaved Pachyneurium seems to be largely a Central American and West Indian one. It is in these regions where some cordate Anthurium species with short stems, short internodes and blades with free-ending veins (a relatively common feature in man date-leaved Anthurium) might be confused with sect. Pachyneurium. Engler placed a number of these species in his sect. Pachyneurium (Appendix 2). These included А. longispathum Carriére (= A. grandifolium), A. grandifolium (Jacq.) Kunth, А. boucheanum K. Koch (= А. cartilagineum (Desf.) Kunth), А. liebmannii Schott (= А. um- Liebm.), А. umbrosum Li al American species of Anthurium are COr- cartilagineum (Desf.) Kunth, А. brownii Masters, . appunianum Schott (= А. cartilagineum), and A. seleri Engl. Except for 4. seleri, most have many of the general features of sect. Pachyneu- rium and are in many ways similar to Central American and West Indies cordate-leaved species of Pachyneurium, except that they lack involute vernation. Thus, it is not surprising that Engler, apparently unaware of the important character of involute vs. supervolute vernation, placed these species in sect. Pachyneurium. On the other hand, using general characters such as leaf shape, short stems, and short internodes alone, many additional species from the West Indies, Central and South America might well have been placed by him in sect. Pachyneurium. In general, South American species are either not subcordate-leaved, or their blades are decidedly elongate with merely the bases being subcordate (e.g., A. fendleri). The remaining sections of Anthurium are not at all similar to members of Pachyneurium and need not be discussed. Моште 78, Митбег 3 1991 Croat Anthurium sect. Pachyneurium MORPHOLOGY OF VEGETATIVE STRUCTURES GROWTH PATTERNS The shoot organization of Anthurium sect. Pachyneurium has been characterized by Ray (1986, 1987, 1988). Though appearing un- branched, the stems of most anthuriums are in fact highly branched with a growth pattern described as triphyllous sympodial (Ray, 1988). Each article (or segment) of the shoot includes a sylleptic pro- phyll (P) and mesophyll (E) (both of which are cataphylls) as well as a sympodial leaf (S) (a foliage leaf) and a solitary, terminal inflorescence (I). Veg- etative buds are formed only on the sympodial segment and on the peduncle base (which is axillary to the mesophyll and the foliage leaf). This growth pattern can be summarized as follows (after Ray, loc. cit.). Studies by Tom Ray (pers. comm.) have defined the shoot organizations and growth patterns. In this regard sect. Pachyneurium is identical to that of all species of Anthurium with the exception of sect. Polyphyllium. It is the sylleptic prophyll which is termed the cataphyll throughout this work and previous works by this author on Anthurium (Croat, 1983, 1986). The stem thus consists of a series of units (articles), each of which comprises a branch with a single foliage leaf subtending (eventually) a continuation shoot and a terminal inflorescence. STEMS The stems of Anthurium sect. Pachyneurium are characterized primarily by being generally short and densely rooted. While most members of other sections of Anthurium (excluding Xialophyllium and Tetraspermium) have short to moderately short internodes as well, few sections ever have species with such extremely short internodes. Typically, those of sect. Pachyneurium are several times broader than long. Moreover, the caudex differs in being much more densely rooted than in other sections. Depending on the size of the stem and especially the size of the petiole bases, the stem diameter is highly irregular from node to node with the petiole bases conspicuously swollen while the very narrow intervening areas between them are devoted entirely to cataphylls and roots. The ab- scission scar of the petiole is generally broader than high (i.e., broadest laterally, perpendicular to the axis), though the shapes of the scars are usually not as variable throughout the section as is the range of variation of petiole cross sections. Petiole abscission scars are generally rounded abaxially and broadly rounded adaxially. The abscission scar is typically inclined inward to the axis at an angle of 110-140? from the axis of the stem. Unlike many Anthurium species with longer internodes, Pachyneurium has leaf scars that are usually not easily visible but remain obscured by the persistent cataphyll fibers and roots (Fig. 7). The stems of Pachyneurium are generally short, usually less than 20 cm long, and rarely more than 30 cm long except in very large old plants. Stem diameter varies from as little as 1.5 cm in some of the smaller species to about 7 cm diam. in the larger species, generally averaging 3-4 cm diam. Stem diameter is in part also a matter of age, but mostly increases only slightly during the life of the plant. Usually the oldest part of the stem rots away except for that portion contained in the living root mass. Older parts of the stem are frequently at- tacked by root borers, which enter through the softer central portion of the stem and bore up into the younger portions. However, perhaps because of the seasonally dry habitats and the generally massive root ball, which provides protection to the stem, sect. Pachyneurium exhibits fewer cases of infestation by root borers than perhaps any other section of Anthurium. rarely do Pachyneurium species have no- ticeably elongate internodes. Some species, such consobrinum, tend to grow rapidly when young, producing internodes to several centimeters long, ing to a slower growth with short, broad internodes. Some species may revert to a growth pattern with slightly longer internodes when disturbed or dislodged from their positions in trees, especially when they land on their sides or land upside down. This reversion to longer internodes, so common in other genera and even in other sections of Anthurium, is not common in Pachy- neurium, and even when reversion to production of longer internodes does occur, the duration of long internode production is always very brief. Of the two series of Pachyneurium, series Mul- 546 Annals of the Missouri Botanical Garden tinervia has a stronger tendency to produce longer internodes. This no doubt owes to their generally more mesic habitats at higher elevations, where the closer, more compact stems have less advan- tage (and perhaps some disadvantage since the more typical plants in series Pachyneurium would probably become waterlogged and rot in such cool- er, wetter environments). Pachyneurium stems are tough and moderately inflexible, perhaps owing to the presence of nu- merous coarse fibers. Only the central or older parts of the stem are soft, offering one of the few avenues for predation (the other being through the soft area of the abscission scar where the old pe- duncles have rotted away). Pachyneurium stems are typically erect, wheth- er habit is terrestrial or epiphytic. In a few rare cases stems may be repent, e.g., in А. lindman- ianum. This is presumably an ecological adapta- tion, providing the paot with a better means of rvival nthurium l lly úccuIl in seasonally dry areas as a terrestrial шг along the edges of watercourses. In cultivation, the spe- cies will grow well in a standing pool of water on rocks or bricks above the water level. Its roots extend into the water and the typical rosulate habit is still maintained because the petioles are turned sharply at right angles to the stem and are held erect. Despite the obvious ecological advantage of re- pent stems in drier areas, this habit has not evolved frequently in Pachyneurium. More commonly, even in extremely dry habitats, the stem and leaves are stifly erect. Stems may be horizontal to almost pendent in A. pseudospectabile, which grows suspended from a few roots. Potted plants of this species may grow in a curve until the apex is again directed more or less downward. ROOTS Perhaps more than any other section, Pachy- neurium shows considerable variability in root dis- position. Depending on the habitat and the species, epiphytic species in mesic habitats that experience a pronounced dry season, at least the uppermost roots are directed upward and are frequently nar- rowed and pointed at the ends as well (Fig. 8). This provides a clear advantage for collecting falling debris and precipitation around the roots (Figs. 7, 13). These uppermost roots frequently even grow into the basket often formed by the rosulate habit and the uli ide leaves (Figs. 7, 8, 14). The uppermost roots are the most active in terms of growth, and pide act as feeder roots. ven in such situations, still other roots, usually those further down on the stem, act more for an- chorage and support. These are also generally lon- ger, stouter, and less well ordered, being instead more frequently intertwined. They are nevertheless equally capable of entrapping debris and are, at least near the periphery of the living root mass, debris-infested. Because long-petiolate species are less likely to be efficient at catching debris, they are more likely to have their roots directed down- ward rather than upward (Fig. he root masses of most species are extremely compact, the roots generally being completely con- tiguous with one another near the surface of the stem in most species of series Pachyneurium (Fig. 10). Some species, especially many members of series Multinervia, have less densely aggregated roots. This compact nature of the root system not only enables the plant to adapt well to dry condi- tions, allowing it to absorb quickly virtually all of the rainwater that passes through in brief rain showers, but also is effective in keeping phytoph- agous insects, especially root borers, away from the stem. Other Pachyneurium species, especially the ter- restrial ones, frequently have their roots directed downward along the stem and into the ground. These species, despite the fact that they too often have the typical rosulate habit with short stems, are frequently growing under conditions where they are less likely to accumulate debris from falling leaves, etc., and often grow out in the open or under low forest, which does not yield such high quantities of debris. Some terrestrial species, such as А. atropurpureum var. atropurpureum, A. pachylaminum, or A. bonplandii, often occur on white sand soils, which are notoriously poor in nutrients, and are either found growing in open areas or growing beneath typically evergreen trees on white sand soils. In such situations, roots di- rected into the soil appear to be an adaptation to take better advantage of nutrients present in the soil, which might not be available from accumulat debris. On some species, e.g., are much reduced and directed upward rather than downard. The roots themselves, usually 3-8 mm diam., are initially covered with a dense layer of root hairs, so dense as to appear as a smooth and con- tinuous layer. Artificial drying of the stems, which Volume 78, Number 3 1991 Croat Anthurium sect. Pachyneurium causes a shrinkage of the roots and root hairs, reveals their true nature, wherein the roots gen- erally appear woolly-pubescent. At least one ex- ception to the typically smooth roots on fresh plants is seen in 4. johnsoniae, which has fresh roots with a somewhat irregular warty surface. In addition to obvious functional differences, the roots have taxonomic importance, especially the color, thickness, degree of spreading, and length of the uppermost feeding roots. CATAPHYLLS The cataphyll is a modified leaf that protects the newly emerging leaf. It is termed the sylleptic prophyll by Ray (1986, 1987, 1988). Once the leaf emerges, the cataphyll has no functional sig- nificance in remaining photosynthetic (for ecolog- ical significance see below) and promptly dries up. Though the cataphylls of Anthurium до not соп- tribute as many useful characters as do those of Philodendron, the cataphyll in sect. Pachyneu- rium is more diverse than in any other section of Anthurium. There are two basic types of cata- phylls, with one type subdivided: (1) cucullate (hood-shaped) (Figs. 15, 46) (2) narrowly triangular or lanceolate (a) straight (Fig. 11 (b) hook-shaped (Fig. 16). The former type is rare, known for certain only in A. salviniae and А. barclayanum. It is inter- esting to note that the cucullate cataphylls (and sometimes also hook-shaped cataphylls, e.g., А. loretense) also have leaves which are circinate in bud in addition to having involute vernation (Fig. 17). Upon unrolling, the young leaf first unrolls lengthwise, then unrolls toward both margins in the usual manner of sect. Pachyneurium. The second cataphyll type, narrowly triangular or lanceolate, is overwhelmingly more common, and most of these are straight and erect. In South America some species have narrowly triangular cataphylls that are hooked (the phenomenon is absent among Central American species). Plants with typically hook-shaped cataphylls also have sharply D-shaped petioles in cross section (Fig. 16). Blades emerge in the same fashion in both types of narrowly triangular cataphylls, but the hook- shaped type are usually smaller than the straight ones and thus the emerging leaf is usually smaller before expansion. Nevertheless, the overall size of the fully expanded leaf blade does not seem to correlate with the size of the newly emergent but unfolded leaf blade, since some species with hook- shaped cataphyls- have among the largest leaves in the section, e.g., А. loretense. Fresh cataphylls of Anthurium are uniformly green (unlike Philodendron, where they may be colorful), and except for length and shape they are monotonously similar. Moreover, the cataphylls dry promptly after the emergence of each new leaf and thus usually no more than one fresh cataphyll is visible. The old, dried cataphylls are more signif- icant taxonomically, perhaps more characteristic than those of other sections of the genus. Although some species, e.g., А. exim ‚ A. luteynii, A. upalaense, and 4 ‘willifordii have cataphylls that dry persistent and intact, most persist in a weath- ered state (Fig. 12). The degree to which weath- ering takes place varies substantially. While some weather to disorganized and inconspicuous fibers that are mostly obscured by the roots and petiole bases (Fig. 14), other species, especially those with longer cataphylls, dry with their fibers in an or- ganized ог semiorganized intact fashion, e.g., А. luteynii (Fig. 18) and A. pendulifolium (Fig. 9); others dry intact, and then quickly weather to fibers (Fig. 19). In addition to the length of the cataphyll, which may vary from 3 to 40 cm, there are dif- ferences of color and thickness, which аге taxonom- ically significant. Unfortunately, like many other characters in Pachyneurium, these differences are difficult to quantify as key characters. Though of minor importance taxonomically, cataphylls have an important function ecologically. They generally weather into fibers which, in con- junction with the roots, aid in entrapping debris. LEAVES Petioles. Тће petioles of sect. Pachyneurium аге typically much shorter than the blades and this, coupled with the short internodes, provides the basic ingredient for the rosulate habit and *'bird's- nest" appearance. Some species of Pachyneurium, even the more typical members of series Pachy- neurium, may have long petioles, but they are less effective as debris catchers and even tend to have leaves spreading laterally, e.g., А. llewelynii, or pendent, e.g., А. pendulifolium, A. pseudospec- tabile, and A. spectabile. Petiole cross-sectional shapes of sect. Pachy- neurium are diverse, perhaps more so than in any other section of Anthurium. Pachyneurium spe- cies have at least some of the cross-sectional shapes in each category from A through E (Fig. 21). Cross- sectional shapes in Pachyneurium are mostly **D- shaped" (category B), “markedly angular” (D) and "markedly ribbed adaxially" (E) with fewer species Annals of the Missouri Botanical Garden having petiole cross sections in category А, "te rete." The U-shaped cross-sectional shape is rare in sect. Pachyneurium. Many species of Pachy- neurium have C-shaped petioles (A 6–8), and pet- iole shapes that are more or less trapezoidal to D-shaped with prominent abaxial ribs are especially common (E 1-3) The petiole base is usually briefly sheathed (Figs. 14, 16) and encircles the inflorescence (the latter may abort). The sheath is usually inconspicuous and short, only rarely extending up to as much as 8 cm long. In most cases the sheath is restricted to a very small percentage of the total length of the petiole but may, for example in А. consobri- num, A. fatoense, and А. harlingianum, have sheaths that extend almost to the base of the blade. The petiole in all Pachyneurium species terminates a conspicuously swollen geniculum, which fa- cilitates leaf blade orientation. Blades. Vernation—The nature of the devel- oping blade is highly significant taxonomically, es- pecially at the sectional level in the case of sect. Pachyneurium. Anthurium sect. Pachyneurium is the only section of Anthurium known to have involute vernation of the developing leaf blades. Indeed, only one other genus in the Araceae, Lag- enandra from the Old World Tropics, has involute vernation. With involute vernation the developing leaf (in bud) has both margins of the blade rolled inward toward the midrib (Fig. 1), whereas super- volute vernation, the common type of vernation for Anthurium, has one margin rolled inwardly to the midrib and the other rolled around the coiled up inner blade margin, as well as the midrib (Fig. Although the | caeci of involute vernation is t plants, juvenile plants of Расћу- neurium species may have supervolute vernation or a transitional stage of development where o a part of a leaf blade (usually the widest part) has involute vernation. Richard Sheffer (pers. comm.) reports observing a series of leaves from a single plant where the vernation of successive blades pro- gresses gradually from wholly supervolute blades when juvenile, to partially involute blades to wholly involute blades in adult plants. In series Multiner- via, which has in general more slender blades, the tip of the blade is frequently supervolute while the remainder of the blade has typical involute ver- nation. This failure of the apical portion to be involute is probably due to the difficulty in rolling up an increasingly narrower section of leaf tissue. General features— Although species in sect. Pachyneurium typically have thick, oblanceolate to obovate leaf blades with short petioles, the group is actually quite variable in blade shape with blades ranging from linear-oblong to ovate-triangular to ovate, with bases acute to subcordate or cordate. The less typical species, such as А. schottianum, A. standleyi, and A. watermaliense, have mod- erately thin blades with distinct posterior lobes. Typically, blades are more or less elongate, broad- est at or above the middle of the blade and gradually tapered to the base, but frequently the blade ends abruptly and is often rounded at the base or even subcordate (e.g., А. fendleri) despite its generally oblanceolate or obovate overall shape. Leaf blades may be quite variable at the base, even on the same plant varying from shallowly cordate to acute ig. 20). While the blade of most Pachneurium species extends to the base of the midrib and ends im- mediately above the geniculum of the petiole, some species have a large section of the midrib com- pletely naked, so that it appears to be a part of the petiole proper. This is perhaps most extreme in А. oerstedianum, but other species, such as А. guanchezii, A. pranceanum, and A. remotigenic- ulatum, also display this feature. While this feature exhibits some taxonomic value, it does not always represent a character warranting specific recog- nition. This is true, for example, in А. bonplandii subsp. bonplandii, which sometimes has the geni- culum appeanog remote еи а анк centimeters, but does not otherw ar rom other plants of Mea bx am Schultes & Cabrera 14083). Venation is relatively uniform considering the wide range of blade shapes. Usually the major veins are prominently raised on both surfaces, with the midrib varying from slightly convex to angular above and typically more prominently convex be- low. Frequently, if the midrib is angular above, this is owing to a continuation of a medial rib on the adaxial surface of the petiole, but more commonly the midrib is flat at the very base on the adaxial surface. From this point the midrib may become progressively more raised and more acute toward the apex, the most acute portion being in the distal 2/3 of the blade. From this point the midrib com- monly becomes diminished and is usually even weakly sunken at the very tip. This is generally true also when the midrib on the adaxial surface is merely rounded or obtuse rather than acutely raised. On the lower surface, the midrib may be convex or angular, but when it is angular it is usually due to an extention of the ribs onto the petiole. In this case, the ribbing on the abaxial surface of the petiole is often present on the geniculum as well, e.g., in A. luteynit. For typical members of sect. Pachyneurium Моште 78, Митбег 3 1991 Croat 549 Anthurium sect. Pachyneurium (always true for series Pachyneurium), the pri- mary lateral veins are prominently raised on both surfaces (hence the sectional name), most com- monly departing from the midrib at 30—55? in the middle of the blade and sometimes at broader an- gles, approaching 90? near the base of the blade. arely, as in A. reflexinervium, A. paraguayense var. coroicoanum and А. latissimum, the veins may arise at an angle of 90? or more near the base of the blade and the primary lateral veins may actually be retrorse in such cases. Blades with rounded, cordate or subcordate bas- es tend to have an aggregation of veins at or very near the base of the blade. For those species with rounded or subcordate bases, all but the lowermost are considered primary lateral veins, whereas those species with cordate bases generally have a series of basal veins (as defined in Croat & Bunting, 1979). These basal veins in all species with well developed posterior lobes are generally directed downward into the lobes and may be in part united into a collective vein (Croat & Bunting, 1979). Pachyneurium species rarely have interprimary veins or secondary veins (i.e., major branches off of the primary lateral veins), both of which are often well developed in other sections of Anthu- rium. Tertiary veins on the other hand are usually well developed, and reticulate veins are often well developed and conspicuous at least after drying. In some species, most notably А. crenatum, the reticulate veins are quite conspicuous even on fresh leaves. The largest tertiary veins are generally very conspicuous in most species and are generall weakly raised on the lower surface and moderately obscure on the upper surface. However, some spe- cies, such as А. jenmanii, A. bonplandii, and А. atropurpureum, have many of the tertiary veins sunken or etched into the upper surface on fresh leaves. Blade margins of most Pachyneurium are entire (though somewhat sinuate in А. barclayanum and A. paraguayense) and may be flat to markedly undulate, such as in А. affine. Typically, blades are cuneate (gradually and evenly narrowed toward the base so that the margin in the lower part of the blade is straight). However, in some cases, e.g., A. latissimum, A. palenquense, A. remotigenic- ulatum, the blades are attenuate with concave bas- al margins. Features on blade surfaces are noticeably absent in most species, though many of the species in the Amazon basin, especially those occurring in areas plandii, have weakly raised plate-shaped, glandlike structures which may be darker than the surface and visible to the naked eye. These are especially conspicuous on А. bonplandii subsp. guayanum and A. xanthoneurum. Rarely, blades may be glau- cous on the lower surface as is sometimes the case in 4. caucavallense and А. glaucospadix, but even in these species not all plants possess glaucous blades MORPHOLOGY OF REPRODUCTIVE STRUCTURES INFLORESCENCES As in other Anthurium sections, members of sect. Pachyneurium have a single inflorescence per leaf axil arising from the sheath. Commonly, the inflorescence emerges fairly promptly from the leaf axil, requiring one to six weeks to fully emerge. Although every leaf axil has the potential to pro- duce an inflorescence, the inflorescence may abort, and usually does so before the plant reaches a certain age. Usually, the inflorescences emerge one at a time from the axils of the uppermost leaves (but rarely from the youngest, i.e., most recently emerged leaf). In this manner a plant may have several inflorescences present, but rarely as many as four per plant. Usually a plant has one or two inflorescences, each in a different stage of devel- opment. The rate of the emergence of the inflo- rescences usually precludes the possibility of self- pollination because they do not reach anthesis at the same time (see section on Flowering Behavior). In unusual cases, the inflorescences may emerge from several leaf axils within a short span of time, e.g., in A. schottianum. In this case the inflores- cences are still not of the same age, but they are more nearly alike in age than is normally the case. This also has an important bearing on the phe- nology, since this species tends to flower in short, brief bursts rather than periodically throughout the year (see section on Phenology) Pedu ncles. Peduncle length in Pachyneurium is evertheless, the petiole-peduncle length ratios are often highly significant taxonomically, at least at anthesis. The peduncle is generally terete and green, or occasionally with an acute angle on one side of the peduncle which extends along much of its length, or rarely with several ribs around the circumference of the peduncle. Most species have peduncles that vary from somewhat longer to sev- eral times longer than the petioles, and only rarely is the peduncle shorter than the petioles (e.g., А. curtispadix, A. plowmanii, and sometimes A. er- nestii and A. purpureospathum). Typically, the peduncle is moderately stiff and is held erect or 550 Annals of the Missouri Botanical Garden semierect at least during anthesis. After anthesis, the peduncle either begins to turn yellow and shriv- el, in the event that pollination has not occurred, or it becomes heavier with the development of fruits and the weight of the resulting infructescence caus- es the peduncle to bend outward or even downward. A few species, such as А. cataniapoense, А. salviniae, A. pendulifolium, and often A. solitar- ium, have long, slender peduncles with the inflo- rescences pendent at anthesis. Regardless of wheth- er the peduncle is thick, stiff, and erect or slender and pendent, it becomes thicker in the fruiting stage to compensate for the added weight of the berries. The net weight of the inflorescences may increase during maturation by a factor of 20 or more. Because of the heavy weight of the infruc- tescence, very few species, even those with erect inflorescences, bear their fruits in an erect position. Among those that do are: А. cubense, А. plow- manii, А. purpureospathum, and А. superbum, all of which have very short peduncles. Spathes. Although sect. Pachyneurium has spathes that are not so typically colorful or as taxonomically useful as in other sections of Anthurium, especially sect. Calomystrium, the spathe still carries con- siderable taxonomic weight in the recognition of species. The typical spathe of Anthurium sect. Pachyneurium is oblong-lanceolate, acuminate at the apex and with the distal margins often slightly inrolled. Typically, the spathe is inserted at about a 45? angle on the peduncle. The base is generally acute to attenuate and is often decurrent onto the peduncle for a short distance. Usually somewhat coriaceous or subcoriaceous, the spathe completely envelops the spadix only when the latter is very young. The spathe is generally loosened and free from the spadix before the pistils of the flowers are visible. Ordinarily, the spathe is initially erect for a time after it opens, the natural result of its erect position while in bud, but in most cases the spathe soon spreads and may ultimately be erect-spreading or spreading, or it may become spreading-reflexed to reflexed and even caducous (e.g., А. fendleri). These spathe positions are not always taxonomi- cally significant, and a single spathe may go through a series of positions throughout the course of its development. Still, many species have spathes that are characteristically positioned in a particular manner, usually comprising no more than two of the above positions for one taxon. The spathes of A. bradeanum, A. sarukhanianum, and А. sal. vadorense remain erect and may enshroud the spadix in the two latter species. The spathe is usually curved along most of its length after spreading so that, strictly speaking, no directly measureable angle is formed with re- spect to the peduncle. In addition, the spathe may be variously undulate, inrolled, or twisted. In some cases this feature may be extreme, e.g., іп А. colonicum, А. pseudospectabile, and А. specta- bile, where the spathe becomes very much con- torted. While the spathes of Pachyneurium are typi- cally thick and persistent, some species, mentioned above, have thin spathes which begin withering almost from the moment they open, and generally wither off and fall free in time. Even in some species with typically thicker spathes, such as А. jenmanii, the spathe withers, or at least yellows, shortly after anthesis. In these cases the spathe, though with- ered, usually does not fall free, but remains on the spadix in its withered state. onec are el about as long as the spadix but may b tially shorter at anthesis or even longer than the spadix as in the case of А. eximium, A. galactospadix, and A. spathiphyllum. While the absolute lengths of the spathe and spadix are variable throughout the course of a plant's devel- opment and are thus usually not so valuable as key characters, the relative length of the spathe in relation to the length of the spadix may provide a useful character, especially when studied at an- thesis. The proportional differences in the size of both the spathe and spadix during their develop- ment need not remain constant. For example, the spathe does not greatly elongate after anthesis for some species, such as 4. fendleri, whereas the spathe of А. salviniae increases in length after anthesis. Spadices. The spadix shape of Anthurium can vary from long-tapered to oblong (Fig. 59) or spin- dle-shaped or even obovoid or subglobose. The spadix of most Pachyneurium species is typically short-tapered (Fig. 232) or long-tapered (Fig. 44). А few species, such as А. cubense, А. superbum, А. bradeanum, A. affine, and A. spathiphyllum, have spadices that are oblong-cylindric or clavate, while А. sarukhanianum has a narrowly obovoid spadix (Fig. 277). Spadix color, though variable, is not very useful for Pachyneurium because so many species have spadices that are green or pur- plish, with only a relative few being red (sometimes in A. ernestii), white (A. superbum), yellow, or other colors. In addition, spadices of some species (e.g., А. barclayanum) may vary from green to purplish. Anthurium ernestii is markedly variable according to herb labels, with spadices at least varying from pinkish to magenta to purple, dark Volume 78, Number 3 1991 Croat 551 Anthurium sect. Pachyneurium red, or green. Overall spadix size in Pachyneurium is variable and depends on the size and age of the plant and its condition. As indicated above under ““Peduncle,”” the relative length of the spadix com- pared to the peduncle may be important taxonom- ically. However, even this character is sometimes highly variable and may vary depending on the age and health of the plant. Flowers are arranged in close spirals (Croat, 1980) with mostly more than five flowers visible per spiral from any angle. The number of flowers visible per spiral, though somewhat variable from spadix to spadix for any species, may also be tax- onomically significant, with the number ranging from as few as 3 or 4 flowers visible per spiral for A. vaupesianum and A. uleanum var. nanayense to as high as 28 flowers visible per spiral for А. galactospadix, and 24 for A. superbum. Descrip- tions of the number of flowers per spiral refer always to the number of flowers visible from any angle rather than the number actually needed to make a complete spiral (the reason being that the reader could not easily interpret the number of flowers on a herbarium sheet). As is true of all Anthurium species (Croat, loc. cit.), Pachyneurium flowers have four tepals, which are mostly hidden from view without dissection. They are arranged symmetrically on four sides of the pistil. Tepals are narrow, thin, and generally colorless except at the apex. Tepals are markedly cucullate at the apex and mostly cover the apex of the pistil. The apical portion of each tepal is more or less 3-sided, with the inner margin (that portion directed inward toward the stigma) thin and sometimes turned up against the pistil, especially when the pistil is exserted, as in А. consobrinum. “Тера! shape" when referred to throughout this work refers only to the apical, visible portion of the tepals. In the same way “flower shape" refers only to the apical portion of the flower. Flower "length" refers to the width of the apical portion of the flower in the longitudinal dimension of the spadix, while flower **width" refers to the width of the apical portion of the flower in the lateral di- mension (i.e., perpendicular to the axis of the spa- dix). The actual length of the tepals, including the hidden portion which extends into the axis of the spadix, is never measured. Tepal shape may be important taxonomically. Most are basically triangular, with the inner margin almost straight to rounded to broadly rounded. The outer margin, though commonly two-sided, is more variable. Some species, e.g., А. barclayanum, А. curtispadix, and А. glaucospadix, may have tep- als which have the outer margin 3- or 4-sided. In the latter case the tepal may be referred to as ““shield-shaped.” Androecium. Like all Anthurium, sect. Pachy- neurium has four stamens, each of which is affixed at the base of a tepal. Stamens are tightly com- pressed between the tepals and the pistil and are not visible without dissection prior to anthesis. The filament is slender (typically much narrower than the hidden portion of the tepal), conspicously flat- tened, and generally colorless and translucent or sometimes whitish. Fresh filaments are usually somewhat succulent with a thin epidermis. Shortly before maturity of an individual stamen, the anther is pushed out from between the tepal and the pistil by elongation of the filament. Once exposed, both thecae of the anthers promptly open by longitudinal slits. Dehiscence is extrorse. Thecae are generally more or less oblong or ovoid, and unlike some sections e.g., Porphyrochitonium, are generally only slightly divaricate (e.g., with the two thecae diverging away from one another toward the base). For further details of stamens see section on Flow- ering Behavior. Gynoecium. Typical of other sections of Anthu- rium, the pistils of sect. Pachyneurium are biloc- ular, more or less oblong-ovoid, and are tightly compressed in between the four tepals. A well- defined style is always lacking. The pistils are visible only at the apex and sometimes are almost com- pletely covered by the tepals. Generally, only the stigma is visible. Each locule contains a single ovule with axil placentation. For additional details of the gynoecium see section on Seeds. Stigmas consist of an elongate slit (Croat, loc. cit.). Generally those of sect. Pachyneurium have the stigmatic hairs contained within the slit, densely covering both sides of the inner surface. Осса- sionally, e.g., 4. palenquense, the stigmatic hairs may be long-exserted, forming a brushlike stigma. Stigmatic fluid is produced by the stigmas and generally accumulates to form conspicuous drop- lets. For additional details, see section on Flowering Behavior. FLOWERING BEHAVIOR Long-tapered spadices generally have flowers developing from the base of the inflorescence and progressing toward the apex. Those with oblong or narrowly obovoid spadices generally have the first flowers to reach anthesis at the middle or lower one-third of the inflorescence with development progressing toward both ends. For a review of flowering behavior of Anthurium see Croat (1980). 552 Annals of the Missouri Botanical Garden All other features of flowering behavior are iden- tical to those described in that paper. Typically, Pachyneurium species have spadices at anthesis for long periods. While most species have spadices in anthesis for about a week, some, such as 4 luteynii, may have spadices in anthesis for several weeks. As in other species of Anthurium, flowers are usually markedly protogynous, with the pistil- late phase clearly separated from the staminate hase. During the pistillate phase most species in as well. In natural situations these droplets are rarely allowed to accumulate, apparently being re- moved by pollinators, but in cultivation the droplets may accumulate to such an extent in some species, e.g., А. upalaense, A. validifolium, and especially A. lutyenii, that the nectar drips off the spadix (Fig. 332) As in other species of Anthurium, the staminal development of Pachyneurium generally begins with the lateral stamens (Croat, 1980) followed by the alternate pair of stamens with the anterior stamen (referred to throughout this paper as the third stamen) developing before the posterior sta- men (referred to throughout this paper as the fourth stamen). In most species of Pachyneurium, lateral stamens emerge well ahead of the first alternate stamen, varying from about 3 to 20 spirals ahead of the third stamen (usually the anterior stamen), and the first alternate stamen usually also emerges several spirals ahead of the fourth stamen. Thus, the available fresh pollen is commonly scattered over 6 or more of the spadix. When the stamens first emerge, the anther is frequently slightly ex- serted on a fleshy, thickened filament and is fre- quently colored. Anthers usually open promptly after emerging from beneath the tepals. Most fre- quently the pollen color is yellow, orange, or pur- plish, but is often white. The exserted anther lasts generally less than one day, and the anther is then generally retracted so that it is positioned at or near the level of the tepal (Croat, 1980). The retraction is caused by the loss of water in the thin filaments and the resulting shrinkage. Usually by the end of the first day of emergence of the stamens and always by the next day, the color of the pollen and generally that of the anther has faded (probably ue to oxidation) and usually turns white. Some species, especially certain members of series Mul- tinervia (e.g., obscuriner- vium, and А. oxyphyllum) have stamens that are prominently long-exserted and remain exserted af- ter anthesis. acutissumum The fact that pollen has faded in color does not necessarily mean that it is not viable, but fresh pollen is most effective in pollination. Generally, individual flowers of Pachyneurium species are so markedly protogynous that stigmatic droplets have dried up on all flowers by the time the stamens emerge on the lowermost flowers; how- ever, in some cases the stigmatic droplets are pres- ent on the uppermost flowers of the spadix, and at the same time the stamens are beginning to emerge on the lowermost flowers of the spadix, thus gei- tongamy is possible. In fact, it is common practice for plant breeders to manipulate the spadix of a single inflorescence to attempt self-pollination when no other pollen is available. While not always suc- cessful, this method will sometimes work on some species. MORPHOLOGY OF FRUITING STRUCTURES INFRUCTESCENCES The fruiting spadix of most species of sect. Pachyneurium is much expanded in size, the extra girth coming from both an expansion of the pithy axis of the spadix and a general increase in the length of the tepals. The expansion accommodates the greatly expanded size of the developing berry. Berries. In most Pachyneurium species the berry begins to emerge above the tepals and becomes colored well in advance of its actual maturation, i.e., before the seeds are matured sufficiently to germinate. Alternatively, the enclosed seeds usu- ally are capable of germination somewhat before the berry becomes fleshy, loose, and obviously ma- ture. Berries are also capable of maturing to some extent even on severed inflorescences. This is im- portant when collecting seeds before the berries are fully mature. The berr fructescence generally emerge in sect. Pachyneu- rium in a scattered fashion over a relatively long ries of the mature in- period of time, but some species, such as A. spa- thiphyllum and A. consobrinum, have fruits that emerge more nearly simultaneously. Typically, the berries are weakly exserted and do not fall free for display. Berry shape is typically obovoid to obovoid-el- lipsoid and generally rounded but sometimes acute to narrowly acute at the apex. Berry shape may e variable even within a single population, e.g., in A. schlechtendalii, where berries may be acute to rounded at the apex Berry color. Berries of sect. Pachyneurium are typically concolorous and are summarized in Tables 1–3. Exceptions, with bicolored berries, include: Volume 78, Number 3 1991 roat 553 Anthurium sect. Pachyneurium (1) Anthurium consobrinum, with white fruits and red to purple apiculate tips; (2) 4. superbum, with fruits colored lavender at the base and white at the tip; and (3) А. pendulifolium with fruits colored purplish lavender to reddish purple at base and white at the tip. In general, fruit color (yellow, orange, red, or purple) remains distinct and permanent throughout their maturation. The purple-fruited species, how- ever, are an exception and are much more variable in color during development. An example is А. pendulifolium, where pure lavender to reddish vi- olet berries may become almost completely white at maturity. Included in the “purple” family, men- tioned in detail in Table 1 are lavender, violet- purple, violet, purple, purplish violet, reddish vi- olet, and maroon Berry color is not only of great taxonomic sig- nificance in terms of species recognition, but it is also important in defining relationships between different groups of species. Fruit colors by per centage are summarized in Table 3. In Central America and the West Indies, = orange, or yel- low fruits predominate, while in South America, purple to purplish fruits predominate (Tables 2, 3). For example, for species with known fruit color in Mexico and Central America (excluding Costa Rica), 84% have red, orange, or yellow fruits (fruit color is not yet known for А. sarukhanianum, but its apparent relatives have fruits in this color range). In Costa Rica and Panama, this percentage is even higher (88%). Of these colors, orange predominates in Panama and Costa Rica, with 60% of the total. The orange-fruited species are in many ways atyp- ical, often having leaves with long petioles and cordate to truncate or rounded bases. The red- fruited species, on the other hand, are more typical, with tight rosulate habits, short petioles, and gen- erally acute leaf bases. Fruit color is still unknown in two Costa Rican or Panamanian species, i.e., А. spathiphyllum and А. oerstedianum. While the former is probably white-fruited, it is difficult to predict the fruit color of А. oerstedianum as it is seemingly unrelated to anything else. Central America has only one species with purple fruits, A. fendleri, which is a northern South American species barely reaching Central America (rare, in Panama o In contrast to Central America, South America, especially the lowland Amazonian basin, is rich in purple-fruited species (or fruits with at least some purple present), with 94% of the total in this cat- egory. The eastern slopes of the Andes are almost as rich in purplish fruited species, with 90% of the total being purple or with some form of purple. The western slopes of the Andes have a significantly lower percentage of purplish fruits (60%). This is largely due to the presence in the central Andes of series Multinervia, which has a large percentage of its species having orange fruits. The western slope of the Andes is more diverse in fruit color than the eastern slope, with fruits in the orange, red, and purple color groups. In contrast, the east- ern or Amazonian slope has only the red and pur- ple-fruited color groups. Of these color groups, purple still remains the most predominant, with the red to orange group having only 10%. With the exception of the western Andean slopes and the far northern region of South America, purple fruits predominate in all other South Amer- ican geographical areas. A total of 7596 of the species of the Amazon Highlands and eastern Brazil are purplish fruited, while the comparable figures for the Guiana Highlands and the Andean High- lands are 100%. On the other hand, for species restricted to far northern South America (define here as species with a range that includes lowlands with a Caribbean drainage), only 33% are purplish fruited. Indeed, of all species occurring in far north- ern South America (including those red-fruited spe- cies entering from Central America, А. concolor and А. salviniae) and in the West Indies (А. си- bense), only 2776 are purplish fruited. The only orange-fruited species not occurring in the central Andes or Costa Rica- Panama, A. johnsoniae, oc- curs in this region. If all species with known fruit color occurring anywhere in South America are considered (in- cluding A. affine, which is also reported as some- times red-fruited), а total of 46 out of 62 (74%) have purplish colored fruits, 12% have red fruits, and only 876 have orange fruits. Comparable fig- ures for all species with known fruit color for Cen- tral America are 3% purple fruits, 29% red fruits, and 48% orange fruits. While the separation be- tween purple-fruited species and those with red or orange fruits seems very clear, it is less obvious that there is a strong separation between red- and orange-fruited species. Though red and orange would intrinsically appear to be closely related col- ‚ fruits of species in these color groups are и conspicuously prenga or conspicuously red. However, there with А. cubense and А. carchiense, both a ‘which have berries orange-red, and A. purpureospathum, with berries reported orange to red-orange or red. Both A cubense and A. purpurpeospathum otherwise ap- pear to be close relatives of red-fruited species of series Pachyneurium, while A. carchiense is oth- erwise Closely related to other orange-fruited mem- 554 Annals of the Missouri Botanical Garden TABLE l. Fruits, sorted by color LOC Species Primary color* Second color CRP consobrinum white greenish white SBR coriaceum grayish white lavender MCA nizandense greenish white CRP acutifolium greenish yellow CRP bradeanum greenish yellow white CRP fatoense greenish yellow orange, pale MCA halmoorei yellow WA acutissimum yellow-orange orange WA a orange? CRP bren orange CRP an orange CRP cotobrusit orange WA holmnielsenii orange NSA johnsoniae orange CRP luteynii orange CRP nervatum orange WA oxyphyllum CRP prolatum orange? CRP protensum orange CRP pseudospectabile orange CRP purpureospathum orange red-orange CRP ranchoanum orange CRP ѕеірегій СКР spectabile orange? CRP standleyi CRP validifolium WA carchiense orange-red WCS cubense orange-red MCA salvadorense orange-red CRP schottianum pinkish? CAR venosum red-orange PSA concolor re NSA crassinervium red CRP eximiu red EA fasciale red LA plowmanii red CSA lvinia red MCA schlechtendalii red WA sparreorum red CRP upalaense red NSA anorianum red, dark CAR crenatum red, bright BH ајћпе ге purple maroon WA lennartii red purple, dark CSA wagenerianum EA latissimum maroon red-violet, deep EA ala ka d var. paraguayense maroon reddis GH wurdac maroon LA ernestii var. ernestit red-violet violet LA uleanum var. uleanum reddish violet AH сае за reddish purple WSA jenm reddish purple, pale EA iio etm reddish violet WA napeaum burgundy purplish violet LA willifordii magenta LA superbum lavender Volume 78, Number 3 1991 Croat 555 Anthurium sect. Pachyneurium TABLE l. Continued. LOC Species Primary color* Second color LA pendulifolium lavender to white LA bonplandii subsp. bonplandii urple EA llewelynii purple WA obscurinervium purple LA oxycarpum purple red WA pallatangens purple EA paraguayense var. coroicoanum purple WA simpsonii purple BH solitarium urple LA loretense purple violet LA atropurpureum var. atropurpureum purple, dark red, deep AH leonianum purple, dark BH lindmanianum purple, dark red, dark WA palenquense purple, dark EA tarapotense urple, dark WA ji purple, light EA solomonii urple, pinkish GH bonplandii subsp. иин purple, reddish purplish violet LA remotigeniculatu purple, reddish WA barclayanum urplis reddish brown LA bonplandii subsp. cuatrecasii purplish? EA pachylaminum purplish? LA pranceanum iolet LA uleanum var. nanayense violet LA vaupesianum violet LA atropurpureum var. arenicola violet-purple AH caucavallense violet-purple AH dombeyanum violet-purple LA ernestii var. oellgaardii violet-purple PSA fendleri violet-purple reddish violet EA harlingianum violet-red urple, white WA basirotundum wine red, deep * Primary = most frequently — color. — Central & South America; EA — S. шл МА = Western Andes; WCS = SBR = Indies & South America. bers of series Multinervia. Thus, based on fruit color, there appears to be a clear separation be- tween Central American and South American spe- cies of sect. Pachyneurium. This will be further discussed under breeding behavior (see Cytology section). White or yellow fruits are rare in sect. Pachyneurium. Yellow berries are known only for . halmoorei, an isolated species from western Mexico. Whitish berries are known only for А. nizandense from Mexico, А. bradeanum from Costa Rica and Panama, and А. coriaceum from southern Brazil. None of the above species are closely related. an Highlands; BH = Brazilian млр CAR = Caribbean; СКР = Costa Rica Eastern Andes; GH = С = Mexico & Central America; NSA = Northern South | им PSA = Рап West Indies, Central America, South jee WSA = шапа adi LA = Lowlan ma & South America; est Seeds. Like those of most species of Anthurium, the seeds of sect. Pachyneurium are one per locule, though one frequently aborts, leaving a one-seeded fruit. Seeds are commonly oblong or oblong-elliptic and are weakly flattened with the basal end rounde and the apical end obtusely notched where the funicle is attached. Typically, seeds are attached to the inner surface of the carpel wall by means of a slender mucilaginous strand, and one or both ends may bear a sticky, usually amber-colored appendage, which aids in dispersal. Sometimes the entire seed is enveloped in a mucilaginous envelope, which may protrude at both ends. A more uncom- Annals of the 556 Missouri Botanical Garden штпитпагиг8 рт ¿2SUIQNI asuaqno voueury YINOS ueeqque-) «пиошиәГ штјриәлә шпѕоиәа ueaqqlIeJ-Uno11) A затриј 159 шталги1в5 010 nupwual 4 10]09U09 мајригј UNUDILOUD apiuosuyol еопәшү YINOG илацјом воџошу YING IDIUIA]DS илацјлоџ /еомошу [ециэ’) шттојпрура 1X2]punis 3]10712ads 711494195 UNUDOYIUDA unyiodsosundind a]1q0122dsopnosd штопор *dsqns wnsuajoid :dsqns „мајригј asuapjpdn штвигјола штио оцов штаојола штјралги unmunxa пиХатј «2suaqno 1151190102 asuso mf штэ1ио]02 шпираролд 40]02u02 1152u234Q шпиладозиод штђојутор eureueg Х тогу 21507) nzouiunf -dsqns пүрриәјцәәруэѕ “dsqns пүрриәјуәәруәѕ (eureueg y азиглоррајр eon] 81507) Зитрирохэ) 4,95u2qno asuapunziu 1900] DY вомошу |ециәг) x ODXIJA je[ot4 *uoo1eur рэ1-э8иело ‘Чзтррэл “pay э8ие1о-рэл ‘JULIO шеало “SYM моје *e[dind-po1 ‘әт ‘эиЧм цошаола qsiuao18 *usmoa1^) ‘Чямоад. $1005 ИПЦ uniunauXuy»pnd 'Áy[eoo| Aq wunıınau{y2D J ut $100 ип1] "c ATV] 557 Anthurium sect. Pachyneurium Croat 3suajodn410j ununssi7] штити уаоу әүотәѕо[ вшејипојј зэриу jo зодоје uioijseq штпиттѕәйтра umqaadns шп]пәлиә8туошәл штиргзирла штђојуприга прот эзиэХрити ‘тел штиројт ‘IBA 381231210] 1p1009120 ‘тел 117$2U49 ‘IBA 11389419 11$022470n2 'dsqns DIONUIID ‘тел wunaandandoa;v ‘тел unaındındo11D пииоштоја швед uozeury ривјмој 1ә[оїл ‘иоолеш pes-azueso “ysippal ‘pay о8ивло-рол *eSue() WLI “SYM мо[әх "аја та -рол ‘ajding “SYM qsruoa18 ysiua913 Ҷѕ!чээіо чимојјоХ $10]05 ип штапгикуора Volume 78, Number 3 1991 'penunuo) ‘с 318v] Missouri Botanical Garden Annals of the 558 "ом se э1эЦ juasaid inq ‘uoaa [peorqde18093 1оцјопе шолу Аплештла seroedg y nyoppinn шпиоХопЯ :dsqns npuvjduoq шт141071]0$ штиттитириу su) nuosduns asusndun3o]¡0d asuanbuayjod umaa42auin2sqo штогари 111Duu23] ndun шпритјолл“рд wnuDADIIIDG штп]уХу4!Ххо nuasjaruumoty unıoa11Dds штитио пир ISUINYILDI UNWISSIINID uN2IIDI109 spue[y3y euemg [2819 uymnos pue¡mo”] spuejy3iH ueizejg зитезипой{ sapuy jo sedo[s иләззә M јә[отл *uoo1eur *e[d.nd-po1 ‘ajding рэ1-э8иезо ‘ysippes ‘pay о8ивло-рал 'egue() штвало ‘эиЧм ‘эиЧм цзшоола ‘ystuaalsy моја qsiuaa18 чимоед s10]09 ИП wunanauXyopd грапициој с XI8V[ Volume 78, Number 3 1991 Croat 559 Anthurium sect. Pachyneurium TABLE 3. Fruit color percentages by region. Fruit colors % by location Greenish, eenish urple Yellowish, white, Red, red-purple, Total greenish white, Orange, reddish, maroon, | number ellow cream red-orange orange-red violet of species Mexico & Central America excluding Costa Rica & Panama 17% 17% 67% 6 Costa Rica & Panama 8% 8% 60% 20% 4% 25 Central & northern South America 100% 1 Могћегп Зошћ Атепса 17% 50% 33% 6 West Indies & Caribbean 25% 50% 25% 4 Caribbean South America 50% 50% 2 Lowland Amazonas 6% 94% 17 Eastern Andes 10% 90% 10 Western Andes 27% 13% 60% 15 Brazilian Highlands 25% 75% 1 Lowland southern Brazil 100% 1 Andean Highlands 25% 75% 4 Guiana Highlands 100% 2 Total number of species 3 4 22 21 47 97 mon type of seed attachment is found in 4. bon- plandii and A. wagenerianum, where the seeds are attached by their apical end to the inner carpel wall by a thick strand of fibers running down the carpel wall toward the base of the fruit. FRUITING BEHAVIOR The production of fruits in Anthurium is gen- erally similar for all sections, with the pistil and the tepals gradually enlarging in unison after pol- lination. Some species, e.g., А. consobrinum and A. colonicum, have fruits that develop from an early-emergent pistil. In these cases, the pistil re- mains green until shortly before fruit maturation. In most species the developing fruit remains below the level of the tepals until very near the time of fruit maturation, or they emerge somewhat above the level of the tepals but remain uncolored. Shortly afte ter anthesis the tepals generally begin to lose third of their development, but may proceed slowly or promptly. Generally the berries remain firm even after full coloration, and for weeks before they begin to soften and loosen. In most species the softening of individual fruits is only partial and is widely scattered on the infructescence. The soft- ened and generally plumper berries begin to emerge an even be extruded and supported by threadlike tepalar fibers (narrow strips of epidermis from the inner surface of the tepals). This form of display broadens the visual target of the infruc- tescence for bird dispersers and makes picking off the fruits much easier. Їп other cases, the mature berries are not extruded but merely become loos- ened and, if not removed, simply fall to the ground. Many examples have been seen where animals have fed on the infructescence only to leave the seeds scattered beneath the plant. Because the seeds are sticky, many even stick to the leaves of the same plant or other plants in the immediate vicinity. By the time fruits are mature the tepals may have expanded to more than twice their original length, and the outer surface of the tepal (the only part regularly visible) may be expanded to a lesser extent, but they remain in appearance very much like they were originally. An exception to this oc- curs in A. sparreorum, which develops fruits in an almost unique manner for Anthurium (seen else- where only in an unidentified Ecuadorian member of sect. Belolonchium). In this species, the apical part of the tepal greatly expands and becomes white and fleshy, in reality taking on a role in fruit dis- persal ecology since the fruits, though red at ma- turity, barely emerge above the tepals. Despite the fact that the fruits are not prominently exserted in the manner of other sect. Pachyneurium spe- cies, the red mature fruits are highly visible since they are displayed between the fleshy white tepals. In most Pachyneurium species the infructes- cence is straight at maturity, but some taxa, such as А. atropurpureum and А. pendulifolium, may 560 Annals of the Missouri Botanical Garden have a coiled infructescence when fully developed. This may serve to provide a more compact visua display to better attract dispersers, or perhaps it serves to spread out the time over which mature fruits are available for distribution by providing fewer fruits that are readily accessible at any time. The amount of time required for fruits to develop varies greatly from species to species and may apparently vary considerably within each species as well (at least under greenhouse conditions). For species of sect. Pachyneurium, fruit development time under greenhouse conditions vary from as little as 3 months (for А. llewelynii) to as much as 13 months (for 4. salvadorense). Fruit devel- opment time may vary greatly on successive pol- linations. For example, fruit development time on A. schlechtendalii (see below) varied from 4.5 to 11 months, and on А. upalaense from 6 to 13 months, while А. halmoorei varied from 8.5 to 12 months. Fruit development time Months to maturity (number of inflores- Species cences observed) A. atropurpureum are- 4 (2); 4.5 (45; 5 (1); 5.5 (2) icola A. atropurpureum atro- 9.5 (1) purpureum A. barclayanum A. colonicum 12.5 (1); 13 (1) 11 (2) A. consobrinum 6 (1) A. crassinervium 4.5 (1); 10.5 (1) A. cubense 4.5 (1); 5 (1) A. fendleri 10.5 (1); 11 (1); 11.5 (1) A. halmoorei 8.5 (1); 9 (1); 10 (1); 11.5 (1); 12 (1) А. llewelynü A. luteynii 10 (1); 10.5 (1); 11 (1) А. ратавиауепзе 4 (1) A. salvadorense 13 (1); 13.5; 15 (1) A. schlechtendalii 4.5 (1% 11 (1) A. standleyi 11 (1 A. upalaense 6 (1); 7 (1); 13 (1) PHENOLOG Y A phenological survey of greenhouse collections was undertaken by volunteer researcher Edwina Medlock. The study was carried out over a two- year period beginning November 1979 and ending in October 1981. Plants were observed on a weekly basis, and the initiation of new leaves and inflo- rescences was recorded. A total of 38 taxa of Pachyneurium was studied. The taxa involved in the study and the number of individuals of each species were as follows: Anthurium affine A. nervatum 1 А. angustilaminatum А. oerstedianum 2 subsp. angustilami- A. oxycarpum 1 natum 2 . plowmanii 1 A. barclayanum 1 A. protensum А. bradeanum 1 subsp. arcuatum 2 A. brenesii 2 subsp. protensum 2 A. colonicum 2 А. pseudospectabile 1 А. с 1 A. purpureospathum 2 A. consobrinum 2 А. salvadorense 3 А. coriac 1 А. salviniae 3 A. cotobrusii 2 A. schlechtendalii А. crassinervium 2 ubsp. jimenezii | А. cren 2 subsp. schlechtendalii 2 A. cubense 2 A. schottianum 2 A. eximium 1 A. seiberti 2 A. fatoense 2 A. каше к ыш 2 A. fendleri 2 A. stan 2 A. halmoorei 2 A. ни 2 А. jenmanii 3 A. validifolium 1 А. luteynü 2 А. watermaliense 1 The following is a summary of the results of the survey. Leaf production. Though many factors are in- volved in the production of both vegetative and fertile plant parts, including growing conditions (es- pecially the amount of available light, nutrients, and humidity, as well as the general health of the plant), the collections were in general growing un- er moderately uniform conditions and were housed in the same room. Obviously some plants, owing to their larger size, received more light. The study showed that substantially more leaf production occurred during long daylength time periods (March to September) than during short daylength time periods. Though the greenhouse was heated, the nighttime temperature was allowed to drop to 65°F during the winter months, whereas in the warmer summer months the nighttime temperature often did not drop so low. Equally important, the daytime temperatures in the summer months were often very high, occasionally reaching 80°F, whereas in the cooler months such temperatures would have been rare. During long days (March-September), a group of 71 individual study plants of sect. Pachyneu- rium produced an average of 120 leaves for the group per year, whereas during short days (Sep- tember- February) the study plants produced only 94 leaves. Each of the study plants produced on the average four leaves per year during the two year study, but the range of new leaves produced Моште 78, Митбег 3 1991 Croat 561 Anthurium sect. Pachyneurium was one new leaf per year produced on А. pen- dulifolium to 10 new leaves per year for А. stand- leyi. Plants performed better the first year of the study, producing an average of 4.5 leaves per year vs. 3.4 leaves per year during the second year. This was no doubt due to the increasing size of the plants and the fact that the larger the plants got the more they shaded one another, causing the poorer growers to obtain increasingly less light. The interval between leaf production was highly variable, with the minimum time between the pro- duction of new leaves ranging from 1 to 16 weeks (averaging 3 weeks), and the maximum interval between the production of new leaves ranging from 8 to 46 weeks (averaging 25.4). Production of new leaves was highly variable throughout the year for most species, with major leaf production activity between March and September. The maximum number of leaves produced during the two-year study was during March (a total of 56 over 2 years), and the minimum were produced during February and October (a total of 18 for each month). Fre- quently, most of the leaves produced by an indi- vidual were produced during the 4- or 5-month period following the initial burst in March. Another period that was important for some species was January, with one or more leaves being produced at that time (Fig. 22). The initiation of leaf pro- duction during the early months of the year is probably best explained by the fact that these months have increasingly longer days following a period of decreasing daylight in the late fall. Plants generally go through a period of low leaf production in the late fall when days are getting shorter, es- pecially between September and December. The burst of activity in January is unusual since leaf production dropped off again in February only to increase once more in March (Fig. 22). During the period of maximum leaf production it is not un- common for leaves on a single plant to be produced at the rate of one per week, though a period of two to three weeks between leaves is more common. Leaf production is greatest on younger plants and tends to slow after the plant has reached a more mature size. Leaf production can be markedly spo- radic. For example, one plant of 4. brenesii pro- duced seven leaves in a 5-month period one year (March to July), then after producing one leaf in November, it the following September. Leaf production on some species can be very slow. For example, an indi- vidual of А. pendulifolium (not included in the study) failed to produce a single leaf in more than a year, despite the fact that the plant appeared to be in good health and even flowered during this did not produce another leaf until period. In general, healthy, robust plants produced more leaves than smaller or unhealthy plants, and there was a high correlation between leaf produc- tion and the amount of fertilizer applied, with most plants responding to increased fertilizer by growing faster and producing more leaves. Flower production. Because each leaf axil ulti- mately produces an inflorescence on adult plants of Anthurium, the production of inflorescences on Pachyneurium closely follows the activity de- scribed for leaf production (Fig. 22), with the peak flowering activity also being in March, followed by a marked diminution until about June. During the warmer months flowering reaches an almost steady state, with 13-17 inflorescences produced each month between June and November for the 2-year period. As with leaf production, December and January show a spurt of flowering activity, with a trough in February, followed by a major spurt of activity in March. As with leaf production, the interval between inflorescence production is highly variable but does not necessarily follow the same pattern. Some species produce their inflorescences from the axil of the newly formed leaves shortly after the leaf emerges, whereas other species delay the production of new inflorescences for several weeks or even months after the new leaf emerges. The minimum average interval between the pro- duction of new inflorescences for all species in the 2-year study was 1.7 weeks (ranging from 1 to 14 weeks). The average maximum interval between the production of successive inflorescences was 29 weeks (ranging from 7 to 46 weeks). The duration of flowering for any single inflorescence is also highly variable, with some species flowering for a period of less than one week and others (such as A. luteynii and A. salviniae) flowering for a month or more. For a discussion of the phenology of fruit production, see the section on Fruiting Behavior. CYTOLOGY Ву В. D. Sheffer Species of sect. Pachyneurium are predomi- nantly diploids with chromosome counts of 2n — 30 and with fewer numbers of species being tet- raploids with 2n — 60 chromosomes. This corre- sponds to the most common number in the genus as a whole (Sheffer & Croat, 1983). While not all species have been counted, R. D. Sheffer has made chromosome counts on many of the species not previously covered in earlier publications. He re- ports (pers. comm.) that a few species do not fit the usual pattern. Anthurium ranchoanum, with two collections counted (Croat 49846 and Stevens 562 Annals MUI рм Garden 13980), has а chromosome count of 2n — 44. Anthurium atropurpureum var. arenicola exhibit aneuploidy at the tetraploid level. An analysis done by Sheffer on a country b country basis indicates that there is a statistically significant difference between the number of poly- ploids in Central and South America. South Amer- ica has more polyploids (41.376) than Central America (10.7%). This suggests that the South American species are older and thus South America probably represents the center of origin of the section. This is borne out by the fact that the South American continent (especially the western edge of the Amazon basin and the foothills of the Andes in Ecuador and Peru) is also the area of greatest speciation in the group (see section on Geographical Distribution). BREEDING BEHAVIOR (with R. D. Sheffer) Breeding studies carried out by Sheffer at the Indiana University Northwest and by Croat at the Missouri Botanical indicate that sect. Pachyneurium is probably reproductively isolated. Garden Crossability between sect. Pachyneurium and oth- er sections of the genus is very limited. Thus far there are only two examples of successful out- crossing with another section of Anthurium that resulted in the production of viable seedlings. These crosses were between the Pachyneurium species А. standleyi and A. Шастит, a member of sect. Urospadix as well as A. hacumense (sect. Por- phyrochitonium) and A. glaucospadix (sect. Pachyneurium). Though these were the only ex- amples of crossability which produced viable seeds, studies show that there is potential crossability with other sections, if only fruit-set or seed production is considered as a significant measu relationship. Of the 433 cross-pollinations, 33 (7.4%) resulted in fruit-set and 28 (6.3%) resulted in seed production. These crosses were between sect. Pachyneurium and five ure of species other sections, Be- lolonchium, Cardiolonchium, Dactylophyllium, Porphyrochitonium and Urospadix. All, except the cases mentioned above (sects. Urospadix and Porphyrochitonium), resulted in abortive seeds that failed t с This is turine evidence that || г / sect. P isolated from other Md Besides being reproductively isolated from other sections of the genus, Central and South American species of sect. Pachyneurium are mostly repro- ductively isolated from one another. A total of 747 cross-pollinations within the section resulted in 116 (15.5%) with fruit-set and 88 (12%) with viable seedlings, which are probably hybrids (e.g., not merely cases of self-inducement to set fruits). Most of the pollinations that resulted in viable seedlings were among either Central American species or among South American species, but not between the respective groups. Thus, the Central American and South American species appear virtually iso- lated reproductively. Exceptions to the failure of crossability between Central and South American taxa are between 4. barclayanum (Ecuador) and A. cotobrusii (Costa Rica and Panama), 4. cras- sinervium (Venezuela) and А. bradeanum (Nica- ragua-Panama). Also worthy of mention are two viable crosses between А. halmoorei (Mexico) and A. fendleri, which is largely from northern South America but also ranges into central Panama. However, if production of fruits without viable seed is also considered, then two additional Central American vs. South American crosses become im- portant. These were crosses between 4. halmoorei and А. uleanum and between А. halmoorei and A. latissimum. Other crosses that might prove viable (producing infructescences but as yet without mature fruits) th American species, А. atropurpu- reum var. arenicola and A. harlingianum, bot are two Sou of which seem to have crossed successfully with А. nervatum from Panama. Similarly, A. bon- plandii subsp. cuatrecasii (South America) seems to have successfully crossed with both 4. halmoorei (Mexico) and А. purpureospathum (Panama). Anthurium crenatum, a West Indian species, shows apparent relationship with Central and South America, having been used in partially successful crosses (fruit matured but seeds inviable) with halmoorei (Mexico) and A. fendleri (northern South America and Panama), and also А. glaucospadix and А. paraguayense of South America. Sheffer’s study also indicates there may be lim- ited interspecific gene flow throughout Central and South America, since 36% (27) of the 74 species tested by him have been involved in interspecific crosses that resulted in viable seed (seedlings viable or not). While we have not attempted all possible crosses between different species of sect. Pachyneurium, the lists below summarize the successful crosses of the 226 cross-pollination attempts (including re- he crosses involve 58 different taxa, of which 27 were successfully in- volved in crosses (produced viable plants). peats of the same cross). Successful crosses of Central American species of Pachyneurium A. consobrinum X A. upalaense А. luteynii Х А. colonicum Volume 78, Number 3 1991 Croat Anthurium sect. Pachyneurium A. luteynii X A. purpureospathum : schlechtendalii X A. salviniae . Schlechtendalii X A. luteynü A A A A. standleyi X A. halmoorei A. upalaense X A. halmoorei A. upalaense X A. purpureospathum Successful crosses of South American species of Pachyneurium A. atropurpureum arenicola X А. ernestii А. atropurpureum atropurpureum X А. atro- purpureum arenicola barclayanum X A. atropurpureum arenicola barclayanum х A. glaucospadix crassinervium X A. barclayanum cubense X A. atropurpureum arenicola fendleri X A. atropurpureum arenicola fendleri X A. remotigeniculatum harlingianum X A. oxycarpum llewelynii X A. ernestii paraguayense X А. llewelynii tarapotense X A. plowmanii meom. m. mo om. 2o m. ms < GEOGRAPHICAL DISTRIBUTION The geographical distribution of Anthurium sect. Pachyneurium is summarized in the geographical checklist in Appendix 4. Although sect. Pachy- neurium is widely dispersed throughout the Neo- tropics, ranging from Mexico to the West Indies and south to Argentina (from ca. 22? north to ca. 27° south), there are major centers of speciation. Species diversity is greatest from Costa Rica to Peru, with a major center in Ecuador (36 spp.) and to a lesser extent in Peru (28 spp.), and with minor centers in Costa Rica and Panama (both with 23 spp.), and the Guiana region (14 spp.), mostly in Venezuela. A detailed analysis of the geographical distri- bution of Anthurium sect. Pachyneurium by area follows: Mexico. This area is important as an area of local endemism (see section on Centers of Endemism below), with five of its seven taxa endemic. wide-ranging species, А. schlechtendalii, consists of two subspecies ranging down either coast with an extension of the typical subspecies into Nica- ragua. Another species, А. salviniae, extends from the moist forests on the Pacific slope of Chiapas to Colombia, while the remaining five species are endemic. One, А. halmoorei, is restricted to Mi- choacán, Nayarit, and Jalisco on the Pacific coast, while А. nizandense is found in Oaxaca and Guer- rero. Along with 4. sarukhanianum, endemic to Michoacán, these species occur in dry habitats as does А. schlechtendalii subsp. jimenezii, which is widespread on the Pacific coast. The fifth endemic species, A. machetioides, occurs in the humid Chinantla region of northeastern Oaxaca. Central America (excluding Costa Rica and the wet southeastern corner of Nicaragua). This region is remarkably low in species, with only А. cubense (a species better known from the West Indies and northern Venezuela and inhabiting only the drier areas), А. salvadorense (from dry areas in northern El Salvador and adjacent Guatemala), А. salviniae, and А. schlechtendalii (present throughout most of this range). The latter two are the only wide- spread species. Four additional species occur in Nicaragua, А. consobrinum, A. bradeanum, A. spathiphyllum, and A. upalaense, but only the last extends much beyond the wet southeastern corner of the country. Costa Rica. This country is a minor center of endemism and is rich in species diversity. While most species occur in the wet Atlantic lowlands or at higher elevations in the Cordillera de Talamanca (including А. brenesii, А. bradeanum, А. conso- brinum, A. fatoense, A. protensum, A. ranchoan- um, А. schottianum, А. spectabile, and А. upa- laense), several taxa are restricted to the Pacific coast. Anthurium cubense occupies only dry areas of Guanacaste and Puntarenas provinces. Several species are restricted to the Pacific slope in the southeastern part of the country, mostly in the vicinity of the Osa Peninsula. These include А. acutifolium var. acutifolium, A. acutifolium var. herrerae, А. eximium, А. oerstedianum, and А standleyi. Only A. salviniae is relatively wide- spread on both the Atlantic and Pacific watersheds. In addition to being high in species diversity of sect. Pachyneurium, Costa Rica is also rich in endemism, with 7 of 23 species endemic. The en- demics are А. acutifolium var. herrerae, А. bre- nesil, А. eximium, А. oerstedianum, А. schot- tianum, А. spectabile, and А. standleyi. Panama. While Panama is similar to Costa Rica in having most of its species of Pachyneurium restricted to the Atlantic slopes, more species strad- le the Continental Divide than in Costa Rica. Few species are restricted to lower elevations along the Atlantic coast (4. concolor and A. spathiphyllum), but among those primarily restricted to the Atlantic slope and to the ridges near the Continental Divide on the Pacific slope are А. bradeanum, А. colon- icum, А. consobrinum, A. hammelii, A. pseudo- spectabile, A. purpureospathum, A. ranchoan- um, А. seibertii, and A. validifolium. Three species Annals of th Missouri BOE Garden are restricted to the Pacific slope, А. acutifolium (Burica Peninsula), 4. cubense (lowland Chiriqui), and А. fendleri (Coclé). As in Costa Rica, А. sal- viniae is distributed widely on both the Atlantic and Pacific slopes. Four species are found on both slopes near the Continental Divide. These are 4. luteynii, A. nervatum, А. protensum subsp. ar- cuatum, and А. pseudospectabile. Panama is one of the major centers of endemism in Central America, being about as rich in endemics as Costa Rica, with 5 of 22 species endemic. En- demic species are А. colonicum, А. hammelii, A. luteynii, A. nervatum, A. pseudospectabile, and A. purpureospathum. Colombia. The Pachyneurium species of Colombia are widely dispersed with no particular areas of species diversity. A few species are widespread and range into the country from Central America, the West Indies, or Venezuela. These include А. sal- viniae, which ranges from southern Mexico to the upper Magdalena River Valley in south-central Co- ombia; 4. fendleri and A. crassinervium, which enter from northern and western Venezuela; and A. cubense, which ranges throughout the western periphery of the Caribbean. An additional species, A. concolor, enters the north from Panama. There are two endemic species inhabiting the northern part of the country, А. johnsoniae, from the Sierra de Santa Marta, and А. anorianum from northern Antioquia. Two additional and more widespread endemics occur in the central highlands of the country, А. caucavallense and А. glaucospadix. The majority of Pachyneurium taxa from Co- lombia occur in the Amazon basin in the southern part of the country. These include А. atropur- pureum var. atropurpureum, А. atropurpureum var. arenicola, A. bonplandii subsp. bonplandii, A. bonplandii subsp. cuatrescasii, А. galacto- spadix, A. harlingianum, A. loretense, A. oxy- Colombia, the remainder being relatively wide- spread in the Amazon basin t is interesting that vili Colombia has the greatest species diversity for the genus as a whole, possibly with as many species of Anthurium as all the rest of Latin America put together, it is not particularly rich in sect. Pachyneurium. This owes to the fact that most of the species diversity for the genus is in the wettest part of the country along the Pacific coast, while the more mesic parts of the country have been long deforested and are isolated from the apparent center of species di- versity for sect. Pachyneurium, namely Amazo- nian Ecuador and Peru. Only four species of Pachyneurium occur on the otherwise aroid-rich Pacific slope of Colombia. One of these, А. саи- cavallense, laps over the Divide of the Eastern Cordillera in Valle, occurring only on dry upland slopes, and is a member of series Pachyneurium. Three other species, А. carchiense, А. napaeum, and А. narinoense, all members of series Multi- nervia, occur in wetter areas in Narino. Though it is not unusual that there are no representatives of series Pachyneurium from the wet Pacific slope (since they prefer drier habitats), it is unusual that so few members of series Multinervia (a group preferring mesic habitats) occur there. It is par- ticularly curious that only three of the species in series Multinervia occur in this part of Colombia, since there would appear to be numerous suitable habitats for these mesic taxa in southwestern Co- lombia in Nariño, Cauca, and Valle. Perhaps fur- ther collecting in these poorly known areas of Co- ombia will turn up additional species or records for this group. The area is still largely unexplored. Colombia has relatively few endemic species, with only 6 of 24 endemic. These аге А. anorianum, A. caucavallense, A. glaucospadix, A. johnson- iae, А. narinoense, and А. vaupesianum. Ecuador. Along with Peru, this area represents the major center of diversity for taxa of sect. Pachy- neurium, with species concentrated primarily in the Amazon lowlands of both countries and on bot slopes of the Andes in Ecuador. Considering its relatively small size compared to the surrounding countries, Ecuador is clearly the richest country in species diversity. The country is broken up into several different phytogeographical regions, but each contributes a considerable share of the total number of species. Those taxa that occur princi- pally in the Amazonian lowlands are А. atropur- pureum var. arenicola, A. ernestii, A. penduli- olium, A. superbum, A. tenaense, and A. uleanum. Several other species occur only on the eastern slopes of the Andes but at higher elevations, usually above 500 m. These include А. bushii, А. harlin- glanum, penningtonii, and A. santiagoense. Anthurium bushii and A. santiagoense are re- stricted to the Rio Santiago drainage system. An- thurium oxycarpum is wide-ranging, extending from the lowlands up to m. Only two species, А. dombeyanum and А. leon- ianum, occur in the drier highland regions of Ec- uador. Most of the remaining species of sect. Pachyneurium in Ecuador are restricted to the Pacific slope. In this respect Ecuador is radically different from Colombia, since only three species Volume 78, Number 3 1991 Croat Anthurium sect. Pachyneurium 565 of Pachyneurium occur on the wet Pacific slope in Colombia. Species that occur at lower elevations on the Pacific slope include А. asplundii, A. bar- clayanum, A. linguifolium, A. manabianum, A. napaeum, and A. sparreorum. No species is known to occur on both slopes of the Andes. This is not unusual considering the age of the Andes and the broad stretch of cold, arid, and generally inhos- pitable land lying between the mesic slopes to the east and west. Ecuador has the largest number (36), as well as the highest concentration of endemic Расћупеи- rium species of any country, the endemics con- stituting more than half of the total species. Many of the endemics are members of series Multinervia and include 4. acutissimum, A. angustilamina- tum, A. fasciale, A. holmnielsenii, A. lennartii, A. napaeum, A. oxyphyllum, A. palenquense, A. pallatangense, А. penningtonii, and А. santia- goense. Series Multinervia is almost endemic to Ecuador, with only three species, А. carchiense, А. napaeum, and А. narinoense, occurring in southwestern Colombia, and two species, 4. sou- Кири and А. ottonis, occurring to the south in Peru and Bolivia. Peru. Peru is the second richest country for species of sect. Pachyneurium. If only the members of series Pachyneurium are considered, Peru is ac- tually richer than Ecuador because about one-third of the Ecuadorian members of the section are mem- bers of series Multinervia. Most species of series Multinervia occur in Ecuador, while relatively few occur in adjacent countries. Nearly all of the Pe- ruvian species of sect. Pachyneurium are members of series Pachyneurium. The vast majority of Pe- ruvian Pachyneurium species occur on the eastern slope of the Andes and in the Amazon lowlands. Taxa occurring in the lowlands of the Amazon basin include А. A. atropurpureum var. arenicola, А. galacto- spadix, А. loretense, А. manuanum, А. pendu- lifolium, A. superbum subsp. brentberlinii, A. uleanum var. nanayense, and А. willifordii. Almost an equal number of species occurs in the foothills of the Andes at medium elevations (300-1,000 m). These include А. basirotundum, A. harlingianum, A. knappiae, A. latissimum, chylaminum, dA atropurpureum var. atropurpureum, A. llewelynii, A. manuanum, A. А. reflexinervium, A. remotigeniculatum, an tarapotense. The majority of these species are from the northern half of Peru, especially in the departments of Amazonas, San Martin, and Hua- nuco. Areas of particularly high concentration of Pachyneurium are the areas around Tingo Maria in Huanuco Department, and around Tocache Nuevo and Tarapoto in San Martin Department. The Tarapoto vicinity, for example, has no fewer than five Pachyneurium species. Several other species are restricted to the east- ern slopes of the Andes and in the Amazon lowlands but are more wide-ranging in elevation, occurring in the lowlands and in the foothills of the Andes. These include A. ernestii, A. oxycarpum, and A. plowmanii. Two species, А. dombeyanum and А. soukupii, are restricted to higher elevations in Peru and range from Cusco Department in the south to Ama- zonas Department in the north. While the former occupies mostly dry life zones, the latter inhabits generally wetter sites in various montane forest life zones. One species, А. ottonis, occurs at higher elevations in the far south of Peru in Puyo De- partment, as well as in adjacent Bolivia. Two spe- cles, А. barclayanum and A. simpsonii, are re- stricted to the moist forests of the Pacific slope of northwestern Peru in Tumbes Department. After Ecuador, Peru has the greatest concen- tration of series Pachyneurium taxa. It also has the second highest concentration of endemic spe- cies, with 13 of 28 endemic taxa. These are: А. о > 5 3 2 E E ым; reflexinervium, А. гора DM A. simp- sonii, А. superbum subsp. brentberlinii, A. tar- otense, А. uleanum var. nanayense, and А. ap willifordii. Bolivia, Paraguay, and Argentina. A total of nine taxa inhabit this southern, subtropical and tem- perate region of South America. The only species common to all three countries is А. paraguayense (also in southern Mato Grosso in Brazil), and only Bolivia has endemic taxa, 4. atropurpureum var. thomasii and А. Anthurium рага- gua yense ranges the furthest south of all members solomonii. manii, is common to both Paraguay and Bolivia, occurring in the Amazon lowlands with its southern extension in central Paraguay. Bolivia, the most tropical of the three countries, is the richest in species, with nine taxa. These are: А. atropur- pureum var. arenicola, A. atropurpureum var. thomasii, А. ottonis, А. oxycarpum, А. para- guayense var. paraguayense, А. paraguayense var. coroicoanum, А. plowmanii, A. solomonii, and А. soukupii. In Bolivia (in contrast to its range in Peru), А. oxycarpum occurs only in the lowlands in El Beni Province. Anthurium atropurpureum 566 Annals of the Missouri Botanical Garden var. arenicola is also a lowland species, known from La Paz Department at less than 20 m. An- ar. thomasii occurs at Anthurium paraguayense, entering Bolivia from more temperate Argentina and Paraguay, ranges very widely in Bolivia, occurring from 200 to 1,700 m. The remaining four Bolivian species, А. ottonis, A. paraguayense var. coroicoanum, А. solomonii, and 4. soukupii, occur only at higher elevations, all being known only from the Yungas in the De- partment of La Paz at 800 to 2,200 m Brazil. Brazil, with 18 species, is relatively rich in species, but considering the broad area encom- passed and that the species are not concentrated in any particular area, it is not important as a center of species diversity. The Pachyneurium species of Brazil are generally widely dispersed except in the upper Атагоп region, especially in Acre, where as many as four species can be found in a small area. Elsewhere, in general, no more than one or two species are present at any site and may be scarce when present. The area of greatest concentration of species in Brazil is in the upper Amazon region in the western part of the states of Amazonas and Acre. This is understandable since, in general, species diversity for the family increases in the Amazon basin as one approaches the Andes from the Amazon lowlands and is probably greatest at 500 to 1,500 m on the Andean slopes. Taxa that are restricted to the western Amazon basin in Brazil include 4. atropurpureum var. atropur- pureum, А. atropurpureum var. arenicola, А. er- nestii var. ernestii, A. galactospadix, A. krukovii, A. oxycarpum, A. pranceanum, and A. uleanum. Another species, А. plowmanii, occurs in the same region as well, but extends much further to the east, ranging throughout the Amazon basin to Ma- naus, as well as east to Mato Grosso and south to Rondónia and Acre. Two species are restricted to the upland regions of central Brazil south of Per- nambuco and south of the Amazon basin. One of these, А. affine, ranges from Minas Gerais to Per- nambuco, south to Mato Grosso do Sul and Espiritu Santo, but is centered mainly in Bahia. The other species, А. lindmanianum, ranges from Рага, south to Rondónia, Mato Grosso, and Goiás. Two species, А. coriaceum and А. solitarium, occur in southeastern Brazil, in southern Minas Gerais, Espiritu Santo, Paraná, Guanabara, and Rio de Janiero states. They commonly occur on rocks in seasonally dry areas with scrubby vege- tation. One species, А. paraguayense, occurs in southeastern Brazil in Mato Grosso do Sul near the Paraguay border. Three species occur in northern Brazil. One of these, А. bonplandii subsp. bon- plandii, is wide-ranging, occurring principally north of the Amazon River, and extending from the bor- der with Peru to north-central Pará State, but is also sparsely distributed to the south of the Amazon River in the states of Amazonas and Pará. The other two species, А. jenmanii and А. catania- poense, occur only in northeastern Brazil in Amapá and Pará, respectively. Both range into Brazil from Venezuela or the Guianas. One-third of the Brazilian species are endemic to Brazil. These include А. affine, A. coriaceum, . krukovii, A. lindmanianum, A. pranceanum, Eod A. solitarium. Venezuela. The Venezuelan species can be for the most part separated geographically and can occupy diverse phytogeographical regions including the northern Amazon basin, the Guiana Highlands, the northeastern periphery as well as the Cordillera de la Costa, the Cordillera de Merida, and the Serrania de Perija. With nine of the 14 Venezuelan species occurring in the Guiana region (with the highlands and the surrounding lowlands), that region consti- tutes a minor center of diversity for sect. Расћу- neurium. If the non- Vepenipan parts of the Guia- nas are included, three mor e species would be added, А. cowanii, А. с апа А. та- guirei. Seven species with eight taxa occur in the far south of Venezuela in the Guiana Highlands and the adjacent northern Amazon basin. These are 4. bonplandii subsp. bondplandii, A. bon- plandii subsp. cuatrecasii, А. cataniapoense, А. guanchezii, A. iramirezae, A. vinillense, A. wur- dackii, and A. xanthoneurum. One additional spe- cies which ranges into the Amazon basin is 4. jenmanii. This species has an unusual distribution in that it also occurs in Trinidad and the eastern end of the Cordillera de la Costa bordering the Caribbean, then ranges east and south through the Guiana lowlands to Amapá State in Brazil. It is the only species besides А. fendleri that enters the Amazon basin and also occurs primarily outside of the Amazon basin. Five taxa, А. bonplandii subsp. guayanum, А. iramirezae, A. wurdackii, and A. xanthoneurum, are restricted to the Guiana High- lands. One species, А. wagenerianum, is restricted to the Cordillera de la Costa in northern Venezuela. Two additional species, 4. crassinervium and fendleri, inhabit the Cordillera de la Costa but also vinillense, A. Volume 78, Number 3 1991 Croat 567 Anthurium sect. Pachyneurium range throughout the Cordillera de Merida and into Colombia. In addition, disjunct populations of А. fendleri occur in the western Cordillera and the Choco of Colombia as well as in Panama. Anthu- rium cubense is restricted to the northwestern part of the country in the Serrania de Perijá west of Lake Maracaibo (Zulia State) and to the east in Falcón. This species inhabits generally drier areas around the western and southern margins of the Caribbean. Venezuela has more than one-third of its species endemic, including 4. guanchezii, A. iramirezae, A. vinillense, A. wurdackii, and A. xanthoneurum in the Guiana region. Anthurium wagenerianum is essentially endemic, being otherwise known only from Curaçao. The Guianas. Like much of eastern South America, the Guianas are relatively low in species diversity. The region has only five known species of sect. Pachyneurium, with a sixth species purportedly described from the region but never found (А. martianum). Only one of the five species, А. јеп- manii, occurs in the lowlands ranging from Ven- ezuela, down the coast of the Guianas to Amapá in Brazil. It is the only species known for certain from French Guiana. The highland species are 4. bonplandii subsp. guayanum, known from Suri- nam; 4. cowanii, endemic to Guyana; А. lanjouwii and А. maguirei, endemic to Surinam. Despite the low number of species in the Guiana region, the rate of endemism is high, with three of the five species endemic. If А. martianum (pur- portedly from Surinam) really did come from Guy- ana, the rate of endemism for this area would be higher still. West Indies. Section Pachyneurium in the West Indies is restricted to the Greater Antilles and to some of the smaller islands in the region, notably the Virgin Islands. Only three species occur in the Lesser Antilles, 4. jenmanii on Trinidad and To- Баро, and А. crassinervium and А. wagenerianum on Curacao. A sect. Pachyneurium look-alike, А. hookeri, occurs more widely in the Lesser Antilles, as far north as Montserrat. Cuba, the largest island of the Greater Antilles, has only two species of Pachyneurium, A. cubense and A. venosum. An- thurium cubense also ranges to Central and South America, while А. venosum possibly ranges to Ja- maica (see discussion under 4. venosum for details of differences of Jamaican material). Hispaniola has but a single species, А. crenatum. This is the most wide-ranging species of sect. Pachyneurium en- demic to the West Indies, ranging from Hispaniola to Puerto Rico and the Virgin Islands. Puerto Rico has one species, А. crenatum, and St. John, Virgin Islands has one endemic species, А. selloum. While three of the six West Indian species of Pachyneurium are endemic to the West Indies, only A. selloum is endemic to a particular island. CENTERS OF ENDEMISM (General Comments) It is > illustrative to look at the distribution of en ‚ since this does not always coincide with the iis of maximum species diversity for sect. Pachyneurium. Although Ecuador and Peru again come out on top with 23 and 13 endemic taxa, respectively, several countries have between l4 and 34 4 of their taxa endemic. Mexico has 5 of its 7 taxa endemic, while Costa Rica has 7 of 23, Panama 5 of 23, Colombia 6 of 24, Venezuela 5 of 14, Guyana 2 of 3, Brazil 6 of 18, and the West Indies 3 of 6. In contrast, there are no endemic species in Central America outside of Cos- ta Rica. These patterns in Central America are re- flected in the genus as a whole (Croat, 1987), where Mexico as well as Panama and Costa Rica are shown to be major centers of endemism. There are 79 taxa endemic to individual coun- tries throughout the range of sect. Pachyneurium. Although the rate of endemism, country by coun- try, is not a particularly useful index, since some countries, e.g., Brazil, are vastly larger than others, such as Ecuador, Costa Rica, or Panama, the over- all rate of endemism for sect. Pachyneurium is relatively high. If we consider all the countries in entral America and South America as well as Cuba, Hispaniola, and Puerto Rico, there are an average of 3.7 species endemic for each. Central America considered alone has 19 endemics, while South America has 59 and the West Indies has 3. Endemism, it would appear, is no indicator of the center of origin of sect. Pachyneurium, since endemics appear throughout the range of the sec- tion. Rather, endemism in the section appears to be more a matter of isolation, since for the most part the endemics are most common in areas where the relief is broken, or where there are isolated intermountain valleys, such as in the Andes of western South America or in the Guiana Highlands. Species in series Multinervia, which is particularly high in endemics (with all but 5 of 16 taxa endemic to Ecuador), reflects this well. Most occur in valleys on both slopes of the Continental Divide, where the degree of isolation tends to be high. Other causes for endemism may stem from isolation due to the discontinuous nature of life zones. For example, the Caribbean slope of Costa Rica and Panama is 568 Annals of the Missouri Botanical Garden broken by discontinuous areas of tropical moist forest, premontane wet forest and tropical wet for- est. Anthurium prolatum, for instance, is primarily restricted to areas of tropical wet forest in north- eastern Costa Rica, while А. principally from another wet area further to the east in Panama. These two regions are interrupted by a broad area of tropical moist forest in north- eastern Panama and southeastern Costa Rica. concolor is known Relatively few species inhabiting upland regions are very widespread, but some, such as А. dom- beyanum, are moderately wide-ranging, extending from southern Peru to northern Ecuador. Other wide-ranging species occur mostly from near sea level to middle elevations. Among such species are: A. schlechtendalii (Mexico to Nicaragua), А. sal- viniae (Mexico to south central Colombia), A. fend- leri (Amazonian Colombia to Panama), and А. cub- ense (Cuba to Central America and northern South America). In contrast, the lowland species, such as those inhabiting the Amazon basin, are much more wide- spread, e.g., А. atropurpureum, А. bonplandii, A. ernestii, A. plowmanii, and A. uleanum. Endemic Anthurium species by region are summarized in the following list: West Indies (3) crenatum selloum venosum Mexico and Central America (excluding Costa Rica & Panama) (6) halmoorei nizandense sarukhanianum machetioides salvadorense schlechtendalii Costa Rica and Panama* (25) acutifoli luteynii schottianum radeanum nervatum seibertii brenesii oerstedianum spathiphyllum colonicu ectabile consorbrinum protensum standleyi cotobrusii seudospectabile upalaense eximium purpureospathum validifc fatoense ranchoanum watermaliense hammelii Northern sts America (extra-Amazon) (8) anorian Es wagenerianum с ен m та fendleri** crassinervium ohnsoniae Coastal lowlands of western South America (5) asplundii linguifolium simpsonil barclayanum manabianum Andes, western slope (15) acutissimum holmnielsenii obscurinervium angustilaminatum lennartü oxyphyllum bucayanum leonianum 5 сатри napaeum pallatangense carchiense narinoense sparreorum Andes, eastern slope (21) basirotundum llewelynü reflexinervium ü manuanum remotigeniculatum curtispadix ottonis santiagoense dombeyanum pachylaminum solomonii fasciale paraguayense harliangianum var. coroicoanum tarapotense ] penningtonii tenaense latissimum * Some а into the southeastern part of Nicaragua. ** Rare in Panama *** Counts refer to нон not taxa, i.e., not including varieties and subspecies except where split into separate endemic region: Volume 78, Number 3 91 Croat 569 Anthurium sect. Pachyneurium Guiana arpa (8) bonplandii subsp. guayanum lanjouwii wurdackii cowanii maguirei xanthoneurum iramirezae vinillense Amazon basin (17) м galactospadix plowmanii bonplandü guanchezii pranceanum t subsp. bonplandii krukovii superbum subsp. cuatrecasii loretense uleanum cataniapoense oxycarpum vaupesianum ernestii pendulifolium willifordii Brazilian Highlands of south-central and eastern Brazil (2) lindmani affine Southeastern Brazil (2) сопасешт Temperate and subtropical South America (1) paraguayense var. paraguayense Species not included in areas of endemism (4) concolor, Panama to northern Colombia cubense, circum-Caribbean martianum, location not known salviniae, Mexico to northern Colombia ABUNDANCE Despite the fact that Pachyneurium species are not very wide-ranging in general (see section on Geographical Distribution), species are commonly quite abundant on a local basis. For example, 4. atropurpureum var. arenicola is rare in Brazil in Rondonia, near an extreme of its range, but abun- dant locally around Iquitos. Species such as А. anorianum, А. pachylaminum, А. чей. qur. А. pseudospectabile, A. reflexinervium, А. vadorense, A. schottianum, A. seibertii, A. E leyi, and others that have relatively narrow ranges are nevertheless often very abundant locally. In an area where any species occurs the plants commonly dominate the local epiphytic landscape, due to their generally large size, conspicuous habit, and ability to adapt to a wide variety of local conditions. Typically, epiphytic species adapt readily to other niches such as on rocks or limestone out- crops within a forest understory or on steep road cuts or stream banks. While quite at home in the understory of a dense forest, plants there are usu- ally much less numerous. While being able to with- stand low light levels indefinitely, plants react quickly to openings that increase light levels. Clearing edg- es eventually become favorite sites for Pachyneu- rium species. Their ability to survive in the sunnier habitats, and their avian fruit-dispersal agents, al- ways more prevalent in such habitats, are no doubt chiefly responsible for their much greater success nianum solitarium in such open habitats. A single large plant that has been pollinated may produce thousands of seeds, many of which are scattered locally. Seedlings are in general quite successful, and a local population may soon consist of hundreds of plants, especially in disturbed sites. Predation by phytophagous in- sects seems to be only a minor problem, perhaps due to the tough epidermis and the seemingly lower number of phytophagous insects in the seasonally dry areas where Pachyneurium is most prevalent. The ability to survive in low-light conditions, as well as in high illumination with limited moisture, make Pachyneurium species preadapted for many abuses they might receive under cultivation. In- deed, some species, such as А. schlechtendalii are very common in cultivation. Their adaptations for conditions of seasonal drought apparently make them somewhat cold-resistant as well. A number of species cultivated in south Florida and in south- ern California are capable of surviving night tem- peratures at or near freezing. INTERSPECIFIC RELATIONSHIPS Although mention has already been made of a high degree of endemism within sect. Pachyneu- rium and the great dissimilarity between Central and South American species in the section, there are several important groups of related species. These groups are outlined below: 570 Annals of the Missouri Botanical Garden Group 1: anorianum concolor crassinervium crenatum cubense halmoorei machetioides nizandense purpureospathum “| se salvi ida upalaense nad ac aa и Occurring in Central America and mostly north- ern, extra-Amazonian South America, this group of species consists of typical members of series Pachyneurium, mostly with red fruits, typical rosulate habit, D-shaped petioles, often with abaxial ribbing. Although breeding studies between some of the Central and South American species have shown little relationship, all of the species listed here have probably have had at least ancient con- tact with one another. There are no mountain ranges separating these species and, indeed, two range throughout much of the area, namely 4. salviniae and А. cubense. A. ђгепези A. colonicum Á. cotobrusii Group 2: A. protensum A. pseudospectabile A. ranchoanum A. seibertii A. spectabile A. standleyi A. validifolium This group, known almost exclusively from mie ta Rica and Panama (a few lap over into the we southeastern corner of Nicaragua), is dii Re by its orange fruits. Many members of the group have cordate to subcordate, greenish-drying blades (in contrast to typical Pachyneurium), frequently with the collective veins arising from near the base. Group 3: А. acutifolium A. bradeanum А. consobrinum A. eximium A. spathiphyllum This small group is characterized by moderately thin, greenish-drying leaves, thin cataphylls which fruits, generally yellowish, greenish, (though tipped with red or purple in А. consobri- num and solid red in А. eximium). Fruits are not yet known for А. spathiphyllum, but it is believed to be closely related to А. bradeanum, which has white fruits. In addition, all but А. acutifolium have cylindroid or clavate spadices. Anthurium schottianum and А. oerstedianum stand alone in Central America, seemingly not be- ing closely related to any other species. The former is unusual in having a thin, cordate blade with free- ending basal veins and in its markedly seasonal flowering behavior (several inflorescences appar- ently synchronize to emerge at about the same time). While berry color is described as pinkish, darker toward the apex and possibly red, the spe- cies is not at all related to other red-fruited species in Central or South America. Anthurium oerste- dianum is unknown as to fruit color, but with its strange leaf blades with the geniculum situated at the middle of the petiole, it seems to be unrelated to any other Central American species. Group 4: А. atropurpureum bonplandii jenmanil krukovii lanjouwii lindmanianum maguirei pachylaminum pendulifolium pranceanum remotigeniculatum ~AARARAAARARRARAARR . kcu NEE Basically а group from the Guianas, lowland Amazonia and the Brazilian highlands, this group is characterized by purple fruits and the tendency to occur on granitic outcrops or on sandy soil or in areas of extensive white sand deposits. Many have pustules or glandular dots on the leaves, thin spathes, and minutely papillate tepals. Group 5: A. asplundii ushii A. cataniapoense Volume 78, Number 3 1991 Croat 571 Anthurium sect. Pachyneurium A. harlingianum A. loretense A. vaupesianum A closely related group of species with densely rosulate, mostly large, short-petiolate leaves, thin spathes, and long-tapered purplish spadices with minutely papillate tepals and purple fruits. A. bucayanum A. campii A. lennartii A. manabianum А. sparreorum Group 6: These species all occur on the Pacific slope of the Andes in Ecuador and have large, mostly green- drying leaves with short petioles, numerous pri- mary lateral veins and a collective vein arising from near the base. A. curtispadix A. ernestii A. galactospadix А. manuanum A. tenaense A. uleanum Group 7: A closely related group from western Amazonia, mostly at low elevations, frequently with minutely papillate tepals which dry with a crustose, frostlike appearance (most noticeable in А. uleanum) and mostly short petioles and peduncles. A caucavallense . dombeyanum A. glaucospadix A. latissimum A. leonianum A. llewelynii A. paraguayense A. reflexinervium A. solomonii A. tarapotense Group 8: These highland species with purple fruits share in common a tight rosulate habit, mostly short, D-shaped petioles, thick blades, and purplish spa- dices. Many have a tendency for the lowermost primary lateral veins to extend out from the midrib at a broad angle (most pronounced in 4. para- guayense var. coroicoanum, А. latissimum, and A. reflexinervium). Anthurium johnsoniae per- haps also belongs here. Occurring in the mountains above Santa Marta in northern Colombia, this spe- cies would be unusual for the group in having orange fruits, but it shares many features in com- mon with А. caucavallense, especially the broadly spreading primary lateral veins. Perhaps it is the result of an ancient hybridization between 4. cras- sinervium (currently occurring in close proximity) and А. caucavallense, as it shares characteristics in common with both of these species. A. superbum A. willifordii Group 9: This strange pair of species, no doubt closely related, is characterized by quilted leaves, some- times tinged purplish or reddish below, short-pe- dunculate, stubby cylindroid spadices with many minute flowers per spiral, and lavender or magenta, often bicolorous fruits. Group 10: А. barclayanum . linguifolium A. simpsonii These purple-fruited species all occur along the Pacific coast of South America in Peru and Ec- uador, in generally arid areas (less so for А. simp- sonii). Group 11: A. acutissimum . angustilaminatum A. carchiense A. fasciale A. holmnielsenii A. narinoense A. oxyphyllum Most of these Andean species occur on the Pa- cific slope (А. fasciale occurs on the Atlantic slope) of Ecuador and have orange, red, or red-orange fruits. All are members of series Multinervia with oblong to oblong-elliptic blades and close primary lateral veins. Group 12: А. napaeum A. obscurinervium ottonis palaenquense pallatangense penningtonii santiagoense . soukupii SR RADE All members of series Multinervia, these species have purple fruits or are most closely related to other members of series Multinervia with purple fruits. Many tend to have triangular petioles, some- times sharply so. Group 13: A. knappiae A. oxycarpum This pair of species, though not close geograph- Annals of the Missouri Botanical Garden ically, seem very closely related. Both share mod- erately thin, veiny blades with the collective veins arising from the base. Though А. oxycarpum 18 moderately widespread in the Amazon lowlands and on the eastern slopes of the Andes, А. knappiae occurs only on the eastern slopes of the Andes in San Martin Department. Among the South American species that appear to have no close relatives are А. affine, А. cori- aceum, А. plowmanii, and А. solitarium. Anthurium affine sometimes resembles А. lind- manianum, a species discussed above as having affinity with a relatively large group of lowland Amazonian species common on rocks and sand. Anthurium affine, like А. lindmanianum, is rel- atively isolated in the central highlands of Brazil and differs from all other species in the above group by its unusual spadix, which is generally broadest toward the apex. Restricted to southeastern Brazil and in an area where no other Pachyneurium species occur, are two species, А. coriaceum and А. solitarium. They are apparently totally unrelated to each other. Anthurium plowmanii appears to be quite anomalous, with its petioles narrowly sulcate with blunt margins adaxially and a spadix longer than the peduncle. side from 4. crenatum and А. cubense, which show relationships to other Pachyneurium species from Central America and northern South Amer- ica, the other West Indian species show no close relationship with other members of the section. Anthurium venosum, endemic to Cuba, shows no eatures in common with any other species, and A. selloum (discussed elsewhere as a possible hy- brid) likewise is unique among the West Indian species. TAXONOMIC TREATMENT Anthurium Schott, Wiener Z. Kunst 828. 1829 (3rd Quart.). LECTOTYPE: Anthurium acaule (Jacq.) Schott. Pothos acaulis Jacq.. Syst. Pl. 31. 1760 (vide ING). num. Herbs, usually epiphytic, less often terrestrial or lithophytic, the growth habit monophyllous or rare- ly polyphyllous sympodial; petioles variously shaped, usually conspicuously geniculate at apex; leaf blades simple to palmately divided or compound with the lesser veins reticulate; inflorescence terminal (but appearing axillary), one per leaf; spathe free, per- sistent; spadix uniformly tapered; flowers perfect, densely arranged in spirals, tepals 4, cucullate at apex and usually overtopping pistils; stamens 4; pollen globose, forate or rarely (sect. Polyphyl- lium) inaperturate, the exine basically reticulate; ovary 2-locular, with 1 or 2 ovules per cell, rarely 3 or more; fruit a berry. а sect. Pachyneurium Schott, Prodr. . 1860. Pothos crenata L. Sp. Pl. 2: m 1763. TYPE SPECIES: Anthurium cre- natum (L.) Kunth, Enum. Pl. 3: 75. 1841. grex Oxycarpium Schott, Prodr. 450-451. 1860. TYPE SPECIES: Anthurium | Poeppig in Poeppig 845 & Endl. Nov. Gen & S . T. 293.f.B. Epiphytic, terrestrial or lithophytic herbs, often large; stems usually short, with short internodes; roots numerous, usually dense; cataphylls usually straight, lanceolate, rarely hooked or cucullate, usually quickly weathering and persisting as fibers. Leaves rosulate; petioles mostly short and about уд ог less as long as blades, sometimes elongate but rarely as long as blades, stiff to flexible, weakly sheathed near base, rarely throughout much of length, geniculate at apex, the cross-sectional shape various, often D-shaped, rarely terete, sometimes with ribs abaxially, often broadly sulcate with raised margins and sometimes also a medial rib adaxially; blades coriaceous to subcoriaceous, rarely thinner, mostly oblanceolate to elliptic, rarely subcordate to cordate or with conspicuous posterior lobes. /n- florescence erect to pendent; peduncle usually te- rete, short to elongate; spathe usually spreading to reflexed, sometimes recurled, rarely erect or hood- ing spadix, green, sometimes tinged purple or red- dish (especially adaxially); spadix tapered to cyl- indroid, rarely clavate, usually sessile, green to purplish or reddish, sometimes whitish; flowers truncate and rhombic to square in outline at apex; tepals usually matte to weakly glossy, sometimes with nectar droplets, the outer margin 2—4-sided, the inner margin mostly broadly rounded; stigma sessile, slitlike, lined with papillae, or sometimes brushlike; stamens with filaments rarely exserted, anthers usually clustered around and held tightly over the pistils, rarely long-exserted; pollen yellow or orange, rarely purplish at anther-dehiscence, quickly fading to pale yellow or white. /nfructes- cence mostly pendent, occasionally erect; berries mostly obovoid, sometimes somewhat ellipsoid, acute to rounded or truncate at apex, exserted and by tepalar threads at maturity, variously colored, mostly purplish, red, or orange, rarely yellow or white, 1-2-seeded; mesocarp usually gelatinous, rarely pulpy; seeds mostly oblong-elliptic, usually with а gelatinous appendage. Volume 78, Number 3 1991 Croat 573 Anthurium sect. Pachyneurium KEY TO SERIES OF ANTHURIUM SECTION PACHYNEURIUM Blades usually br панк“ to oblanceolate-elliptic, rarely oblong-elliptic t vate-cordate e А dry e Канар ге eddis h brown, greenish brown, or _ бөр а, гагеју greenish or yellowish green; primary lateral veins usually widely spaced and stout andis i more than 3 cm apart), much conspicuous than the interprimary veins (the pm veins; fruits mostly orange, sometimes purple; mostly eum to Ecuador, also а Colombia to Boliv eries Multinervia Anthurium series Vx pn ia (Schott) Croat, comb. ium grex nov. Anthu Pachyneurium Schott Prodr. 466. 1860. Anthurium series Multinervia Croat, ser. nov. TYPE: Anthurium napaeum Engl. Differt a serie Pachyneurium laminae in sicco viridibus cum nervis primaries lateralibus vix conspicuioribus nervis interprimari es For an indication of species in series Multinervia see the species preceded by an asterisk in Appen- dix 3 Annals of the 574 Missouri Botanical Garden ‘u 006 глоде ‘Buoj wo CZ uet ssa[ Ayjensn xipeds 3unmay eui “Buoj шо QZ 01 'srsatpue је peo1q иеці 198uo[ х QZ чеч ssa] xipeds ‘Buoj шо og uey} ssa] ѕәре ‘491 iexeg ^V ^M Y 18017) unmod ‘р ш 009-005 :eureueq изәјѕәм ut pue vory 21507) Jo edo[s эциеру ‘Buoj шо og uet элош xipeds ZuNMA əy} ‘Buoj шо CQ 01 'peo1q uey; 192uo| x c; -0g Xipeds ‘Buoj шо $8 uey, элош sapelg ‘едт “Ayyetxepe eie[ngue Ápun[q 10 рорипол ‘ajesayqns 20 91819) sajoneg 'qG[ 2 Joyeg ^V `Ҹ 29 1801) 1114291а5 "p ш 000'Z-00€'1 ‘вэчешеет, ep елә[үрлогу ш ey eso) pue ешеиед :A[[eixepe peqqu-c-e *1e|ngue1penbqns sənəd LGI "цувоџод pastes Аизипшола *eAoqe uexuns Ápueurunud ѕшәл [e1oje[ Алеш] “ep 'sisaqjue ye ystumoiq 10 ysydind хредс ‘вет ‘aseq Əy} тези шо UJO ‘эре|4 A? JO $4 ломој Əy} шолу SuISIIE sutaA олцодјој “eZ | '(Sseqo[ лочэзив se Juo] se °И uey} 5521 seqo| 10421504 әці ‘os jt) aseq 1e ayepsoogns Ајолел Á[uo 'aje[oeoue[qo 10 зјелодо о] ondqpe-Suo[qo Ápsow ‘apm uey; 192uo| x злош 10 $ Ápsow *ajeZuo[o зореја 'qc yoy “YN unogas p 7777 ш 00€ uey ssa] (spue¡s] шала *uqof IS) затриј sayy :Чејлоло Surqoeo1dde пало 1элэи $240] (шпојјав "p ut oroqeredqns Á[o1e1) ‘ajoned 1ue1moop ҷим ојеполе 0j ојеполе лејпаџем) snuis (родој Á|deop әлош Ајалел) aseq 18 o1eounuj 1soume 0} 9jepjooqns Ајаош səpejg ‘901 Jayeg ^V "М Y 1801) Мари ү ш 009'1-005 01 #1505 tpeo1q se Buoj se x0z чеч} әлош A[[ensn xipeds tazeaoqo oi eieAe[» snurs Чиэшцовце әјоцәа jo jutod је jsepeouq ‘Buoj шо (jg uey, злош Ајепеп səpejg “qT | Iloxeg ^V Y 29 1901) 115140102 "p ш ().CQg'T-009 :eureueq oi вог iso?) :peouq uey, зәдиој x 07 uey ssa] A[[ensn xipeds tur1ojdouooddrq 10 оцоделед ој ојеполе зпш #9524 Iy} 1вэи 10 је 1sopeouq ‘Buoj шо QG пецу ssay Á[ensn sope[qg ‘eT | "ешеџед pue eorny 21500 *durdde[1e40 ѕәшцәшоѕ saqo] əy} чим ојелодо о} ојејпузеда 10 шлојдолоод snurs əy} ‘aseq е родој Apueuruoud ѕәре[ '201 'epe|q әці jo z4 Jaddn aui ш ѕшәл [e1e1e| Kreurud əy} jo euo urojj Живые $шәл oAnoo[[oo чим зореја ‘46 12017) штјралги "y ходе je ојпов *1uo1aura Aiea sa1119q ї$шәл [elo]e| Атеш јо saved сс-61 Чим ѕәре '48 Iloxeg "Y "Y Y 1801) 1151490102 ү ш 068'1-009 *peo1q uey, 198uo| x0Z-01 'peredei Арәҳ:еш xipeds ‘stsayjue ye xipeds Sutpooq 1элэи *peA1n2o1-Surpeauds ayieds ‘stsayjue злојод 12919 92uoo0se1ogu[ '46 -Bug UNUDOYIUDA p a реола uey} зә8ио] x 9 uet 55ој [eoupui4o xipeds ‘stsayjue ye xrpeds Яшрооц aujeds “sisayjue олојед Зитррои зопаоволојиј '26 "ходе је рорџпој ‘зиэ8лэше Ајлез jou зэы1э4 ‘sutaa pesaje] Атеш jo sired 6-ф Ҷим ѕәре “eg ‘U 00$ чеч sse[ е *spuejsp UBA ‘UYOL ‘35 JO ur QOS eAoge ‘ешеиед pue BIIY 21807) :ejeAe[o 03 ојелодо Jo оцоделед ој ојеполе snurs ‘ql 10421504 јешвлеш е шлој 0} peose[eoo ѕшәл [eseq jo sited јзошломој Əy} 1583] 18 *snoooenuoo səpejg 'q9 “qasuy шпвоига сү ш 00$ uey: sso[ ‘eqn sured 6-с зшал pesaje; Areunid ‘Buoj шо 9p-21 ѕәре 'q/ lexeg "V "M Y 1801) шпитуоуов ‘р 7 ш 001 U?U1 ssa] ‘EOI 21507) u1ejseoqinos “saved 05-6 =ШӘАл [9.1918] Алеша :Зиоү шә 001-$Ф ѕәре “eZ ‘u (Og ueu sso] гревојо оз re[nZueun моеи snuts ‘aseq Iy} 0} 221} |8 SULSA [ESP 991 Чим зореја “eg '(SsutaA perae Kreurud 15ошломој Əy} jo auo uio1j Á[a1e1) ѕшәл [eseq əy} jo auo шоу Sursue зитэл злцодјјоо Ҷим sope|g “eC 'siseujue 193je s[edoj jo [әлә[ 0} pa1oe1ja1 pue pajlasxa Á[oo1eos suourejs ‘SISIYJUL je sjeda}; ^o[eq Ápsour prey $13514 yum xipedg USEN Y Азред хә “110y г5игтошлајот ^p ш 00#‘ 3-0 ‘eIquo[o0) 0} тогу 21505 ‘Utes рие jeg + ѕш8:еш Ҷим ayieds ‘әрим uey} ладиој x 0'£-G’1 $240] 101421504 əy} ‘aseq 18 oje1seq 0} o1ejjides-ojepu1oo A[daap зореја ‘qt эзпегУ ^W шпотиојоо "p TIE ш QG['[-0 ‘eueueq enuan tempun 4үрәҳзеш ueu ayjeds topim se Buoj se 1noqe s2qo] 101191504 ay) ‘aseq 18 ojepuooqns 0} зјезипд ѕәре “ep 'sisoqjue лојје рајавхо Zutureura1 pue ројлосхо- Suo| иәјјо suaurejs *ooue1eadde qnoi e xipeds ay} ЗшліЗ ‘stsayjue 1e sjeda; ay} әлоде [әм ‘зиэйлэше ÁA[1eo ssid Цим xipedg “ee 'seqo| 1o119jue ayi se Зио] se $ uet элош Áppensn saqo| 101421504 əy} (5отриј 1924 шод} шпојјав ‘р ut api« uey, 192uo[ x е 01) әрім uey, 193UO| хс UY} ssa] Á[isour ‘aseq 1e 21e1se 10 ojep1ooqns 'ajep1oo 0} ajenides ‘этело-8и0[40 JO әјело Á[MOlIPU 0} 9]PAO зэре|а “ez "вотриј вод pue *eureueq "воџошу pnu) “OMXIJA jo вогоо с “el d 2 со ИЛПІУПЯМАНЭОРа NOLLOAS WINYAHINY 30 53193945 OL АЯУ Croat 575 Volume 78, Number 3 1991 urium Anthurium sect. Pachyne '(xtpeds se Buoj se oot uey, элош зецмәшоѕ Á[ore1) xipeds se Suo[ se 991] uey; ssa] Ajensn ‘afoned se Зиој se am uey} олош 1олоџ “ajonad uey} 191104$ Á[[ensn oppunpag ‘в/с (eqn) шолу osje 2suaqno сү) eoueury Чтоб илэЧзлои pue вопәшү [едиәгу ‘pastes Арјеом Suríip ‘mojaq pastes Аүувәм 0} jeg 'eAoqe иәзип 20 aInosqo 1әциә зэлез| Цѕәлу uo шел oje[nono1 Чим sepelg 'qoz yqiuny (Ст unnua42 ‘р (spue[s] шала pue “оогу onang *onqndey ueoruruoq) загриј 159 y ‘pastes Ápueururo1d SutK1p *qsa1j иэцм uoo sooepms YIOG uo sno[nururo1d sua әде[поцәг Ҷим sope|g “eoz ‘(ajdind-jajo1a o лорпелвј 41150) siseqjue је цатррол Ајолел 10 ysydand xipedg ‘ecg "duroe| 10 ојррпш ayi олоде wos [sour ‘эре ayi jo 4¢ 1eddn оцу шолу Sursue ѕшәл олцоојој 421 ug ам wnydydiyiods ‘у ш 08£-0 *eureueq оз enZereotN| ‘эоипрэ4 о} e[due „081 `L? 18 рјец *xode рлемој parade] ApySys *j221oqns 10 12919 ‘эицм Ашволо 01 мо[ә^ э[е4 xipeds эзе[попо pue 19018 *oje[oooue| Á[peo1q oj eje[oooue| ayieds tooejims 1э44п uo uexuns Ápueururo1d ‘saed QE-QZ $шәл pesaje] K1eurud *e[qist^ jou s([oo eprydes чим sepe[q :Ajerxeqe peqqu-c-g Á[dreus 10 re[nSueu ѕәјоцәй ‘sseqy 01 Виплоцувом s[udeje7) ‘Bug штипхә ‘р ш 009'1-0 ‘EOI 21505) uiojso^qinos jo edo[s ogioeq te[ounpod оз ә[ив „06 ‘F2 1e рЧ ‘xede рлемој ролодеј jou 'ојелејо 01 рголрш4о ‘11MM Аштволо xipeds грохеџал 'ојело-ојвјооов| эцуе4$ :eoejins 1eddn uo uayuns Ápqeo^ ‘sated 21–01 ѕшәл [21э1е| Алешила Храмр чецм 1589] 18) o[qrst sio eprydes чим зореја :Х|ретхеде peqqu-g оз pepunoz зојоцод ови Sunsisiod syAydeyey ‘epg ш 009'1–0 *eureueq 0} enZereotN ‘PIM шо $ UY) $5] ‹әјело-әје[оәоие[ ISOU 1? “19918 оцједе ‘зэре|4 se Япо se Y, 5, sənəd “qez 18017) пјашшру y | NOMEN ш G/9'[ ‘ешеие илојвом PIM шә л ‘во ‘punos Á[reau ‘ajdind qsippaz x1ep ayieds *sepe[q oui uey, 193u0] sajonag ‘вес ‘BOI 21505 pue engezeotN “Á[[erxeqe peqqu-e 0} pepunoz ‘Suoj шо } uey, олош зэюцэ4 :моДэХ o[ed 1o опцм 01 швало “parade je 1€ 10и 10 Ајаолеов ‘aeae оу Suo[qo жредс 'qzc ephe зартоцгуорш "V ` ш 005‘1-005 :edo[s опиецу uo воехеб) ш oorxe[w peU oj omuepue :A[[rxeqe рорипол *8и0] шо у uey, 559] sajonad ‘Цешәә28 “ралоде)-Зиој xipedg ‘egg ‘(oorxay јо edo[s эуюеа 1104} nzauauní "p ut лоргм ѕәшцәшоѕ) әрім шо p] ULY} ssa pue Suo шо (jg Ue сој АЦепзи sope[q ‘111591121 10 onÁudide 19319 за а 404 a + © Noy 2]1911224$ "p 7 ш 00$ 1-00$ ‘gry ©1507) pue епаелеоцу ‘eoejins лодап uo Ásso[8 ој Ásso[grures ‘apm uey, 12840] x p-G'z sope[q їте[п#ивлрепЬ Á[dreus sajonag '412 1201.) 21qDI92dSOpnasd p ш Q0S'T-OSC'T *eureueq Џлојзам :ooepms 1э44п uo Ásso[grures ој одеш “apim uey, лодиој x g-9 sope|q :poprs-p Ápuniq Ајолел *ojo1ejqns вә[оцә LIZ 'epi^ шо Gg pue Buoj ш | uey, элош Áppensn sepe[q :se1iKqdide зиэриэ4 зи ‘вос "эре Jee] Surpuayqns Ајәјегрәшип штупошов чим гјопод '461 0455 wmup:pojs120 у ` U QOE*T-00€ ‘BY #1500) u1ejse^qinos jo edo[s ogroeq ‘эре jo aseq ^o[aq шо / sea] је ројепив um[noruo2 чим o[onoq '261 'sisaqjue je USIIYM Jo 19918 0} qsiuoa12 хтредс '421 1801) (EPN) 1122и2ш1/ -dsqns yuny пурриәзуәәрцәѕ "p "em оэтхэЙ{ илојзом :xipeds Зитрооц jou 'Zurpeeds ‘эуеоэоие| ayieds •тергогодед 01 1e[nZue1penbqns sajor 1ueutuo.1d әлош yonu *enonotdsuoo ‘sired y 1-6 шәл [e191e[ Атеш toppprur oAoqe [әм 1зэреола * 1e[oooue[ loxeg ‘У ^M Y 1901) usauasq y к= Bory 91505 илецјоџ ш [едиәг) тлзә[їрлогу pue uere[rp елојпрло) :xrpeds Зитрооц ‘эц4а-эзело ој э1ело ayieds :219191 зојоцод ‘эл Алеппла дани uey} snonoidsuos asour Ајеолеов ‘snonotdsuoout Surip ‘sired 91-1 зшал [e191e[ Áreurud :4емріш jsapeoiq 'Suo[qo ззрејд '281 "цеопод pastes Apysys *eAoqe uexuns Апиэииио14 jou ѕшәл элцээ[оэ pue шәл [e1oje| Атеши{ '4%1 19d ‘sura Алешидлојш ay) uey, qo-2uo[qo 01 гјејозопејдо ѕәре[а *qgq 18017) штутолр *dsqns поцос unsuajo4d ‘у "Ut 000'2–001'1 *eureueq илојзом Ápsow ‘Buoj шо 6'9 uey} sse[ Á[[ensn хтредѕ :aje[oooue| Á[peo1q 0} ојело ayieds 291 ÁAq[ensn ‘Buoj шо Qc uey} ssa Á[ensn зореја уеэ] '421 wnsuajoid ‘dsqns цоцэс̧ шпоигјола үу ш 00/‘5-<89 ‘ешеие4 pue eory 21500 :Zuo[ шо 6 uey; әлош Áppensn xrpeds :әје[оәоиеү о} oje[ooouv|-Zuo|qo ayieds ҷиәриәй + Аүепвп ‘Buoj шо (c uey, әлош A[[ensn səpejq јео] ел ү Annals of the 576 Missouri Botanical Garden UMÁRIL) зу jeor) UNUDIPDAQ "p ш 008 01 әле ‘u (сф—-( *eureueq о} епделеотј *edo[s әциецу ‘Ayjerxeqe 1e[n3ue-g Apdieys Aqpenen ‘зэ8рэ pastes цим теп8иезрепЬ Áqdaeus sajonad taym о] ^o[[oÁ qsruoo12 вошлод ‘sisoyjue 1? xrpeds эці Зшрооц pue pedeus-1eoq ayieds topounpad шош e[gue „06 ‘Eo 18 раја лр 'ајррпш əy} олоде угцмошов jsapeoJq *ajeA?[? xipedg “egg (124001110 “y 19} 19зэхэ) xipeds Зшрооц 10 Фшрполцвиз ‘1021a р[әч 'реола uey, 192uo| x Gg uey ssa] "опа ја o} дјело ÁA[peouq ој Ајмоллеџ aad ‘BIE "uonoes $5012 ш [eptozedes Jo 1e[ndueıpenbqns sajoneg “egg 'sisaqjue је цээ18- -морәќ Jo *мо[әА 'зицм ‘Чзтиээ1 ‘иәә18 xipedg ‘qcg Азјешен apnnajps y ш 009'1-0 ‘eIquo[o) 01 OdIXAJA ‘pel вәлә ‘astmy Sue] Зицјолип ѕәлеәј eui ‘pedeys-isy *eie[[nono s[Audeieo :snoone[d Аүҳвәм *ләрцәлеү uoos *Zuno& поцм 19918 o[ed xipedg 'асе -Pug nupwual y (edeury) [izeig 01 e»ueury yinog ulejseeqi1ou *pepruu ‘seq ie әм ѕәшцәшоѕ ‘ysydind satiszeq “(asimyiBua] рэЦол зои) иоцеилел ојепшоло Зигуовј ‘[[AUde1eo шолј ју деце дшВлошо зэлез| Suno& əy} ‘ayejoasury s¡¡4ydejeo :гпоопеј8 [je 38 jou *uooreur x1ep xipedg ‘все чәриәле uoos ‘19918 Jo поолеш sjeda, чиэ8лешо Á[reo [үе 1e Jou 'xede je рорипол spusid шо (| uey, 1919218 ‘рэзэ4е1-8и0] ne “PE 1OYIS umuiiqosuoo "p uei ystyuid Зита ‘aym ystuss13 зједо) зиэ8лэше Areo ‘payutod Ајалеце spusid :8иој шо Qq uey, ssa] ‘plorpuyAoqns xipedg "epe 'K[[eixeqe peqquun pue рэрипол 'uonoes $5012 ut padeys-(] sse[ 10 злош зајоцад ‘465 nouog мејригј y ш 000'1-0 Арзош *eureueq [enue?) ojut Зитриззхэа jnq eououry yinog иләцілои ÁAq[ediouud ‹#шләцуеәм pue шкір uoos “иці ayieds :snonproop 10 siaqy 01 Jo uonpuoo зови -tuas о} Зицэчзвэм uoos “шц Á[aje1epour s[[Audeieo :eoejins лемој uo иээ18 qst^o[[oA aped 8шАлр ‘aseq 1e pepunoi oj esnjqo Á[[ensn sepe[q :Аегхере peqqu-c-[ Аүҳеәм 'repnZueujqns ој рэрипол ‘pedeys-q sejonag “qe€ iexeg ^V ‘М Y 18017) asuanjodn ‘р ш Q0g'T 01 Ајалел) ш 002-0 Арзош ‘teory 21507) pue enZeieotN Jo ado¡s onuepy Човуш Sunsis1od ‘yori Ајојелорош eujeds довуш Вицазлод 'snoooeuoo spKqdejeo :ooejins Jamo] uo uo912 -ke13 oj имо ЗшАлр 'eseq је aynoe soepe[q :Á[[erxeqe peqqu-c-g Ароошшола *re[nguejpuenbqns so[oneq ‘еее 'sisoujue ye реолд иеці 1280] x әлош 10 c *eje[oeoue[ Apsour səyzeds (родлошо әлец ѕиәшедѕ әлош JO Z Jaye asuənjodn ју ut 1idooxo) зјојдолр Surxoe| sjeday ‘stsayjue је peojq uey} 1e2uo| x()g чеч} e1our 4[[ensn ‘рэлэ4ез-8ио| xipedg ° asnely ^w 40]02и02 "p ш 001 Uey, ssa] ferquio[o? ој ешеиед jo 1s?07) oyuepy ‘имо Зи р ѕәреү :sisoqjue је peolq ueui ладиој x uey, sse[ A[pensn *ondmpe-Zuo[qo ој 91e40-2uo[qo ayieds ‘stsayjue 1? ээвулиз May} лоло sjodo1p snonotdsuoo Suureoq s[edoj ‘stsoyjue 1e реола uey, 193uo| x Qz- 7, ЧеэмрицАэ xipedg ‘все "реда Зшрооц л1элэи ‘рәхәрәг (ysydind рәЗиц ѕәшцәшоѕ) us913 ayyedg `q 18017) nukan p ш QC€'[-0£6 :seuoz әу 15210jute1 euejuoura1d pue 39m [eordoı} *eurPUeq “4 1omo[ inoqe је ззореола xipeds :siseqjue је xipeds #шрооц pue eijej[ono 'snooo?uoo “ajdind yep ayyedg ете эчи PA чим peg. ADE азпелу Уу шпуойрита ‘р ш (0£ 7-0071 ‘ешеие4 изәјѕәм pue вот] 21507) изаз5вэ ‘зам [етхеде Zupyoe| әјоцәд '202 "од 10 Á[[erxeqe peqqu-a1our 10 -¢ 10 “uonoas 55010 ur 1e[nZue1puenbqns 1эЧиэ so[onaq “*6z ‘xipeds uey} 192uo[ x элош Jo g Á[[ensn 'ојоцод uey, 192uo| x әзош 10 g Ápsow appunpag 9/5 "Вил asuaqno сү ш 007 -00T 59402 op 1зәло} 19m эивзиошел4 pue још peoidon jo syed 1oup pue 3э10} Алр jeoido ut e[onzeuaA pue *erquio[o?) *eureued ој Ugen g ‘eqn’) ísmsegjue je лэриэле| oj ysniya xipeds :Kjperxepe рәрипог епп гтеја Вие трепђ Ајелпозао о} э1э1214п$ sənəd ‘4985 18017) штутоавог та та Y ш 00S'T 01 јолој eas Чволој јам әиејиошәза pue 189.03 1510ш peordo.] jo sed ледом ш грицј [езизицио”) о} ISO) эциецу “ешти ‘stsoyjue ye ofdind-jajota тер xrpeds А етхеде peqqu-g- | Á[[ensn 1o рэрипол «padeys- q 10 rejpn2uegpenb Ápunjq o1 Á[dreus sajonag “e8gz € = N س $ Croat 577 Volume 78, Number 3 1991 urium t. Pachyne rium sec Anthu 'OQIX9[A илајзом “UMOUNUN JO мо[әА э[е4 simaj te1e[oooue| 5 Ачдејео topim uey} ладиој x С uey} sso[ 'ојело Á[peo1q о ојело ayieds 10919 эоцаозелоци] ‘EGF ‘apm uey} зәдиоү x c'e uey, элош *a[pprur əy} әлоде 1sapeolq “SIPAOGO о} aije[oooue|qo sape]g ‘4/ф 18017) 20424424 ‘1вл [дия uumpofimoo y ш 00$ -00€ ‘eo eso) ‘moped xipeds :sso[ 10 opt« шо | ayieds tury} “пәәл8-мо[әА ојед #шАър ‘apm шо CT uey, sso зареја ‘48$ epnyey әѕиәриргти ‘py ш 058 01 ((e»exe() pue ололлопо)) оотхәр{ мола qsrppa1 Zururooeq ua213 xipeds :opi« шо Ст uey} әлош ayieds :еподовџоо Ајојелорош pue имо14 YSIUISIS oj имола SUIÁIP ‘opm шо сү uey, ојош Á[[ensn заре ‘Eg гаргм uey} лодиој хе uey, Sse] *a[pprur əy} 1noqe ye 1зареоза “оца је Амоллеи оу onde зорејд ‘ELF 'xede рлемој ралодеј-8иој xipeds :xipeds oui se Виој se эо1м1 uey} элош зАем[е “sajonad uer; 1o8uo| цопш £[[ensn вә[зипрәд ‘49$ 893 asuaqno ‘р $9102 afl] 15910} 1stour [во4оц jo syed тәр 10 15910} Asp ¡eordox tpropur4o xipeds txipeds se Suo[ se ao1m] иеці элош Ајолел 'војоцод uey, 1911045 A[isour apunpeg ‘EOF ‘sured ст Uey} sse SUA [e1oje| Атеш :ојејпвол ‘42919 золвој цим тие ‘сф 1801”) 2719D}2adsopnasd y ш 0061 -052'1 ‘ешвив{ илојзом 'seuoz э}| 1зэлодие1 опејпошола 10 yam auvjuoured ‘sured QE uey, олош ѕшәл [елојеј Атеш чиорџод sjue|q ‘BG (spquiwadsopnasd `р ut рэрз-ф Арип ѕәшцәшоѕ) uonoas 55010 ш одолојапв о} рэ4ецз-( + зојопод '492 asnely `y әѕиәозо/ y ш QOZ'[-0 ‘ешеиед 01 BOTY eso” *aguv1o aped 10 мо[[әА ystuaa13 oj weai зип “wep шш | ‘во “ysumya ‘раем4п papap Ápsow $1001 ‘aseq 1е "шер шш ([ uey, ssa] ‘рэлэ4е1-Зио| ир "ФФ PUNY прриг/у22]ц25 E ш 009'1–0 :enZeieotN оу OXW, ‘p21 14844 зипц urep шш g-£ ‘ueg, 0} ystuaa13 ‘8шриәоѕәр ој Зиризозе 51001 :әѕед 18 "urerp шш у | UB} алош 'ралоде) Апиа xipedg "epp ‘ued ur s1eqy IMIYA о} ролоцувом uejjo "ту sp&udeje2 “A ¡eixeqe peqqu Ápueururo1d jou so[onaq ‘qf IMP ^V “Y Y 19017) asuanpodn Y ш 009 01 јалој Bas свог] #1507) 0} enZereoIN гралоцјвом Ајолел Ájuo ‘snonotdsuos *joejut Buysissad s[Kudejeo :A[erxeqe peqqu-g Ápueumuoud soponaq ‘BG '"Kj[erxepe peqqu-¢ Арипја ѕәшцәшоѕ «леп Вие лрепђапа 01 1e[ngue1penb Крит sajonad довуш Ápsour Sunstsed s¡¡4ydexer) "ЧТР пјориг1у221ц28 ^dsqns y1uny пюриәүзәуэ$ “y ш 000'1 чеч sso] Á[[ensn oj [әлә eas :seunpuog 01 омхорј “ado¡s onuepgy “шер шо c[ иецз алош Aqjensn xipeds (23e ш Japun рајјол ѕш8:еш oq1) pexago1 Á¡Buo.ns Ápensn *re[nSueu; Амоллеџ АЦепеп “ој та pasun Апавоц иәцо *snoaoeuoo ayieds ‘sed 21 uey; злош Á[[ensn ѕшәл pelare Атеш toseq 18 esniqo 03 ојпов ‘зпоэовмоо Ајоделарош зэре “шеф шш 01 uey, олош Ápsour *[eprozedez Áppensn зојоцод $4201 uo ѕәшцәшоѕ 10 оцАча до зи "qz p Bug штујојутор p 77777777 ш 006-0 ‘ешвие{ pue BOTY esop ‘шер шш g uey} 559] xIpeds :(pexoga1 4801$ лолоџ чер Зшшешәл) Sutpea1ds ој pexogo1-3utpeo1ds клира 1gəu "пола чуј eyjeds ‘sired z [-g sutea |елојеј Алеша ‘aseq 18 эзепиэце-8и0] ‘из Ајајејорош soape[q “шер шш 6 uey, sse| “remnSue1penbqns sajonad :[е141$э1123 sjue|q “ezp "шер шш 6 uey} sso] pue squ үегхере dieys © Чим 1e[ngue1penbqns 10 "шер шо G'{ uey, әлош pue |ергогодел Ájdieys зәциә вајоцод :51о4у ој Sutiayieam Арашола «у Aydejen сер ОР шшщ 8-2 ‘Japus[s JO оз 51001 :o[pprur Əy} элофе jsepr^ 5Аемүе ‘әрим иеці 192uo| x c'e uey, әлош Á[[ensn 'ојелодо 01 ојејоооџејдо ѕәре {OF ерер 2suapunziu y ш ()Gg 01 (еовхво рие orang) оотхәр{ Jo edo[s oyloeg “Wep шо | 01 “оці Ápsow 51001 :o[pptur əy} пое 1? IS9PIM *әрїм uet 1э8ио| хе ueq ssa] *oudipo-o1e40qo 01 onda ѕәре[ "eor (asuapupziu ‘р 10} әрім uet 19840] х$ 01 ѕәшцәшоѕ) әрім иеці 19840] x c'e uer олош A[ensn ‘гепиец Амоллеи 01 әје[оәоие[ ayiedg "4/0 18017) 124100щ]юу ү ш 000°1-089% ‘Ч®отоҶо pue *oosmef “шекем ‘ONXA илојзом ‘моДэХ э[е4 5911194 *xipeds Зшрпојцеиә зәләи ‘Suipeaids ayiedg 'q6€ 18019 25иглоррајр `Y IA ш 008-006 :edo[s dytoeg ‘e[euaens pue лорелјес [3 грал-адивло зоџлод :xipeds Surpnosysua pue 12219 Аце euiedg ege "1e[ngue1penb Ápgunjq ѕәјоцәа ‘pas 10 мојјоХ зомлод ‘stsayjue је xipeds (Surpoou jou jnq) 2urpno1qsua pue 12919 10 Zutpeaids oujeds ‘yoa1aqns 20 12919 рјец 'eseq әці 1е 1зэрео14 'xede оца рлемој parade, жредс -qge Missouri Botanical Garden Annals of the 578 почцос xe "uduoig шпирХәдшор "y ш Q67'Z-0G6 ‘п1э4 uieuinos о] лорепод ‘apts зә (| UE элош ѕшәл [елојеј Атеш ‘469 воі) пиХјәтәр "p т ш 0S.-0€1 под ‘UNIE иес ‘apis 124 ОТ иеці 1эмэ} ѕшәл [1938] Алешма се "шеш eje[npun Á[snonotdsuoo Цум 'eie[oeou?[qo Апзош Jeo] '429 ш 05+ ‘шәл pesaje, Á1eurud јзошлоддп ay) шоу Sursue 10 Zuryoe] suta eAnoej[oo :o»ejuns 1эм0] ay} uo эуезоип9- -тејпривј8 K[snonotdsuoo sepelg '479 1901) aviddvuy сү ш 005-00? плод ‘шеру ueg ‘эре eui Jo aseq eui шолу Яшѕие ѕшәл олцоојоо tooejins ломој uo ojejoundo sope|g ео '(seuozeury) e[enzeuaA 1o (Uep uec) плод *$шЗтеш jeg ҷим опата Ајрвола оз ondqpo A[peosq зорејд Jeo] “ego 'urejsÁs әдешеір pue шѕед uozeury eui Jo samadg '229 'ојоипрод uo „081 + Pley ‘pexeye 1o pexegep 10и xipedg '419 из © 3unung әѕиәрупла y oitpog un] AydAxo y ш Q0$-00€ ‘1openog ш sapuy jo sedo[s илојзом “appunpad шош e[due „091-081 18 р[әч '*pexega1 oi рәхәрәр xipedg “eg “Ayyetxeqe pe[gue-[ Á[uo ло рарипол Ápsour зојоцод ‘416 1e017) ndum сү ш QGI'z-00z'[ “0zeJoquityr) pue reAgog ‘Buoj шо ОТ uey, элош хтреда :Suo[ шо (p uey; олош apounpad :әрім шо QZ Uey} әлош зорејд уеэ] ‘409 POE) Dp Y ш 007'z-00T'I telog иләјѕәм ‘Buoj шо ОТ uey, ssa] xipeds ‘Buoj шо (ур uey, $39] o[punped гори шо Qg uey, ssa] ѕәреј jes] '209 "Jopeno^ ut sepuy jo sedo[s илодзом *1qgreujs s[[&udeieo ‘apis әй ст uey, олош ѕшәл [елојеј Áreuruq 101°) ISUA]]DADINDI `F "q6c ш 0671-006 'eiquio[o?) [ециәо tpaxooy si[&qdeieo ‘apts Jed ст uey} ломој ѕшәл [елојеј Алеш] ‘266 ‘snoa2er102 0j SNOIIPIIOIGNS Burp sope[q jee] topissasqns Jo әүіѕѕәѕ xipedg ‘486 ш 069-015 ‘лорепэ ut зэриу jo sadojs u1e1so^ fury) ди лр saperq zea tuo /-c'c ojeidus жредс “egg '"Á[eixeqe peqqu-Q[-z Арзош зајоцод ‘в/с ‘Burpuaose [je sutea [e1eje| Kieuruq '496 ug umunsstim] Y 1901) UNUDÁDINQ ‘р 77777 ш 008 1-006 плод [едчзээ taue OEI 01 25104391 01 quptur 0} тејпоприод од ope[q yea] jo aseq p1e«oj suraa [е1э1е| Атеш] “eos ‘govjins 1э44п uo рэзтел Á[dreus pue Ациәшшоз jou ѕшәл |е1э3е| Lewd “q SG 3o1) nuojguuuad сү ш 00S'Z-000'1 ‘d изәцізои pue зорепо ш зэриу фо sedo[s uieisee “(ysaxy џоцм шала) aoejins 1э44п uo pastes А|4тецз pue Арпошшола sura [e1eie| Алешма ‘BGG 'siseqjue 1e snoone[W Jo пәә1З qsm[q зои xipedg ‘4с Siddeoq wndapoXxo сү ш 028-001 *eruozeury илојзом ш pea1dsapi« tury} 8и ар sepe[q Jeo] ‘Buoj uro cg-p xipedg "qpc ш 00U'Z-00S ‘е4шоо’) [enu :snoeo»euoo Зи лр sepe[q уез| ‘Buoj шо gg-p] xipedg "esc 'siseqjue је snoone¡3 Suureedde pue иээ18 Ча xipedg "eec 'sisaqjue је USIIYM 10 43119918 xipedg ‘EGC 'ape[q Jee] eui jo e[pprur əy} ^o[eq 1583] 18 JO aseq Əy} шолу SUISLIB шәл IANIAO) ‘етс 'Poueury YING јо satvedg 41 Áe|surgH anru1aps ‘р ш 00$‘1-0 ‘е4шою) oi Oorxay ¿pal зип :pedeus-1sy *e1ej[nono sqAudejeo topim uey, 1280] x С uey}; әлош *1eaur-e1e[oooue] aqeds :juapued Аепѕп aouaose:ogu[ qo звог) 12400104 “y ^ ur 000'1-оср :ue»eougor[y pue oosqef ҷиеќем ‘dy је Чзтиээ18 ‘moped зип toseq at 1eau 10 је 1sopeoJq *xade p1e4o1 po1ode; xipeds ‘Зио| uro çq uey, элош *eprsur uo a[dınd чим рәЗиц вош је *uoa12 *Surpea1ds ayredg 1э8ееН № 1201) wnupiupyynivs у [9^3[ Bas теэи ‘ивовоЧэтА] 01 этшәриә “UMOUAUN sunaz txade ay) 189u јзореола ‘aeae xipeds ‘Buoj шо 9 uey, $$2[ “apisur uo o[dund этер Áppensn ‘42912 әщефс̧ 1201) х1ро4502тт]8 y а © N 579 Anthurium sect. Pachyneurium Croat Volume 78, Number 3 1991 "Burpueose [үе ѕшәл [e1o1e| Kreurud цим ‘paymb you sope[q jee] ‘962 18017) шталгитхој2л ‘р ona ш 008-059 ‘nieg ‘oonueny +uL „01-08 12 Ајовлодал Zutpea1ds иәці pue Surpuoose вшәл [81э3е[ Алешпла чим ‘(sutaa [e191e| Алеши ay} изэмзэ4 pastes A[snonotdsuoo әпѕѕц yea] əy; Чим “әч) рэциЬ Ájsnonotdsuoo sope[q JF] '26/, "грм шә OZ uey} 559] Арзош зорвја je?[ 'q9, 310456 xa "uduoig шпирАодшор сү ~ appunpad ФОМ № 4204 M шпиртирш су 0 SBUEINS ayi шолу од oj pres ‘имоиҳип ино :гјоцод uey, 1e2uo| xg uey, ssa] олу CL onuns pe ut sapuy jo sedo[s иләдѕвә :Á]peixeqe paj3ue Apemoe ‘remue sajonad гаргм шэ (1 uet эюш sape[q јео] ел опрос 2101250/ "p ~ ш 009'1-026 Ховецивс обли "у шодо у нна I ш Q07' 1-068 *®!їАЦО{ pue плод :j2ejurtures оу зовуш Sunsisiad sAudejeo ‘sajonad eui uey, 1e2uo| x Gg uey, sse sepe]g “ep, "ept^4 шо ()] UY] ssa] ѕәрера јеәт ‘вел ш 008 глоде [sour Bumoso ‘apis sad q] uey, олош Ápisour suraa ¡219381 Áreurud ҷим зорејд јео] “ez, 'eppunped uey, лојлоце xipedg qq, 18017) пиошојов ‘р 770707 S ш Q0t'T1-000'T :erAqog 'zeq eT ‘ajounpad uey, лодиој xipedg ‘ету "urejsÁs ogeureJp pue ulseq uozeury əy} jo ser edg “ep, 'sisaqjue је ysudind 10 ysıppa1 xipedg -qzc (имоиҳип UNUDIJIDU p зо шо) 18017) шталгимповдо сү 70 ш (000'2) 022'1–05 :лорепод илојзом ut pea1dsopt« ‘podeys-q Аүйтецв ѕә[оцәа :ooejins лоддп uo Ásso[Sruros qrensn ‘Surpeasds-yoasa Ápyus sope]g Виз штштввутор y 77 *91919]qns 0] 931919] sajoned teoejns лоддп 2J1 uo Азэл[эл pue одеш “juapuad Á[[ensn sope[g 2 69 12017) s1;u2d.42snq12 ‘dsqns Bus wnipuruvpiisngup y ° UNO MUT UE ш 00S'Z-001'Z ‘Ixedo107) pue o10 [4 ‘sajoned uey, зәдио хе uey} элош apounpag “egg ‘lopenoy ut зэриу jo sado[s и1э1зэм uo Juumoso 5210246 '429 Annals of the 580 Missouri Botanical Garden 180.17) UNPUNJOLISDQ “Y плод ‘Unie ес 'ojodege[ Jesu шолј [uo имоих :зовјлпа 1э44п uo peuoije jou ѕшәл Алецлој tumoJq Ápsow ер Зи ар ‘әүрриш мојәд 3зэреола sope|q јео] 'q06 iexyuof x 1exuof MMNOÍUD] p т ш 058 ‘weung to[punped uo ә[див ¿cp ‘E0 1e рә1ләвш pl че) элош Á[[ensn) “шә 21 uey, e1our “UMOIG-MO[[9Á ој иәә:8-мо[әќ Yep Zup sepelg '416 ura md пао» "IBA олпавуј 29 зоупцос y umaundandoa;p сү ш 091-031 пад '019107 pus ‘[Izeig *e12y ш urseq uozeury jo шалеш илојзом ‘әјәипрә uo әјир ,06 `L? 1? рај Ца taseq 1e зиэллтоэр ÁApueuturoud sea] 1? 10 ојпов Á[MOJIBU ‘apim шо 21 иеці 5581 Aqyensn * әрїм шә $] иеці ssa] 'uao18-Áe18 ajed 8шАлр зореја '216 ‘Ul (91 мо[әд ‘ешоғешуұ илојгомудлои ‘Цѕәлу uaq^ eoepms әдіп uo рацоја ѕшәл Алецлој :19918 o[ed Зишлр 'ејррпш теэи јгореола sope[q јео] '206 'duo| ur» Qp uey, ssa] sepe[q је] ‘488 18017) поршоу) “IRA IIMBRIA 29 senuog "up шполпалпаоло сү (зјоцод ouo jualimoap Á[oinoe ѕәшцәшов) aseq je рэрипо1 'оцацје-ојело зэреа 'q6g 801°) n]021u234D стел әлп8еүү 7p sei[nuog y umaundandoaj» ‘р ЕЕЕ авва уе зјпов Á[are1 Jo зјепподе-8иој ‘ae[oaoure[qo Ајрвола оз Аүмолтеи Áj[ensn səpejg '268 эш QOS мо[әд Арзош t*eruozeury илајзам ‘Buoj шэ (ур uey, әлош Апош sepe[q јео] “egg 'ejoure1 Sutieadde jou *ope[q jo aseq 1e un[norua 4/8 18017) Naoynsy “Y [izeig ‘seuozeuly ‘шо [[—p Aq эре jo aseq шо зјошол Suureadde шгпүпошәсу '2/8 "Адолјал 10 payinb jou sape[q jee] ‘Buoj wo с uey, олош xrpeds ‘stsayjue је 8иој шо ()[ uey, элюш apounpag 'q9g 12017) прлојупт сү ш 005 MO[9q пад “o1a107 'uresnong org uroij [uo имои :Ајалјел pue payinb səpejq уез| :Зиоү шо с uey, ssa] xipeds :Зиоү шо ()[ uey, ѕѕәј әјәипрәд “egg ‘peoig uey} 192uo| xg pue p usam]aq Ацѕош sape[q уеэ] suyung пузррлпа p т ш 628-081 “ejenzauap 'seuozeury *eueoedej 01197) рие ешдем 01197) ‘эре eui jo e[pprur ay} әлоде Jo тези [эм шалеш əy} Яшшо( 1sourieddn əy} jseo[ је ‘sred p-e ѕшәл [eseq “aseq уе ојерлоо Ајмојјец5 sepe]g 'qcg зәҳио[ № iexuo[ nanolun] p ш 058 ‘шешіпс 'surejunour ешш ‘peq 941 jo % Jamoj ey} ш шлеш ay} qua Зи лэш ‘sated z Ајалел ‘лей | Á[[ensn ѕшәл [eseq “aseq те әјвәипо 10 рорипол 01 osniqo sope|g “ecg рео uey} 1а8иој xZ= 'eje[oooue[ Á]peo1q 10 ondqo-o1e40 Á[peouq оз эп4е ѕәре :Sutpea1ds 10 19319 зэлеэ] ‘+9 ug ам штуојутрига ү ш 055-06 *eiquio[or) pue nieg зизэвре pue лорепол :peouq ueui 192uo| хф 1940 ‘ae[oaoue[qo oi Zuo[qo зэре tiuepued-Surpeaids зэлеэ] ‘epg 'peoiq uey} 192uo| x ZS JO рео uey} зәдиој хф 1940 sape[q јео] ‘egg "Suipuaose [је ѕшәл [e1a1e[ K1eurud цим sepe[q уеэ] 'qcg ug зиме Y эш 008' 1-006 ‘әд [ENUI {95104491 01 фирии 0} 1e[aotpuad.ied aseq рлемој ѕшәл [e1eje[ Á1eurud цим зорејд јео] '228 ‘U 008 мојед Арзош Bumoso ‘apis Jed 1949] JO ст зшал [е1э18| Áreurud Чум sape]q jee] 'qz/ NOYIS хә 'uduoig шпирАддшор сү ш 09) ‘Z-0S6 ‘neg Ulayinos oj Jopenog ‘ajoned uey, 198uo| xg uey, әлош ајоџпрод :eie[oeoue[qo Арзош səpejq јеәт '418 олрос 2]D19SDÍ “y ш 0091 -016 :10peno ш sapuy jo sedo[s иләјѕвә ‘ә[оцәй uey, зәдиоү x С uey, ssa] appunpad “3uo¡qo sepejq уеэ] '218 'eoejins 1э44п ио pastes Adreus pue Апизипио14 you ѕшәл [2223] Атешма ‘908 18017) пиојбишиәа ‘р ш 00S'Z-000'1(008) әд Ulay110u pue лорепол ш sapuy jo sadojs илэззвэ (8шАлр uo usas) зоврлп5 1eddn uo pastes Ајатеце pue Ápusumuoid ѕшәл Аеш '208 581 rium sect. Pachyneurium Croat Anthu Volume 78, Number 3 1991 BOI) shine p а - — — ul 000* 1-006 :®14ш0[07) "ошле ‘aseq зе 9jenuojje-Zuo| sape]q Jeo] tori пецу 19840] x QE Uey, злош xipedg се) (01 "e[rssesqns “ә[отл Ҷѕтррәл о} e[dund улер жредс 4901 E01) шпитдрирш ср 7777777777777 ш (Gf :лорепод ‘ч4вией ‘wo с̧с ajeudos мола ysydind хредс '2901 'epouad иецз 198409] х@-@ ‘aseq 12 зјепподе 01 ојпов sope[q уеэ] '4201 Чох Y Yoy ^W шпироулош ‘р 77 Пе uMouxun што ‘apis 19d злош Jo QZ SUISA [елэуе[ Алеша '4201 18017) asua]]naponpo "p 777 ш (067*[-006 *etquro[o?) fenua ‘рэхооЧ sp&udereo :Азлем jou *juaoseqnd зоол tep[dmd-jopora зипа "901 jeor) apiuosuyol сү 7777777777 E j BOUT теәи ‘erquo[o’) *euepepZeyy *3qgrezis spáydereo :Азлем A[snonotdsuoo mq Ájəmuwu pue sno4qe[2 $1004 ‘әЗие20 31114 "ego ‘apts ad 05 uey} ломој ѕшәл [e191] Алеши ‘egol ‘pedeys-q Ajyensn sajonag '4001 ‘Buy шпәрари б re ys аа ааа ааа ш 008'1-005 *oseq рлемој хэлиоэ JO 1482.35 SUIS 1eui чим Sape[q jes] `4501 18017) әѕиәпһиәрраі : раро охе Ш 00-08 *oseq рї®мо] BVABIUOD K[snonordsuoo sutd.1eui чим sope[q je?'] ‘GOI "eiquio[o7) pue Jopenoy ш sapuy jo sedo[s и1эзэм “aya Áp simaj :Zurpeaads ој рэхэрэл ‘apm шо c'z—$'() eyreds ї1еїп#итц вәүоцә "101 199MG ш ио( *£) UUNIIDILOS сү ~ E MEE а кесара ас аса a ш 062-0 *[ze1g [e1seoo uiejseeqinos ‘zapora э[е4 oj oj Азир зип 199.19 ‘PIM шо ep-c'z ayieds :919191 sajonag "e[0T ‘912193 10 1e[nZu?eu зајоцод '2001 apm шә ст uey} әлош Ацепзп зарвја јео] '426 уголо огидтолро "p 7 ш009'2-002 1 :erquio[o?) ‘OUNEN *10peno'] ‘yore «(ојд ла QstuMOJqQ 0] UMOIG “UMOIQ -овивло) YSIUMO.G 0} поолеш xipeds ‘sajoned uey} 192uo| x c^ - cc ‘Buo шо (ус uey} ssa] Apsour sepe[qj??T 'q66 umppurumpijsngup ‘dsqns Вид unjpuruvpiisngup tee ш 002°] ‘вә :лорепол "ецошуога геј та млер xipeds ‘sajoned uey, зә8иоү x 9:9- c 'Suo| шо QG uey} әзош Апзош зарејд јео] '266 ‘aseq 18 зјеппоце 0} ƏMƏL зорвјд је] ‘4986 ‘Pug шпграри сү PUES ш 008‘1-005 *1openo, ut sapuy jo sedo[s илојгом ‘Buoj uro 91 uey} элош Ápsour *1e[ngueuj ÁA|dreus so[onaq “8/6 ‘aseq рлемој хэлиоэ JO juZre1js зш8леш Чим зарвјд је] '496 12017) asuanbuajod сү pron ш 00-051 :лорепод ш sapuy јо sedo[s илојвом taseq premo, алвоџоо Á[snonordsuoo suiZ1eur чим зареа jee] '296 “aoejans лодап uo pastes K[dreus jou sutaa perae Атеш] '426 "Вил аса un pud ү eee i i tti e rm t i e i i it Ds ш 000'£-000'c *10peno ш sapuy jo sedo[s илојвом (цчзәл} џацм џала) ooejins 1oddn uo pastes Á[dreus вшол [91918] Áreuniq “ecg леда шо с uey} ssa] Арвош эре Jee] јо ојррпш је ѕшәл [e1eje| Алешы{ '426 18017) nuosduns ‘р © ш 008-009 ‘әд 'sequin әри шә / uey, әлош *aje[oooue[qo-Suo[qo sepe[q zea condydidy ‘qp6 qug uwegofinbug р не Jopenog [ејзвоо ‘әрим шә / ueuj ssa] “леошј-Виојдо sape[q jeo| ensar L "v6 "jede шо g uey} олош epe[q Jeo] јо әүррїш reou ѕшәл perae Алеш ‘веб ‘apm шә GI uey} ssa] Ájensn sope[q J2] '226 "urojsÁs әдешелр 10 шева uozeury əy} ut виштооо 100 "[Ize1g [635802 uiejsea Jo плод [21509 илоцјлоџ pue Jopenoy ut sepuy jo sado]s плојзом *erquro[o7) [21509 ошова pue [едиәо pue илоцјлоџ Jo seredg -qg) Missouri Botanical Garden Annals of the 582 иәцм) aoejins 1eddn uo одеш Áppensn sape[q jeo[ :Zurpea1ds-j2e1e e»ueoseiogu[ 'egz[ 'eje[nono Jo paxooy sp&udeje?) '2221 'авлодај 1элэи 'Sutrpueose [е 5шәл [e1918] Атеш '4121 (Hug umuissim] p ш 008'1-005 ‘məd [p11u29 ‘ашрии 01 „Of [ 01 9$101121 01 1епотриэ41э4 эре yea] jo aseq рлемо) suraa [елојеј Атеш “BZ шер шш c Uey} элош UYI “отці ULY 12840] х p uet әзош й 10 отці ивці 192uo[ x (ур ue зә xipeds '4021 18017) шпиттѕәйпра y (seuozeury *sadneA) eiquiojo?) uro1j Кио UMOUY (рәр иэцм) Авмрии "urerp шш фе “оці uet 193uo[ x OF чецу элош xipedg '2021 '198uo| Á[ISOUI ѕәреү ze] ‘X011 ueui 198uo| x ст uey, элош xipeds uat *e[pprur ^o[oq Jo 1eau 4зэрео14 jt 10 'ојррпш əy} eAoqe 35эреол4 әре јео] "q6TT 12017) umpunjoaispq "y nieg ‘Une чес 'ojodere] jo Ayunta шолу Á[uo uMOUY “отці uey} 1ә#цо| x сү uey, ssa] xipeds ‘Buoj шә үф 01 dn “apppru ^o[eq 10 1e 159peo.q эре Jeo] '2611 Бе одаја pue әцу ои Á[ensn ‘snouea $1244 ¡¡4ydexeo *e4 peorde ш рэлэцим jou Кепѕп зоџаовајоплјш :o[punped uey, 1231043 xipedg '4811 jeo1) пиоштоја y ш 006-06 *[ze1g pue ‘Aendeieg ‘euog 01 плод мола. чатррол *axrreq *aug злоду ПАЧЧезео гролоцим 54 “eo [eotde цим aouoosojonugur tepunpad uey; 1euo[ xipedg '2811 "919191qns 1элэи jnq uoroes 55010 ut padeys Ајеполел sənəd '4211 12017) шпиргзирла сү [zeig “91819 әләү fur 05-Ф Aq әреү sea] Jo eseq шолј әзошәз Suureadde umpnoruad ‘Buoj шә cz uey, злош sajonag '4/11 8017) азиаХрити стел и шпират ү З ш 091-001 “пә *озәлот ‘әре Jeo] јо aseq 18 шпүпотиә8 ‘Buoj шо GG uey, sse[ ѕәјоцәд "e/ TT ‘u 00$ ueui ssa] turseq uozeury Jaddn :oseq је әјепиәце ѕәре ‘4911 1әуио[ N Jayuof птпоЃит] PY ш 058 *ureuung гојепиојје IJAJU “aseq уе зјезипо ѕәшцәшоѕ JO рэрипот 0j әѕпідо зореја '2911 "uo| шә (С uey} 5591 a[ounpad {510998405 А]эзетэрош 0] зпозовјлецо ‘aseq je esnjqo 0j зјепизуе sope[q jes] ЕТІ Suuung myoppinm сү ш с$78-<с/ 'seuozeury *ejonzauaA ‘шо c'/-z 10} 1ue1moep эцуе4$ tayejounda *әрїм uey, 1e2uo| х2'2-6'1 *ondmqpe-a1e40 ѕәреү ‘сү Зиципд әр2ә1шоп "p ш 0се* :seuozeury 'ejonzauaA ‘шо [ uet sse[ 10} уполтоор oujeds ‘aoejins Jamo, uo ojejound-re[npue[g гер ‘pim uey} 19guo| x7'[-G'|[ ‘91840 әре "?GITI 'epe[q əy} Jo aseq ayi uro1j ојошал Suueoadde jou шпјпошов ayi ‘aseq је ajeydys jou эре уеэ] "qr TI Зиципя пгаузитпа y 777 - ш QG[- 0T ¿seuozeury ‘ejanzauaA "Шә 5 -G'[ 19} aseq 18 po»[eu ѕшәл [21938] Алеша јаошламој ay} pue зиэлиоэр jsee[ 12 эре əy} JO ‘(э10шэ1 Зиџеодде шпјпошов ayi) шпүпошәЗ oui #шцәвәз э10}э4 шә с 0} dn jo ooueisip е 10} эре ay} jo eseq aui puoAaq paxeu qupru əy} *ajeitdus эре јеә ‘етт 'Suo| шо € uey, әлош e[punpad :snoooeuoo [Xo ‘aseq је ejepiooqns зорвја fea] LEII 'Uonoes 58010 ur 9je1ejqns Ápsour sənəd ‘BG [ [ 'peouq uey} 192uo| хє> зореја јео] “el TT (еициәЗіу ‘AenSeieg ‘eog 'spue[qZrq uemze1g) шә1545 одешелр pue ulseq uozeury jo ser»eds '2011 ‘gjoned uey, зә8ио] x с uey, 5591 э|оипрэ4 “*601 ‘a[dınd-1a[ora 0 ysydind 10 ‘uooseur *qsippo1 “pax xipedg '2801 ‘juasqe 10 эре Jeo] jo e[pprur 1eau JO олоде шолу Fuse шәл 9AHn29[[O) ‘416 вог) uma402440ds ү - ш 067-015 ‘2орепоя :әѕед Ie ojeeuno 10 esnjqo 0} ƏMƏL зореја jeo[ ‘Jory, uey, 192uo[ х 9g uet 5521 жредс '4201 583 Anthurium sect. Pachyneurium Croat Volume 78, Number 3 1991 сорепэя pue плод ut Bumoso Á[isour ‘ondydida Ápsour уе :so[nisnd 10 ѕиоцејэипа тејпривја јпоцим sope[q jeeT 'qcc[ Pug шпитиршри y [zeig ut eruopuoy рие ете илоцупов ‘sgrog ‘osso1s) отер :A[eiuozuou #шЧәәлә шојс 4821 18017) (Sunung) urmnupÁpn8 -dsqns Sunung приојаиод y ЗЕЕ ш QOT'T-007T “ejenzaus A u1eqnos *ureuung ‘әјоцәй ay} se 8иој se отму 1noqe Á[[ensn :e1eAoqo 0} oje[ooout[qo ÁKrensn sope[q jeaT `40©1 npuvjduog -dsqns Sunung npunjduog ‘р 77777 ^ ш OOS Mojeq Ápsour “гел иләззәмцүлои *e[onza о ш 088-61 ‘веп2эиэл ш лелцод pue 'seuozeury "equio[o? uiejseeqjnos :2uo| шш (ф-с] edus :qsmuooi8 3ulAIp Арѕош ‘aseq je ојеоиплапв ој papunos sepe[q je9T '2621 '(sunueoo[ огу e[pprur) seuoz ешү ory əy} jo ymos Ајалел *urseq uozeury uieqiiou 40919 SWS '282 1 'siseqjue је paieui jou ayjedg ‘4951 (Зид тиршиәГ y 00 ш QOS мо[әд Арзош телу ш едешу 'веџето) *e[onzauaA *oZeqo[ pue рершиј :Зио шә (5-11 ‘зпоэпрео jou ‘snoaoeiioogns оцједс -q) 7] 10495 мојригј “y | ш (02%'2) 000'1 91 [әлә[ eos *eureueq *e|onzeuaA илојзом “етфшојог) UJ9IS83YIIOU 01 илојзом ‘Buoj WO c'[[-G'c 'snoonpeo “umi eyed '2/21 ‘sIsoyjue је ролоцим Á[[ensn ayiedg '2921 ‘SBUBIN<) ay] pue “eponzaua A ‘erquojog pue [zeig ш реолдварим ‘oryiytde 20 үетцзәллә} Арзош иаеч (Аја "Xeqe 15291 18) зојгдепа 10/pue suonejound геприе ҷим АпиэпЬэлу ворејд jeoT '2071 'eje[oeoue[ yens syAydeyey '4221 18017) ә$иә]әло] ү 77700000 M M m лорепо pue ‘erquo[o) плод (seuozeury) [12218 илојвом “urseq uozeury 1odd() ‘sisayjue је "wep шш TT -9 '1uepued-Zuipee1ds ој #шрвәлд *oponad uey; 192uo| xz[-y epunpeqd "ФТ 18017) asuaodpiupjDO 7M [291] Ul 31819 етед uioqiiou *1eAmog pue seuozeury ut ®эп2эиэл ‘з1вэцзие ye "ureip шш c-e 'juepued Aporys *e[oned uey, 1e8uo| x9-c e[punpeq ‘ефе ‘u 00$ ^o[eq *eruozeury илэЧзлои pue ulojseA :(qsa1p Udy) aovyins 1eddn uo Ksso[2tures sope[q jeo| tiuopuod-Surpeouds o1 зиэриэ4 o2ueosei1ogu[ '4221 VOID штиво итату ју 7o Me Jopenoy pue eqwuojog ut sapuy jo sadojs u1ojsea uo (ш QQg‘| o1 dn) w 00$ элоде :(uso1j Annals of the 584 Missouri Botanical Garden шпирэп тєл аи шпиоэт y ш (СС мојед :eruozeury и1э15эм ‘pends 1эЧиэ ш o[qrst4 SISMOY Q ueqi AO "epp ‘snoone|s Suureodde pue э[е4 xipedg ‘EOF [ ‘эюцэ4 uey, ладиој x ¢ uey, злош эоипрэ ‘6£ [ "Зшриәэѕе pje ѕшәл үрләр] Алеш ‘[ ug штшл пој y ш 008'1-00$ плод [едчәә :ope[q Jeo] јо aseq premo, Á[uo ѕшәл [81342] Á1eurud овлодал Á[snonotdsuoo Цум “sow ye aje[[nq Аруеэм ‘aseq 1e ојеппоце 0] ојпов 'аргм шо COE 01 sopetq Jeo] "ВЕТ 18017) umaa2uixajfa4 '*Y ^ ul 008 -099 плод '(oonueng) eue оЗшу jo Anunta шолу Ајио имоим *Y13U9] ләці поцЗполцу sutaA [елоје] Алециа4 ојеполе “әвло.дәл Ajsnonoidsuos цим ‘paymb Á[snono:dsuoo ‘aseq ye эзерлоэ Á[Mo[[eus 01 asn1qo ‘эрмм шо 81 01 sope[q уеэ] “*g£l "y gue| sir 1104810141 10 ape[q јеә[ jo aseq рлемо} os101191 ѕшәл [eJoje[ Атеш “peo1q uey} 193U0] х € uey} злош Á][ensn səpejq j**'I “ТТТ (Вил asuajodpip] сү + ш 00/-09$ ‘зә ‘UNIE URS :peouq uey} 19840] xg]-9 'snoone[8 Зиыеэ44е jou xipedg ‘qog шпирәјт ‘тел и шпират сү ОЕ ш (сс мо[әд ‘zeig pue “Jopenoy *eiquio[o? зиәов(ре pue ‘плэ{ ‘01э10’] :peouq ueui лодиој х0<-81 Арзош *snoone[2 Suueoadde xipedg ‘воет "силе dieys цим Апахере aeons А]рео44 01 хэлиоэ sajoned гојошој Suureadde jou шпүпошәсу 46е 1 уколу штрјпотиг Jow сү ш (091'1 -O€S ‘плэ4 'seuozeury pue unie иес ‘зи8леш jun[q чим Аегхере o1eopns A[osniqo pue Ајмоллеџ sajonad turo QT ој dn Áq эре jo aseq шо әзошәз Zuureodde шпүпошәсу egg] ‘apts sad 01-© Арзош sura [елојеј Алешма ‘ее олрос шпиттиогј “Y ^ ш 00S'1 -006 :1openo3 ‘eimqequiy ‘ajdind тер xipeds ‘Buoj wo 16-06 əpunpəd :Апетхеде peqqu-g-c зајоцод :[eujse1je], "е 12015) шпир иль ‘р 07070007707 ш 008'1-ОФФ :лорепол pue вашою’) ut sapuy јо sedo[s u1aj1seo “apdimd-1a[ota ajed о Xutd-uouies xipeds ‘Buoj шо 06-12 eppunped :Aqpeixeqe peqqu-c-¢ 10 peqquun pue рэрипоз sajoned :on&udidq ‘вет ‘apts sad $5-01 зшол [еләдеү Алеш "eel "ustMoj[[3Á JO Yystumoiq ѕшәл [e1oje| Алеица yum “umo1q ZutKip Арзош səpejq jeo| ‘snonotdsuos -ut Ајолцејал злоду eui цим 'имола pue jo?jurtures Sunsis1od sp&qdeien) 4121 1$2u42 ‘тел Вид и15аилэ p 7707 ш QOS мојед ‘ешоғешұ ul9]s9M ut peaidsaptm :Зиоү шо cz-g Ajpensn әјәипрә 12919 зопеозалојиј] '4221 18017) прлор8 10 "тел "Su тгила ‘р ш 00$ :лорепод ‘ezeyseg шоьл} Аүио имоиҷ 'Виој шо £c-gz гјоипрод ‘juapuad-Zutpeaids зопоовалоиј “ezÉl “ysiBuezo A[snonoidsuoo ѕшәл үезәдеј Атешила Чим ‘19918 Вир Арзош səpejq yea] :ѕзәду зјед ‘auy jo uojajays e se Sunsisiad spjáydeep °1 £ 'q2£T LET Croat 585 Volume 78, Number 3 1991 Anthurium sect. Pachyneurium 1.01) прироЯао ‘тел "Hu тузадила ‘р 7700007 ш 00$ :лорепол “ezejseg шолу Ајио UMOUN *onÁudide :3uo| wo g-, xipeds ‘Buo шо Qc uey, ѕѕәј o[punpaq LECI ‘apis 1од 91-01 =шәл [елојеј Алеши 'ezcT "Аетхеде peqqu-(g)c-e(T) sejoneg ‘9871 прирјдиод ‘dsqns Зиципя mpunjduog ју ~~ Шр рне сак ата ш QOS мојед Ápisour “геа uyou *e[onzouo A UJaY NOS *etquio[o) uiejseegjnos ‘Buoy шш (z-g ednus ‘цемоцәќ 0} имола ЗшАлр Апзош ‘aseq ie әјепиәце о} əmə зорејд Jeo] ‘атс 1201.) перодлурто -dsqns Sunung пригаиод ко ` Ui 088-52 *e[onzaeuoA ш телпод pue seuozeury *erquio[o7) изәцѕвәцјпоѕ ‘Buoj шш (y - cq adns {45119918 Зи Алр Á[isour *oseq ye ојеоппдапе ој рорипол sope[q јео] сес] ‘apts 19d 01-6 ѕшәл [елэзе| Алеши ‘96$ 10496 хә 'uduo1g шпирХәдшор сү 77m nun ш 97'5-0G6 ‘п1э4 uieqinos 0} лорепод [ециәо рео uey, 192uo| x С uet ssa] Á[[ensn sape[q Jee, ‘Buoj шо Qe-([ зојопод "4021 18017) nysng р O ш 00е“ :лореп -9Y 'овециес-еполој ш NININA) эр ешеллос ƏY} шош Ајио UMOUN ‘peoiq uey} 1e2uo| x p'G ‘во səpejq jeo| ‘Buoj шо g “eo sajonag ‘вост ‘apis ad [4-01 зшал [еләјеү Áreuruq ‘61 "Aqpeixeqe pepunoi sajonag ‘8F [ "Qse1j пәцм ‘әлповцо Арѕош “payoja Ајепопогдеџоо jou ѕшәл Алецлој Чим ope|q јео] "921 18007) шпитшюХ py A аннан ш 022-03€(0bz) плод [Fuso {4594} поцм aoejins 1э44п uo payoja A[snonotdsuoo ѕшәл Алецлој чим эре јез] ‘LPI 'e1e[oeoue| ‘148ел4$ siÁudeie?) "Т 8017) ISUIOİ DIUDI DI "p 700 ш 00$ моЈәд zeig ш еле ‘в|эп2эиэд ‘телцод pue seuozeury ш {Ҷѕәлу usym) аовјлпе 1э44п uo Ksso[2turos ѕәреүа jeo[ ҷиәриәа a»uoose:ogu[ “q9p1 18017) шпили 1Dy О C C D MM E ш 008‘1-0pp *1openo; pue etquio[o) ut sopuy jo sedo[s u1ejseo ломој (чвәл} џецм) aoejans 1eddn uo одеш Ájpensn заре yea] :Surpeo1ds-15219 ээцээзэлоци] “GOP 'eje[[nono 10 padeus-xoou sp&udeje) ‘ест 'sisaqjue је рэлэцим jou ayiedg {Bug а ТТ Що ш 006 мо[э4 Арзош тела ш едешу "веџето *odeqo[ pue рершај, ‘Buoj шә (c-[[ 'snoonpeo jou 'snoooeuooqns eujedg ‘фт nouog мејригј ‘р ш (02%'2)000'1 01 [9491 eas teureueq *ejonzauaA UJ9]S9M “BIQUIO[O”) u1ojseat1ou 01 [ециәо-иләјѕәм “SUO| uro [ [- GG *snoonpeo ° ung ayiedg ‘фет "Чзэ поцм a2eJıns teddn uo роцоја Апизипио/4 jou SUI9A Алецлој, "ЧЕРТ 18019 0]021u24D стел олпавуј 29 sei[nqog "wp шпапатаолјро Yo ш (096-001 *eruozeury иләјѕәм :qseip usym ooejms 1eddn uo роцоје Ápuoururo1d ѕшәл Алецлој “eepl 'siseqjue је рэлэцим Á[pensn ayredg 'snoone[3 Zuuoedde jou ‘э[е4 you Ájensn AA "YI xipeds 4071 Missouri Botanical Garden Annals of the 586 телцоя рив seuozeury “BIQUIO|O”) илојевоцупов ‘Buoj шш (ус [ edus *qsruao18 Зи тр Арѕош ‘aseq 18 ојеоиплапв оз pepuno: sope[q jeeT ‘ео (Апет -xeqe sea] 18) сиоцејошпа лејпривја snonotdsuoo Чим 4епѕп sope[q n 'е291 'Suo| шо og uey, ssa] Ápsow xipeds 0] шо 0$ uey ѕѕәј Арвош зајоцод :Surpea1ds-15219 о} 12919 зопаоволорјш pue SIALIT '4291 ug ам wnyofynpuad у ш 66-06 *etquiopon) pue niaq iuooeípe pue лорепод ‘Buoj шо 05 uey, әлош xrpeds ‘Buoj uro 0$ uey} элош АЦепзи ѕәјоцәй ‘yuapuad-Surpeesds оу yuapuad o»uoosorogur pue зэлвэ] ezo] "увалу поцм amosqo Азош *eAoqe pou1a Á[snonoidsuoo jou sutea Алецлој Ҷим греја jeeT '4191 12017) шпититјАуора сү ш 011-05$(0$5) ‘әд [enua ‘4894; иэцм ooepms 1eddn uo payoja Ајепопотавиоо ѕшәл Алецлој чим әре јеәт ‘191 "ад лец рие әш jou Aqjensn *snourA $14944 [¡4ydexeo ‘$4 једе ut рәзәцим jou Ajyensn a2uo2sojonugut {121104$ цопш Áppensn “appunpad uey 191104$ xipedg '4091 12017) пирштоја T ^u 006-06 ‘zeig pue AenZeseg ‘euog 01 глад :uwoJq чаррол 'exrprreq ‘guy злоду || Ачдфејео ‘рэлэцим 54 “eo [eotde цим eoueosejonujur :гјошпрод uey, 1e2uo| xipedg '2091 "ваше је рәләцим jou “әгі Ајејелорош ayyedg -qgc] HOYOS мајригј p 77—77 ш (0ер'2)000' 1 91 [әлә[ Bas *eureueq ‘e[anzaua A џлојзом *erquio[o-) 0195293100 0j илә}зәм їцвә1} поцм aoepms 1eddn uo роцојо Ариәшгшол4 jou ѕшәл Алецлај "9651 ]e01) nj021u24D ‘тел ame 29 seynuog ^y шпапатаоло "p ш 096 мојод teruozeury изәјѕәм *цѕәлу иэцм aoejms 1eddn uo payoja Ápueururod sutea Алеплеј ^e6G[ "води је рэлэцим Ayjensn * шщ; Алол ayiedg ego] ‘snoone[s Suureadde 10и ‘oped you Аепѕп xtpedg "сс̧Т 3o1) 2SUIÁDUDU ‘лвл "dup шпиројт "p ш 091-001 “әд ‘032107 “ends лациг ut o[qist^ $1әмор 9 UY 1эмэ4 “G) CT шпитэ]п ‘тел ‘Bug шпират y ш (сс мојод 'ешогешу илојвом tends yyə ut o[qtst^ зломој 9 иеці BOP в/с] "Адалјол 10 paymb jou saaea] ‘Buoj шо ст-с xipeds ‘Buoj шо gg-p] apunpag ‘49961 yeory праојупт ^p пад ‘03э10] 'uesnong оту шол Ао uwoux «Азолјал pue рәт ѕәлеә[ ‘Buoj шо ;^p-7'[ xipeds ‘Buoj шо 6-S'z e[ppunped '9961 'snoone[8 Зимеэ44е pue зјед xipedg ecc] *ejonad uey} 19840] x gs e[punpeq “q6€1 прит]4иод :dsqns Зиципа tpupjduoq y Еак ш QOS мојәд Apsour {геі иләцугои *e[onzeuaA илоцјпов 'eiquio[o-? uiejseauinos :мојод suonejound геприе цим Á[pensn "оца је рео o1 ondqpe Ápsour *enooo»euoo Адзош sepe|q уеэ] "qpc[ азиэАрпЯолю4 ‘тел ¡Bug огиоХоптолра сү ш 00/'1-0S1 ‘випчэ8лу uio -you *Kengeieq *erAmog ‘zeig изәціпоѕ < *suonejound зе[прие[8 yno -Чим *ajepoooue[qo-Zuo[qo 01 e1e[oeoue[qo *snoooeriooqns зореја јео “eFC ‘apis 19d ()[-p ѕшәл [e1e1e| Áreurud ‘9561 o1pog шпиртиогј “Y касе ш 00S'1-006 о ‘einqequiy шолу Ајио имоих [23591123 ‘Buoj шә OZ | uo Qc uey} лош o[punpaqd 'qec[ Croat 587 Anth Volume 78, Number 3 1991 urium t. Pachyne urium sec Аојешен готлајов y ш 009'1 01 [әлә[ eas 'suotdo1 zəm 01 pumy :епаелео oj е4шорю”) uieqjiou ‘snoone[3 ‘зэриэле| e[ed xipeds чиорџод Ajjensn oouoosoeiogut tÁ[rerxeqe pepunoz se[onad :әде[прип Ардеом Ао pue ojenuis [үе је зои зи леш əy} ‘әрим uey} 19800] xp uey} ssa] Á[ensn ѕәре[ '42/1 ‘uq UNUDÁD]I1DQ V 777 U 068 01 [9۸2] eos *suotgo1 Arp Ápsour тад pue Jopenoy [eiseos *snoone[ jou Чајом ysydind oi ystpper xrpeds ‘Sutpeasds-joasa A[ensn e2ueose1ogur :A[[eixeqe peqqu-p- 1 Á[[ensn sajonad :әје[прип Арэхлеш pue ayenuts Áj[ensn sui леш əy} ‘әрім uey, 193u0| xp uey, e1our Ацепзи заре °32 1 'ejejnono syAydeiey ‘етут "pedeus-xooq Jo ejejnono s[[Ayde1e) '20/1 "uiseq uozeury əy} ш Вишлтооо jou ‘zerg [vjseoo ulejseeqinos *e[anzeuaA pue etquio[o-) илацјлоџ ‘sapuy jo sedo[s usaysam jo ser»edg 4011 ug ei groom "exec cat PPS BIUQPUOY “ете “seros) 05501) ove ш [zeig :onÁudido Ајолел отиде 10 [2143591124 :A[[erxepe ојвојпе Ánugis од jeg sejoned :Surdoo1 pue [ejuozioy Аүепвп vag ‘9691 18017) asuanuay p 777007 ш 009 ‘open -94 ‘обем шолу A[uo umoux ‘onAydide : (үет -Xepe peqqu-[ Á|[erpeur sajonad “19319 WAG L691 ‘gjoned uey, лавиој xg uey sse[ e[punped ‘apis sed Q[-9 ѕшәл [елојеј Алеши '4891 NOIS xe ‘иЗиол шпирАодшор сү 7 و‎ ш 09/'@-066 ‘әд UJSYINOS 01 лорепол [ециәо ‹әјоцәа uey, зәЗио x / -£ ajounpad ‘apis 19d 05-01 $шәл [e1eje| Алеииа '2891 тлед pue 'iopeno ‘[izeig :Á[perxeqe pepunoi ѕәјоцәд '4/91 asuaKpngv4vd стел [Ju asusLonSoiod у ~ CT DR ш 002'*1-0S1 *eunue21y иләцулои *Kenge -ed ‘еглцоя ‘zeig u1eqinos :A[rerxeqe peqqu-e-[ ѕәјоцәд '2/91 ‘э8ивло Buap jou ѕшәл [еләде Алешпла :влоду osreoo Apsour “поп -otdsuoout Ayjensn Чим имола pue зэеуш-тшәѕ Sunsts1od sAydeyey :q991 111$әилә ‘тел Зи 1759419 ‘р ~ МЕЕ ш OOS Mo[eq ‘етиогешу изәјѕәм ш peardsopim ‘Buoj шо 05 uey; ssaj Ajfensn apounpad “19319 эизозэзоци] 9991 18019 np4p02]]20 стел “|Buy 11152и4э ‘р 77 5 7 lopeno, 'ezejseq шолу Аүио umouy ‘Buoj шә QZ чеч әзош әјоипрәа Зиэриэ4-Зитреэ14$ agusosalopu] ‘e99 [ 'eoepins ламој uo әйиело A[snonotdsuoo SurAip Хепѕп шәл [е19]8] К ud: 1 кеде Suns sp 4ydexey “eco 'suornejound Je[npue[3 зпоцим зорејд јео] '4291 притјдиод ‘dsqns 3unung приојдиод y ~~ cp MC S ш 005 мојед Арѕош “геі илэЧулои *e[onzaua A uleQjnos 'etquio[o) uiojseeuinos ‘Zuo, шш (07-е edus ‘USIMO][[9A Jo umoıq 2utKip Á[sour ‘aseq je әјепиәце 01 ƏMƏL sape[q јеә ‘от 8019 11$222470n2 *dsqns Sunung npupjduoQ cp 777 ш 088-61 ‘ејәпгәџәд ш AJ l JJ t Annals of the 588 Missouri Botanical Garden ‘горепэя eseo pue etquio[o?) uieqi1ou ш Surumooo seeds 4281 12027) 2521240] “Y Jopenoy pue еташојог) зизов(ре pue плод ‘019107 '(seuozeury) [zeagp ‘швед uozeury saddn *juepuod-Surpeauds oi Surpee1ds e»uooseiogur шюр шш 11-9 *ejoned uey, 1e2uo| х [-y әјочпрәд ‘4681 18017) ггигод тип 02 сү ejanzauaA ‘eaog pue seuozeury “uspuad Ароше аопаоволојш “wep шш p-g ‘әјоцәй uey, 192uo| x9-c гррипргд “*cgl ‘U (се ^o[eq teruozeury илэциои pue илајзам ш *Чѕәлу uaq« aoejms Jaddn uo Аззојвишог sope|q jeo| tiuapued-Surpeaids oj juepued aoueose1ogu] '4р81 gir] IINE ид ш ОФФ злоде Адзош :лорепол pue eiquiojo?) ш sapuy jo sedo[s изәјѕвә uo *цзәл} иэцм ooejans зәп uo одеш Ацепзи səpejq Jeo] :Surpee1ds-j2a19 eouaosarogu[ 'ер81 ‘sapuy jo sadojs иләдѕеә uo pue шѕед uozeury ayi ш Zutumooo $91946 '2281 'eje[[nono 10 padeys-yooy spudeie?) '2281 'ejoued uey, 1e3uo| x € uey, элюш әјәипрәд "q60T asnely `y 10702102 "p eureueq 01 (02045) etquio[o?) изәцізои ‘tends 1эциэ ш @[qISIA s19M0 01 ue зәмәу tA[[exeqe peqqu-g—¢ Apueururoud ѕәјоцәй ‘Buoj шо 9g иеці элош apounpag '4181 "Bug esuaqno y 777 LIWY [едиәг) 01 ‘eqn) *e[anzeuaA pue etquio[o)) u1oqj1ou ‘ends зәциә ш a|qIsIA SISMOY Q1 ueui a1our :[erxeqe рэрипол Ápsour зајоцод ‘Buoj шо Og uey, ssa] appunpag ‘2181 "xong uey, 1280] хс чеч sse[ хредс 4081 nouog (boe) umaaauisspa2 "p ur 008'1 91 [әлә[ eas (овде 'etquio[o^) илә}зәмцулоц 'ејаптопод uyou ‘отці uey, 19840] хс uey, лош хредс '2081 "A]exeqe родам-јелалов–1 10 18y А|елоцов *Kqpeixepe ѕш9:еш dieys цим ојеојп5 Ајмојјеца Jo pausyey Арзош ‘pədeys-q Ajjensn *pedeus- Jou зајопод ‘4621 1OYIS (ouo PA) 2742071708 ^F ^ ojueg olds pue олоџеј әр огу ‘91е195) SEUI ut zeig ‘Апетхеае рэрипол *Ajpeixepe suig1eur jun[q Áppensn чим ajeo[ns A[doop pue Ајмоллеџ *pedeus-) sajoneg “*6L1 "тела [e1seoo ‘eqn *oueury jenug oj еоџошу YINOG uyou мола Яшќір Апзош sape[q jea] ‘эр 1ad / [-с шәл [81918 Kreuniq '49;1 18017) npunjdsp Re peer M M CE EN QC DM ын ш (OF 01 [әлә eas :10penog “eyouryolg pue sory SO] “ѕер[еәшѕ ‘apm шо PE-OZ ‘Buoj шо 651-06 sepe|q Je] ‘uoj шә сј-) sənəd '48/1 12017) штлогалра“ сү ш 065-015 :10peno ‘Ixedojor) pue sory вот ‘apm шә 05-01 ‘Buoj WO 06-9ф зореја Jea ‘Buoj шо ge-c] вә[опцәд '28,1 телф sad apqusta зломој 9 uey} олош Áppensn ‘padeqs- 10 -q ѕә[опәд ‘4221 -Bug шпграри сү = ш 001-005 “ends sed ajqista зломој 9-ф :лерааџеш Á[peo1q sənəd ^?7/T "Jopeno ш sapuy jo sado¡s uzajsam ‘19918 Апоицяф ЗшАлр sepe[q јез ‘apts Jad Q£-9] suta [елајеј Алешма '29/1 'sisaujue je рэлэцим jou ayiedg ‘фу [ -Pug пиошиог сү = ur 006 mojaq Арзош [zzezg ш edeury ‘seuems *oSeqo], pue рершиј, :8и0] шо ()g- [| 'snoonpeo you 'snoooeuooqns э41е45 "qc? | 310496 мајригј ү ш (0£t'c)000'I 0} јалој VƏS te[onzaua A и1э1зэм 'PIquio[o)) UJ9ISBIYILOU O] UJ9ISIM 'Pureueq ‘Buoj шә [[-G'G ‘зпоэпрео ‘шц ayiedg ‘ec [ '(sueurtoeds рәр uo 1$вә[ је) ѕ:ѕәціив је рэлэЧим Ajyensn ayiedg ‘еру | ‘sens syAyderey ‘QOL TI 12017) npunjdsp ү ш OOF 01 1919] Bas :лорепол [esLos :эюцэ4 uey; 12840] х9-$ appunpad :ој8ив „0G-OF 1e qupru Зиціедәр ѕшәл [елоје] Атеш "qe; | 801°) asuaj]paponpo ‘р ш QG7'I -006 *eiquiojo”) enua? ‘ә[оцәй uey, ловиој x Z-| appunpad ‘ajue ¿06-02 1e quprur Zun1edop ѕшәл |елојеј Алешма ‘LELI ‘padeys-yooy sp£qdeie) “q121 589 Anthurium sect. Pachyneurium Croat Volume 78, Number 3 1991 WOR) Spider 7ш (OF оз [әлә] eos глорепод [e1se09 19918 0] u«oJq USIMO[[SÁ pue спозовџооапв #шАър sapeqq јео] “e00z ‘sisoyjue уе ощ: Ue} 192uo| x()gz uey, олош xipedg "9661 340496 xe "uduoig WNUDÁSQUIOP “Y ои ш 09/'c -0S6 məq uieqinos oj лорепод :smeqjue је xong uey, ло8иој x (02 uey sso| жредс ‘E66 [ ‘sgjoned uey, 193uo| x 9 uey; sse Ánsour эрипрэ{ '4961 1201) HUDMOD “Y о ш Q7 $ ‘eueAns) ‘nead Majarey ‘ор ur g'[ uey} олош ‘8ио| шо 6 uey; олош oyieds ‘uo c'z-c'[ ојендуе xipedg 4861 току мною er qaqe ш 067-00€ ‘quo ‘embonuy :opi шә c'[ uey, ssa] “Suo шо 6 uet ssa] ayieds 9559$ xipedg '2961 "етђоциу uıaqı1ou 10 ѕриецӣ!Ң euer) :әѕед үт эзепизце 0 әјпов 'ојелодо о} a1e[ooour[qo ѕәреү ‘9/61 18017) пзршоц ‘тел әлп®ерү 29 зоцпцос “y unaandandoap ју 7 ш (GL ‘Uog UJ9jseo ш eoeqoueng op PIUBIIAS taseq ye asniqo Á[peoJq оу рарипол *ondipa-o1e4o səpejg '2/61 ‘sajoned uey; 193uo| x 9 uey; әлош ә[зипрә{ ‘96 | Bug mane p тата Jopenoy [215202 Ap topim шә ; uey} ssa] 'леоиц-Виојдо səpejq јео ‘ефт ‘sisoyjue je ралоцим jou “ysidind ој uoo13 aqiedg '4261 ‘Buy пиошиәГ ју 777777777 ш 005 Uey, ssep (edeury) jzezg 01 pepu, :siseqiue је мојјех Эшит pue рајоцим aqiedg '2261 'suonejound 1e[npue[a snonoidsuoo јпоцим sope[q }®әТ “9681 е5 ш 00S'1-00? "вјопхопо A u1i9ujnos pue шешлте гајелодо 01 әје[оәәие[ао Á][ensn зорвја je? “9161 приојаиод :dsqns Buyung n"pupjduoq A 41843] Tay) $4 uey, олош ou роцјеоце ‘Buoj шо (1 uey} злош se[orjad pim шо cq uey, олош ѕәре '4261 Чим saoepmns әре :Хемртш ој payieays ‘Buoj шо gg иц ssa] sajonad :әрім шо QT uey; 5591 səpejg ‘Z6 [ ‘Ul QOS ^o[eq Арзош ‘zeig ureqjiou 'e[onzouaA urəymos *etquio[o) uiojseaqjnos :ondmpe A[peoiq oi onde A[[ensn sopejq уеэ] '2161 ‘uoj шш (5-1 edus їчєїмо[әА 10 имола Зи р Apsour ‘aseq 18 esniqo 10 зупов oj әјепиәце sope[q jeeT “9061 180.1.) 11$D22470n2 ‘dsqns Sunung прирјаиод p — эш 088-61 'вјепзопод ш лелцод pue seuozeury "eiquro[o) uiejseaqijnos :Suo| шш ()p-Gq adus ‘ystuees3 #шАтр Арзош ‘aseq је 9je»unijqns oj рарипол sapeq jes] '2061 ‘(Ayjerxeqe 1593] 18) suonejound repnpuej8 snonordsuoo yum Аүепвп sop?[q јез] '2681 "Jurpueose вшәл |елојеј Атеш чим ‘paymb jou ворејд јео] "1881 8017) wuma42auixajf24 сү Еше Е ш 008-059 ‘п1э4 *oonueng :ојеполе-овлодал Á[snonotdsuoo ѕшәл [e1eje| K1eurud Ҷим 'роутђ Ајепопогдеиоо зэре јео] '2881 'eje[oeoue[ ‘348$ spíudeje?) :q281 шпирАоолрд p 7777777777777 ш 068 01 [9aa] ees :лорепод [ejseoo Алр :surd1eur ојејприп Аү#иод$ Аүепвп цим зорејд zeal ‘АЦегхее peqqu-p- [ Ајјепеп зојоцод “Surpea.rds-199.8 Аујепеп допаовалоји :гпоопеј8 jou ajdind-jajora оз ustippa1 xipedg ‘4181 КәјѕшәҢ әртитароѕ y 770078 ш 009'1 01 [әлә Bas 'suorga1 Jam о ргшпц :оотхәу{ oi erquiojo?) илоцјлоп *suid1eur oje[npun Apysys Чим sope[q jeo| :Á[[erxeqe рэрипол sajonad зиэриэ4 Ápensn аопаоволојш ‘snoone[3 ‘әриәлеј e[ed xipedg ег gT 'eje[nono spÁudeie?) ‘4981 POOR) принос oerte то ш (Ot 01 [әлә[ ees :eie[oooue| *padeus-xooq sp&udeie?) '2981 Missouri Botanical Garden Annals of the 590 sajonad təjdınd 10 pas peur jou ‘әрім шо 81 Uey, олош sape[q јео :xipeds uey, 1о8иој ayieds '4212 180.7) 11114201u21Q ‘dsqns попрвуј umqaodns сү ш 00$-005 'n1ed *seuozeury “воле edaua”) огу шо} Á[uo UMOUY *peo1q uey} 192uo| x p Uey} a1our зорвја уеэ] "qp 1с unq4adns пн uosipe]q wnq4adns "p ш 005 ‘:орепэд ‘одем шод Á[uo UMOUY :peo1q uey} ловиој хр uet ssa] sape|q jeeT ‘ес "Suo| шә ()| uey} 5521 sajoned •мојед ajdind ој pas pagun ‘apim шо 81 иеці 5591 sape|q Jeo] :xipeds uey; 1ојлоц5 ayiedg ‘вета "Чем xipeds ‘yestds зәҷиә ш эл SISMOY ZI чеҷі 310 "?Z IZ '(ssa| цопш Ápsour) 32141 uey, 1280] x 0g uey, sse[ xipeds ‘sajoned uey, 19840] xp UY) әлош Арѕош sape[q јеә ‘4115 шпитәрт стел [dup шпирәт y АЕ xong uey, 1280] x 02 uey, әзош xipeds ‘sajonad uey, 19840] xp uey, ssa] Апзош ѕәреј Jeo] “BIZ ‘ayqey pue auy jou АЦепеп “пошел 51оду |АЧФетео ‘54 pearde ur рәзәцим jou Áppensn аопаовојопауш te[punped uey, 1911045 xrpeds {013 12015) пироштоја сү ш 006-09 ‘22:9 pue Kengereq *eragog 01 плод мола чаррал *axr[req ‘ouy 5лоду || Ачдејео гролоцим $4 ‘во pearde Чим әоиәоѕәјәплуш toppunpad uey, 1280] xipeds '2012 брез» uey} 19840] xz uey, ssa] әјоипрәд '?60c "еоџошу пос ш peaudsopta ‘рии ay] 01 ә$ло.цә1 10 лејпогрпод од jou *Surpueose [үе ѕшәл [e1eje[ Алеши 4504 ‘Bug 17155110] y ш Q08'T-006 плод enU *3uo] шо (ус uey, олош ѕәре[ јео] ‘4905 19017) urnun0210102 стел ‘Зи osua&nngvand "p ~~ ш 006-002 ‘uyog ‘zeg e :Зио шо ср uey, ssa] зэре4 јео] ‘F903 "ОСТ 01 25лодол 01 qupiur оу тепотриэ4лэ4 әре jeo| Jo aseq рлемој ѕшәл [елоје] Áreuruq “ecOz ‘qsHIJM 01 48119918 xipedg ‘9801 asuaKnnSv1vod ‘ел BUJ гзигАоптролра ‘у n ш 0021-0851 teuguedg1y илэЧзлои pue ‘Kengereg ‘uyog '[zzeg илоцупоз :Зио шо 9 uey, элош ayieds :eie[oeous[qo-Zuo[qo 01 eie[|oooue[qo sepe|q уеэ] “qvOz 12015) umpunjoaispq “Y пад ‘unıe ueg :Suo| шо 9 uey; ssa] ayieds :әцац|ә-әдело sope[q Jeo] ‘FOG "Aqerxeqe рәдд -£- T Ápsour ‘Ацехере зш8леш ojno? ҷим ajeo[ns Ајмојјец5 10 peueijeg Арзош sənəd ‘4505 jeor) шпитиојАуора p ш 0/7-0c£ ‘әд [e1uoo “appunpad uo jue1inoop &jsnonoidsuoo jou *enoaoeuooqns 'pexogei Á[[ensn ayieds '4202 nogog ('опођ aA) WNIUDINOS "p ~ узела ‘ueg oiutdeg pue олаџеј эр ory ‘әјәипр -əd uo jualinoap Ajsnonaidsuoa *snoooeuoo 'Surpueose-Surpeeids Ayjensn ayiedg “eg gz (¡penceqe рәр [Ex ep? sut junjq qua aeons Аүдәәр рие Ајмоллеи Ápsourse[onaq “ezoz RB uey} 12840] x()p uey, ssa] xipeds *qsmu«o1q 01 usa13-mo[paÁ ‘иээ18 Адзош Sutíip sepe|q јеәт '4102 18017) umunisadnpa y ш OSZ-0€1 под зазоре “eiquio[o”) ш seuozeury pue задпед ‘yor иец: одиој xop uey, әлош xrpeds мола ust&e18 01 ЧА е18 Suiap зарејд уеэ] '2102 ‘apts 1од 01-$Ф Арзош suta [еләјеү Kreurud ‘9661 jeo1) UNUDIFUNADY p ш Q0g' 1-007 :лорепод pue eiquiojo?) ut sapuy jo sado[s илајзво ‘UMOIG x1ep pue сподовџоо #@шАър ѕәреүа уеэ] ‘4005 Croat 591 Volume 78, Number 3 1991 Anthurium sect. Pachyneurium ‘Bug 2гигјодолт y сане ш 002-09€ ‘әд “UNIR иес ‘вәјоцәй uey, 19800] xg uey, e1our e[punped “199.19 WAS “eczz 'peouq uey} 192uo| x С uey} ssa] зореја J9] '4222 esuaKpngpapd ‘тел Bug osua&nnSpavod y 7777 ш 0021-0851 ‘еициә8:у изәцізои pue ‘enog 'Kengereq '[zeig uieqinos ‘Ayerxeqe peqqu-e-[ ‘Аегхере qu [erpeur 1иәшшоз Чим ѕәјоцә '"qpzc in пит] Y эш Q67-0£[ под ‘ицтей иес ‘Ацехеде рорипол 'Aqjeixepe хэлиоо Ápgugms ој је sajonag 'epzc 'peouq uey} 192uo| x G ueuj әзош Ацвош ѕәрер јеә ‘вес 'аовјлпв ломој uo 91e3jound-1e[npue[ jou зорејд zeo '4127 увол”) 11$022470n2 'dsqns Buyung приојаиод y Zuo] шш с uey, олош edus цим xipeds {шо cz uey; элош Аепеп) uoj шо 91 uey, элош ‘pim шо c uet олош sope|q Jeo] 'qzzc SIAMPH ‘у 1242n8vui сү weung :Suo| шә с uet $53] adus чим xipeds :Suo[ шо 91 uet ssa] ‘әрім шо c UY) ssa] sope[q Jeo] "ezcc 'аовјлпв ломој uo ojejound-1e[npue|g sope[q јео] ‘LIZZ 'seuoz ој] 1эыр Ápsow ut опирао ло [euso119) A[[ensn ‘snoone[s Suueodde you jnq зјед ѕәшцәшоѕ хгредс -q) [Z 1e017) 2S5UIÁDUDU `Іел "du шпират Y ш 091-001 “әд ‘019107 ‘fends лоцио ш e[qrst4 зломој 9 Ue} 19494 407 UNUDIM хел Bug шпирот ‘р ш (сс мојод :eruozeury илојзам ‘ends лацио ш эл зламој 9 ULY} SOJA 'LOZZ "(довјлпв 1eddn əy} uo лалди 1563] 1? 10) sno[nururo1d jou ѕшәл Алецлој цим “UMOIQ USIPp31 jou 114 “UMOIQ 01 идола Ц е43 ој qsiuao12 SutK1p ‘(apm шо 9[- TT Aqjensn uay} andre J1) asua&punu стел шпиројт ју ut оца а Ајреола о опау је Ájere Jo ә[ррїш ay) әлоде [әм 1sepeo1q pue aie[oooue[qo Á[[ensn *snoooeuooqns sape|q јео] ‘4615 12015) огигрид “Y ш 009 ‘лорепэ ‘одем шолу Ао UMOUY ‘səsezins од uo зпопипиол ѕшәл Алецлој оцу Чим имо ysippal Зи лр *a[pprur ay} 18 ззэрвоз4 ‘aptm шо c'6 01 *ondqpo-3uo[qo K[Mo1reu ‘зпоэовоэ sope[|q Jeo] ‘в6 тс `эЗиело шір зои зшал [21918] K1eurud Чим *UMOIq Zup sape[q јез] мола *1oe1ur-rures Zunsts1od 5[[АЧЧетег) '4812 твдила ‘тел Ви 17s2u42 p ur QOS мојед Арзош *eruOZPUIV и19}59м ш peoidsapim :snoone[8 Zuueodde jou xipeds :ooejins Jamo] uo agueo K[snonotdsuoo Фиш лр Á[[ensn ѕшәл [21918] Kreurud ҷим ‘ysruaa13 Зи лр A[[ensn sope[q језј :s1equy eyed ‘suy jo уломјаџ зови ue se Surnsts1od ичије) '2812 "ешогхешу ривјмој ‹вәцо?2 ә} 159.10] 19M 01 ош ш onÁudido :ooejmns Jamo] uo э8иело Á[snonotdsuoo #шАър шәл [елојеј Атеш pue злоду o[ed ‘auy jo Хломјоџ зовзит ие se BuNsisiod s[Áudejeo uou; *snoone[8 3uneadde jou jl Jo *snoone[8 Suueodde Áj[ensn xipedg '2/12 "реда uey; 12840] (x С uey} эюш Kpsour) xg uey} әлош әјәипрәд ‘9605 jeor) xippndsuano "p 7777777777777 ш 000'1 ‘вә :лореполј 'ezeiseq 12919 xipeds 'Surpeouds e[punped ‘э8иезо Зи ар jou ѕшәл [21218] Kreurud мола *jejur-rures 944515194 spyder 4912 1)52u42 ‘тел [Bu 11959419 "p ш QOS м9[э4 teruozeury илајвом ш реэл4зэрим 0919 xipeds pue ajounped :SutKip uo мојед эЗиело A[snonordsuoo suta [eloje| Áreunid ‘$1944 э[е4 ‘auy jo uoje[oxs зови ue se Sunsisiad ПАЧ”) '2912 'snoone[8 Suueodde jou xipedg ‘qc {Zz 12017) шпирпирш ‘р ш 002 плод ‘зо әр APP “snoone¡3 2uueadde xipedg -ec{z "[sruao18 Ápsour xipeds ‘ends зәцџә ш эл зломој 21 uey, Jamey Азор] “AZ Ic 1201) хлррасо125128 сү речи ш 00$ чеці ssa] teiquiojo”) рие плод зизов(ре рие [ze1g ‘элоу ‘Buoj шо ()| uey, олош Annals of the 592 Botanical Garden issouri M “Zuo, шш 05-1 edus цаморад 10 цзшмола ZutKip Арзош ‘aseq je asniqo JO ƏMƏL ој әјепиәце зорејд јео] 'qgec 18017) 118D92430N9 *dsqns Sunung прирјаиод сү ш 088-61 :e[onzeueA Ulayinos ‘eIqUIO[OT) илојзвоцупов ‘Buo шш OF —СТ adns {5112918 шір Арзош ‘aseq ye ejeounujqns oj pepunoi Апзош səpejq јео] “egez ‘uoj ur 91 Ue} a1our *apt« шо G иеці әлош sape|q јео] '4/22 SIAMBH сү NIDU p 770700007 ureuung :guo| WO 91 ueui ssa] ‘pim шо c иеці SSA| sepe[q јео] 'e/ez 'eoejins ламој uo зјејоипда-лејпривј8 зорејд јео] ‘FOGG 'snoone[ Zutreadde jou xipedg ‘ес 12017) әѕиәХрири ‘тел Ви шпиротр ш 091-001 әд ‘012107 “¡ends 1эциэ ш эл ѕЈәмор 9 ueuj 194234 'qcez штиројт стел Ви шпитат Y ш OSS м9[э4 *eruozeury и1э]зэм ‘pends зәцџә ш эл SISMOY Q ueui e1oJN "ecec 'snoone[8 guueedde xipedg -epez 'exreu pue auy jou Á[pensn ‘ѕпоцшел 5лоду [[Audeieo {54 ‘eo [eotde ut ролоцим jou эоцэозазоплуи te[punped uey, 19110qs xipedg 4222 18017) nupumojd y ш 006-06 тела pue “Xfen3eleg ‘euog 01 плод “UMOIQ Ysippes *exr[req ‘ouy злоду [| Аучдејео ‘рэлэцим $4 ‘eo [eotde Чим eoueosejonagur tepounped uey, 1e2uo[ xIpedg ‘вес "шәјѕ4ѕ ogeureip pue шѕед uozeury əy} Jo загоодс ‘есес "атвјозопеј “1y3rens syAydeyey "qozc ‘Bug ununÁD]91D0q p uu EC IE ш 068 03 [әлә] eas плод pue 1openo; [e1seoo Алр :sutgieur e1e[npun Á[guouis Á[[ensn цим sope[q zea *Kqpeixeqe paqqu-p-1 Ájensn sajoned :итрвэ14$-12э1э Аепѕп aduedsasogut ‘snoone[3 10u *e[dund-1o[orA ој ystppas xipedg '4122 Ke|sureH 27071а]0$ y ш 009°Т 91 [әлә[ eas *suoiga1 19^ 0] рпшту “O9IXAJA о} ешојог) u1aujjou зи ле eje[npun Аца цим saperg уез|[ ‘АПегхеде рарипол sajoned ‘juapued Á[pensn зопаоволојрш :snoone[8 ‘sisoyjue је 1эриэле| зјед xipedg ‘егес "e1e[oooue| [үе 1? jou ‘peoiq Ајојелорош 'eie[[nono s[[AJde1e) ‘4067 18017) ггигурраропро ‘р ш 05/*1-006 'etquio[o7) :e1e[oooue| pue padeys-yooy sp&udeie?) '2022 'pexoou Jo eie[[nono sqíudeie?) 2657 ‘gjoned uey; 192uo| x gz epunpeq ‘4105 nouog (boef) uma42auisspa2 y ш (008'1 03 [әлә[ eos :oe5eun7) *erquio[o") и1э15эмЧзлои “вјопхопод uiaujiou ‘Buoj UO се -с1 xipeds :Зиоү uro c'z [-0:g 9uieds {Зи 1р uo поцеошивеш мој 1e ajqista saprydes aed зпоцим горејд јео '4822 ayonog № 4204 “Y шпириәиәбрт p ш 009 01 [919] Bas ‘oedeın) “epanzaua A uieuiiou ‘Buoj wo с-ф xipeds ‘Suoj wo 9-с ayieds :Zutkip uo џоцеошидеш мој је e[qrst4 зэрщ4ел әүей цим зореја уеэ] “egzz 'eseq тези jsapeoiq xipeds :зјејприп Аццәшшоз jou ѕш8:еш чим səpejq je] ‘4155 noqog гие ‘р ш 00$°Т оз јела eas 12218 иләјѕеәціпоѕ pue uiejsee :xede 1еэи jsepeouq изо xipeds ‘эзетприп Apueururo1d зшвлеш qim səpejq јео ‘ес "уцате 45 $ПАЧАе1ео ‘э8ие ,0/ -0g 1е quprur Зипледэр SUISA [81938] Алеш ‘4955 18017) ISUI]DADINDI ‘р = ш 0€L° 1-006 *?tquio[o?) enua ‘padeys-yooy зПАЧАево ‘a[3ue ¿06-02 1e диырпи Sunsedap ѕшәл [елојеј Алеши ‘955 "шава uozeury əy} ut 8ии11290 jou ‘zeig u1ejseeuinos ‘в|эптэцэл 'etquio[o-) uyou pue |ецизо jo ser edo ‘9805 -Bug wnupiupwpuy y ешориоу рие '*eieq ‘ѕт10‹) “о$$олсу OJE ш ее ‘ѕәјоцәй иеці 1eguo[ xg uey, 559] Á[sour e[ounped ‹#шЧфәәлә '[ejuozuoq Á[[ensn шојс ‘9655 Croat 593 Volume 78, Number 3 1991 Anthurium sect. Pachyneurium 110496 (-boef) шталәи1$$юлә у 5 ш 008'1 91 [ede] Bas ‘oeSeın) ‘ашоо’) UJ9ISIMYIIOU :в[эп2эиэд илэЧзлои ‘Buoj шо се-71 xipeds ¿3uo| uro с'21–8 эЧ3е4$ "Bu ip uo uon?ogruZeur мој 18 эл зэрщ4ел aped упоцчим sope[q уеэ] 'q/ pc eqonog № 4203 ^M шпиомгигбот сү —— ш 009 01 [әлә ees toe3em;) ‘e[anzeua A uyou ‘Buoj шо zT -p xipeds :Suo| шо 9-7 oqıeds :uikap uo uoneogrugeur мој 1€ ajqista saprydes o[ed цим sope[q Jeo] ‘в/с 'eseq 1eou jsepeoJq xipeds :ојејприп Ápueururo1d jou зшвлеш ҷим зореја уеэ] '49%2 110426 әиур ү ш QO I 01 [949] ees zeig uıaseaqınos рие ulojsea :xode Jesu jsopeojq uayjo xipeds :әјејприп Ápueururoid зш8леш цим ѕәреү јео '29%2 ‘snoone[s Suueadde 10 uəə13 qsm[q jou xrpeds :peo1q uey, ладиој x 9-9'[ ayiedg ‘сұ 18017) х1ррѕото]8 сү ш 001'2–005 шоо?) [ециәо :snoone[8 Suureadde pue ueo18 чүч xipeds греола uey, 198uo[ x()[ uey, әзош ayiedg ‘ес 'urojss ә#ешелр 10 uiseq uozeury əy} ш Zut1mooo jou pue '[zeig изәцѕвәціпоѕ ‘е[әпгәџәл иләцулои *еідшоүог) uyou pue [ецизо Jo зогоодс 'qzec ‘Bug әѕиәјоӣріт) y ш 007-09£g ‘әд *иплер{ иес ‘ѕш8:еш oje[npun ApYysys ҷим 'реола uey, 192uo| хе uey, Sse] зореја јео] "qppc asuadonoind ‘rea Bug 2suaKnngpavd сү эш 056-051 ‘еициәдіу иәцігои pue ‘AenSeieg ‘eog Peig uleyinos ‘sulgiew oje[npun Apuoururo1d Чим ‘peoig uey} 192uo| x С uey, әлош sope[q јео] ‘врс ‘apis Jad ОТ uey, Jamey ѕшәл [e19]e| Алеш ‘фс nouog xo 'uZuoig шпирАгдшор ‘р ш Q97'z-066 ‘плэ4 uieqinos oj лореполј :реола uey} 12800] xG uey, $39] sape[q Jeo, 211 uey, зәЗио xog uey: ssa] жредс -qepz 8017) 114/574 “Y ш QOg'[ :1openo 'nonin? әр ешеллос əy} шош Апо UMOUY ‘peo1q иеці 192uo| x С uey, әлош sope[q zea ‘X011 uey, зәЗиоү xog чеці элош xipedg ‘вефс ‘apis 19d ОТ uey, әлош ѕшәл |елојер Атеш ‘EZ; "мојед a3uelo шір jou ѕшәл үеләје Атеш ‘имо pue joejur-ruros Buysisied syAydeyey 4192 пуздила ‘тел Ви тузгила "p ш 006 мо[әч Арзош teruozeury илојзом ‘мојәд ogue1o A[snonotdsuoo SutK1p uajjo ѕшәл [елојеј Атеш ‘ssaqy e[ed ‘auy jo Яломзэи зовии ие se Зицалод syAydeyey ‘ер 'eoejins 1940[ uo 9jejound-1e[npue|2 10u sope|q јео] 'qogc jeor) (3unung) wunuvAon2 -dsqns 3unung прирјаиод y peup UIYA pastes [үе е jou ло Á[92189s sutaA ојејпоцол чим sooejans IPP] “SPIM шо Qq иеці олош ѕәреү ‘0 Buyung әѕиәјриа p 770007 pester Апизипио2 4 SutÁ1p “шәл ојејпоцај чим saoejins эре ‘Авмрии ој payieays зэюцэ4 ‘apm шо (01 uey, ѕѕә[ səpejg “eobz ‘U 00$‘[-00$Ф “epenzaua A uieuinos 'ureuung ‘ajue элош 10 „(уф 1€ qupr Zur 1edop ‘Зитризозе £[da21s jou suraa [еләдә] Алеш] “q6Ez Buyung шпапдиоујирх сү IRI NEUE ш 066 ‘эп2эиэл 'seuozeury ‘эдоАццелу 0119) шолу Á[uO имои “a¡3ue „ур иеці $89] зе qupru Вицледор “Burpuaose 414э91$ ѕшәл pesare Атеиил “e6£z 594 Annals of the Missouri Botanical Garden Anthuri tifolium EngL, Bot. Jahrb. Syst. base, 3-4 mm diam. near apex; flowers square to 25: 365. 1898. rhombic, 1.9-3.4 mm long, 1.7-2.7 mm wide, the a. Anthurium acutifolium var. acutifolium. TYPE: Costa Rica. Puntarenas: Rio Ha Buenos Aires (SE of San Isidro del General) elev. 250 m, Pittier 6539 (lectotype, B; iso- lectotypes, BR, CR). Figures 23, 27. Anthurium scopulicola Standley & L. O. Williams, Ceiba 3: 105. . TYPE: Costa Rica. Puntarenas: vic. Palmar Norte, near sea level, P. Allen 5788 (ho- lotype, EAP; isotype, F). Terrestrial or often epilithic; stem 1 ст diam.; roots descending, greenish, smooth to velutinous, elongate, blunt, coriaceous, lanceolate, 6-10 cm rounded to obtuse at apex with a subapical apic- ulum, dark green, drying dark tan, persisting semi- intact, then as linear fibers, eventually deciduous. Leaves erect to spreading, occasionally pendent; petioles (2)6–22 cm long, (3)6-9 mm diam., sharp- ly D-shaped to subquadrangular, flattened to broad- ly sulcate with the margins sharply raised adaxially, sharply 3-ribbed abaxially, surface sparsely pale- speckled; geniculum slightly thicker than petiole, 1.4-2.5 cm long; blades subcoriaceous, elliptic to oblanceolate to broadly oblanceolate, gradually acuminate at apex (the acumen apiculate), atten- uate to narrowly acute at base, (11)25-62 cm long, (3.5)5.5-27 cm wide, broadest at or above the middle, the margins weakly undulate; both sur- faces semiglossy to matte, medium green above, paler below, drying greenish to greenish brown, matte; midrib flat at base, becoming acutely raised toward the apex above, sharply 2-ribbed at base, becoming sharply raised toward the apex below, paler than surface; primary lateral veins 8-12 per side, departing midrib at 40—45? angle, broadly arcuate to straight, acutely raised above, convexly raised below; tertiary veins weakly raised to sunken above, raised and darker than surface below; re- ticulate veins weakly visible above, obscure below; collective vein arising from about the middle of the blade or near the apex, sunken above, raised below, equally as prominent as primary lateral veins, 3— 5 mm from margin. /nflorescences erect to erect- spreading; peduncle (17)24—53 cm long, 3-5 mm diam., 3-6 x as long as petiole, angular; spathe reflexed to reflexed-spreading, subcoriaceous, green, linear-lanceolate, 5-12 cm long, 0.6-1.1 cm wide, broadest just above the base, acuminate at apex (the acumen inrolled); spadix green to white to yellow, sometimes tinged with red-violet, sessile, long-tapered, rarely stipitate, to 2 cm, occasionally cylindroid, (4)7-16 cm long, 6-8 mm diam. near mm diam.; cataphylls sub- long, narrowly sides smoothly to weakly and jaggedly sigmoid; 5— 6 flowers visible in principal spiral, 5-7 in alternate spiral; lateral tepals (0.5)1 —1.7 mm wide, the inner margins straight, the outer margins 2—3-sided; stig- ma ellipsoid, 0.3-0.5 mm long; stamens held above tepals in a circle around the pistil, persisting; an- thers 0.4—0.5 mm long, 0.6-0.8 mm wide; thecae oblong-ovoid, divaricate. Infructescence 1 cm diam., bearing berries in the lower part of the spadix; berries greenish yellow, obovoid to globose, round- ed to somewhat mammilliform at apex, 3.8-4. mm long, 3.8-4 mm diam.; seeds 2 per berry, yellow-orange, oblong-ovoid, flattened dorsally, 2.9— 3.5 mm long, 1.8-2.3 mm diam., 1.5-2 mm thick. This species is known from Costa Rica and Pan- ama principally on the Pacific slope in tropical moist, premontane wet, and tropical wet forest life zones, mostly from sea level to 900 m (rarely to 1,500 m). The species has been collected in Pan- ama on the Burica Peninsula in Chiriqui Province. There are two varieties. The typical variety occurs throughout the range of the species, whereas the variety herrerae is restricted to drier parts of trop- ical moist forest. Anthurium acutifolium var. acutifolium is dis- tinguished by its terrestrial or epilithic habit, by its thin, green-drying, oblanceolate to broadly lanceo- late leaf blades that are attenuate at the base, by its slender, long-tapered, usually green to yellowish spadix, and by its greenish yellow berries. It is similar vegetatively to А. consobrinum, but does not have the early-emergent pistils charac- teristic of that species. See variety herrerae for comparison with that species. OSTA RICA. PUNTARENAS: cabecera del Bkis, Pittier 11129 (BR, CR, US); Boruca, Pittier 4656 (BR); Buenos Aires, Pittier 6539 (BR); Rio Carias- Buenos Aires, Pittier 3862 (BR); Buenos Aires, Ujarrás, 300 m, Ocampo 2818 (MO, CR); Las Cruces- Villa Neily, Fila de Cal, 500-600 m, Gómez 19633 (MO, CR); Mellizas, 1,500 m, 8%54'N, 82°46'W, Romero & Fallas 7 (CM, CR, MO); along road to microwave tower above Golfito, 90 m, Croat 67615 (MO) Palmar Norte, Croat 32962 (MO); Refugio de Fauna Silvestre, Peñas Blancas de Esperza, 10?7'50"N 84?40'25"W, 1,000-1,400 m, Herrera et al. 280 (B, CR, К, MO); W of Rincón de Osa, along abandoned “high road,” 250-540 m, 8°42’N, 83°31'W, Croat & Grayum 59853 (МО); Burica Peninsula, S of Puerto Armuelles, Quebrada Macho, 100-200 m, Croat 22127 (F, K, MO); Osa Peninsula at Sirena, Liesner 2921 (MO); Quebrada Palito, Croat 22614 (MO). SAN JOSÉ: Quebrada Micos, 8.5 km W of Ciudad Colón, 600 m, 9%55.5'N, 83°17.5'W, Grayum & Sleeper 6098 (B, CR, К, M, МО); 700 m, Burger et al. 11938 (F, MO); El General basin, Skutch 749 (МО); Pacaca, Pittier 4099 (BR); El Rodeo, Hun- Volume 78, Number 3 1991 Croat 595 Anthurium sect. Pachyneurium newell 16560 (СН); 800 m, 9°55' М, 84*16'W, Bar- S of bridge over river, E Hosen & nou 6210 (ОМО); Rio Uruca near Santa Ana, 900 m, Burger & Liesner т 72 (МО). dp CHIRIQUÍ: Puerto Armuelles, Wood- on & Schery 904 (MO); W of Puerto Armuelles, Croat 21938, PU ov) 22474 (MO, PMA, ЕЗА, US), Liesner 49 (MO, PMA). b. Anthurium acutifolium herrerae Croat, var. nov. TYPE: Costa Rica. Puntarenas: Esperanza, Macona, finca of José Herrera C. Miramar, 10?01'40"N, 84*36'33"W, 300- 400 m, Herrera 372 (holotype, MO 3587626; isotypes, CR, K, US). Figure 351. var. Planta terrestris: internodia brevia; cataphyllum per- sistens semi-intactum; petiolus 7-9 cm longus, ca. 3 mm sulcatus; lamina oblongo- -elliptica aut elliptica, 21- 33. 5. ст longa, 7.8-13.5 cm lata, i sessilis, luteus, 2.5- 5 cm longus. Based on dried material only. Terrestrial; stem short; internodes short, 1-1.5 cm diam.; roots moderately dense, whitish, woolly-pubescent, to 10 cm long, 5 mm diam.; cataphylls narrowly long- acuminate, 5-6.5 cm long, drying brownish, per- sisting semi-intact, eventually fibrous at least ba- sally. Leaves + erect; petioles 7-9 cm long, ca. 3 mm фат., + C-shaped, sulcate adaxially, round- ed abaxially; geniculum slightly paler than petiole, less than 1 cm long; blades subcoriaceous, dr ing moderately thin, oblong-elliptic to elliptic, gradually long-acuminate at apex, attenuate at base, 21– 33.5 cm long, 7.8-13.5 cm wide, broadest at middle; both surfaces semiglossy; midrib drying sharply acute above, several-ribbed and paler than surface below; primary lateral veins (9)10-14 per side, departing midrib at 50—55° angle (sometimes to 75° angle nearer the base), moderately straight to the margin, then arcuate toward apex; tertiary veins obscurely visible, some weakly raised below; collected vein usually arising from about the middle of the blade, less commonly from the lower third, flat above, weakly raised below, 3-10 mm from margin. Inflorescences erect; peduncle 11.5-28.5 cm long; spathe spreading to reflexed-spreading, green, linear-lanceolate, 3.8-5.8 cm long, 8-10 mm wide, broadest near base, inserted at 45-60? angle on peduncle, narrowly acuminate at apex, narrowly acute at base and weakly decurrent; spa- dix yellow, sessile, moderately tapered, 2.5-5 с long, 3-5 mm diam. near base, 2-3 mm diam. near apex; flowers 4-lobed, 2.3-2.5 mm long, 2.2- 2.7 mm wide, the sides jaggedly sigmoid, 5-6 flowers visible in principal spiral, 9-10 flowers vis- ible in alternate spiral; tepals semiglossy, lateral tepals 1.3-1.4 mm wide, + inequilaterally shield- shaped, the inner margins broadly rounded, the outer margins 3—4-sided; pistils not emergent but pushing up against the tepals and inclining their margins upward; stigma 0.4 mm long, 0.3 mm wide, brushlike, with medial separation, weakly ex- serted; stamens emerging in a regular sequence, the laterals preceding the alternates by ca. 5 spi- rals, the 3rd stamen preceding the 4th by 3-4 spirals, held tightly aggregated over and obscuring pistil; anthers 0.4—0.5 mm long, 0.7-0.8 mm wide; thecae ovoid, moderately divaricate. Infructes- cence not seen. Anthurium acutifolium var. herrerae is endem- ic to Costa Rica, known only from the type col- lection in a dry area of tropical moist forest at 300-400 m It differs from the typical variety by its leaf blades, which are more elliptic and dry yellow- green rather than green. In addition, the dried epidermis of the blades is smooth with a distinct areolate cellular pattern on the upper surface at higher magnifications. In contrast, the typical va- riety dries dark green and has the surface rough and granular on high magnifications, with no sign of an areolate pattern. The variety is named in honor of Gerardo Herre- ra, an outstanding Costa Rican collector working on the Costa Rican Flora Project with B. Hammel and M. Grayum of the Missouri Botanical Garden. CosTA нна PUNTARENAS: Esperanza, Macona, finca of José Herrera C. Miriamar, 300-400 m, 10°01’ '40"N, 8436/33^W. ри 372 (CR, К, MO, US). Anthurium acutissimum Engl., Bot. Jahrb. Syst. 25: 398. 1898. TYPE: Ecuador. San Nicolas, Sodiro s.n. (holotype, B). Figures 24, 25. Anthurium acutissimum var. maius Sodiro, Anales Univ. ntr. Ecuador 22, по. 162: 267. 1908. TYPE: Ecuador. Sodiro s.n. (n.v.). Epiphytic, epilithic or terrestrial; stem short, 3-4(10) cm diam.; roots forming a dense mass, de rH drying whitish gray, somewhat pubes- cent, elongate, 2-4 mm diam.; cataphylls linear- lanceolate, (7)1 2-17 cm long, linear-elongate, dry- ing brown to reddish brown, persisting semi-intact or as coarse linear fibers or as a reticulum of fibers Leaves spreading to pendent; petioles (7.5)20-50 cm long, 5-12 mm diam., terete to subterete, somewhat flattened adaxially, rounded abaxially, 596 Annals of the Missouri Botanical Garden surface pale-speckled; geniculum flattened, slightly darker and thicker than petiole when dried, 1—2 cm long; blades usually pendent, coriaceous, oblong to oblong-elliptic, acuminate at apex, acute to nar- rowly acute at base, 60-118 cm long, 7.5-14 cm wide, broadest at or near the middle, the margins flat; upper surface matte, semivelvety, lower sur- face glossy, both surfaces dark to medium green, drying green to yellowish green, matte; midrib con- vexly raised, paler than surface above, prominently convex, much paler than surface below; primary lateral veins numerous, more than 25 per side, departing midrib at 50-70? angle, straight to the collective vein, flat and pale above, obscure below, slightly raised above when dried, more so below; interprimary veins moderately numerous, drying almost as conspicuous as primary lateral veins, raised above and below; tertiary veins obscure; collective vein arising from near the base, weakly raised above and below when dried, equally as prominent as primary lateral veins, 4-17 mm from margin. /nflorescences spreading to pendent; pe- duncle 22-64 cm long, (2)3-4 mm diam., equally or to 1.7 X as long as petiole, terete; spathe spread- ing, subcoriaceous, green to green-tinged with pur- ple at margins, oblong-lanceolate, (5.5)8-21 cm long, 1.1-2.4 cm wide, broadest near the base, acuminate at apex (the acumen inrolled, to 25 mm long), acute at base; stipe 7-20 mm long in front, 0-4 mm long in back; spadix green to yellow- green, subsessile, erect, straight to curved, held at 160-180? angle from peduncle, (6.5)10.5-21.5 cm long, (3)6-8 mm diam. near base, (2)4-5 mm diam. near apex, broadest at the base; flowers square to rhombic, 2.2-2.8 mm in both directions, the sides smoothly sigmoid, (3)4-5 flowers visible in principal spiral, 6-9 in alternate spiral; tepals matte, pale-punctate when dried; lateral tepals 1.3— 2 mm wide, the inner margins convex, becoming turned up against the pistil, the outer margins 2- sided; pistils emergent, raised, same color as Mm stigma slitlike, becoming ellipsoid, 0.5-0. long; stamens emerging from the base, held well above the tepals, persisting, laterals emerging al- most to apex before the alternates emerge in basal third, held in a circle around the pistil; filaments white, flattened, exserted, 0.7-1 mm long, 0.6– 0.7 mm wide; anthers drying yellowish to brown, 0.5-0.8 mm long, 0.4-0.7 mm wide; thecae ob- long, not divaricate. Infructescence о spathe persisting, green or withered; spadix ca. 23-30 iam., with berries ا‎ throughout; berries yellow-orange becoming or- ange, ovoid to broadly ellipsoid, **hollow" when rehydrated, acute at apex, (6)8.6–9.5 mm long, cm long, cm 4—4.5 mm diam.; pericarp thickened, with some raphide cells; seeds 2 per berry, yellowish, Е ovoid, flattened, 3.7-4.8 mm long, 2.3-2.4 т diam., 1.3-1.6 mm thick, with a transparent, ge- inoue appendage at both ends. A member of series Multinervia, Anthurium acutissimum is known from the provinces of Pi- chincha and Los Rios in Ecuador where it occurs at 650-1,200 m in premontane moist and pre- montane wet forest life zones. This species is characterized by its very elongate, mostly pendent, coriaceous, green-drying leaves, which are narrowly acute at the base and long- acuminate at the apex, by its large cataphylls which persist semi-intact as a reticulum of pale fibers, by its long-pedunculate inflorescence which has a ta- pered, green spadix with anthers exserted on long, whitish filaments and by its ovoid to broadly ellip- soid orange berries. Anthurium acutissimum is most closely allied to А. holmnielsenii, which has generally more erect- spreading leaves, D-shaped petioles, which are sul- cate adaxially with prominently raised, sharp mar- gins and broader blades (averaging 5.1 vs. 8.6X longer than broad). Leaves of both species dry dull green and matte. The label data of Madison 3826 mention that the plant was “ant inhabited" and that the stem was “massive, 10 cm thick," perhaps owing to its being myrmecophilous; the manner in which it was inhabited by ants is not mentioned, and root mass probably accounts for this exceptional measure- ment. ADOR. LOS RÍOS: 12 km E of Patricia Pilar, 650 Road, Alluriquin—Chiriboga, 950-1,100 m (originally 2094) (MO, SEL), Madison 4075 (MO, SEL); 2-3 km from main Aloag-Sto. Domingo de los Colorados rd., 890- 1,010 m, 0?18'13"S, 78°54.5'W, Croat idu (MO, USy; 12 km E of Patricia Pilar, 650 m 3828 (SEL); Rio Pilatón, Sodiro 20 (B); Quito- “Santo Ingo Domingo de los Colorados, 44 km E of Santo D ; 1, ae m, 0°23'S, 78°50'W, Hammel & Wilder 16081 (B, МО); Rio Pilatón-Río Toachi ence, 800- con m, Madison 4030 (K, QCA, SEL), Sparre 18453 (S). Anthurium affine Schott, Oesterr. Bot. Wo- chenbl. 5: 82. 1855. TYPE: illustrated by Schott Aroideae 465 (lectotype). Figures 26, 28, 29. Terrestrial; stem short, 1.5-4 cm diam.; roots moderately numerous, descending to spreading, greenish, smooth to weakly pubescent, thick, mod- erately elongate, 3-8 mm diam.; cataphylls sub- coriaceous, broadly lanceolate, 6-20 cm long, acu- minate to narrowly rounded and prominently Volume 78, Number 3 1991 roat 597 Anthurium sect. Pachyneurium apiculate at apex, light green, drying brown, per- sisting semi-intact, eventually dilacerating into coarse linear fibers. Leaves erect to spreading; petioles 5-28 cm long, 7-16 mm diam., U-shaped to quadrangular, shallowly to prominently sulcate, sometimes with a medial rib adaxially, prominently and sharply raised marginally, usually (1)3-5- ribbed, or sometimes rounded abaxially; geniculum thicker and slightly paler than petiole, becoming fissured transversely with age, 1–2.5 cm long; blades moderately coriaceous, obovate to oblanceolate to X elliptic, obtuse to short-acuminate or rounded at apex (the acumen apiculate), acute to obtuse to rounded or shallowly cordate at base, 33-95 cm long, 10.5-32 cm wide, broadest at or above the middle, the margins markedly undulate; upper sur- face matte to semiglossy, dark green, lower surface semiglossy to glossy, paler, both surfaces drying olive-green to yellow-green; midrib flat and 1-ribbed at base, becoming obtusely angled toward the apex above, slightly paler than surface, prominently thicker than broad and 2-3-ribbed at base below, becoming convexly raised toward the apex, paler than surface; primary lateral veins 7-14 per side, departing midrib at 40—60° angle, slightly arcuate to the margin, prominently and convexly raised above and below; interprimary veins absent; ter tiary veins obscure to weakly sunken above, slightly raised and darker than surface below; reticulate veins not visible to prominulous when dried, col- lective vein arising from near the apex or absent, less prominent than primary lateral veins when present. /nflorescences erect, somewhat shorter than the longest leaves; peduncle 31—78 ст long, 4-15 mm diam., 2-9X as long as petiole, green or green-tinged with purple at apex, subterete; spathe reflexed to recurled, moderately thick, pur- ple to green tinged with purple, ovate to broadly lanceolate, (3.5)5-10 cm long, 1.7- cm wide, acuminate at apex, acute and usually decurrent to 1–3(6) cm at base; stipe 0.5-5 cm long, 6-18 mm diam., pale green to purple; spadix olive-green to yellowish, cylindroid to clavate, subsessile to long-stipitate, erect, straight, 4.5-17 cm long, 5- mm diam. near base, 6-10 mm diam. near apex, broadest near the middle or near the apex; flowers + square, 1.6–2.2 mm in both directions, the sides straight to sigmoid; 7-14 flowers visible in principal spiral, 4-10 in alternate spiral; tepals slightly roughened, pale-punctulate (when dried); lateral tepals 0.7-1 mm wide, the outer margins 2-sided, the inner margins convex to rounded; sta- mens emerging in a regular sequence from the base, semi-erect in a tight cluster above the pistil, arching over and obscuring it; anthers 0.5-0.7 mm long, 0.7-0.8 mm wide; thecae somewhat divaricate; pollen creamy. Infructescence erect; spadix ca. m diam. (not including exserted berries); berries red to rich purple-maroon or pur- ple-maroon at apex becoming red below, whitish toward the base, obovoid to oblong-ellipsoid, trun- cate at apex, 7-13 mm long, 4-7 mm diam.; seeds oblanceolate, 5-8 mm long, 2-3 mm diam. Anthurium affine is found in eastern and south- eastern Brazil in the states of Mato Grosso, Minas Gerais, Goiás, Bahia, and Pernambuco from sea level to 1,300 m on rocky slopes, along streams, and in open areas. Heringer 7801 notes the veg- etation type as “‘cerrado mata ciliar." This species can be recognized by its large, co- riaceous leaf blades, frequently with strongly un- dulate margins, and its relatively short spadix, which is sometimes broadest near the apex. Because of its fruit color, spadix shape, and geographical isolation, А. affine is not easily con- fused with any other species. The species was placed into synonymy with Ап- thurium solitarium by Engler, but that species has an elongate, tapered spadix and an elongate spathe, and is primarily restricted to southern coastal Bra- zil. Engler's drawings of A. affine are, in fact, A. solitarium. Confusion arising from Engler's treatment of A. affine as a synonym of A. solitarium has permeated subsequent publications and label data in Brazil, Europe, and the United States, where it has been cultivated to a limited extent for many years. To this day, any member of sect. Pachyneurium from eastern Brazil is usually called Anthurium solitar- ium in botanical gardens and on herbarium labels, and annotations on the latter reflect the confusion between the two epithets. Anthurium macedoanum, an unpublished name attributed to A. D. Hawkes, was based on an un- usually broad-leaved form of A. affine (Macedo 2185) from San Vincente in the state of Minas Gerais. Confusion has also resulted from the Field Mu- seum photograph (FM 11838) of Glaziou 9040 from Rio de Janeiro. This photograph is actually A. solitarium, not A. affine as annotated by Engler, nor a type of А. affine as suggested by the pho- tograph. BRAZIL. ВАНА: Blanchet 1055 (С); Hwy BA:052, Cha- pada da Diamantina, E of Morro do Chapeü, 900 m, Davidse et al. 11902 (K, MO); N of йар, Salvador, 0 m, 12%58'/5, 38237, Plowman & Almeida 10046 (F, K, MO); Mato Grosso-Serra do Rio Contas, N of Vila do Rio do Contas, 980 m, 13°33'5, 41°49'W, Harley et al. 19973 (К, МО); 3 km М of Milagres, BR-116, Hatsch- 598 Annals of th Missouri Ed Garden bach 45069 (К); Morro do Chapeú, 900 m, Hatschbach 42428 (К); Mucugé, 2 km along Andarai road, 850 m, 12°59’S, 41°21'W, Harley et al. 20615 (К); Hatsch- bach 47955 (K); Pasto Guanabara, Faz Morro de Pedra- Itaberaba, Ferreira 191 (К); Rio do Contas- Livramento, cachoeira near road, Storr 194 (K); Rio Cumbuca, N of Mucugé on Andarai Road, 850 m, Harley 16000 (K, 05); Rio Ferro Duido, 19.5 km SE of Morro do Chapeú on ВА 052 road to Mundo Novo, 900 m, 11*38'S, 41902", Harley et al. 19254 (K, MO), (К); Ко Itapicuru, Jacobina, 450 m, Martinelli 5143 (RB); ndarai-Mucugé, ВА Mucugé, Pirani et onfim- > = Estiva, 850 m, Harley 16589 (K); Serra das Almas, NW of Rio de Contas, 1,000-1,200 m, Mori & Benton 13536 (NY) Serra do Jatobá, Morro do Couro or Morro Sào Cristóvão, 500-600 m, 12%54'5, 39°52'W, Harley 19427 (K), 19429 (K, MG, MO). Serra do Rio do Contas, Rio do Contas- Mato Grosso, 12-14 km N of Rio do Contas, 1,200 m, 13928'5, 41%50'W, Harley 15196 (К); Serra do Sincorá, 5 km S of Andarai, road to Mucugé near bridge over Rio Paraguacu, 400 m, 12%50'S, 41?19'W, Harley et al. 18594 (K, MO); Serra dos Lençois, Seabra- Itabera, 7-10 km, 1 gesinho, 12%28'S, 41?26'W, Harley 22695 (К); Mpo. inm ere Jaraguara-Milagras, Rodovia BR116, Silva al. 6 (K 2А b Lençois, NW of Lengois, Barro Shae ud 12°32'$, 41?20'W, Lewis et al. 920 (K, МО); {сены Кодома BR242, 720-760 (К); Mpo. Livramento do ach & Kummrow 49742 (MBM, МО); road to Diamantina, SW of Меп- danha and Rio Jequiti, 1,150 m, Anderson 8841 (F, NY, US); NE of Diamantina, on road to Mendanha, 1,300 m, Irwin et al. 22630 (US); Rio Abaite, Rodovia Brasilia, Heringer 7801 (MG); Rio Jequiti, E of Diamantina, 790 m, Irwin et al. 27430 br е NY, SEL, US); Мро. Dia- mantina, Biribiri, Hats 52 RB) vic. Joao Pessoa, os da Costa, (us K). PERNAMBUCO: Tapera, Pickel 2373 (US); eue "d o Gua 35 km Recife, 10-30 m, Tsugaru e 7 (MO, 00M); io jen Horto PM p» e Falcao et al. 775 (MO, R Anthurium angustilaminatum Engl., Bot. Jahrb. Syst. 25: 411. 1898. a. Anthurium angustilaminatum subsp. an- gustilaminatum. TYPE: Ecuador. Pichincha: Gualea, Sodiro s.n. (holotype, B; isotypes, P, Q). Figure 31. Anthurium angustilaminatum var. albidum Sodiro, Апа les Univ. Centr. Ecuador 22(156): 21. 1906. TYPE: Ecu Mie Imbabura: Guallupe, Sodiro s.n. (n.v.). qu пи minatum var. brevipes Sodiro, Апа v. Centr. Ecuador 22(156): 21. 1906. TYPE: between Paramba and Gual- lupe, Anthurium angustilami atum var. crassum Sodiro, Апа- les Univ. Centr. Ecuador 22(156): 20. 1906. TYPE: a. Pichincha: Nanegal, d s.n. (n.v.). Anthurium angustilaminatum var. gladiatum Sodiro, Anales Univ. Centr. Ecuador 28156): 20. 1906. TYPE: Ecuador. Imbabura: between Coajara and Pa- ramba, Sodiro s.n. (n.v.). Based on dried material only. Terrestrial; stem short, 1.8-2.5 cm diam.; roots moderately merous, grayish brown, puberulent, slender, elon- gate, 2-3 mm diam.; cataphylls 9.5-20 cm long, acuminate to short-acuminate at apex, light brown, persisting semi-intact as fine linear fibers; petioles -15 cm long, 5-8 mm diam., D-shaped to sub- triangular, broadly rounded to sulcate adaxially, the margins thin, prominently raised, appearing winged, rounded and acutely to obtusely 1-ribbed abaxially; geniculum slightly thicker than petiole, (0.5)1.5-2 cm long; blades coriaceous, narrowly to broadly oblong-elliptic, acute, sometimes shortly acuminate at apex, acute to narrowly acute (rarely rounded) at base, 40–94 cm long, 6–12 cm wide, broadest at or near the middle, the margins nar- rowly undulate; upper surface green, slightly paler below, both surfaces matte to semiglossy, yellowish brown, sometimes greenish; midrib convexly raised above, prominently raised and l-ribbed at base, becoming narrowly acutely raised toward the apex below; primary lateral veins darker than surface, 25-35 per side, departing midrib at 30-55? angle, straight to weakly arcuate to the collective vein, prominently raised above; interprimary veins nu- nu- merous, almost as conspicuous as primary lateral veins; tertiary veins prominulous; collective vein arising from the base, raised above and below, equally as prominent as primary lateral veins, 5— 10 mm from margin. + кайын apparently spreading; peduncle 15-55 cm long, mm diam., 2-3.1x as long as petioles, green, sharply angular, rarely terete; spathe erect to reflexed, subcoriaceous, yellowish green, lanceolate to ovate- lanceolate, 5-9 cm long, 1-1.3 ст wide, acumi- nate at apex (the acumen inrolled), narrowly acute at base; spadix dark purple, cylindroid, subsessile, 6-12 cm long, 5-12 mm diam.; flowers rhombic, 1.9-2.4 mm long, 1.5-2.1 mm wide, 4-6 flowers visible in principal spiral, 9-12 in alternate spiral; tepals minutely papillate, lateral tepals 1.4-2 mm wide, the inner margins straight, sometimes broadly convex, occasionally turned up against the pistil, Volume 78, Number 3 1991 599 Anthurium sect. Pachyneurium the outer margins 2-sided; pistils slightly raised; stigma slitlike, 0.5-0.7 mm long; filaments prom- inently exserted, holding anthers above tepals, the exserted part 0.5-0.7 mm long; anthers yellow, .4-0.5 mm long, 0.5 mm wide; thecae oblong- ovoid, not divaricate. Infructescence with spathe persisting; spadix са. 15 cm long, 1 cm diam.; berries drying 2.5-4 mm diam., probably orange. А member of series Multinervia, Anthurium angustilaminatum is comprised of two subspecies, the typical one known only from the slopes north of Volcán Pichincha in Pichincha Province in northern Ecuador, and subsp. cibuserpentis, from Cotopaxi and Е Oro further to the south. Subspe- cies angustilaminatum occurs in a premontane wet forest life zone at ca. 1,200 m The species is distinguished by its coriaceous, oblong-elliptic leaf blades with numerous primary lateral and interprimary veins, and its dark purple spadix. The typical subspecies appears to be most closely allied to А. soukupii, which ranges from southern Ecuador to Bolivia, and is very similar morpholog- ically. Anthurium angustilaminatum subsp. an- gustilaminatum occurs on the Pacific slope of Ecuador at ca. 1,200 to 1,600 m in a premontane wet or lower montane moist forest life zones, where- as А. soukupii occurs on the other side of the Andes at over 2,000 m in various montane forest life zones. Neither species has been studied in cul- tivation, and for the present, it is felt that the two taxa should be recognized as distinct based on phytogeography, and that study of further collec- tions can be expected to reveal more definitive morphological features for both species. Both sub- species of А. angustilaminatum differ from А. soukupii in having short, compact (rather than elongate) stems. The habit is epiphytic in at least ubsp. cibuserpentis, rather than terrestrial as for . ѕоикири А new subspecies, subsp. cibuserpentis, is de- scribed in the present paper and is distinguished from the typical subspecies mainly by its smaller overall size and abruptly obtuse to truncate leaf bases. It occurs further to the south in Cotopaxi and El Oro provinces, at higher elevations in drier life zones. Like the typical subspecies, it is known only from the Pacific slope of the Andes in Ecuador. Three of the four varieties described by Sodiro (1906) were collected in the same area in Imbabura Province, not far from the type locality of var. crassum and also the typical subspecies. All col- lections are from the same life zone. Although the type specimens of all four varieties described by Sodiro have not been seen and probably no longer exist (except for a photograph of the type of var. crassum), the differences given by Sodiro are slight, and no key is given to distinguish between the varieties. In view of the fact that Sodiro's species concept was often quite liberal, and that the type of var. crassum can scarcely be distinguished from material of the typical variety, it seems best to reduce all four varieties to synonymy. ECUADOR. PICHINCHA: are ca. 1,200 m, Sodiro 18 (B, P); Sodiro s.n. (G, MO, SI, US); Parroquia p 78*30'W, 1,600 m, Ceron et al. 4763 (MO, ОСМЕ). | b. Anthurium angustilaminatum subsp. ci- buserpentis Croat, subsp. nov. TYPE: Ec- uador. Cotopaxi: Pilalo, Km 80 on Quevedo- Latacunga road, 2,500 m, Dodson 15397 (holotype, MO 3247505). Figure 32. Differt a subsp. typicum petiolo breviori; lamina о longa, basi rotundata vel truncata, nervis primariis latera- lis numerosis. Based on dried material only. Epiphytic; stem short, ca. 2 cm diam.; roots dense, numerous, ascending to spreading, whitish, velutinous, short, tapered, ca. 3 mm diam.; cataphylls subcoriaceous, to at least 8 cm long, yellowish tan, persisting as fine linear fibers. Leaves erect; petioles 3.5-7 cm long, ca. 3-6 mm diam., D-shaped, flattened with the margins raised adaxially, rounded abaxially; geniculum 0.3-1 cm long; blades subcoriaceous, oblong-lanceolate to oblong-elliptic, acute at apex, truncate to narrowly rounded at base, 32-61 cm long, 5-9 cm wide, broadest at or near the middle, the margins slightly undulate; both surfaces matte to semiglossy and yellowish green; midrib convexly raised, becoming acute toward the apex above, higher than broad at base, becoming convexly raised toward apex below; primary lateral veins numer- ous, to more than 20 per side, departing midrib at 20-70? angle, + straight to the collective vein, raised above and below; interprimary veins nu- merous, almost as conspicuous as primary lateral veins; tertiary veins raised; collective vein arising from near the base, equally as prominent as pri- mary lateral veins, 3-8 mm from margin. Inflo- rescences erect; peduncle 17-23 cm long, ca. 3 mm diam., 3.9-4.2X as long as petioles, terete; spathe spreading, subcoriaceous, green, at least 7 cm long, 1-1.5 cm wide, acute at base; spadix (post-anthesis) green, slightly tapered, ca. 8 cm long; flowers apparently + square, 5-6 flowers visible in principal spiral, 6-9 flowers visible in alternate spiral, lateral tepals 2-2.5 mm wide, the 600 Annals of the Missouri Botanical Garden inner margins + straight, becoming turned up against pistil, the outer margins 2-sided; stigma ellipsoid, 0.4-0.5 mm long; anthers 0.5-0.6 mm long, 0.5-0.7 mm wide; thecae oblong, slightly divaricate. Infructescence erect; spathe persisting and withered or absent; spadix ca. 12 cm long, ca. 1.2 cm diam., with berries scattered throughout; berries greenish (immature), oblong, mammilliform at apex, 5.5-6 mm long, 3.2-3.6 mm diam.; peri- carp thickened, with raphide cells present; seeds 2 per berry, orange-brown when dried, ca. 3 mm long, 2.2-2.4 mm diam., ca. 1 mm thick. Anthurium angustilaminatum subsp. cibuser- pentis is known from the provinces of El Oro and Cotopaxi in Ecuador, where it occurs at 2,135 to 2,500 m in lower montane moist and lower mon- tane dry and/or montane moist (undeterminable from map) forest life zones. The subspecies is distinguished by its short pet- ioles and more or less oblong leaf blades with nu- merous primary lateral veins, which are rather abruptly and shortly rounded to truncate at the base. The color of the spadix at (or even near) anthesis is unknown. Anthurium angustilaminatum subsp. cibuser- pentis differs from the typical subspecies princi- pally by its abruptly ending leaf bases and by oc- curring in drier life zones at higher elevations (2,135-2,500 vs. 1,200-1,500 m) to the south, rather than north, of the equator. The new sub- species is smaller in overall size as well. See dis- cussion under subsp. angustilaminatum for distin- guishing both subspecies from 4. soukupii. he name is taken from ““cibus”” (Latin, meaning food) and “зегрепиз” (Latin, meaning snake) in reference to label data stating that the fruits of the plant are eaten by snakes. ECUADOR. COTOPAXI: Km 80 on road Quevedo-Lata- cunga, 2,500 m, Dodson 15397 (MO). EL око: Pampa de los Cedros, S of Cerro Chivo- Turco, 2,135-2,285 m, Steyermark 53764 (US, NY). Anthurium anorianum Croat, sp. nov. TYPE: Colombia. Antioquia: between Segovia and Canaverales, vic. Rio Bagre, 300 m, Croat 56757 е МО 3116672. 74). Figures 30, 33, Planta epiphytica; internodia brevia, 2-4 cm diam.; eolatum, p late sulcatus; lamina late oblanceolata, basi attenuata, 36- 100 cm longa, (12)20-60 cm lata; spatha lanceolata, 7.5-9(35)em longa, 1-1.3 cm lata; spadix purpureus, leniter contractus, 14-23 cm longus, 5-6 mm diam Epiphytic; stem short, 2-4 cm diam.; roots nu- merous, very dense, spreading, green to whitish, grayish brown on drying, moderately thick, 6-24 cm long, 3-5 mm diam. when fresh, drying 1-2 mm diam.; cataphylls sul 3 cm long, Soule at apex, green, drying brown (B & K yellow-red 4/10), persisting semi-intact. Leaves spreading; petioles 4—9 cm long, 9-17 mm diam., obtusely D-shaped, broadly convex adaxially with the margins shortly and sharply raised, round- ed to 3-5-ribbed abaxially, surface pale-speckled, reddish when young; geniculum thicker and slightly paler than petiole, 1-1.5 cm long, prominently and transversely fissured with age; blades subco- riaceous, broadly oblanceolate, acute to acuminate at apex (the acumen apiculate), attenuate at base, 36-100 cm long, (12)20-60 cm wide, broadest above the middle, the margins broadly and shal- lowly undulate; both surfaces glossy to semiglossy, medium green above, slightly paler below, drying brown to greenish brown; midrib flat to obtusely raised at base, becoming acutely raised toward the apex and slightly paler than surface above, obtusely raised below and slightly paler than surface, pale short-lineate; primary lateral veins 12-16 per side, departing midrib at 40—55? angle, straight or weak- ly ascending to the margin, sunken above, prom- inently convex and slightly paler than surface be- low; tertiary veins obscure above, prominulous and slightly darker than surface below, drying raised on both surfaces; collective vein arising in upper third or fourth of the blade, sunken above, raised below, less prominent than primary lateral veins, raised above and below on drying, 4-20 mm from margin. /nflorescences probably erect-spreading; peduncle 69-105 cm long, 5-15 mm diam. when dried, 6-9X as long as petioles, green, drying brown to pinkish brown, terete; spathe spreading, slightly recurled, subcoriaceous, green, tinged with purple near the apex, lanceolate, 7.5-9(35) cm long,1-1.3 ст wide, broadest near the base, acu- minate at apex (the acumen inrolled), acute to almost decurrent at base; stipe 1.5-1.8 mm long in front; spadix purple to dark purple, long-tapered, 14-23 cm long, 5-6 mm diam. near base, 3-4 ‚ lanceolate, 2— 9 flowers visible in principal spiral, 7–8 in alternate spiral; tepals smooth, pale-punctate, lateral tepals 2. m wide, the inner margins straight to broadly convex, the outer margins 2-sided; pistils green; stigma ellipsoid, 0.4-0.6 mm long; anthers 0.3-0.4 mm long, 0.4-0.5 mm wide; thecae ob- long, not divaricate. Infructescence with the spathe persisting withered; spadix (12)18-41(100) cm Volume 78, Number 3 1991 Croat 601 Anthurium sect. Pachyneurium long, 1.4-1.6 cm diam., bearing berries in the basal portion only; berries dark red, probably acute at apex, drying 4.8-5.5 mm long, 3-3.2 mm diam. Anthurium anorianum is endemic to Colombia in the Department of Antioquia at 300-750 m in tropical moist and tropical wet forest life zones. All collections to date have been made from the north- ern drainage of the Rio Canoa. This species is characterized by its short petioles, its subcoriaceous, broadly oblanceolate leaf blades which are attenuate at the base and acuminate at the apex and dry tan to brownish green, and by its slender, long-tapered, dark red to purplish-brown spadix. Especially distinctive is the usually pinkish coloration of the peduncle and primary lateral veins on drying. Anthurium anorianum is similar in appearance to A. vaupesianum, also from Colombia, with which it shares a slender, long-tapered spadix. Anthurium vaupesianum differs in being smaller in overall size, in having obovate leaf blades, and in occurring in the Amazon basin. The species is named after the Municipio Anori, where most collections were made. COLOMBIA. ANTIOQUIA: Rio Bagre, Segovia—Canaver- ales, 1 km NE of village of El Rio, 300 m, 7*10'N, 74*38'W, Croat 56757 (COL, К, МО, US); Segovia- Antioquia, village of El Rio, 750 1 Rentaris et al. 1576 (HUA, MO); Mpo. Anori, Buenos Aires Corregimento de Providencia, 500-700 m О); Mpo. Cáceres, El Doce Baja Cauca, 7°27'N, 75?20'W, Callejas 210 (HUA); Mpo. Valdivia, Corre- gimiento Puerto Mni 5 km from Puerto Valdivia at Rio Pescado, 410 m, 7°20'N, 75?20'W, Callejas et al. 3447 (HUA. MO, E Anthurium asplundii Croat, sp. nov. TYPE: Ec- uador. Los Rios: Rio Blanco, across river from ila Hermosa, 3 km S of Sto. Domingo- Esmeraldas hwy. at Km 24, 00905'5, 79?15'W, Croat 50688 (holotype, MO 3313109-10; isotypes, К, ОСА, US). Figures 8, 10, 35-37, 43 Planta epiphytica; internodia brevia, 3.5-5 cm diam.; cataphyllum lanceolatum; petiolus 7-15(25) cm longus, c mm diam., adaxiale complanatus cum costa me- a, abaxiale rotundus; lamina coriacea-subcoriacea, late oblanceolata-elliptica, acuta apud apicem, acuta usque late rotundata ad vadosa cordata apud basim, (60)90-125 26 utroque. Pedunculus (20)33-60 cm longus; spatha а viridis, aliquando purpurea tincta, (9)12-30 cm longa, 1-1.7 cm lata; spadix purpureus extendens ad шора. Жай contractus, 12-23 cm longus, 3-6 mm diametro. Baccae ignotae Usually epiphytic, sometimes terrestrial; stem to 60 cm long, 3.5-5 cm diam.; roots very dense, spreading-ascending, greenish to white, smooth, ca. 7-15 cm long, ca. 5 mm diam.; cataphylls sub- coriaceous, lanceolate, straight or hooked, to 14 cm long, acuminate at apex, drying whitish to reddish brown, persisting as linear fibers. Leaves erect-spreading; petioles 7-15(25) cm long, ca. 10 mm diam., sharply D-shaped, somewhat flattened with a medial rib and sharply erect or acute margins adaxially, rounded and sometimes weakly 1-2- ribbed abaxially; geniculum moderately thicker and slightly paler than petiole, 1-3 cm long; blades coriaceous to subcoriaceous, oblanceolate to broad- ly oblanceolate-elliptic, acute to abruptly and short- ly acuminate at apex, acute to broadly rounded or shallowly cordate at base, rarely narrowly acute, -125 cm long, 20-34 cm wide, broadest above the middle, the margins moderately undu- late; upper surface semiglossy, medium green (B & K yellow-green 2/7.5), lower surface weakly glossy, moderately paler; midrib convexly raised to prominently 1-ribbed at base, becoming acutely raised toward the apex above, moderately paler than surface, convexly raised and conspicuously paler than surface below; primary lateral veins (11)16-26 per side, departing midrib at 40-50(60Y angle, straight-ascending to very near the margin, then arcuate, frequently congested at base, con- vexly raised above, less conspicuously so below, paler than surface; interprimary veins usually pres- ent toward the apex and base of the blade; tertiary veins rather obscure above, less so below, drying conspicuously raised on both surfaces; reticulate veins prominulous on drying; collective vein arising from near the apex or absent, rather obscure on both surfaces, to 8 mm from margin. /nflores- cences erect; peduncle (20)33-60 cm long, 3-5 mm diam., 3-6 х as long as petioles, terete; spathe reflexed, subcoriaceous, green, sometimes tinged with purple, oblong-lanceolate, (9)12-30 cm long, 1-1.7 cm wide, broadest near the base, acuminate at apex, obtuse at base; spadix dark to dull purple to maroon, sessile or stipitate to 4 cm, cylindroid, slightly tapered, somewhat curved, held at 30—50° angle from peduncle, 12-23 cm long, 5-6 mm diam. near base, 3—4 mm diam. near apex, broad- est at the base; flowers rhombic, 1.2-2.5 mm long, 1.2-1.6 mm wide, the sides smoothly sigmoid; 6— 9 flowers visible in principal spiral, 4—5(8) in al- ternate spiral; tepals muricate; lateral tepals 0.7— m wide, the outer margins 2-sided, the inner margins nearly straight to rounded; pistils yellowish when dried; filaments (dried) exserted for 0.2-0.3 mm, 0.3-0.7 mm wide; anthers pinkish, 0.3-0.4 © 602 Annals of th Missouri а а Garden mm long, 0.4-0.5 mm wide; thecae ovate, slightly or prominently divaricate. Anthurium asplundii is endemic to coastal Ec- uador in Esmeraldas, Los Rios, and Pichincha prov- inces, where it occurs from near sea level to 400 m in tropical dry, tropical moist, and premontane wet forest life zones. This species is distinguished by its short petioles, its large, oblanceolate leaf blades which are fre- quently rounded to subcordate at the base and often dry green (sometimes yellowish brown), and by its purple to maroon, often stipitate spadix. Though sometimes terrestrial, this species is usu- ally epiphytic and in coastal Ecuador can only be confused with А. barclayanum. The latter occurs further to the south in mostly drier life zones, and the more narrowly oblanceolate leaves are held in an open rosette, whereas in 4. asplundii they may overlap in the lower portion and form a tight “‘bird’s nest" rosette. In addition, the blades of 4. bar- clayanum are usually strongly undulate, much more so than in the present species, and the cataphyll is cucullate (vs. lanceolate and straight or hooked in А. asplundii). The spadix color of A. asplundii is dark purple to maroon, and in А. barclayanum it is usually greenish. A species which is remarkably similar in overall appearance to А. asplundii is А. harlingianum, from the eastern foothills of the Andes, mostly in Ecuador. It differs mainly in its spadix color (dark salmon-pink to pale violet-purple vs. maroon to dark purple in А. asplundii) and in having leaves that dry mostly dark brown rather than greenish. The disjunct distribution suggests that the two are remotely related. An interesting collection is Dodson & Embree 13097, from Borbón in northern Esmeraldas, for which the spadix color is reported as “‘gray-white.”” In the dried state it is brown, not dark maroon as in other specimens. It otherwise agrees well with A. asplundii. Another somewhat aberrant plant is Harling 4465, from the Rio Blanco in Pichincha, with the geniculum 3 cm remote from the base of the leaf blade. The species is named in honor of Erick Asplund, the first Swedish botanist to collect for the Flora of Ecuador project and the first to collect А. as- plundii, in 1955. ECUADOR. ESMERALDAS: without locality, less than 300 m, Cobb 44 (MO); 20 km S of Esmeraldas, Hacienda = ayas, Sparre 15366 (S); Esmeraldas airport-San Ma 4.1 km N of bridge over Rio Esmeraldas, near sea E 0°54’N, 79*38'W, Сой 55586 (СМ, ОСА, МО, US, Му); и Sua, 6.4 km N of Rio Sua, 29.4 km SW of jct. with road to Esmeraldas, 40 m, 0°52’N, 79%52'W, Croat 55598 (MO, aes Santo Domingo de los Colorados-Esmeraldas, 62 km М of Quinindé, 80 km SE of Atacames, 270 m, 0935'М, 79*33'W, Croat 55580 (MO, QCA, NY, RSA); 8.8 km NW of Quinindé, 270 m, 0°26'N, 79°03’ W, ree 55554 (MO, QCA); 4.4 km SW of Goan coastal dip k j raldas, near vel, 0%51'N, 55608 (MO, QCA); a sea level, Christenaón 1521 ; Ri E of San ‚ Kvist & Asan 40886 (AAU); o Esmeraldas, opposite و‎ (Rosa Zarate), А5- plund 16346 (S); Rio Dodson & Embree 13097 (SE 300 m, 1%00'N, 7 Rio Blanco, Santo a de los Colorados—Esmeraldas, 3 km S of Km 24, across river from Villa Hermosa, 250 m, 0°05'$, 79°15’W, Croat 50685 (К, MO), 50688 (MO, К, QCA, US), 50693 (B, MO). PICHINCHA: 10 km NW of Santo Domingo de los Colorados, road to Esmeraldas, Rancho Brahman, 400 m, Sparre 15210 (S); below con- fluence with Rio Toachi, 200 m, Harling 4465 (S); con- fluence of Rio Napo and Rio Pilaton, Sodiro s.n. (Q). Anthurium atropurpureum R. Schultes & Ma- guire, Bot. Mus. Leafl. Harvard Univ. 16: 60. a. Anthurium atropurpureum var. atropur- pureum. TYPE: Colombia. Amazonas: Que- brada Agua Preta, Black & Schultes 46-375 (holotype, GH). Figures 38, 44. Terrestrial or rarely epiphytic; stem to 30 cm long, 1.5- mostly ера на 3-30 cm long, 2-4 mm Hide cataphylls su riaceous to coriaceous, 1-7 cm long, ен acute at apex, drying pale brown (B & K yellow 5/2.5), persisting + intact, sometimes as a retic- ulum of fibers with the apex remaining intact. Leaves t erect; petioles 2-25 cm long, 2-8 mm diam., bluntly D-shaped or terete to subterete, obtusely or narrowly sulcate, sometimes weakly flattened adaxially, rounded abaxially, surface sometimes pale-speckled; geniculum thicker and paler than petiole, concolorous when dried, becoming fissured transversely with age, occasionally to 2.5 cm re- mote from the base of the blade, 0.5-1.5 cm long; blades subcoriaceous to coriaceous, mostly elliptic, often narrowly so, acuminate at apex (the acumen flat), attenuate or acute at base, 11-75 cm long, 3)5-14 cm wide, broadest at or near the middle, the margins flat to undulate; upper surface semi- glossy to glossy, medium green, lower surface matte to ео sometimes paler, drying pale green; y raised to angular at base, becoming narrowly raised toward the apex above, broadly acute at base below, becoming conspicuously and acutely raised to narrowly raised toward apex, paler than surface when dried; primary lateral veins 5- lam.; roots “~ Volume 78, Number 3 1991 Croat 603 Anthurium sect. Pachyneurium 11 per side, departing midrib at (30)50—60? angle, mostly arcuate-ascending, rarely straight to the collective vein, raised above, prominently convexly raised below; interprimary veins rarely present, etched above, flat below; tertiary veins visible, usu- ally weakly etched above, prominulous and darker than surface below; collective vein arising from near the base or near the apex, sunken above, raised below, 5-10 mm from margin. /nflores- cences erect, shorter than or equaling leaves; pe- duncle (8)15-47.5 cm long, 2-5 mm diam., 1.3- 4X as long as petiole, terete; spathe spreading to reflexed, withering at anthesis, subcoriaceous, streaked with raphides when dried, green heavily tinged with purple at margins to purple-red throughout, lanceolate to linear-lanceolate, 2-15 cm long, cm wide, broadest near base, inserted at 90? angle on peduncle, acute to nar- rowly acute at apex, acute at base; stipe 3-15 mm long in front, 2-8(12) mm long in back; spadix usually deep purple (B & K purple 2/7.5), some- times maroon to brown, cylindroid to long-tapered, erect, 2-14 cm long, 3-8 mm diam. near base, 2-4 mm diam. near apex, broadest at the base; flowers rhombic or 4-lobed, 1.6-2.2 mm in both directions when fresh, 1.8-2.4 mm in both direc- tions when dried, the sides += straight or smoothly to jaggedly sigmoid, 4—8(10) flowers visible in ei- ther spiral; tepals matte, muricate, weakly to con- spicuously and minutely papillate, the papillae ma- roon (B & K red-purple 2/2.5); tepal surface olive- green (B & K yellow 6/10); lateral tepals 0.6– 1.5 mm wide, the inner margins straight to con- cave, the outer margins straight to weakly 4-sided; pistils emergent, raised, the exposed portion + square, 0.6-0.7 mm in both directions, dark purple to green tinged with purple; stigma ellipsoid to oblong, 0.5-0.6 mm long, 0.2-0.3 pillate, densely brushlike; stamens emerging rap- idly from the base, in a + scattered pattern, the laterals preceding the alternates by ca. 8 spirals, the 3rd stamen preceding the 4th by ca. 2 spirals, held against the pistil; anthers greenish yellow, 0.4–0.6 mm long, 0.5-0.8 mm wide; thecae el- lipsoid, slightly divaricate; pollen yellow-orange, creamy when dried. /nfructescence with spathe withered; spadix 4-13 cm long, the stipe 0-5 ст long; berries dark purple to deep red, globose to ellipsoid, 5-6 mm long, 4-6 mm diam.; pericarp dry, thickened with numerous raphide cells; seeds 2 or frequently 1 per berry, oblong-ellipsoid, 3.6— 4.4 mm long, 1.8-2.4 mm diam., 1.2-1.4 mm thick, with a gelatinous appendage. mm wide, pa- Anthurium atropurpureum ranges from south- ern Colombia to Brazil (Acre) and in western Bo- livia, at low elevations (100 to 160 m) on white sand soils. This species is distinguished by its leaf blades, with the tertiary veins conspicuously etched above (when fresh), by its adaxially sulcate petioles, and its purple, scarcely tapered spadix subtended by a purple-tinged spathe that withers at anthesis. Anthurium atropurpureum is most easily con- fused with Anthurium bonplandii. А general dis- tinction for dried material is the color of the leaf blades: yellowish/brownish and mostly concolorous for А. bonplandii (except subsp. cuatrecasii) vs. + green and often bicolorous for А. atropurpu- reum. See further discussion under А. bonplandii. Anthurium atropurpureum comprises three sympatric varieties. Anthurium atropurpureum var. arenicola is a forest epiphyte or an understory terrestrial herb and greatly exceeds the typical variety in size. Variety thomasii is endemic to the Serrania de Huanchaca in Bolivia (see that variety for differences). Anthurium atropurpureum var. atropurpureum ranges from southern Amazonas Dept. in Colombia to Maynas Province (Dept. Lo- reto) in Peru, as well as Acre and the eastern edge of Amazonas in Brazil, at low elevations in a tropical moist forest life zone. It is distinguished from var. arenicola by its much smaller leaves with the col- lective veins arising from near the base instead of from the upper 4 and to % of the blade. The varieties atropurpureum and arenicola share sim- Џаг soil preference and geographical ranges, al- though the former is more often terrestrial. BRAZIL. ACRE: Cruzeiro do Sul, near airport, Ramos 136 (INPA). AMAZONAS: Rio Tontantins at Villa Velha, m, Schultes 6905 (GH); Agua Preta, Black & Solus 46 -375 (GH); La Chorrea Ко Igara- Parana (affluen utumayo), Sastre 3232 MEXU, MO, TEX); е а ПАР, Vásquez & Criollo 5759 (M, МО); 1 an p Werf et al. 10212 (MO); Iquitos- sd pun 32. m, 4?10'S, 73°30'W, Vásquez & aran 7847 К. МО, 05); 130 m, Vasquez & Jaramillo 10558 (KRAM, a Iquitos, Carretera Quisto Cocha- Varillal, ca. 160 m, McDaniel et al. 27447 (IBE, MO); Quisto Cocha road, near airport, 120 m, Gentry & Daly 19118 (F, MO); 12-25 km S of Iquitos, Croat 18279, 18405, 18183, 18665 (MO); Carretera Quisto Cocha-Nauta, 160-170 m, McDaniel et al. 27605 (IBE, MO); Iquitos at Рећа Negra, Davidson 3659 (RSA); 100 m, Ayala 1821 (AMAZ); Iquitos-Santo Tomas, S of Iquitos, Croat 19104 (CAS, COL, F, MO, NY, RSA, SEL); Varillal, road to Nauta, 130 m, 3?45'S, 73°21'W, Vásquez & Jaramillo 1197 (К, MO), 1198 (МО, RSA); Roca Fuerte (Momón), 160 m, 3?40'S, 73°30'W, Vásquez & Jaramillo 5232 (MBM, MO); Саг- retera de Zungaro Cocha, near Shushana, 160 m, Ri 604 Annals of the Missouri Botanical Garden machi 6566 (IBE, МО); near Villa Angel, 140 m, Ri- machi 5888 (IBE); Iquitos Region, Iquitos, Mishuyacu, 120 m, Killip & €: 29887 (NY, US), Klug 316, yc 207 (F, . 120 m, Revilla 4295 (AMAZ); vic. Quista 20-140 m, Croat 51180 (GH, M, A, MO), Gentry 207 19 (MO), Gentry & Ayala 15510 ‚ Gentry & Jaramillo 22335 (MO), McDaniel 13604 E d e Amazonas, ы о no, Quebrada Үа- amo 0 m, 3°24'S, 72:49", ur 56927 (DUKE, MO, A 61761 (AMAZ, МО); Ко Nanay, Picuruyacu, 160 m, McDaniel & Rimachi 26371 (IBE); 30 km SW of Iquitos, Mishana, Estación Biológica Callicebus, 150 m, Gentry et al. 15782 (B, О, Sy Mishana, Rimachi 1239 (IBE); 30 km SW of Iquitos, Mishana, 120-150 m, Foster 4217 (MO), Gentry et al. 31671 (BM, К, MO), 36536 (MO, RSA), Solomon 3548 (MO b. Anthurium atropurpureum var. areni- cola Croat, var. nov. TYPE: Ecuador. Napo: road between Lago Agrio and Rio San Miguel, 3 km М of Lago Agrio, 0%5'N, 76°50'W, 450 m, Croat 50303 (holotype, MO 2828986; isotypes, AAU, B, BM, CAS, CM, COL, C, СН, IBE, INPA, К, NY, QCA, RSA, SEL, US, W). Figures 39, 40-42, 45. Differt a var. typica pro ratione amplioris et distinctis habitationibus suis; et nervo collective prope basim exo rient. Terrestrial or epiphytic; stem to 30 cm long, 1–3.5 cm diam.; leaf scars obscured by root mass and cataphylls, 1.5-2 cm wide; roots numerous, dense, mostly descending, uppermost ascending, gray or green, smooth to densely pubescent, elon- gate, blunt, 5-30 cm long, 3-6 mm diam.; cata- phylls moderately thin to subcoriaceous, bluntly 1- гфђед near the apex, 5-12 cm long, narrowly acute to acute at apex, tinged with red, drying tan (B & K yellow-red 7/2.5), persisting with apex remaining intact, eventually deciduous. Leaves erect-spreading, rosulate or scattered along stem; petioles 4-63 cm long, 3-12 mm diam., erect, slightly thicker than broad to D-shaped, shallowly sulcate adaxially, sometimes obtusely V-sulcate, the margins moderately sharp, rounded abaxially, the surface pale-speckled, tinged with red near the base; geniculum markedly thicker than petiole, slightly paler, 1-3 cm long; blades coriaceous to subcoriaceous, narrowly to broadly oblanceolate, less often elliptic, acute to abruptly acuminate, sometimes gradually acuminate at apex (the acu- men to 4 cm long, inrolled), long-attenuate or rarely acute at base, sometimes making geniculum seem remote, occasionally acute to obtuse, (16)45- 80(100) ст long, 8-25(41) cm wide, broadest at or above the middle, the margins undulate; upper surface weakly glossy to glossy, medium green, lower surface semiglossy to matte, only slightly paler and mostly eglandular, rarely weakly pus- tulate, both surfaces drying green and paler below, matte; midrib flat at base, becoming acute-raised to convexly raised toward the apex above, higher than broad to prominently acute-raised below; pri- mary lateral veins 5-13 per side, departing midrib at (30)40-65(70Y angle, usually weakly arcuate- ascending to the collective vein, sunken to convexly raised above, prominently and convexly raised be- low, paler than surface; interprimary veins etched above, weakly raised below, prominulous when dried; tertiary veins conspicuously etched above, flat to prominulous below; collective vein arising from about the middle or in the upper Уз of blade, sometimes arising in lower 4, sunken above, raised below, flat above when dried; antimarginal vein visible on live plant, running close to margin from the base; 151 collective vein 9-12 mm from margin. Inflorescences erect to spreading, shorter than or equaling leaves (rarely longer than leaves); pedun- cle (12)20-73 cm long, 2-8 mm diam., 1-4X as long as petiole, green to green heavily tinged with purple, terete; spathe spreading to reflexed and recurled or twisted, withering at anthesis, subcor- iaceous, matte on both surfaces, pale to medium green, heavily tinged with red-violet, sometimes maroon to purple, lanceolate to narrowly lanceo- late, 4-22 cm long, 0.6-2.5 cm wide, broadest near the base, inserted at 45? angle on peduncle, long-acuminate at apex, broadly acute, decurrent to 2 cm at base, the margins meeting at ca. 45? angle; spadix green (B & K yellow 5/7.5) heavily tinged with violet-purple, to dark purple or maroon, cylindroid to tapered, sessile or stipitate to 16 mm, erect, 4-26 cm 3-5 mm diam. near apex, broadest at the base; flowers square to rhombic, 1.5-2.6 mm long, l- 2.5 mm wide, the sides straight to sigmoid, 5-8 flowers visible in principal spiral, 5-10 in alternate spiral; tepals matte, weakly to densely and minutely papillate, muricate, lateral tepals (0.5)0.8-1 mm wide, the inner margins pale, thin, pinkish, straight to weakly convex, the outer margins 2-3-sided; long, 3-8 mm diam. near base, pistils weakly emergent before stamens emerge, + square, green, developing a maroon tinge, some- times somewhat papillate; stigma ellipsoid, 0.2-0.6 mm long, depressed medially, lightly or densely papillate, droplets appearing 7-11 days before an- thesis; stamens emerging in a regular sequence from the base, lateral stamens followed by alter- nates in rapid succession (7 days), the laterals preceding the alternates by 5—10 spirals, held over and obscuring pistil, 3rd barely preceding 4th; an- thers yellow, 0.2-0.6 mm long, 0.8 mm wide, held Volume 78, Number 3 1991 Croat 605 Anthurium sect. Pachyneurium at edge of the pistil, only partially exserted; thecae oblong-ellipsoid, 0.2-0.4 mm wide, not divaricate or only slightly divaricate; pollen yellow fading to cream, faintly sweet- or yeasty-scented. /nfruc- tescence spreading to pendent; spathe withered or absent; spadix 14-35 cm long, 1-2.5 cm diam., bearing berries in the basal portion only, the stipe to 1.5 cm long.; berries white to greenish white basally, violet-purple in apical V6 (B & К purple 2/7.5), subglobose, rounded at apex, 6-8 mm long, 4-6 mm diam.; pericarp somewhat thickened, with raphide cells; mesocarp with raphides, thickly gelatinous; seeds 1-2 per berry, pale green be- coming purple (B & K purple 3/2.5), oblong- ellipsoid, 3-6 mm long, 2-3 mm diam., 1.4-1.6 mm thick, with a basal mucilaginous appendage at radicle end. Anthurium atropurpureum var. arenicola ranges from southern Colombia (Putumayo and Amazonas) to western Brazil (Amazonas, Acre, and western Rondónia), Ecuador (Morona Santiago, Napo, and Pastaza), Peru (Amazonas, Loreto, and Madre de Dios), and Bolivia (La Paz), at 100-560 m. It is principally found in tropical moist forest life zones, in mature forest on white sand soils (hence the name); the variety is both terrestial and epi- phytic. Anthurium atropurpureum var. arenicola is distinguished vegetatively by its mostly oblanceo- late leaf blades, which dry greenish and often bi- colorous and are usually long-attenuate with straight margins toward the base and acuminate at the apex. The spadix may be violet-purple or maroon, or green tinged (usually heavily) with these colors, and the spathe is withered at anthesis. This variety is distinguished from the typical variety by its larger size and occurrence in the forest understory, commonly as an epiphyte, rather than as a terrestrial plant in open areas. It may be confused with the typical variety and var. tho- masii (see discussion under both varieties). It may also be confused with А. bonplandii subsp. bon- plandii; see discussion under that subspecies. Variety arenicola is also similar to А. fendleri, but differs in having more coriaceous blades when dried, mostly long-attenuate leaf bases and long- acuminate apices, whereas А. fendleri is rounded or short-acuminate at the apex and mostly rounded to shortly cordate at the base. BOLIVIA. LA PAZ: Prov. Iturralde. 2 Beni, pe 180 m, 13%05'S, 67?15'W, Haase 343 (MO). BRA ACRE: Сено do Sul, Estrada Alem id Б et e 11819 (INPA, NY); near airport, Montei Damião 315 (INPA), NW of Cruzeiro do Sul pee Barao do = Branco, 4 km NW of Sao Francisco, 150 m, 7?36'S, 2°37'W, Croat & Rosas 62694 (AAU, BM, CM, INPA, 1 NY); Mutis 48 (US); Cruzeiro do Sul- Rio Branco, 72°35'W, nad 62330 (INPA, MO, NY); vic. airport, 150 m, 7?38'S, 72*36'W, Croat & Rosas 62642 (CM, INPA, MO, NY). amazonas: Ca- sauari, NW of Rio Jurua, Silva et al. 392 (NY, 05); Rio Aripuana, Nova Prainha, SB20-ZD Point 5, Mota s.n. INPA) Rio Javari, Estiraó de Equador, Lleras et al. 17265 (INPA, ); Paumari, above Atalaia, Prance et al. 23788 (INPA); Rio Urubu, above Iracema, Prance et al. 4969 (INPA); Ed Humaitá, Humaitá- Porto Vel- ho, Km 60, 08%05'S, 63°W, Texixeira et al. 187 (INPA, NY); Rio Madeira, Tres Ceres, Krukoff 6549 (NY); Mpo. Novo Aripuana, 30, Rodovia Transamazonica, at km 300 from Humaitá, 4?33'S, 60*40'W, Ferreira 5711 (INPA, NY); Rio Madeira basin, Rio Livramento, Liv- ramento, Krukoff 677 1 (NY). RIO DE JANEIRO: (cultivated by T Burle- мо San Carlos, Croat 53698 (CM, MO, GUA, NY). RONDONIA: Porto Velho-Ariquemes, 4 km N of Arimuetes, Grea 53582 (MO); Serra dos Tres Irmáos, opposite Mutumparana, N bank of Rio po ince et al. 5535 (COL, INPA, NY, US). CoL tye a ~ 02?20'S, Shiki RBAE207 (NY). маро: Lago Agrio-Baeza, 43 km E of Lago Agrio, km 43, 560 m is М, 77915", Croat 50441 (МО); Lago Agrio-Coca, 6 m 5 of Lago Agrio ferry crossing over Rio Coca, 200 m (ed 12558 (MO); Lago Agrio- Rio San Miguel, 3 km N of Lago Agrio, 76*50'W, Croat 50303 (AAU, B, BM, US, W); 21. 2 km N of Lago Agrio, road near Rio San Miguel, 470 m, 0?08'N, 76°50’W, Croat 50358 (AAU, M, К, KYO, L, LE, MBM, MO, QCA, RSA, SEL, US); Puerto El Carmen de Putumayo—Lago Agrio, 76 km E of Lago Agrio, Tarapoa, 240 m, 0%07'N, 76°23’W, Croat 58630 (MO); Cantón Aguarico, Parque Nacional Yasuni, lagunas de Garza Cocha, 200 m, Cerón & Gallo 5006 (MO, QCNE); Río Aguarico-Coca, 8 km S of Rio Aguarico, E of Lago Agrio, Croat 58514 (CM, МО, ОСА); Rio Aguarico, 300 m, 091 76°21'W, Brandbyge et al. 33207 (AAU); У oi Putum mayo, Dureno, 500 m, Plowman et al. 4048 (GH); Santa Cecilia, 220 m, 0?4'S, ОСА); Lagunas de Cuyabeno, 300 m, 0%01'S, 76°11'W, Brandbyge et al. 33975, 36010 (MO, AAU); Reserva de Producción Faunistica Cuyabeno, 265 m, Poulsen 78992 (AAU) S of Laguna Garzacocha, 76911! W, 00°01'S, Balslev et al. 84689 (AAU); Laguna Canan- gueno, 76?11'W, 00%01'S, Balslev et al. Ce M MO); Rio Napo, 20 km downstream from Coc Taracoa, 250 m, Besse et al. 1962 (SEL); SW wir Bherto Francisco de Orellana, 300 m, 0°32'$, 77°0'W, Balslev & Madsen 10621 (AAU); Lago Agrio-Coca, 15 km М of Coca (Francisco de wap 350 m, Croat 50411 (IBE, MO, P, U); Rio Suno, 1 mi. above confluence with Río Napo, 290 m, Whitmore 860 (K, NY); Coca, Texaco bois strip, 300 m, Foster 3557 (F, MO); Limoncocha, ‚ Madison et al. 5330 (MO, SEL); Rio Yasuni, 606 Annals of the Missouri Botanical Garden 80 km upriver from Rocafuerte, 225 2 Foster 3735 (F, MO, y Lagunas Jatun-Cocha, 200 m, Cerón Gallo 5145 (MO, QCNE), Cantón F Agrio, Dureño, 350 m, 0%2'S, 76°42'W, Cerón 346 (MO); Reserva In- digena Cofan-Dureño, Cerón & Cerón 3152 (МО); Es- tación Biológica Jatún Sacha, Río Napo, 8 km E of Mi- sahualli, 01%4'S; 77°36'W, 450 m, Cerón 1729 (МО, CNE). PASTAZA: Laguna Garzayacu, 250 m, 1?29'S, 76°39'W, Neill & Palacios 6707 (MO); Rio Сов Lorocachi, 200 т, 1?38'S, 75°5 ] al Quebrada Caterpiza, 200 m, Huashikat 2216 (MO). LORETO: Corrientes, upriver from Tnte. Lopez oil camp, almost at Ecuador border, 280-300 m, Gentry et 75 (MO) Dtto. Pebas, Río Yaguasyacu, Bora native community of Brillo Nuevo, 150 km ENE of Iqui- 0-160 Ер Carretera Peña ra, 4k beyond Quisto Cocha, 50, McDaniel & Rimachi 25741 (IBE), Rimachi m (IBE, MO): Nina , 130 m, McD al 27489 (IBE); Iquitos Region, 1 7 km SW of Iquitos, Croat 18393, 18506 )), 18460 (МО, USM); Quisto Cocha, 120 m, 3?47'S, 73?24'W, Croat 51179 (С, K, MISSA MO, SEL), 51180 ve GH, M, MO, US), 51182 (MO), Rimachi 8086 (IBE) road to Picuruyacu, 160 m, McDaniel et al. 24770 (MO), Rimachi 5465 (IBE); Rio Itaya, Cahuide, Vasquez & Jaramillo 5644, 5686 (MO), below mouth of Rio Ucayali, 130 m, 4?20'S, 72°30'W, Gentry et al. 30041 (MO); San Antonio, 150 m, 4?10'S, 73°20'W, Vásquez & Jaramillo 3537 (MO); Puerto AI- mendras, 120 m, 3?48'S, 73*25'W, Croat 19048 (MO), Croat 51216 (BM, CM, MO, US, W); 12 km SW of jig te Croat 20093 (MO); 130 m, van der Werff et al. 9778 (MO); 122 m, 3°48'$, 73°25'W, Vásquez & Ja- ramillo 4595, 4949, 6099, 6849, 11485 (MO); 5 km Dod Vb 18560 (F, MO, NY, USMy 150 m, 4?10'S, 73?20'W, Vás- 4 Biologica Callicebus, 130 m, 3°55'S, 73°35'W, Vásquez et al. 643 (MO), Vásquez & Jaramillo 8934 (МО); Indiana-mouth of Río Napo, Explorama Tourist Camp, Yanamono, 130 m, 3?28'S, 72?48'W, Croat 61761 (AMAZ, MO), Gentry et al. 31422, 39695 (MO), quez et al. 5170, 5179 (MO); Rio Momón, Rimachi 5815 (IBE); Santa Maria de Nanay, Quebrada Yarina, 150 m, 3955'5, 73*40'W, Vásquez & Jaramillo 12221 (MO); Rio Nanay, та Мапа de Nanay, Maynas, 140 m, 3*52'S, 73°30'W, Gentry & Emmons 38783 (Е, MO), Gentry et А 39345 (MO). Iquitos-Sta. Maria de o, 110 pend 11689 (MO); Rio Yaguasyacu, tributary of Rio Ampiyacu, Brilla Nueva, Gentry & Revilla 20433 (МО); Prov. Ramón Castilla, Rio Ampiyacu, ep e Pucaur- quillo, upriver from Pebas, 140 m, 3?15'S, 71?56'W, Stein et al. 4006 (МО); Prov. сна пена ен 170 m, 4%50'S, 73°45'W, Vásquez et al. 2154, 11865 (MO) 140 m, van der Werff et al. 10015 (MO). MADRE DE DIOS: serrania across Rio Alto Madre de Dios, Pantia- colla, 480-840 m, Gentry et al. 27357 (MO). Anthurium atropurpureum var. thomasii Croat. var. nov. TYPE: Bolivia. Santa Cruz: Prov. Velasco, Serrania de Huanchaca, Par- que Nacional “Prof. Noel Kempff Mercado,” 13%59'S, 60?43'W, rocky hills Е of airport, ca. 750 m, Thomas et al. 5696 (holotype, MO 3583031; isotypes, K, LPB, NY, i- versity G. N. Moreno, Santa Cruz), Figure Differt a var. typico lamina ovato-elliptica, rotundata ad basim Terrestrial; stem short; internodes short, to 2 cm diam.; roots densely compacted. Leaves erect- spreading; petioles 7-20 cm long, to 6 mm diam. (dried), subterete, sulcate adaxially, rounded abax- ially; blades moderately coriaceous, ovate-elliptic, rounded to obtuse and apiculate at apex, rounded to broadly obtuse at base but sometimes ending abruptly on the petiole and weakly decurrent, 35- 50 cm long, 15-27 cm wide, broadest at or below middle; both surfaces drying matte to weakly glossy, pale yellowish green to gray-green, inconspicuously pustular but not glandular; midrib convexly raised on both sides; primary lateral veins 6-10 per side, departing midrib at 35-70? angle, at least those near the base of the blade arising at an acute angle with the midrib, arcuate-ascending to margin in a broad curve, most of them merging with the margin above the middle of the blade; interprimary veins lacking; tertiary veins prominulous on both sur- faces; collective vein usually lacking, sometimes arising from below the middle of the blade and loop-connecting the primary lateral veins, 12-20 mm from margin. /nflorescences + erect, equaling or much longer than leaves; peduncle 52-73 cm long; spathe spreading to spreading-reflexed, thinly coriaceous, soon drying, green, linear-lanceolate, 6.5-11 cm long, (0.9)1.2-1.6 cm wide, broadest above point of attachment, inserted at ca. 70? angle on peduncle, acuminate at apex, acute at base and sometimes decurrent on the peduncle to 1.3 cm; spadix brownish purple, sessile to subsessile, mod- erately tapered, erect, 6.5-9.5 cm long, (drying) 5-6 mm diam. near base, 4 mm diam. near apex, broadest at the base; flowers square to rhombic, 1.6-1.8 mm wide, the sides straight parallel to spiral, smoothly to jaggedly sigmoid per- pendicular to spiral, 9-11 fl 1.7-2 mm long, owers visible in prin- cipal spiral; tepals matte, minutely papillate, dense- ly covered with a frostlike grayish coating which Моште 78, Митбег 3 1991 Croat 607 Anthurium sect. Pachyneurium can be scraped off; lateral tepals 1-1.2 mm wide, the inner margins rounded, often turned up on drying, the outer margins 2-sided; pistils not at all emergent, obscured by anthers; stamens emerg- ing in a regular sequence from the base, lateral stamens emerging up to 10 spirals before alternate stamens emerge, the 3rd stamen preceding the 4th by ca. 5 spirals; anthers 0.6 mm long, 0.7 mm wide, contiguous, inclined over pistil; thecae ovoid, moderately divaricate. Infructescence to 14 cm long, 1.5 cm diam., bearing berries in the basal portion only; berries purple on upper half, white below, obovoid-ellipsoid, rounded at apex, 6.5-8 mm long, 4-5 mm diam.; seeds 1 per berry, dark brown on rehydration, slightly flattened, 4.5-5 mm long, 3.5 mm diam., lacking any obvious append- age. Anthurium atropurpureum var. thomasii is en- demic to the Serrania de Huanchaca near the east- ern border of Bolivia, known only from the type specimen collected in Santa Cruz Department in a region of tropical moist forest at ca. 750 m ele- vation. The type locality is an outlying upland area of the Brazilian shield with “cerrado” and “campo rupestre" vegetation, most closely related to that of adjacent Mato Grosso (W. Thomas, pers. comm.). The population of plants from which the type was collected was growing on steep, dry slopes in a relatively mesic area near a stream in both sunny areas and in partial shade (W. Thomas, pers. comm The taxon is distinguished by its narrowly ovate- elliptic blades which dry pale green and are rounded basally. Also distinctive is the long-pedunculate in- florescence with a lanceolate spathe which is three or more times broader than the purplish spadix. The typical variety as well as the var. arenicola are both distinguished from var. thomasii by typ- ically having attenuate to narrowly acute blade bases. e new taxon is named in honor of Wayt Thomas, New York Botanical Garden, who made the only known collection of this taxon and who supplied valuable data on the ecology. BOLIVIA. SANTA CRUZ: Velasco, Serrania de Huanchaca, Parque Nacional **Prof. Noel Kempff Mercado," 750 m, 13*59'S, 60%43"W, Thomas 5696 (K, LPB, MO, NY). Anthurium barclayanum Engl., Pflanzenr. IV 23B(Heft 21): 79. 1905. ТУРЕ: Ecuador. Manabi: Isla Salango, Barclay 646 (holotype, BM). Figures 14, 46-49, 59. Terrestrial or epiphytic; stem to 40 cm long, 3-5 cm diam.; roots dense at the nodes, spreading- ascending, green to greenish white, fuzzy, ca. 7 cm long, 4-8 mm diam.; cataphylls subcoriaceous, 3.5-4.5 cm long, cucullate, rounded at apex, dry- ing tan to brown, persisting + intact or as fine linear fibers. Leaves erect-spreading; petioles 5- 2] cm long, (5)7-15(20) mm diam., = D-shaped, flattened or convex with sharp, erect margins and an acute or obtuse medial rib adaxially, obtusely angled, usually 1-4-ribbed abaxially; geniculum slightly paler and thicker than petiole, 1-7 cm long; blades subcoriaceous, narrowly oblong-oblan- ceolate to oblanceolate-elliptic, obtuse to acute and apiculate at apex, narrowly acute, sometimes abruptly obtuse, less often broadly rounded at base, (40)70-137 cm long, (7)11-32(48) cm wide, broadest above the middle, the margins strongly to moderately undulate; upper surface matte to semiglossy, medium green, lower surface matte to semiglossy, scarcely paler; midrib obtusely raised or flattened with a blunt medial rib at base, be- coming acute toward the apex, same color as or paler than surface above, broadly and acutely raised at base, becoming convexly raised toward apex below; primary lateral veins 11-25 per side, de- parting midrib at 40—60° angle, straight- or ar- cuate-ascending to the margin, obtusely raised and nearly same color as surface above, convexly raised and paler than surface below; tertiary veins slightly raised, same color as surface above, slightly raised and darker than surface below; collective vein aris- ing from near the apex, prominulous on both sur- faces, ca. 5 mm from margin. /nflorescences spreading, sometimes erect; peduncle (20)40-80 cm long, (3)5-7 mm diam., 3.5-6(9)X as long as petioles, green to green heavily tinged with maroon, terete; spathe spreading to reflexed-recurled, co- riaceous to subcoriaceous, green (sometimes tinged with violet-purple) to dark violet-purple or maroon, narrowly ovate to oblong-lanceolate, (4)7-24 cm long, 0.6-4 cm wide, broadest near the base, in- serted at 30-70? angle on peduncle, acuminate at apex (the acumen inrolled), acute to subcordate at base; spadix usually brownish or olive-green to greenish yellow, sometimes reddish to violet-purple, sessile, cylindroid, slightly tapered, erect, usually somewhat curved, held at 140? angle from pedun- cle, (5)8-18 cm long, 5-10 mm diam. midway, 3.5-5 mm diam. near apex, broadest at the base; flowers 4-lobed, 1.2-2.5 mm long, 1.6-2.5 mm wide, the sides straight to sigmoid, 7-11 flowers visible in principal spiral, 6–7 in alternate spiral; tepals matte, weakly to moderately papillate; lateral tepals 0.8-1.4 mm wide, the inner margins straight to broadly rounded, the outer margins irregularly 3-4-sided; pistils emergent or not, reddish violet 608 Annals of the Missouri Botanical Garden to dark purple or green; stigma linear to elliptic, 0.6-1 mm long; stamens emerging in a regular or scattered pattern, the laterals preceding the alter- nates by ca. Jê length of the spadix, the 3rd stamen preceding the 4th by ca. 15 spirals, arching over and obscuring pistil; anthers purplish violet to pink- ish tan, 0.6-0.8 mm long, 0.8-0.9 thecae oblong-ellipsoid to ovoid, slightly divaricate; pollen pale yellow to creamy, fading to white, spicy- mm wide; scented. Infructescence spreading; spadix (6)14— 28 cm long, 2 cm diam., greenish when young; mature fruits unknown; berries reddish brown, ob- ovoid, 8-9 mm long, 4-5 mm diam.; seeds 5-6 mm long, 2.5-2.7 mm diam., 1.3-1.5 mm thick, with a gelatinous appendage at each end. Anthurium barclayanum occurs in coastal Ec- uador in El Oro, Manabi, and Guayas provinces, and in adjacent Tumbes Dept., Peru, from sea level to 890 m, in premontane moist, premontane dry, and very dry tropical forest life zones, as well as in tropical desert scrub. It is expected to be in tropical dry forest. This species is characterized by its large, usually strongly undulate leaf blades, the usually greenish spadix and by its occurrence in dry habitats. An- thurium asplundii, из closest ally, occurs in ad- jacent Esmeraldas and Pichincha provinces in less dry life zones. Another close ally is A. linguifolium, which is sympatric with А. barclayanum in at least one locality, where А. barclayanum is epiphytic and А. linguifolium terrestrial in open areas. See those species for distinguishing characters. n the southern part of its range, in El Oro Province, the spadix color of А. barclayanum is sometimes violet-purple, but these plants (e.g., Croat 50711, 50712; Thompson 362) otherwise agree well with other specimens. Àn intergradation with A. asplundii (with spadices maroon to purple) is conceivable, but the aberrant specimens of 4. bar- clayanum occur in the part of its range furthest from A. asplundii, well separated by typical ma- terial with greenish spadices. One collection (Пиз & Mohr 234) shows the odd character of having the tepals colored like the pistils, shriveled and very inconspicuous (compared to the pistils) in early fruit, whereas in all other specimens at apparently the same stage (or very near) the tepals are pale and conspicuous and not as shriveled. ECUADOR. EL ORO: Avenillas-Piñas, 03?40'S, 79%45- 55'W, Thompson 362 (CAS, CM, GH, MO, RSA, UCLA); Machala-Loja, 1 km SE of La Avareada, 18 km NW of Saracay, 80 m, 04?25'S, 79*55'W, Croat 50711 (C, МО, Р, QCA); 25 km SE of jet. to Pinas, 890 m, 04?15'8, 79945'W, Croat 50712 (BM, M, MISSA, MO, OOM, nada, Haring. et al. 207 01°36'S, 80°52'W, Barclay 646 (BM); picis Puerto López, 5 km S of Punta Cayo, 150 m, 01?25'S, 80?42' W, M, is Hda. Napo (Hotel de la Playa), CAS, CM, F, GH, IBE, K, MO, OOM, Mohr 234 (WIS, QCA); MANABÍ-GUAYAS: Puerto López- Rio Ayampe, 01*33-45'S, 80°40'W, Thompson 373 (МО, SAR). PERU. TUMBES: Prov. Zarumilla, Dtto. Matapalo, Cerro Tres Picos, El Caucho- Figueroa, 3-4 km on trail from Figueroa, Campo Verde road, 550 m, Simpson Schunke 394 (F). Anthurium basirotundum Croat, sp. nov. TYPE: Peru. San Martin: Tarapoto, above Hotel Tur- ista, 400 m, cultivated at SEL (# 86-1976- 2), Plowman 59804 (holotype, MO 2727371; isotype, SEL). Figures 50- Internodia brevia, 2-3 cm diam.; cataphyllum lanceo- latum, persistens fibris subtilibus et reticulatis petiolus 5- 24 cm longus, 3-5 mm diam., D-formatus aut plus min- usve triangulatus, adaxialiter ари marginibus acute elevatis; lamina ovata-elliptica, (15)19-41 cm lon- a uta, obtusa, rotundata vel sub- data; nervis primariis lateralis (3)4-8 utroque. Pedun 6 cm longus; spatha oblonga- Па 2 cm longa, 1.2-1.8 ст lata; зрафх longus, violaceus ad purpurascens. Bacca oro as. 4.5- 5 mm longae, vinosae Stem to 12 cm long, 2-3 cm diam.; leaf scars obscured by cataphylls, 0.8-1 cm high, 1.4-1.7 cm wide; roots numerous, descending to spreading, green, scurfy, weakly pubescent, slender and elon- gate, 10-11 cm long, 3-5 cm cataphylls subcoriaceous, lanceolate, 3-7 cm long, rounded diam.; at apex with a subapical apiculum, green tinged reddish, drying tan to brown (B & K yellow 6/5), persisting as a reticulum of fibers with apex re- maining intact, eventually deciduous. Leaves erect- spreading; petioles 5-24 cm long, 3-5 mm diam., D-shaped to somewhat triangular, flattened, some- times with a medial rib adaxially, the margins sharply raised, rounded to obtusely 1-ribbed abaxially, sur- face sometimes minutely pale-speckled; geniculum slightly thicker than petiole, ca. 0.7 cm long; blades subcoriaceous, ovate-elliptic, acute at apex (the acumen slightly inrolled), acute to obtuse to usually semirounded to subcordate (sometimes acute to obtuse) at base, (15)19-41 cm long, 6-17 cm wide, broadest at or below the middle, the margins broadly undulate; both surfaces matte to semi- glossy, dark to medium green above, somewhat paler below, brown to green on drying; midrib acutely raised at base, becoming flat toward the Volume 78, Number 3 1991 Croat 609 Anthurium sect. Pachyneurium apex above, prominently convex at base and slight- ly paler than surface below; primary lateral veins (3)4–8 per side, departing midrib at 45—50° angle, weakly arcuate-ascending, slightly acutely raised above, somewhat convexly raised below; interpri- mary veins obscure; reticulate veins obscure; col- lective vein arising from near the base or in the upper Уз of blade, weakly sunken above, raised below, 5-10 mm from margin. Inflorescences erect to erect-spreading; peduncle (17)26-36 cm long, (2)3-5 mm diam., 1.5-7.7X as long as petioles, green tinged with red-violet or purple, terete, some- times sulcate; spathe reflexed-spreading or some- times recurled, subcoriaceous to coriaceous, deep red-violet to dark purple (B & K purple 2/7.5), oblong-lanceolate, 2–6.2 cm long, 1.2—1.8 cm wide, broadest near the base, inserted at 30—35? angle on peduncle, abruptly acuminate to cuspidate at apex (the acumen inrolled), rounded to obtuse at base, the margins meeting at 180? angle; spadix red-violet to violet-purple (B & K purple 2/7.5), tapered to somewhat cylindroid, sessile, erect, straight, 3.5-9 cm long, 5-9 mm diam. near base, 3-4 mm diam. near apex, broadest at the base; flowers + square to rhombic or 4-lobed, 2.2-3.2 mm long, 2.1-3.5 mm wide, the sides straight to smoothly sigmoid, 5-6 flowers visible in principal spiral, 8-9 in alternate spiral; tepals smooth to weakly papillate, matte, weakly punctate; lateral tepals 1-2 mm wide, the inner margins + straight to convex, the outer margins 2-sided; pistils emer- gent, dark red-violet; stigma ellipsoid to linear, 0.5-0.7 mm long, brushlike; stamens emerging in a regular sequence from the base, laterals emerging almost to the apex before alternates emerge; fila- ments exserted, ca. 0.5 mm, 0.6 mm wide; anthers reddish to purple-violet, 0.4-0.8 mm long, 0.6-1 mm wide, inclined over the pistil; thecae ellipsoid to oblong, mm wide, not divaricate; pollen cream-yellow (B & K yellow 9/2.5), fading to white. /nfructescence with spathe persisting; spadix 4.3 cm long, 0.9 cm diam.; berries deep wine-red, = ovoid, rounded at apex, 4.5-5 mm long, 3-3.5 mm diam. Anthurium basirotundum is known only from Peru in San Martin Department in the vicinity of Tarapoto in a tropical dry forest life zone at 40 m. This species is characterized by its small overall size, leaf blades which are broadest generally below the middle and usually rounded to subcordate at base, and by its long-pedunculate inflorescence with a tapered, red-violet to violet-purple spadix and deep wine-red berries. Anthurium basirotundum is probably most closely related to А. tarapotense, which occurs in the same area. The latter differs mainly in its elliptic-oblanceolate leaf blades which are attenuate to long-attenuate (rarely acute) at the base The species is named for the leaf blades, which are usually rounded to subcordate at the base. PERU. SAN MARTÍN: Tarapoto, above Hotel Turista, 400 m, vua at SEL (#81-1976-2), Plowman 5980A (MO, SEL). Anthurium bonplandii Bunting, Acta Bot. Venez. 10: 267-268. 1975 a. Anthurium bonplandii subsp. bonplan- dii. TYPE: Venezuela. Amazonas: Dept. Atures, Rio Orinoco near Siquita, between Isla Cas- tillito and San Fernando de Atabapo, 100- 140 m, Bunting et al. 3676 (holotype, MY). Figures 53-55. Anthurium atropurpureum var. apertum R. E. Schultes, ot TYPE: Colo Amazonas: Jerijerimo, Río Apapo- s, Schultes 12094 (holotype, СН). dinum bonplandii Bunting subsp. rionegrense Bun- ting, Phytologia 64: 459, figs. 1, 2. 1988. ТУРЕ: Venezuela. Amazonas: Dept. Río Negro: San Carlos o Negro, 125 m, Steyermark & Bunting 102741 (holotype, MO). Terrestrial, rarely epilithic or epiphytic; stem 10-20 cm long, 1-4 cm diam.; roots descending and spreading, sometimes ascending when epi- phytic and forming a globose “ant garden," grayish brown when dried and with raphide cells, reportedly with velamen, 5-28 cm long, drying 3-5 mm diam.; cataphylls subcoriaceous, 2-13 cm long, acute at apex, green, drying pale brown, persisting intact or as weathered fibers, once reported to be deciduous. Leaves erect-spreading; petioles (6)10— 35 cm long, 3-20 mm diam., D-shaped, often broader than thick, flattened to broadly concave and occasionally with a medial rib adaxially, round- ed to 4-ribbed abaxially; geniculum slightly thicker than petiole, 0.5-1.5 cm long, sheathing in lower % to в of the petiole; blades coriaceous, elliptic to broadly elliptic, rarely somewhat oblanceolate, acute to acuminate at apex (the acumen + flat or slightly inrolled), usually acute to attenuate (some- times barely rounded) at base, (10)30-75(100) cm long, (5)10-30(40) cm wide, broadest at or near the middle, the margins usually flat; upper surface glossy to semiglossy, dark green, lower surface + matte, paler, usually pustular or glandular-punc- tate; midrib prominently convex above and below, 610 Annals of the Missouri Botanical Garden hat raicad f. | : both when dried, gloss А paler than surface below; primary lateral veins (6)7-10(15) per side, departing midrib at (25)45— 70? angle, usually arcuate-ascending to the margin, sometimes loop-connecting (especially in the upper l4 to % of the blade), raised on both sides; inter- primary veins very few, obscure when dried; ter- tiary veins sunken above, raised on both surfaces when dried; collective vein arising from about the middle to the upper % of blade, 3-15 mm from margin; antimarginal vein arising from the base and continuous with margin. Inflorescences + erect, equal to, shorter than, or longer than leaves; pe- duncle (18)30–90 cm long, 2-15 mm diam., 1.5- 9(12)X as long as petiole, green, + terete or slight- ly flattened on one side; spathe spreading-reflexed to reflexed, subcoriaceous, pale green, often with purple nerves or suffused violet-brown, linear-lan- ceolate to lanceolate, 5-16 cm long, 0.5-2.5 cm wide, broadest near the base, narrowly acute to acuminate at apex (acumen 10 mm long), decur- rent at base; stipe 4-23 mm long in front, 3-20 mm long in back; spadix dark purple to reddish at grayish green to brown post-anthesis, mostly cylindroid, slightly tapered at apex, erect, (3)5-15(17) cm long, 1-4(6) mm diam. near apex, 3-9 mm diam. near base; flowers rhombic to 4-lobed, (1.8)2.2-2.8 mm long, (1.6)1.8-2.4 mm wide, the sides + straight or somewhat jaggedly sigmoid, 3-7(8) flowers visible in principal spiral, 5-7 in alternate spiral; tepals pale-punctate, smooth to minutely papillate; lateral tepals 1.2-1.8 mm wide, the inner margins rounded, the outer margins + straight, 2-sided; pistils emergent, somewhat papillate; stigma squarish to rounded, 1.2 mm long, depressed medially; lateral stamens preceding the alternates by 6-20 spirals; anthers yellowish white, 0.6-0.8 mm long, 0.7-1 mm wide, inclined over and obscuring the pistil; thecae oblong to somewhat triangular, slightly divaricate. Infructescence with spathe withered or absent; spadix (5)8-25(35) cm long, 1-2 cm diam., bearing berries in the basal portion only, stip white at base, obovoid, obtuse-truncate at apex, 5-7 mm long, 5-6 mm ened, with raphide cells; seeds 1-2 per berry, 4— 5 mm long, 2-4 mm diam., with mucilage on one anthesis, stipe 5-15 cm long; berries purple, diam.; pericarp very thick- end but hanging on to carpel wall by thick band of fibers. Anthurium bonplandii is known from sea level to 1,400 m throughout much of the northern Am- azon basin, ranging from southeastern Colombia to southern Venezuela and to northern Brazil. It oc- curs on sandstone rocks of the Guiana shield as well as on granite boulders or granitic outcrops (materials originally underlying the Guiana shield) and on white sand deposited by eroded Guiana sandstone on the lowland plateau of the Tertiary Amazon Lakebed (Guerra, 1959). This species is recognized by its distinctly long- pedunculate inflorescence (the peduncle typically being 1.5-7X as long as the petioles), by the glandular punctations on the underside of the leaf blade, by the tough, leathery pericarp and meager mesocarp in rehydrated berries, and by the seeds attached to the apical end of the inner carpel wall by a thick strand of fibers running midway from the apex to the base of the fruit. Anthurium bonplandii may be confused with A. atropurpureum, which has leaf blades of similar shape and which also occurs on white sand deposits. The latter differs in having generally smaller, eglan- dular leaf blades which dry greenish (vs. brownish in A. bonplandii), as well as a smaller inflorescence and different berries. The berries of А. atropur- pureum typically have a soft, pulpy mesocarp on rehydration and lack a fibrous attachment to the carpel wall, having instead a mucilaginous ap- pendage, the typical condition for the section. Ап- thurium bonplandii is also similar to A. guan- chezii, A. lanjouwii, and A. xanthoneurum. See those species for a discussion of the differences. Anthurium bonplandii is variable in its overall morphology, as well as in ecological preferences. There are three subspecies: iramirezae, A. Subspecies cuatrecasii ranges across the north- ern fringe of the Amazon basin from southeastern Colombia to the western half of Amazonas in Ven- ezuela at 75-38 from the other subspecies in its typically rounded or subtruncated leaf bases, longer peduncles and elongated stipe. Subspecies guayanum occurs in the Guiana Highlands of Venezuela, Guyana, and Surinam, usually above 500 m, and differs from the other O m, primarily on rocks. It differs subspecies in its generally much larger, promi- nently dark, glandular-punctate (abaxially) leaf blades Anthurium bonplandii subsp. bonplandii rang- es from southeastern Colombia (Vaupés, Meta, Guainia, and Amazonas) east to Venezuela and Brazil in the middle and upper Rio Orinoco Basin, the upper Rio Negro, and the Rio Trombetas, and south to the states of Amazonas, Рага, and Roraima in Brazil at usually less than 500 m (rarely to 825 m). It is predominantly terrestrial in sandy areas or, less frequently, is found growing on rocks or in open savanna areas, or in primary forest. Їп Volume 78, Number 3 1991 Croat Anthurium sect. Pachyneurium Brazil, it has been reported from “‘caatinga”’ forest or ““campina abierta." In Venezuela it occurs most- ly in a tropical moist forest life zone. Subspecies bonplandii is recognized by its elliptic or rarely oblanceolate leaf blades which are acute at the base, and by its raised, caviform stigma. It is highly variable in size, leaf shape, and in the degree and type of glandlike structures on the lower blade surface. It overlaps geographically with the other two subspecies; see the discussions under the latter for further comments. Particularly noteworthy in subsp. bonplandii are a series of collections from southwestern Ven- ezuela in Amazonas along the upper Rio Orinoco Basin and the intervening river system connecting the Rio Orinoco with the Rio Negro, e.g., the Rio Casiquiare (including the Rio Yatua, one of its tributaries). These collections are notable for their large size, with blades ranging from 63 to 98 cm long and 30 to 48 cm wide. Despite the fact that these collections differ rather radically from most others, they differ little from the largest plants found in populations around nearby San Carlos de Rio Negro, where a wealth of collections from a small area enables a more clear assessment of in- terpopulational variation. Of these, the largest leaf blades observed were those of Liesner 3778, which are 85 cm long and 32 cm wide, with 10 primary lateral veins per side. The smallest plants from this area have blades 28 x 11 cm and only 8 pairs of primary lateral veins per side. These smaller plants differ little from collections made along the Upper Rio Negro in Brazil. At the western edge of its range, the leaf blades of many collections of subsp. ically more narrowly oblong-oblanceolate than else- where, being up to 5X longer than broad. On average, blades are 2.6 х longer than broad and infrequently are more than 3x longer than broad. The blades of most collections from the upper Rio Negro are 2.5-4X longer than broad, while col- lections from the middle Rio Negro and the upper Rio Orinoco drainage range from 1.7 to 3 x longer than broad. The proportional length of the blade vs. the petiole is highly variable, ranging from only slightly longer to 14.4X longer. Blade shape appears to be in no way correlated with blade size, though the generally more narrowly oblanceolate blades of populations in eastern Colombia (Vaupés) have a more narrow range of variation, with blades 1.2— 4.7 X longer than broad and averaging 2.8 x longer than the petioles. Populations in the northern ex- treme of the range in Amazonas in Venezuela have blades ranging from 1.1 to 8.3 x (averaging З х) bonplandii are typ- longer than the petioles, while those in the middle of the range (upper Rio Negro in Brazil) are less variable, with blades ranging from 2.1 to 6.7 x (averaging 4X) longer than the petioles. Popula- tions from northwestern Pará in Brazil have pro- portionately shorter petioles, with blades ranging from 4 to 14x (averaging 7.9 x) longer than the petioles. An unpublished name, 4. disparile Schott, rep- resented by Schott Drawing 315 (New York Bo- tanical Garden negative #3951 and microfiche 25: C3) based on a collection made by Linden on the Ко Orinoco is A. bonplandii Bunting subsp. bon- plandii. BRAZIL. AMAZONAS: и Republica, Rio Curu- San Gabriel da Cachoeira, Morro dos Seis Lagos, Lago do Dragão, 400-450 m, Farney et al. 1714 (МО); Iga- гаре do Вшао, Manaus, Rodrigues & Chagas 2010 (INPA); Manaus-Caracarai, km 26, Prance et al. 3073 (INPA, NY, US); Manaus-Porto Velho, trecho Castanho- Tupana, Silva et al. 200 (INPA), Manaus-Sào Gabriel, 00°35'S, 64?40'W, Alencar 252 (NY); 53 km W of Rio Aripuana, Transamazon Hwy., Calderon et al. s (INPA, MO, US); Santa Isabel do Rio Negro, Campin e Tamendaui, Madison et al. PFE 39 (INPA), 241 INPA): 0°23'S, 65°2'W, Maas 6679 (F, MO); km 795- 790 Transamazonica Hwy., Rodovia do Estanho, 7°35'S, 62°30'W, Vieira et ^ 187 (INPA, MC, MO, ы Pico Rondon, vic. Km 211 of Perimetral N ox 700 m 1°32'N, 62°48'W, Ра et al. 6658 (М O, NY), Prance et al. 28768 (INPA, MO, NY), 28785 (MO, Las NY); Reserva Biologica do INPA, Manaus, BR-174, Km 45, 500 m, Anderson 318, 342 (INPA); Mpo. anes Ma- naus- танина. Campina ice Km 45, less 100 m, 1?40'S, 60°05'W, Croat 62207 (INP : Reserva Florestal Ducke, Manaus, Ferreira 5877 (INPA); Rio Brancinho, Rio Cuieras, Prance et al. 17809, 17858 (INPA, NY); Rio Negro, Maia et al. 689 (INPA); Bar- celos- Tapurucuará, Nascimento 627 МУ); Ia- uarete, Coelho & Francisco 203 (INPA); Ilha das Flores, 50 m, Cavalcante 655 (GM); Rodrigues 905 (RB), 906 (MO); Manaus-Sào Gabriel, Morro dos Seis Lagos, 80 km N of Sao Gabriel, 100 m, 0°20'N, 66?45'W, Poole 2055 (INPA, MG, MO, NY); Serra de Uaupes, Rodrigues & Coelho 1466 (MO), Rodrigues 10718 (NY); Tamen- daui, ca. 40 km below Tapurucuará, Madison et al. 6241 (MO, SEL), e et al. 252 (INPA); Tapurucuará-Mi- rim, Rio Marie, Nascimento 656 (MG, МО, NY); Projecto Flora Epifita . 311 (MO); Rio Uaupes, ore Luetzelburg 23787 (M, R); Taracua, Cavalcante 747 (GM); Rio dos Pombos, 9 km W of river, ca. 1.5 km a of Igarape dos Pombos, ca. 64 km E of the Aripuanã, 7*10'S, 60*00"W, Calderon et al. 2606 (INPA), Serra Araca, 0°48’М, 63°18'W, Pipoly et al. 6782 (NY); Central Massif, 700 m, 0%50'N, 63°17'Ұ, Prance et al. 29620 (MO, NY); Mpo. Barcelos, Rio Araca, Jauari, afluente de granito, 00°25'N, 63°25'W, Cordeiro 242 (MO, NY); Mpo. Ma- naus, Rio Negro, NW of Manaus, Madison et al. 6311 (K, SEL, US), 6393 (SEL), Projecto Flora Epifita 39 (INPA). РАКА: Cerro de Cupaty, Caqueta, Lucke 12298 (GM); Rio Jaramacaru, Óbidos, Egler 282 (MG, RB); Rio == Annals of the Missouri Botanical Garden Negro, Cavalcante 3125 (MG, NY); Mpo. Altamira, Igar- apé Ipixuna, affluent of Rio Xingu, 5 km 5 of settlement, Arawete Indian Reserve, 4?49' 5 5993] 1'W, Balee & Ri- beiro 1901 (NY); Mpo. Oriximina, Rio Mapuera, 30 km from Cachoeira Porteira, Cid et al. 1208 ОХРА, NY); ca. m upriver from Cachoeira Porteira, Davidson & Martinelli 10603 (INPA, MG, MO, US, NY); Rio Parú, Cachoeira Chuvisco, Cid et al. 2228 (INPA, MG, NY); Rio Trombetas, 4 km from margin of Lake Maecue, Cid et al. 1499 (INPA, NY); 7 km below Cachoeria Porteira, Cid et al. 1298 (INPA, MG, MO, NY, US); Mpo. dom Represa Tucurui-Breu Branco, BR 422, Km 25, 90 п 3552'S, 40*44"W, Plowman et al. 9564 (MG, MO, NY) теа 100-200 m, Lisboa et al. 2307 (МС). RORAIMA: Manaus-Caracari, Km 515, along bank of Igarapé Dias, сао et al. 131 (NY, INPA); Rio Uraricoera, Canal Maraca, Cachoeira Menori, 61°55'W, Pires et al. 16792 (INPA, MG); Mpo. Sao Luiz do Anaua, киеш 9233 (МО). COLOMBIA. AMAZONAS: Rio Apapor Jirijirimo, 250 m, Schultes 12094 (COL, CH), 5 Mataveni (confluente of Rio Orinoco), 100-140 623 (NY). AMAZONAS-VAUPÉS: Rio pad res 250-300 m, Garcia-Barriga 13751 (COL, US), Schultes & Cabrera 12370 (AMES), 12436 КЫ GH), 13466, 13517 (СН), 13501, 15952 (AMES), 14641 (AMES, GH), 14051 (GH, U), 14571 (COL, US); Jinogoje, at mouth of Вю Piraparana, са. 350 m, 0%15'S, 70°30'W, Schultes & Cabrera 17891 (GH). GUAINÍA: Rio Negro, junction of Rio Negro, Brazo Casiquiare and Río Guainía, 65 m, 2?00'14"N, 67907 М, Davidse & Miller 26577 (MBM, MO, VEN); across river from San Fernando de Atabapo (Venezuela) 120 947" 67°06'W, Gentry & Stein 46412 (B, К, MO, VEN), 46413 Sa META: Llanos Orientales, La Macarena, Rio Guayabero, Sabanas de Arenisca, 235-700 m, Gar- cia- Bactra 17031 (COL, GH, NY, US). vaupés: Cerro de Circasia 300-400 m, Cuatrecasas 7178 (COL); Rio Guainia, Puerto Colombia, opposite Venezuelan town of Maroa, 250-300 m, 2°40’М, 67°30'W, Schultes & Ca- brera 18242 (AMES); Ко Kananari, Cerro Isibukuri, 0?15'N, 70°35'W, Schultes & Cabrera 15082 (СН); affluent of Rio Apaporis, Schultes & d 13422 (GH); Rio Kuduyari, trib. of Rio Vaupes, -300 m 1915' АА 70°05'W, ultes & Cabrera 17877 (CH. MO, X Kuduyari, Sabana Yapo 350-400 m, Sc le 22677 (ECON); Rio n San Felipe, below confluence of Ríos Guainia and Casiquiare, Schultes & Lopez 9330 (GH, NY, US); Río Parana- Pichuna, Мий, Zarucchi 1348 (COL, МО); Rio Piraparana, tributary of Rio paporis, Schultes & Cabrera 17475, 17523 (СН); i varete, Schultes & Cabrera 19266 Circasia, Миа, Janice 2047 (COL, O, US). VENEZUELA. AMAZONAS: 200 m, 5?05'N, 65°35'W, Colchester 2043 (K); Cerro Moriche, Rios Manapiare and Ventuari, 150-250 m, Maguire & Ma- guire 35533 (NY, Map Cerro Neblina Region, Cerro Neblina, Rio Mawarinuma, 140 m, 0%50'N, 66?10'W, Croat 59326 (MO, RSA), 59388 (CM, MO), 59627 (B, K, MO, RSA, TEX); Loma de las Pinas (Pineapple Ridge), 1.5 km S of Neblina Base Camp, 150-215 m, 0%49'15N, 66?9'40"W, s 30554 (MO, NY, VEN Rio Mawar- inuma, Cañón Grande, 350 m, 0%50'N, 66%02-06'W, Davidse & Miller. 27132 (VEN), 27174 (MO); 0.5-1.5 km SE of Cerro Neblina, 140 m, igh 66°10'W, Liesner 16147 (MO); Rio Baria, 80 m, 1?27'-1?10'N, 66°32'-66°25'W, Davidse 27683 (VEN); mouth of Rio Pasimoni-junction with Rio Baría and Rio Yatua, 80 m, 1253-27 N, 66°35-33'Ұ, Toup 27780 (МО); Cerro Yapacana, 125-400 m, 3*45'N, 6°45'W, Steyermark & Bunting 86762, 103060 cad 825 m, Steyermark ii Bunting 103169 (VEN); Cerros del Casiquiare, 120 Vareschi 6626 (VEN); Rio Negro, Piedra de Cocui, Schultes & d 9423 (CH, 05); slope of Cerro Ara- camuni, 600 m, Liesner & Carnevali 22311 (МО); Rio Orinoco, Gran Laja near Nericagua, Morillo 7451 (VEN); 5 km above junction with Сапо Cotua, SE of Cerro Ya- pacana, 100 m, 3?45'N, 66°50’ ы о & Вип- ting 102991 (MO, VEN); Cer e 3°45'N, 66°50'W, 140 m, Sytsma et al. 510 WIS); Rio Para- ua, vic. “Minas de Manaima,” оч 63945 МУ, Ster- gios 10289 (MO, PORT); 6?22'N, 66°22'W, Stergios 10129 (MO, PORT); Саћо Cupaven, opposite mouth o Rio кзы 125 m, Wurdack & Adderley 42828 (NY, US, VEN); Sanariapo-point 3 hr upriver by 33 n ou 4255 (MO, MY); Dpt. Casigquiare, Pn i 25- of Cerro Arauicaua er mark & Bunting 102532 (MY, VEN), 102543 pipe Piedra Arauicaua, 110-550 m, Maguire et al. 416 (NY); San Carlos de Rio Negro Region, 3-4 km N y San Carlos i Río Negro toward Solano, Plowman 13536 (MO) Km 11 NE of San Carlos de Rio Negro on road to Solano, ns , 1953'М, 67%02'W, Davidse & Miller 26556 (MO, VEN, W), 26559 (VEN); 115 m, Berry 1366 (MO, VEN); Stergios & Aymard 7742 (PORT); near airport, Croat 59258 (AAU, MO, NY, US), 59260 (B, CAS, F, M, MO, NY); Steyermark & Bunting 102741 Ca 79037 , Liesner 3778, 6309 (МО), 6137 (VEN), 3277 AMO. SEL), 6402, 6568, 6569 (MO, VEN), Morillo & Hasegawa 5093 (VEN); San Calais Solano, 2 km N of San Carlos, 100 m, Morillo & Villa vidse et al. 17270 (МЕМ); Cano Caname, Curital de Caname, 100 m, 3°40'N, 67°22'W, Davidse et al. 16966 VEN); Dpt. Atures, Isla de aton, Carmen, 5?08'N, 67°54’W, Carnevali et al. 1527 (MO); Rio Orinoco, La Esmeralda, Croizat 208 (NY); Siquita, 100-140 m, Bun ting et al. 3676 (MY); Puerto Ayacucho-Sanariapo, Bunting 4279 (NY); Sanariapo-San m оны 4239 (NY), Serrania de Yutaje, 0-1.5 km m Coro, km N of settlement of Yutaje, 270. m, 5°40'N, 66°7'30"W, Liesner | Eo 21263 ен | ER W of Rio Coro-Coro, W of settlement of Yutaje, 700-1, 2 m, 594 ms 6699 W, Liesner & Holst 21816 U, B, MO). APURE: Piedra de San Vincente, 80-160 zalez 14537 (MO, NY, US, VEN); ejo, of Paso de Cinantico, 60-65 ek ыны е 12467, 12612 (MO, VEN); Rio M "Río Cinaruco, Cario El Caballo, 75 m, Davidse & Gonzalez 15887 (MO, VEN), 13945 (MO, VEN); at jct. of Cano Siriaco, 70 m, Davidse & Gonzalez 14147 (MO, VEN); Dtto. San Fernando, 60 m, Davidse & Gonzalez 13895 (F, MO VEN). BoLÍvAR: Mpo. Raul Leoni, 64 km SE of Pijiguaos, 550 m, 06?09'N, 66?23'W, Delgado 298 (MO). — Volume 78, Number 3 Croat 613 Anthurium sect. Pachyneurium b. Anthurium bonplandii subsp. cuatrecasii Croat, Aroideana 91 –4): 11-14. 1986. TYPE: Dept. Atabapo, vic. Puerto Ayacucho, along rd. from Puerto Aya- cucho to Sanariapo, Croat 55065 (holotype, MO 2934844; isotypes, B, DUKE, GH, NY, RSA, US, VEN). Figures 56-58. Venezuela. Amazonas: Terrestrial or epilithic, о epiphytic; stem often caespitose, 1— iam.; roots dense, mostly descending, the nine ascending, gray- ish when dried, slender and elongate, to 25 cm long, 2-5 mm diam.; cataphylls coriaceous, 4-14 cm long, acute to weakly acuminate at apex, per- sisting intact, eventually dilacerating into fibers at base. Leaves erect; petioles 4–30(40) cm long, dryin C-shaped or nearly D-shaped, broadly and sharply sulcate to nearly 3-11 mm diam., flattened, with the margins prominently raised adaxially, rounded abaxially; geniculum 0.5-2 cm long; blades coriaceous, oblanceolate, often + el- liptic to slightly ovate-elliptic, rarely oblong-elliptic, acute to weakly acuminate at apex, narrowly rounded to truncate at base, sometimes obtuse, (16)25-55(70) cm long, 4–25(32) cm wide, broad- est at or above the middle, the margins weakly undulate; upper surface semiglossy to glossy, me- dium to dark green above, the lower surface se- QUNM puts pustulate or glandular-punctate be- low; y raised at base above, becoming convexly raised bed the middle, prominently convexly raised below; primary lateral veins 5-10 per side, departing midrib at (30)40—55(75)° angle, usually weakly arcuate-ascending to the margin, loop-connecting in the upper / of the blade, con- vexly raised; interprimary veins occasionally pres- ent; tertiary veins weakly etched above, slightly raised or flat and darker than surface below, prom- inulous when dried; collective vein arising from near the apex, less prominent than primary lateral veins, often absent, 2—10 mm from margin. /nflo- rescences erect; peduncle (19)30-100 cm long, drying 5-10 mm diam., 1.3-7 x as long as petiole, terete; spathe reflexed, usually twisted or recurled, membranous-chartaceous to subcoriaceous, yellowish green, sometimes suffused with reddish purple, linear-lanceolate, 5-20 cm long, 0.5-2.5 broadest near the base, acuminate at apex (the acumen ca. 5 mm long), decurrent at base; stipe 15-40 mm long; spadix pale green, coming pink to maroon at maturity, drying green, tapered, erect, (6)10-25 cm long, 5-15 m near base, 3-6 mm diam. near apex, broadest at the base; flowers rhombic to 4-lobed, drying 1.6- pale cm wide, mm diam. 2.6(3.0) mm long, (1.2)1.5-2.6 mm wide, the sides smoothly sigmoid, (4)5-8 flowers visible in either spiral; tepals smooth to somewhat roughened sometimes whitish muricate; lateral tepals 1.2-1.8 mm wide, the inner margins convex, the outer margins 2—3-sided; pistils + square, barely or not at all emergent; stigma oblong, + rectangular, 0.6 mm long, densely papillate; filaments exserted ca. 1 mm; anthers 0.4-1 mm long, 0.4-1.2 mm wide, contiguous, surrounding pistil; thecae drying + ellipsoid, divaricate. Infructescence with spathe withered or absent; spadix (5)10-22 ст long, 0.8— 2 cm diam., bearing berries in the basal portion only; berries (immature) green, obovoid, + trun- cate at apex, 4-6 mm long, 3-5 mm diam .; peri- carp thickened, with raphide cells; seeds 2 per berry, with raphides on the surface, broadly ellip- soid, 4-5 mm long, 1.8-3 mm thick, 1.2-1.4 mm diam., attached to apical end of carpel by strand of fibers. Anthurium bonplandii subsp. cuatrecasii rang- es from southern Colombia (Mitü in the Serrania de la Macarena) northeast to the drainage of the middle and upper Rio Orinoco in Venezuela (Ama- zonas) at 75-380 m. Principally found on granite boulders, it often has a caespitose habit, growing in humus tufts either on granite outcrops or ter- restrially on soils with a granitic base. It occurs in the tropical dry, tropical moist, and premontane wet forest life zones. Subspecies cuatrecasii differs from the other two subspecies of 4. bonplandii in having leaves which mostly dry green or greenish and are usually rounded to narrowly truncate at the base, and by its long-pedunculate inflorescence with a usually long-stipitate spadix. The species is difficult to separate from subsp. bonplandii in regions of geographical overlap such as the Мип area of Vaupés in eastern Colombia or the Middle Orinoco River Basin. Schultes & Cabrera 19709, from the Rio Vaupés area near Миа in Colombia, with both long- and short-stip- itate spadices, exemplifies such taxonomic diffi- culty. It is possible that this collection, and others within this range of sympatry, are the result of local hybridization involving subsp. bonplandii. COLOMBIA. GUIANÍA: Rio Orinoco, S of Casuarito, Lejas, 90 m, 5?40'N, 67°37'W, Davidse & Miller 26370 (MO); Я феі Nacional Sierra de la E of Refugio, 410 m Callejas & Arias 6475 (HUA, MO); Rio Сова 10 km below Cano Lozada, 500 m, Pinto & Jaramillo 208 614 Annals of the Missouri Botanical Garden (COL), 250 m, Echerverria 2054 (COL). vAUPÉs: Cerro Миа, 380 m, Cuatrecasas 6885 (COL, US), Schultes et al. 24202 (AMES), Schultes 22702 (GH); Cerro Mitú- Río Vaupés, 200 m, Soejarto & Lockwood 2495 (F, K MEDEL); 350 m, Maguire et al. 44103 (NY, US); Río Vaupés, Mita, 250 m, Schultes & Cabrera 13912 (GH, U); Mitü-Javarete, Агага Cachwira, Schultes & Cabrera 19397 (СН); Raudal de Yurupari, near Миа, 0°40’М, 70°30'W, Schultes & Cabrera 19709 (AMES, GH, МО, NY, U); Mita, Zarucchi 1676 (COL, K, MO, US); 350 m, 69°45’ W, Ferndndez 2156 (COL, F, US); Rio Karuru (tributary of Rio Vaupés), Mesa de Yambi, 300-350 m Schultes & Cabrera 19147 (AMES), 19194 (AMES, СН, Uy Río Kuduyari (tributary of Rio Vaupés), Yapoboda, 230-270 m, 1?15'N, *05'W. Schultes & Cabrera 14257 (GH), 14273 (GH, U), Schultes et al. 18456 (AMES), Garcia-Barriga et al. 158234, 15823 (COL, ЕДД bis W, Zarucchi & Barbosa 3416 (MO). VENEZUELA: NAS: Bogner 960 (M); Сапо Mosquito, Cario leta, nee 266 (K); Puerto Ayacucho Region, vic. ы: Ayacucho, Huber 612 (M, MO, VEN), Plowman 7742 (F, К, MO, SEL), Wessels Boer 1904 (MO, U), о 18379 (VEN); Puerto Ayacucho-Sanariapo, Rio niapo, near airport, 100 m, Croat 55065 (B, MO, NY), Gentry & Berry 14438 (MO), Trujillo & Pulido 14958 (MY), Williams 16037 (Е); З km М of Puerto Ayacucho, 80 m, Morillo 6753 (MY, VEN); 8 km S of Puerto Ayacucho, Estación de Piscicultura, 85 m, А 67937 МУ, Davidse & Huber 14894 (MO), Davidse & Miller 26343 (MO); 20 km S of Puerto |“ N of Caro Garcitas, 100-130 m, 5?33'N, 6 , Carne- 100 m, Williams 13446 (US, VEN); pe Cataniapo, 80 m from bridge to Sanariapo, 6%25'N, 67°25'W, Castillo 1237 (MO); Rio Guayapo, 10-15 d dh mouth o Río Guayapo, tributary to Rio Sipapo & Ishikawa 3432 (VEN); Rio с middle an 100 m, Меш et al. 36194 US, МЕМ); La Esmeralda, Medina а, От 66°01'W, Liesner 18654B (MO); „Ventuari, 20 km Е of Río Orinoco junction, 98 m, 4?03'N, 6649", Huber 1853 (VEN). BOLÍVAR: mouth x the EE Laja del Zamuro, 150 m, Trujillo 10756 (MY); vic. Minas de Manaima, middle Rio Paragua, 6°7'N, 63?45'W, 300 m, Stergios 10291 (PORT); Rio Orinoco, rocky outcrops on Isla Sta. Elena, 100-150 m, Wurdack & Monachino 39874 (Е, NY, US, VEN); Rio Paragua, Badillo 1451 (VEN); Dtto. Cederio, 35 km SW of Caicara del Orinoco, 100-300 m, 7*30'N, 66°20'W, Ste» rk et al. 131274 (MO) c. Anthurium bonplandii subsp. guayanum (Bunting) Croat, comb. nov. Anthurium guay- anum Bunting, Acta Bot. Venez. 10: 268. 1975. TYPE: Venezuela. Bolivar: Chimantá Massif, NW part of Abacapa Терш, 850- 1,100 m, Steyermark 75100 (holotype, VEN 2005409; isotypes, F, MO, NY). Figures 60- 62. Anthurium corocoroense Bunting, dba, e 64: 462. 1988. ТҮРЕ: Venezuela. Amaz : Dpt. Atures, Cerro Yutaje, a ridge, in mo a ane woodla nd, 1,500 m, Maguire & Maguire 35425 (holotype, NY ) Anthurium к inimae Bunting, Phytologia 64: 462. 1988. T PE: Venezuela. Bolivar: Dtto. Heres, sum- mit of (те Guaiquinima, 1-2 km river above Salto Szczerbani (Río Caparo), 5°44'04”М, 63%41'08"W, 750 m, Steyermark et al. 113258 (holotype, MO 2581469; isotype, VEN). Terrestrial or occasionally m stem to ca. 20 cm long; internodes short, cm diam.; roots dense, spreading, the uppermost ones ascending, thick and blunt at apex, 5-12 cm long, 1-5 mm diam.; cataphylls subcoriaceous or fleshy, 2.5-12 cm long, acute at apex, tinged with red, persisting as coarse linear fibers at base. Leaves erect to arching when large; petioles (3)6-30(42) cm long, 8-15 mm diam., C-shaped or sharply D-shaped, sulcate to flattened with the margins sometimes weakly raised and a broad medial rib occasionally present adaxially, rounded abaxially, sometimes 1— 2-ribbed; geniculum slightly thicker than petiole, 0.5-1 cm long; blades coriaceous, oblanceolate to obovate, rarely broadly elliptic, acute to obtuse to rounded at apex, long-attenuate to acute or obtuse at base, 30-80(1 40) cm long, 20–40(65) cm wide, broadest above the middle, the margins undulate; upper surface semiglossy, dark green, lower sur- face semiglossy to matte, paler and with pustules and glandular-punctations; midrib flat to broadly sulcate at base, sometimes with faint medial rib, becoming prominently convex toward the apex above, higher than broad at base, becoming con- vexly raised toward apex below; primary lateral veins 6-9(13) per side, departing midib at 40-60? angle, ascending and weakly arcuate to the col- lective vein, loop-connecting in the upper Y3, raised in grooves above, raised and paler than surface below; tertiary veins sunken above, raised below; collective vein arising in the upper / of blade, almost as prominent as primary lateral veins, or often absent, 5-1 cences erect-spreading, nearly as long as or longer than leaves; peduncle 24-96 cm long, 4-10 mm diam., (1.5)4—5(7)X as long as petiole, green tinged with maroon or purple, terete; spathe spreading- reflexed, sometimes inrolled, coiling-recurled or re- mm from margin. /nflores- flexed, subcoriaceous, pale to medium green, some- times suffused with purple, oblong-lanceolate, (4)8— Volume 78, Number 3 1991 Croat 615 Anthurium sect. Pachyneurium 18(24) cm long, 1-3.5 cm wide, broadest near the base, acute to shortly acuminate at apex (the acu- men ca. 10 mm long), obtusely clasping at base; spadix greenish when young, becoming reddish to purple or purple-brown at anthesis, sessile or sub- sessile, * cylindroid, shortly to gradually tapered at apex, erect, 5-33 cm long, ca. 5-7 mm diam. near base, 2-3 mm diam. near apex; flowers (dried) rhombic, (1.6)2.8-3.4 mm long, (1)1.8- wide, the sides straight parallel to spiral, smoothly and weakly sigmoid perpendicular to spiral; 3- 7(9) flowers visible in principal spiral, 4—6 in alternate spiral; tepals minutely papillate, sometimes whitish muricate, matte; lateral tepals 1.4-2 mm wide, the outer margins 2-sided, the inner margins broadly convex; pistils squarish, weakly emergent; stigma rectangular, 0.6 mm long, appearing papillate; sta- mens emerging in a regular sequence from the base, the laterals preceding the alternates by 7- 10 spirals, the 3rd stamen preceding the 4th by 6–8 spirals, arranged in a circle, inclined over and obscuring the pistil; anthers yellow, 0.4–0.6 mm long, 0.5-0.8 mm wide; thecae oblong, 0.2-0.3 mm wide, slightly divaricate; pollen white. Infruc- tescence with spathe withered, eventually decidu- ous; spadix 8-35 cm long, to 2 cm diam., bearing berries in the basal portion only, the apical portion sometimes weathering away; berries reddish pur- ple, white in basal 4, obovoid, 6-10 mm long, 5- 8 mm diam.; pericarp thickened, with raphide cells; seeds 1 or 2 per berry, reddish violet, ellipsoid, weakly flattened, 3-6 mm long, 2-5 mm diam., 1-2 mm thick, attached to carpel wall at radicle end and by a thick strand of fibers. Anthurium bonplandii subsp. guayanum oc- curs in the Guiana Highlands of Venezuela, Brazil, Guyana, and Surinam, most material being from Bolivar State in Venezuela, at (100)400 to 1,500 m on sandstone outcrops, sandstone boulders or in pure white sand deposits in open areas and in partial shade. It is ecologically variable, occurring in pre- montane moist, premontane wet, premontane rain- forest, and tropical wet forest life zones. is subspecies is characterized by its usually large, often broad, coriaceous leaf blades which have unusually dark and conspicuous glandular punctations on the lower surface. It is distinguished from subsp. bonplandii main- ly by its oblanceolate to obovate (vs. mostly elliptic) leaf shape, and from subsp. cuatrecasii by the leaf blades being usually acute at base and drying brown (vs. greenish) and by the sessile to subsessile (vs. usually markedly stipitate) spadix. Both of these subspecies occur at lower elevations than subsp. guayanum, usually below The subspecies guayanum may be easily con- fused with the other two subspecies of А. landii in areas where their ranges idée a es- pecially in the upper Rio Orinoco drainage and along the rivers of southern Amazonas, Venezuela, that flow into Rio Negro. Two species described recently by Bunting (1988) are synonymized here. Both are compared with only А. bonplandii. Anthurium corocoroense dif- fers in no way from typical А. bonplandii subsp. guayanum, and А. guaiquinimae differs only in a minor way. Characters which are stressed as important are the broad spathe (to 2.5 cm wide) and the abruptly rounded leaf base. The spathe of subsp. guayanum may range to 3.5 cm wide and the abruptly rounded leaf base is an occasional feature on some material of subsp. guayanum (see for example Davidse & Miller 27174 from Cerro Neblina and Geyskes 16 from Surinam). BRAZIL. RORAIMA: Serra dos Surucucus, NE of Mission Station, 1,000-1,400 m, 2*42'N, 63?33'W, Prance et al. 10010 (INPA, NY). GuvaNa. MAZARUNI-POTARO REGION: upper Mazaruni River Basi above Maipuri Falls, 550 m & Alfred 7635 (US). SURINAM. Arrowhead basin, Taf- elberg, 500 m, Maguire yocp Y, US); Emma Mts., W slope, 600 m, Boer 1481 (NY); Gonggrijp top, 950 m, Daniels & Jonker 1 (U); Gonggrijp top-Hendrik to "ud sy l top, 700 m, Daniels & Jonker 807 (NY). 8 km N of Yutaje, 650-750 m, Marahuaca, 1,000-1,200 m, 3°34’N, 25547 (F, DUKE, MO, US) Serrania del Parü (Aroko), 1,100 m, 4°31'N, 65°35'W, Huber 4345 (VEN); Dpt. Rio Negro, Cerro Neblina, N of Puerto Chimo, 500-700 m, 0?50'N, 66? 06' W, و‎ rA & Miller 27384 (МО); along Rio Mawarinuma, Cañón Grande, 300 m, 0°50'N, 66?02-06'W, Шыны & Miller 27245 ege BOLÍVAR: 0-6 km S of El Puaji, 800-900 m, 4?30'N, 61°35'W, Liesner 19718, 19782 (MO), Liesner x А 18799 (МО); Esmeralda Ridge, Esmeralda, 100 m, Tate 189 (NY); Santa Elena-Icabarü, 106 km m of Santa Elena, NE of Icabarü, Esmeralda, 750 m, 4?19'N, 61?46'W, Croat 54062 (B, F, GH, MO, US); Canaima, 800 m, Ferrari 1106 (MY), Trujillo 6015 (MY), Bunting 4346 (NY); ); Quebrada Los Brasileros-road to Playa Blanca and Pa Ua aiparú, SW of Icabarü, 450 m, 4%20'N, 614 8'W, Steyermark et al. 117747 (MO, VEN); Cerro Bolivar, Piar, 750 m, Aristeguieta 2275 (NY); Cerro Guaiquinima, 300 m, Maguire 32743 (NY, VEN); Salto del Rio а (Rio Сагаро), 1-2 km above Salto Szczerbanari, 750 m, 5?44'N, 63°41'W, Steyermark & 616 Annals of the Missouri Botanical Garden Dunsterville 113258 (F, MO, U, VEN), Steyermark et al. 117258 (MO, VEN); Cerro Marajanu, Alto С a- tepuí, 850-1,100 m, Steyermark 75100 (F, MO, NY); Auyan-tepui, Guayaraca, 1,000 m, Bogner 1534 (M, MO), Steyermark 94210 (VEN); Dtto. Piar, Pie de la Roca-Guayaraca, Auyán-tepui, 600-850 m, 5?43'N, 62*31'W, Davidse & Huber 22658 (MO); lower part of Río Caroni, Sabana de Arekuna, 6?31'N, 62°53'W, Prance d uid 28334 (МО); Río Icabarú and Rio Hacha, 450- m, Bernardi 2827 (NY) Río Paragua, Salto de jen 275 m, Killip 37358 (US, VEN); Sierra Ichun, of Salto María Espuma (Salto Ichun), near Rio Ichun (tributary of Río Paragua), 625-725 m, 4%46'N, 63°18'W, ы Mal 90240 (BH, US); Dtto. Roscio, 1 km N of 600-800 m, 4%25'N, 61°37'W, Holst & Liesner 2479 (MO); 7 km NW of Icabarú, 400 m 4°20'N, 61°51'W, Steyermark et al. 127301 (B, мо) Mpo. Urucuru, Avequi, “Ciudad Muerte," 500 m nardi 1687 (NY); Mpo. Urucuru, Rio Hueque, Corto del Papelou, 700 m, Bernardi 164.1 (NY). Anthurium bradeanum Croat & Grayum, sp. nov. TYPE: Panama. Cocle: Cerro Moreno, Mo- lejón-Coclecito, 13 km NW of Cascajál, 130- 250 m, Davidse & Hamilton 23715 (holo- type, MO 3039173; isotypes, В, K, RSA). Figures 63, 64. =. Planta e cataphyllum lanceolatum, natum, persistens semi-intactum, demum fibris linearibus findens; lamina si aea ad oblanceolato- ат yel ovato- elliptica, (20)25-60 cm longa, 8.5-21 cm lata; nervis diu edm lateralis (5)10-15 utroque. PS 21-58 ongus, angulatus; spatha ovata, deflexa, cucullata, m lata; spadix flavo- m diam. Fructus luteus ad albus, ovoideus. Epiphytic; stem very short, 0.5-1.5 cm diam.; roots moderately dense, descending, green, pubes- cent, moderately slender and elongate, tapered, 2~ 4 mm diam.; cataphylls subcoriaceous, lanceolate, 6–9 cm long, acuminate at apex, light green, dry- ing tan (B & K red 9/10), persisting semi-intact, eventually as linear fibers. Leaves erect-spreading to spreading; petioles 6-32(40) cm long, 5-7 mm diam., + quadrangular, occasionally D-shaped, flattened to rounded adaxially, the margins sharply raised but not winged, sharply 2—3-ribbed abaxially (sometimes rounded), the surface pale-speckled; geniculum paler and thicker than petiole, becoming fissured transversely with age, 0.8-2.5 cm long; blades subcoriaceous, elliptic to oblanceolate-ellip- tic to obovate-elliptic, less than 5 х longer than wide, long-acuminate at apex (the acumen apicu- late), acute to abruptly attenuate at base, (20)25— 60 cm long, 8.5-21 cm wide, broadest at or above the middle, the margins weakly undulate, both sur- faces glossy to semiglossy, medium green above, paler below, drying green, semiglossy to matte; midrib flat at base, becoming convexly raised to- ward the apex, pale-speckled, paler than surface above, sharply 2-ribbed and higher than broad at base, becoming convexly raised toward the apex, paler than surface below; primary lateral veins (5)10-15 per side, departing midrib at 30—40* angle, + straight to the collective vein or weakly arcuate-ascending, moderately sunken, becoming weakly sunken toward the margin above, promi- nently and convexly raised below, paler than sur- face above and below; tertiary veins weakly visible above, slightly darker than surface below; collective vein arising from near the base or about the middle of the blade, sunken above, raised below, less prom- inent than primary lateral veins, 1-3 mm from margin. Inflorescences erect-spreading to spread- ing, usually shorter than leaves; peduncle 21-58 cm long, equaling or 2—3 longer than petioles, several-angled to subterete with prominent sharp rib abaxially, conspicuously pale-speckled; spathe boat-shaped, about as long as spadix and hooded over it (directed at ca. 180° angle to peduncle), coriaceous, dark green, somewhat discolored along margins, ovate, cucullate, 2.5-6(9) cm long, 2.1- 4 cm wide, inserted at 50—70° angle on peduncle, abruptly to gradually acuminate at apex, rounded to truncate at base, decurrent on petiole; spadix deflexed, pale yellow-green to lemon-yellow to creamy-white, (occasionally brownish), sessile, cla- vate, 1.9–5 ст long, 8-10 mm diam., directed at ca. 90? angle from peduncle; flowers 4-lobed, 1.6– 2.2 mm long, 2-2 i sigmoid; 16-20 flowers visible in principal spiral, .5 mm wide, the sides jaggedly 13-16 in alternate spiral; tepals matte, conspic- mm wide, the inner margins straight to convex and turned uously papillate; lateral tepals 0.6–0.9 up against pistil; pistils emergent to 0.5 mm, white to pale greenish white; stigma 0.2 mm long, brush- like, droplets appearing 2-3 days before stamens emerge, persisting for 2 weeks; stamens emerging in a complete sequence, emerging well above tepals then retracting; lateral stamens soon followed by 3rd and 4th above pistil; flaments whitish, exserted, 0.3-1 mm long, 0.2-0 0.4 mm long, 0.5-0.6 mm wide; thecae ellipsoid; stamens, arranged in tight cluster .3 mm wide; anthers pale yellow, 0.3— pollen pale yellow fading to white, sweet-scented. Infructescence spadix 5-11 cm long; berries green- ish yellow to white (also reported as reddish), ob- ovoid, sharply pointed at apex, 4-4.3 mm long, 2.5 mm diam.; seeds 2 per berry, yellowish brown, Volume 78, Number 3 1991 Croat 617 Anthurium sect. Pachyneurium oblong-ovoid, flattened, 3.2-3.5 mm long, 1.5- 1.9 mm diam., 1.1-1.2 mm thick. Anthurium bradeanum ranges from Nicaragua to Panama in tropical wet forest life zones from near sea level to usually 450 m, rarely to 800 m. This species is distinguished by its long, sharply margined petioles, its more or less elliptic leaf blades, its cucullate, coriaceous, ovate spathe, its deflexed, short, clavate, lemon-yellow to white spadix, and yellow to white fruits. Anthurium bradeanum may be confused with Anthurium spathiphyllum where they occur to- gether, such as at the La Selva Biological Station in Costa Rica. Both species have a short clavate spadix, but 4. spathiphyllum differs in having narrower leaves with 20-30 prominently sunken primary lateral veins (vs. 5-15 scarcely sunken veins in A. bradeanum). In addition, А. spathi- phyllum commonly has three-sided petioles and a lanceolate spathe that is about twice as long as the spadix. In A. bradeanum, the spathe is ovate and as long as the spadix. The species is named in honor of Alexander Curt Brade, who made the first collection of the species in Costa Rica during 1908-1910. The species was not recollected until 1970, when I collected it in Panama at Guásimo in Colón Prov- ince The species was mistakenly called 4. cuneatis- simum (Engl.) Croat in the revision of Anthurium for Central America (Croat, 1983, 1986). That name is now considered synonymous with А. con- sobrinum Schott. CosTA Rica. oe 15 km air, 1 km W of Jaillos WNW of Quesada by air, 4 km W of Muelle San Carlos, 10%28'N, 84°30'W, Liesner 14112 (CR, K, MO); San Carlos, vic. Florencia, 0 m, Haber et al. 1854 (MO); between San Miguel and El Pilon, 600 m, 10?43'30"N, 85?00'30"W, TOR KE, (MO), i dem 36275, 36296 (MO) 13, 8 km N of Bijagua, di 0 m, Croat 36454 (B, C, DUKE, F, K, M, MBM, OF Pn S, U, US; NNE of Bijagua, 200 m, Croat ‚м (MO); Rio Peje; Ciudad Quesada-Florencia, 1 km E of Florencia, 250 m, 10%21'N, 84228 W. Hammel et al. 14032 (MO, US): Monteverde Reserve, Penas Blancas river valley, 800 m, 10?20'N, 84°40'W, Haber et al. 5116 (MO). HEREDIA: Puerto Viejo Region, Rio Sucio, 20 m, Croat 35751 (CAS, inni San Joseé- Puerto i i. М of Cariblanco, 100 m, Croat. 68361 (MO); Finca La Selva, Choistensun 1532 (MO). LIMON: between Barra del Colorado and ocean beach, 0-2 m, 10?47'N, 83°35'W, Stevens 24092 (MO); 3.5 km S of Islas Buena Vista in the Rio Colorado, 16 km SW of Barra del Colorado, 10- 120 m, 10?39'N, 83?40'40"W, d idse & Herrera 83°02'W, Grayum et al. 5810 (MO); 100-250 m, Gó- mez et al. 23663 (CR, MO); Rio Colorado, 2 km upstream from downstream branch of Caño Bravo, 5 m, 10943'N, 83°40'W, Stevens 24018 (MO). NICARAGUA. ZELAYA: 1 km N of Barra de Punta Gorda, 10 m, 11°31'N, 83°46'W, Moreno 13274 (МО); E Punta Gorda, Atlanta, landing field, 10 m, 11%34'N, 8 o & Sandino 13117 (МО); Monkey Point, 1.5 km у Сайо Е 2 10 m, 11?35'N, 83%42'W, 2 (MO); Cano Monte Cristo, Las -60 m, 11%36'N, 83*51'W, Moreno 14810 m of Co E. Serrano, 70-80 m 11°34’N, 84°21 -22'W, Sandino 3446 (MO). PANAMA. COCLÉ: Cerro Moreno, Molejón-Coclecito, 13 km NW of Cascajal, 130-250 m, 8?46'44"N, 809?31'54"W, Da- vidse & Hamilton 23715 (B, К, MO, RSA); El Cope Region, N of El Copé, Limón, Folsom 5815 (МО, PMA). COLÓN: Rio Miguel, Guásimo, Croat 9902 (MO, PMA, SEL). Anthurium brenesii Croat & R. A. Baker, Bre- nesia 16(Suppl. 1): 28. 1979. TYPE: Costa Rica. Alajuela: along Hwy. 15 between Na- ranjo and Quesada, 3.2 mi. N of Zapote, 1,560 m, Croat 46923 (holotype, MO 2682420; isotypes, COL, CR, DUKE, F, GH, K, MEXU, NY, P, PMA, RSA, SEL, U, US, VEN). Fig- ures 65, 67-69 Epiphytic; stem usually pendent, less than 30 cm long, 1.5- m diam.; roots sparse, descend- ing, green to brown, pubescent, thick, elongate, slightly tapered, 5-7 mm diam.; cataphylls mod- erately coriaceous, weakly ribbed near apex, 10- 25 cm long, acuminate at apex, drying light brown, persisting semi-intact, eventually as a reticulum of bers. Leaves spreading to pendent; petioles 11.5— 45 cm long, 3-6 mm diam., terete, rounded to obtusely and obscurely sulcate adaxially, the sur- face pale-speckled; geniculum somewhat thicker than petiole, 1.5 cm long; blades coriaceous, ob- long-elliptic to oblong, acuminate at apex (the acu- men minutely inrolled), acute to obtuse to narrowly rounded at base, 16-68 cm long, 4.3-10.3 cm wide, broadest at or near the middle, the margins weakly undulate; both surfaces matte to weakly glossy, dark green above, paler below, drying yel- lowish green; midrib bluntly angular-raised above, pale-spotted in lower half, prominently convexly raised, much paler than surface below; primary lateral veins 13-16 per side, scarcely more con- spicuous than interprimary veins, departing midrib at 35-45? angle, broadly arcuate, weakly sunken above, weakly raised and darker than surface be- 618 Annals of the Missouri Botanical Garden low; tertiary veins obscure above and below; col- lective vein arising from the base, slightly sunken above, raised and darker than surface below, equal- ly as prominent as primary lateral veins, 2-5 mm from margin. /nflorescences shorter than leaves; peduncle (10)28-43 cm long, 4-5 mm diam., 2- 3/4 х as long as petiole, terete; spathe erect and somewhat hooded over the spadix, moderately co- riaceous, medium green, sometimes tinged with red or violet-purple, ovate to ovate-elliptic, 5.5-8.5 cm long, 3.5-5.5 cm wide, broadest just above the base or near the middle, inserted at 25-40? angle on peduncle, acute to rounded at apex (the acumen apiculate), cordate and somewhat clasping at base; spadix green to red-brown to red-violet to purple, short-cylindroid, somewhat stubby, 3.2-8 cm long, 6-8 mm diam.; flowers + square to weakly 4-lobed, 1.7-2.5 mm in both directions, 6-10 flowers vis- ible in either spiral; lateral tepals 0.8- 1.2 mm wide, the inner margins straight to broadly convex, the outer margins 2—3-sided; pistils scarcely emergent; stigma oblong, 0.5-0.6 mm long; stamens held just above tepals; filaments short, translucent, soon re- tracting to hold anthers at level of the tepals; an- thers orange, (0.4)0.6 mm long, (0.5)0.8 mm wide; thecae ovoid-ellipsoid, slightly divaricate; pollen or- ange, faintly spicy-scented. /nfructescence spread- ing to pendent; spathe usually dried; spadix to 9 cm long, 2 cm diam.; berries orange (B & K yellow- red 7/5), obovoid, rounded at apex, 9-11 mm long, 5-7.5 mm diam.; pericarp thin, lacking ra- phide cells; mesocarp juicy, with sparse, thick scler- eids; seeds 2 (rarely 1) per berry, light brown, 3- 4.5 mm long, 2.2-3 mm diam., 1.8-2.2 mm thick, enveloped by an almost translucent envelope ex- tending to 10 mm long and as wide as seed. Anthurium brenesii is known from 800 to 1,560 m on the Atlantic slope in the Cordillera Volcánica and the Cordillera Central in the provinces of juela and Guanacaste in a premontane rainforest life zone. This species is distinguished by its oblong to oblong-elliptic leaf blades with rather obscure pri- mary lateral veins and the collective vein arising from the base, and by its broad spathe, cylindroid spadix, and orange berries. Anthurium brenesii is related to А. protensum, with which it shares similar leaf texture and cata- phylls and a similar altitudinal range. It differs from that species by its abruptly acuminate leaf blades, its short, broad spathe which is subcordate at the base, and its shorter, broader spadix. lt is also similar to А. prolatum, but differs in having much shorter, broader leaves and a shorter, stubbier spa- dix and an erect, more or less ovate spathe. In contrast, 4. prolatum has a spadix many times longer than broad (more than 50 x ) and а linear- lanceolate spathe. Costa RICA. ALAJUELA: between Naranjo and Quesada, 3.2 mi. N of Zapote on Hwy. 15, Croat 46923 (COL, CR, DUKE, F, GH, К, МО, NY, P, PMA, RSA, SEL, U, US, VEN); La Palma de San Ramon, Brenes 5738 (Е); Reserva de San Ramón, 800-1,000 m, 10?12'53"N, 84*36'28"W, Herrera & Mora 180 (CR, MO), Herrera et al. 411 (B, CR, МО); Cordillera Central, ca. 15 km N of Zarcero, Williams et al. 29034 (F). GUANACASTE: Santa Elena-Laguna de Arenal, 4 km N of El Dos de Tilarán, 1,000 m, 10%23'N, 84°53'W, Hammel et al. 15134 (MO), Haber & Bello 6549 (MO); Parque Na- cional Rincón de la Vieja, SE slopes of Volcán Santa María, above Estación Santa Maria, 900-1,200 m, 10947'N, 85?18'W, Davidse et al. 23448 (CR, K, МО); 1,000 m, Baker R20 (BM); 900-1,000 m, Herrera 1609 (MO); 1,100-1,300 m, Herrera 1405 (MO). Anthurium bucayanum Croat, sp. nov. TYPE: Ecuador. Guayas: Rio Chimbo, 0.9 km E of junction to El Triunfo at edge of General Eli- zalde (Bucay), 2°12'S, 79°05'W, 510 m, Croat 61597 (holotype, MO 3422020-24; isotypes, AAU, B, BM, CAS, CM, СН, К, NY, ОСА, RSA, SEL, US). Figures 66, 70-72 Planta epiphytica aut terrestris; internodia brevia, 1.5- 4 cm diam.; cataphyllum lanceolatum, persistens semi- intactum; petiolus (5.5)9-14 cm longus, 4- 10 mm diam., acute D- aut C-formatus; lamina late oblanceolato-elliptica ad oblanceolatae, (38)50-112 cm longa, (10)15-26.5 cm lata; nervis primariis lateralis (13)18-26 utroque. Pedunculus 34-48 cm longus, ca. 4 mm diam.; spatha oblongo-elliptica, 13.5-15 cm longa, 3 cm lata; spadix cylindricus infirme contractus, 8.5-10 cm longus, 6-8 mm diam. prope basim, 4-5 mm diam. prope gp CUP Baccae 4.5-6 mm longae, 3-3.2 lat Epiphytic or terrestrial; stem to ca. 60 cm long, 1.5-4 cm diam.; roots dense, ascending, pale greenish, conspicuously pubescent, short to mod- erately elongate, bluntly pointed at apex, 2-5 mm iam.; cataphylls membranous, lanceolate, ribbed, 2.5-12 cm long, acute at apex with a subapical apiculum, green, drying light brown with dense raphides, persisting semi-intact, eventually deciduous. Leaves erect to spreading; petioles (5.5)9-14 cm long, 4-10 mm diam., acutely D-shaped to C-shaped, broadly convex adaxially with the margins sharply raised and slightly turned inward, 5-10-ribbed abaxially, the surface some- what pale-speckled; geniculum slightly thicker and ип- paler than petiole, fissured transversely with age, 1-2 cm long; blades subcoriaceous (very thin on drying), broadly oblanceolate-elliptic to oblanceo- late, acute to shortly acuminate at apex (the acu- Моште 78, Митбег 3 1991 roat 619 Anthurium sect. Pachyneurium men weakly inrolled), narrowly acute to acute to narrowly rounded at base, (38)50-112 cm long, (10)15-26.5 cm wide, broadest at or above the middle, the margins broadly and shallowly undu- late; upper surface glossy to semiglossy, medium green (B & K yellow-green 4/10), lower surface semiglossy; both surfaces drying semiglossy to weakly glossy, dark green to yellow-green; midrib broadly convex to obtusely flat-raised and obscure- ly sulcate above, higher than broad and bluntly 3- 5-ribbed at base, becoming prominently convexly raised toward the apex below; primary lateral veins (13)18-26 per side, departing midrib at 55-75? angle, straight (sometimes weakly arcuate) to the collective vein, prominently convex above and be- low; tertiary veins weakly visible when fresh below, prominulous when dried; collective vein arising from near the base, weakly sunken above, weakly raised below, drying weakly raised above and below, less prominent than primary lateral veins, 3-10 mm from margin. /nflorescences erect to erect-spread- ing; peduncle 34—48 cm long, са. 4 mm diam., —4.2X as long as petiole, drying blackish brown, terete; spathe spreading-recurled, subcoriaceous, green, oblong-elliptic, 13.5-15 cm long, 3 cm wide, acute at apex; stipe 5.5-7 cm long in front, 5.3- 6 cm long in back; spadix dark violet-purple, black- ish when dried, weakly tapered, + erect, curved, 8.5-10 cm long, 6-8 mm diam. near base, 4-5 mm diam. near apex; flowers square to rhombic, 1.5-2 mm long, 1.2-1.9 mm wide, the sides + straight; (5)6-9 flowers visible in principal spiral, (8)9-12 in alternate spiral; tepals matte; lateral tepals 1.3-1.5 mm wide, the inner margins straight to concave, the outer margins 2-sided; pistils weak- ly exserted; stigma slitlike, ca. 0.3-0.4 mm long; anthers ca. 0.4-0.5 mm long, (0.4)0.5-0.6 mm wide; thecae oblong, not divaricate. /nfructescence with spathe withered; spadix 17.5 cm long, ca. 1.2 cm diam.; berries (rehydrated) oblong, apparently rounded at apex, ca. 4.6-5 mm long, 3-3.2 mm diam.; seeds 2 per berry, yellow-brown with darker punctations, oblong-ovoid to somewhat broadly el. lipsoid, slightly flattened laterally, 2.9-3.5 mm long, mm diam., 0.9-1.5 mm thick. Anthurium bucayanum is endemic to Ecuador, in the provinces of Guayas and Cotopaxi at 510- 690 m in lower montane wet forest and premontane moist forest life zones. This species is characterized by its epiphytic habit, its large very thin-drying, broadly oblong- oblanceolate to elliptic leaf blades which dry mostly dark green and have a collective vein arising from the base and running close to the margin, and by its prominently stipitate, curved spadix. Anthurium bucayanum is similar in appearance to А. сатри, with which it shares general blade shape and size and overall spadix appearance. Ап- thurium campii differs in having much more co- riaceous leaf blades and a sessile, green spadix, and occurring at higher elevations (1,200- 2,000 m vs. 510-690 m). Anthurium bucayanum also resembles А. sparreorum (also from lowland coast- al Ecuador) in overall appearance, but the latter differs in having thicker leaf blades, a subsessile spadix and perianth which becomes fleshy and con- spicuous in fruit. The species is named for the town of Bucay, near which it was first collected. ECUADOR. COTOPAXI: 2 km N of Pucuyacu near bridge over Rio San Francisco, 690 m, 0941'5, 79*06'W, Croat 57082 (MO). Guayas: Rio Chimbo, Bucay-Río Bamba, Bucay, 510 m, 2?15'S, 79*5"W, Croat 50912 (МО); 900 m E of junction to El Triunfo at edge меге Elizalde (Bucay), Los Ríos, border of Boliva d Chimborazo provinces, 510 m, 2?12'S, 79*5"W, Croat 61597 (AAU, M S, CM, GH, K, MO, NY, QCA, RSA, SEL, Anthurium bushii Croat, sp. nov. TYPE: Ecua- dor. Morona-Santiago: Cordillera de Cutucú, along trail from Logrono to Yaupi, Madison et al. 3285 (holotype, SEL 014964). Figure 73 Planta epiphytica; internodia et cataphylla ignota; pe- tiolus 8 cm longus mm diam., D-formatus; oblanceolata, ca. 100 cm longa, 18.5 cm primariis lateralis ca. 21 utroque. Pedunculu dicla ca. 2 mm diam.; spatha verisimiliter lanceolata, m longa, ca. 1.3 cm lata, purpurea; spadix oblongo- ока 12 ст longus, medio ca. 4 mm diam., colore ignoto. Baccae ignotae. Ф Description based оп dried material ошу. Ер:- phytic; stem and cataphylls unknown; petioles 8 cm long, 7 mm diam., D-shaped, with medial rib and the margins sharply raised adaxially, probably rounded abaxially; geniculum conspicuously thick- er and somewhat darker than petiole, 0.7 cm long; blades subcoriaceous, oblanceolate, narrowly acute at apex, very long-attenuate at base, ca. 100 cm long, 18.5 cm wide, broadest above the middle, the margins moderately undulate; both surfaces matte, grayish green, slighty paler below; midrib + acutely raised above, slightly paler than surface, broadly convexly raised and brownish below; pri- mary lateral veins ca. 21 per side, departing midrib at 40—50° angle, usually straight, sometimes weak- ly arcuate to the margin, conspicuously raised above and below, the same color as the midrib; tertiary 620 Annals of the Missouri Botanical Garden veins weakly visible, flat to weakly raised above and below; collective vein absent. /nflorescences erect; peduncle 36 cm long, ca. 2 mm diam., 4. as long as petiole, drying the same color as n. terete; spathe spreading-ascending, subcoriaceous, purple, drying brown, lanceolate, 8 cm long, ca. cm wide, acuminate at apex (the acumen in- rolled, 13 mm long), narrowly acute at base; spadix drying brownish, oblong-cylindroid, slightly curved, 12 cm long, ca. 4 mm diam. midway; flowers square to rhombic, ca. 4-7 flowers visible in prin- cipal spiral tepals + smooth; lateral tepals 4 mm wide, the inner margins broadly convex, the outer margins 2-sided; anthers 0.4–0.5 mm long, 0.6 mm wide; thecae slightly divaricate. /nfructescence not seen. Anthurium bushii is known only from the type collection from the Province of Morona-Santiago in Ecuador, where it was collected in the Serrania de Cutucú at 1,300 m in a premontane moist forest life zone. This species is characterized by its very long, oblong-oblanceolate blades which dry grayish green and are very long-tapered in the lower third, and by its purplish spathe and slender, cylindroid spadix which dries brown. This species is probably most closely related to A. harlingianum, which differs in having propor- tionately broader leaf blades, which are mostly truncate to shallowly cordate at the base and usu- ally dry brown. The species is named in honor of E. O. (Mike) Bush Ш, former superintendent of Selby Gardens, who accompanied Michael Madison when the type collection was made. He is now at Bermuda Bo- tanical Gardens. ECUADOR. MORONA-SANTIAGO: Cordillera de Cutucú, W opes, Logrono- Yaupi m, 2?46'S, 78%'W, Mad- ison et al. 3285 (SEL). Anthurium campii Croat, sp. nov. TYPE: Ec- uador. Chimborazo: along road between Alausi and El Triunfo, 6.9 km W of Huigra, 2?19'S, 79°0'W, 1,350 m, Croat 61562 (holotype, MO 3420768-71; isotypes, B, CM, K, NY, QCA, RSA, US). Figures 74-76, 79, 80. decis terrestris aut epiphytica; cataphyllum 22-40 cm lon anceolatum, persistens intactum vel semi- ни: eis (3)1 7-23 cm longus, (6)10-15 mm diam., D-formatus, trapeziformis vel C-formatus; lamina late ia vel oblonga- elliptica, (46)90-145 cm cm lata; nervis primariis lateralis 14-40 utro oque; nervis collectivis de basi exorientibus. Pedun culus (36)50-85 cm кебе spatha oblonga- pice (7)13-20 cm longa, 1.3-2 cm lata; spadix 12-23 cm longa, 5-9 mm diam., viridis. Baccae purpureae Terrestrial or epiphytic; stem short, 3.5-5.5 cm diam.; roots dense, upper ones ascending, lower ones descending, whitish to grayish, woolly-pubes- cent, moderately tapered, са. 3-12 cm long, 4- 5 mm diam.; cataphylls thickly coriaceous, lan- ceolate, 22-40 cm long, obtuse green, drying medium brown, persisting intact or semi-intact, eventually as coarse linear fibers. Leaves erect-spreading; petioles (3)1 7-23 cm long, (6)10—15 mm diam., thicker than broad to D-shaped to trapezoidal or C-shaped, sharply to broadly V-sulcate adaxially with the margins raised, sharply 2-9-ribbed abaxially; geniculum conspicuously thicker than petiole, 1-2.5 cm long; sheath 9-1: cm long; blades coriaceous to subcoriaceous, broad- ly oblanceolate to oblong-elliptic, rounded to obtuse or sometimes acute at apex, acute at base, (46)90— 145 cm long, 23-37 cm wide, broadest above the middle, the margins slightly undulate; upper sur- face matte to semiglossy, dark green, lower surface semiglossy to glossy, paler; midrib broadly convex at base, becoming flat-raised toward the middle and narrowly convex at the apex above, trapezoidal at apex, medium and 3-ribbed at base, becoming acutely angled and then convex toward the apex below; primary lateral veins 14-40 per side, departing midrib at 50—70° angle, straight or slightly arcuate to the collective vein, convexly raised above, less conspicuously so below; interprimary veins sometimes present, in conspicuous to almost as conspicuous as primary lateral veins; tertiary veins darker than surface and flat below, moderately indistinct; reticulate veins prominulous above and below (or not visible) on drying; collective vein arising from or near the base, equally as prominent as primary lateral veins, 1-13 mm from margin. /nflorescences erect; pe- duncle (36)50-85 cm long, (3)5-6 mm diam., 2.5-3(6)X as long as petiole, terete to bluntly or sharply multiribbed; spathe spreading, with the margins curled under, subcoriaceous, green, becoming yellowish, oblong- Е (7)1 3-20 cm long, 1.3-2 cm wide, broadest near the base, inserted at 30? angle on peduncle, shortly acuminate at apex, acute at base; spadix medium green, sessile or shortly stipitate, cylindroid to long- tapered, erect to curved, ( -23 cm long, 5-9 iam. near base, 3-4 mm diam. near apex; flowers rhombic, 1.7-2.6 mm long, 1.1-1.9 mm wide, the sides straight to smoothly sigmoid; 6-13 flowers visible in principal spiral, 5-7 in alternate spiral; tepals matte; lateral tepals 1-1.3 mm wide, the inner margins rounded, erose, the outer mar- mm Volume 78, Number 3 1991 Croat 621 Anthurium sect. Pachyneurium gins 2-sided; pistils weakly emergent; stigma ellip- soid to broadly ellipsoid; anthers tan, 0.5-0.6 mm long, 0.6-0.7 mm wide; thecae oblong, not divar- icate; pollen yellow fading to white. /nfructescence (immature): spreading(?); spathe reflexed and per- sistent, green; spadix 22-30 cm long, 1.4 cm iam.; berries light purple; pericarp with raphide cells; sends unknown. Anthurium сатри is endemic to Ecuador in southern Bolivar and southwestern Chimborazo provinces at 1,200 to 2,150 m as an epiphyte or terrestrially in either lower montane dry or pre- montane dry forest life zones (life zone uncertain). This species is distinguished by its very large, broadly oblanceolate to oblong-elliptic, coriaceous leaf blades which dry green or yellowish green, its very long, thick cataphylls, and by the collective vein which arises from the lowermost portion of the blade. Its closest ally, and the only species with which it might be confused, is А. sparreorum, which occurs at 210-250 m in Cotopaxi and Los Rios provinces and has shorter cataphylls (9-18 cm vs. 22-40 cm long), a much shorter, stubby, reddish violet spadix on a short peduncle, and a perianth becoming fleshy in fruit and extending beyond the red berry. Anthurium сатри is also similar in overall appearance to 4. bucayanum, which has very thin leaves and a long.stipitate, dark violet-purple spadix. Anthurium сатри 18 named in honor of Wen- dell H. Camp, who collected during 1944 and 1945 in Ecuador under the auspices of the New York Botanical Garden, and who was one of the first collectors of the species. Occasional herbarium specimens of a different species have been seen bearing the name “сат- pi" as proposed by A. D. Hawkes but never published. These specimens are attributable to А. andicola Liebm., an unrelated species of sect. Be- lolonchium from Mexico, and are not to be con- fused with this new Ecuadorian species. EcUADOR. pape Valle de Tablas, 1,200 m, Acosta Solís 6075 (MO, F). cHIMBORAZO: Alausi Triunfo, 6.9 m W of Maa. 1,350 m, 2*19'S, 79°0'W, Croat 61562 (B, CM, K, 1 , ОСА, RSA, US); Ri anchan, km N of Ншрга, 1,650-2,150 m, Camp 3428 (MO), E-3360, E-3460 (NY). Anthurium carchiense Croat, sp. nov. TYPE: Ecuador. Carchi: vic. Maldonado, 1,500- 1,900 m, Madison 3998 (holotype SEL- 0277098; isotype, MO-2925092). Figures 77, 78, 81 Planta terrestris aut epiphytica; internodia 6-8 cm onga, 0.8-2 cm diam.; cataphyllum anguste lanceolatum, persistens fibris зона et reticulatis; petiolus teres vel D-formatus 7-17(30.5) cm longus, 4-5 mm diam.; lamina oblonga aut oblanceolata, = cylindricus, purpureus ad porphyreus, 2.3-6.3 mm ; stamina exserta. Baccae aurantiaco-rub- rae, 6-7 mm а 3.2-3.4 mm diam Terrestrial or epiphytic; stem ca. 6-8 cm long, 0.8-2 cm diam.; leaf scars obscured by cataphyll fibers, 5-6 mm high, 1-1.3 cm wide; roots mod- erately numerous, generally descending, pale gray- ish white, drying tan, smooth, short to elongate, blunt at apex, to 14 cm long, 1-4 mm diam.; cataphylls subcoriaceous, narrowly lanceolate, un- ribbed, 5-9 c drying tan to dark brown, persisting as a reticulum of fibers. Leaves spreading; petioles 7-17(30.5) cm long (mostly less than 16 cm long), 4-5 mm diam., terete to D-shaped, rounded to bluntly and narrowly sulcate (rarely broadly and sharply sul- cate) adaxially, rounded abaxially, the surface somewhat pale-speckled; geniculum slightly paler and thicker than petiole, 0.7-2 cm long; blades subcoriaceous to coriaceous, oblong to oblong-el- liptic, sometimes somewhat lanceolate or oblan- ceolate, long-acuminate (rarely acute) at apex, nar- rowly acute at base, 30-57 cm long, 2.7-6 cm wide, broadest at or slightly above the middle, the margins flat; upper surface glossy to semiglossy, m long, acuminate at apex, green, surfaces greenish to yellowish on drying; midrib acutely raised above, prominently and sharply acute below, slightly paler than surface; primary lateral 13-16 per side, departing midrib at 50-60? angle, + straight-ascending to the collective vein, flat above, obscure and very weakly raised below; veins interprimary veins numerous, obscure above and below, drying raised and almost as conspicuous as primary lateral veins; tertiary veins visible when dried; collective vein arising from or near the base, weakly sunken above, raised below, raised above and below when dried, 3-8 mm from margin. /n- florescences erect to spreading; peduncle 17.5-3 cm long, drying 1-4 mm diam., 1.9-5.5 x as long as petiole, green (brown when dried); spathe re- flexed, subcoriaceous, pale green tinged with purple at margins, oblong-lanceolate to lanceolate, 3-6.5 cm long, 0.6-1 cm wide, broadest near the base, inserted at 45? angle on peduncle, acuminate at apex, acute at base, the margins meeting at 60— 622 Annals of the Missouri Botanical Garden 80? angle; stipe 2-6 mm long; spadix maroon to orange-brown to brownish purple to brown (B & K yellow 4/5), cylindroid, shortly stipitate, straight or sometimes curved, 2.3-6.3 cm long, 3-5 mm diam.; flowers = rhombic, 2-2.4 mm long, 1.5- 2 mm wide, the sides straight to smoothly sigmoid on drying; 4—6 flowers visible in principal spiral, 7—9 in alternate spiral; tepals purple-red, papillate, matte; lateral tepals 0.8-1 margins convex, the outer margins 2-sided; pistils .5 mm wide, the inner scarcely emergent at to dark purple; stigma slitlike, 0. forming droplets before stamens emerge; stamens emerging in a regular sequence, held well above tepals, the laterals followed by alternates in rapid succession, the laterals preceding the alternates by 3—4 spirals, arranged in a circle around the pistil; filaments translucent, exserted 0.7-0.8 mm, 0.2 mm wide; anthers orange to pink (B & K red 8/10), 0.2-0.3 mm long, 0.4 mm wide; thecae ovoid, 0.6–0.4 mm wide, slightly divaricate; pollen white. /nfructescence with spathe persisting; spadix 3.5-10.5 ст long, 1-1.7 cm diam., with the = ries scattered throughout; berries orange-red, ellipsoid to ovoid, acute at apex, (dried) 6-7 mm long, 3.2-3.4 mm diam. anthesis, raised, purple-red –0.6 mm long, A member of series Multinervia, Anthurium carchiense occurs in Ecuador in the Province of Carchi (hence the name), and in Colombia in Narino at 1,200-2,600 m in premontane wet, premontane dry, and premontane thorn forest life zones; it presumably occurs also in premontane moist forest as well. This species is distinguished by its small size, more or less oblong leaf blades which are long- acuminate at the apex, maroon to brownish spadix, flowers with exserted stamens, and red-orange ber- ries. A species similar in overall appearance, А. ob- scurinervium, differs in having a green spadix and D-shaped petioles, and generally occurs at lower elevations. COLOMBIA. NARINO: Cerro Gualcala, W slopes, vic. of Tuquerres, 2,400-2,600 m, Lehmann 5331 (F, K); La Planada, 7 km above Chucunes on road Tuquerres-Ri- 0 m, Croat 69571 (B, COL, K, M, MO, ); near о 1,800 m, Gentry et al. 5968 1 (МО). ECUADOR. CARCHÍ: Rio San ace Chical, 12 km below Меза, 1,200 m, 1?4'N, , Madison et al. 4762 (K, SEL), 4812 a LUN 1,200-1,250 m, 0*56'N, 78°11'W, Thom mpson & Rawlins 995 (CM); Gualpi Chico area Hon Awá Ethnic encampment, 1,300 m, 0°58'N, 78°1 , Hoover Е је 2752, 3212 (МО); pines in чо. a Quinchul, 0%58'N, 78°12 mpson & | mb (CM); Maldonado, 1,500- T 900. m, Madison 3998 (MO, SEL), 4264 (SEL); Rio Chilma, Machinas- Planada де Chilma, 2,350-2,450 m, 0%52'N, 78*3'W, Stein 2913 (MO); Rio Verde, Rafael Quindis mountain finca, 1,890 m, 0%52'N, 78?08'W, Hoover 1986, 2265 (MO); 1,870-2,400 m, 0%52'N, 78907 МУ, Hoover 2265 ( Anthurium cataniapoense Croat, Aroideana 9(1-4): 18-20. 1986. TYPE: Venezuela. Ama- zonas: Dept. Rio Negro, vic. Cerro Neblina base camp, Rio Mawarinuma, 140 m, 00°50'N, 66°10'W, Croat 59319 (holotype, MO 3401082-3; isotypes, AAU, B, CAS, COL, F, INPA, K, M, MY, US, VEN). Figures 82- 85 Epiphytic; stem 1.5-2.5 cm diam.; roots dense, spreading-ascending, ca. 3-12 cm long, ca. 3-4 mm diam.; cataphylls hook-shaped, to 6 cm long, acute and shortly apiculate at apex, drying dark brown to reddish brown, persisting as linear fibers. Leaves erect-spreading; petioles 6-13 cm long, 11-14 mm diam., D-shaped, flattened when young, broadly sulcate and with a prominent medial rib adaxially, 1 -3-ribbed or rounded abaxially; genic- ulum paler and thicker than petiole, 1-1.5 cm long; sheath to 5 cm long; blades subcoriaceous, broadly elliptic-oblanceolate, acute to obtuse at apex (the acumen inrolled), narrowly acute to rounded at base, 60-117 cm long, 20-50 cm wide, broadest above the middle, the margins broadly undulate; both surfaces semiglossy, green ( yellow-green 6/5); midrib acutely raised, becoming higher than broad toward the apex above, concol- orous with the surface, acutely raised and paler than surface below; primary lateral veins 11-16 per side, departing midrib at 40—60° angle, ar- cuate-ascending to the margin, weakly rounded- raised and concolorous with the surface above, obscurely raised and paler than surface below; ter- tiary veins flat and darker than surface below; collective vein arising from near the apex or absent, less prominent than primary lateral veins, 5-13 mm from margin. /nflorescences pendent; pedun- cle 32-53 cm long, 3-4 mm diam., (3)5-6(8)x as long as petiole; spathe spreading, subcoriaceous, green, ially, lanceolate-oblong, 9-26 cm long, 1-2.5 ст sometimes tinged brownish or purplish adax- wide, acuminate at apex, acute at base; spadix maroon to dark purplish violet (B & K red-purple 2/2.5), sessile, tapered-cylindroid, slightly curved, 11.5-24.5 ст long, 3-5 mm (6.5) diam. at an- thesis midway, 3 mm diam. near apex; flowers rhombic, 1.5-2.3 mm long, 1-1.6 mm wide; the sides sigmoid; 8-12 flowers visible in principal spi- ral, 6-13 in alternate spiral; tepals matte, smooth; lateral tepals 0.5-1 mm wide, the outer margins Volume 78, Number 3 1991 Croat 623 Anthurium sect. Pachyneurium 2-sided, the inner margins straight to rounded; stigma linear-elliptic, 0.4-0.5 mm long; anthers 6-0.8 mm long, 0.6-0.8 mm wide; thecae ob- long-obovoid, slightly divaricate. Fruits purple (ob- served by Anibal Castillo, Puerto Ayacucho). Anthurium cataniapoense is endemic to the northern Amazon basin in Venezuela, in the basin of the Rio Negro and the Rio Orinoco in Amazonas and Bolivar at less than 300 m. A specimen from northern Pará in Brazil, in the Tumucumaque Mountains, also appears to belong here. The plants are epiphytic in the understory of mature, well- shaded forest, occurring on white sand soils in a tropical moist forest life zone. This species is characterized by its broadly ob- lanceolate-elliptic, short-petiolate leaves which dry greenish brown to brown below and grayish brown to almost black above and have one to three pairs of primary lateral veins arising within the lower 1 cm of the base of the blade. Further characterizing the species is the slender, long-pedunculate, pen- dent inflorescence and the obtusely D-shaped pet- ioles which аге obtusely 1-ribbed adaxially and 1- 3-ribbed or rounded abaxially. This species is closest to А. loretense from the western Amazon basin in northeastern Peru, with which it shares similar leaves, hook-shaped cata- phylls, and a pendent inflorescence. That species differs, however, in having a much stouter spadix and a longer, stouter peduncle. An aberrant specimen from Bolivar (Delascio & Lopez 2812) is tentatively placed here. It differs in having relatively small leaves (ca. 44 cm x 11- 12 ст vs. 60-117 cm x 20-50 cm) and a shorter spadix. BRAZIL. PARA: Rio Part de Oeste (Irepecuru), Sampaio 5158 (RB). VENEZUELA. AMAZONAS: between -— and San P NY); Rio O m darrer dz 4248, 4250 (MO, NY); Dpt. Atures, Puerto ucho Region, road to Gavilan, vic. TI at dioc Pi Diablo, 20 km SE of Puerto ze ,.593 °27'W, Miller et al. 1596 (MO, VEN», San Pedro d кыйыгы = Milagro, Puerto Aya- cucho, 90-100 m, 6?25'N, °25'W, Castillo 2229 ; Rio Cataniapo, Las Pavas, 37 m, 06%25'N, 67°25’W, Castillo 1650 (MO); 3 km downstream from damsite, 45 km SE of Puerto Ayacucho, 200-300 m, 05%35'N 67°15'W, Steyermark et al. 122191 (VEN), 122268 (MO, VEN); Puerto Ayacucho-Sanariapo, 12 km E of highway, less than 100 m, Croat 55038 (МО); San Fer- dinando de Atabapo ое a Braun 5 (VEN); Dpt. Rio Negro, Cerro Neblina, , 00*50'N, 66?10'W, Croat apr B, CÁS, COL. F, INPA, K, M, MO, MY, VEN), Gentry & aig: 46520, 46862 (MO), Eo 15660 0 (MO), Plowman & Thomas 13672 (MO); side o oe Neblina, 200 1 m, 0949'25"N, 66%9'45"W, Nee 30826 (МО); Rio Вага, 80 m, 1?05'N, 66?25'W, Davidse & Miller 26842 (VEN); Rio Mawarimima, 0- 2 km М of Cerro Neblina Base Camp, 140 m, 0*50'N, of Rio Mawarinuma, 7 air km ENE of Puerto Chimo, 300 m, 0%50-51'N, 66?2-6'W, Da- vidse £ Miller 27244 (MO). BOLÍVAR: Río Suapure, Pica Caicara del Orinoco-San Juan de Manapiare, 202 km S of Caicara, 100-200 m, Delascio & López 2797, 2812 (VEN). Anthurium caucavallense Croat, sp. nov. TYPE: Colombia. Valle: between Cartago and San José del Palmar, 10-15 k nuevo, М 56729 (holotype, MO 3107354; isotypes, CAS, COL, JAUM, К, NY, ОСА, US; live at MO). Figures 86, 87, 91, 92. Е epiphytica aut ea шии brevia, 1- m diam.; ا‎ пен т њен uncatum; rim (9)15-34 cm longus, - сүн m diam., plus iusve. D-formatus, po ler нин tus ad | leniter sul catus, шан elevatus, abaxiler triplinervis- -quinquene subcoriacea, oblanceolata a sminusve cllipticae, (25) -90 cm longa, 0-30 cm lata, nervi primariis lateralis (9)11- troque; nervo collectivo e nervis infernis primariis lateralibus exoriens. Pe 3-53 cm longus, 1-2-plo bis longiorus quam petiolo; spatha oblonga- S2 (8)10-23 cm longa, (1)1.5-2 cm lata; spadix violaceo-purpureus, vel e violaceo-pur- Le viridis, leniter ii d an (3.7)8-21 cm longus, -10 mm diam. Baccae violaceo-purpureae. Epiphytic or terrestrial; stem to 20 cm long, 1- 4 cm diam.; roots dense, mostly descending, as- cending at uppermost nodes, pale green, densely pubescent when fresh, moderately thick, tapered, to 5 mm diam.; cataphylls coriaceous to subcori- aceous, lanceolate and hook-shaped, 7-10 cm long, acuminate at apex, olive-green turning reddish pur- ple. heavily short lineate throughout, drying brown, persisting as a reticulum of fibers. Leaves spread- ing; petioles (9)15-34 cm long, 4-10 mm diam., erect to spreading, + D-shaped, flattened to weakly age, 1-2.5 cm long; blades subcoriaceous, oblan- ceolate to += elliptic, acute to acuminate at apex (the acumen apiculate), narrowly acute to sub- truncate at base, (25)50-90 cm long, (6)10-30 cm wide, broadest above the middle, the margins undulate; upper surface glossy to semiglossy, me- dium green, lower surface semiglossy to matte, sometimes weakly to conspicuously glaucous, pal- er; midrib flat at base, becoming acutely raise toward the apex above, higher than broad at base below with 1-2 blunt ridges, becoming convexly raised toward the apex; primary lateral veins (9)1 1— 624 Annals of the Missouri Botanical Garden 20 per side, departing midrib at (45)70—90° angle, + straight, convexly raised above, more promi- nently so below; interprimary veins few, almost as conspicuous as primary lateral veins; tertiary veins flat to sunken above, weakly raised below; collective vein arising from near the apex, occasionally in the lower third or near the base, flat to scarcely sunken above, weakly raised below, 3-5 mm from margin. Inflorescences erect to spreading or pen- dent; peduncle 23-53(83) cm long, 4-6 mm diam., 1-2(5)X as long as petiole, green often tinged with maroon, terete; spathe spreading to reflexed, sub- coriaceous, olive-green tinged with maroon, oblong- lanceolate, (8)10-26 cm long, (1)1.5-2 cm wide, acuminate at apex (the acumen 5-10 mm long), acute to obtuse at base; spadix violet-purple or green tinged with gray-purple, tapered, sessile or stipitate to 7 mm long in back, (3.7)8-24 cm long, –10 mm diam. near base, 3-5 mm diam. near apex; flowers square to 4-lobed, 2-2.7 mm in both directions, the sides straight to jaggedly ru 2-8 flowers visible in principal spiral, 13-14 i alternate spiral; tepals matte, conspicuously 7 late and with droplets; lateral tepals 0.7-1.3 mm wide, the inner margins broadly convex, the outer margins 2-3-sided; pistils emergent at anthesis, densely ¡ales green; stigma slitlike to ellipsoid, 3 m long; stamens emerging in a regular sequence rae the base, the laterals preceding the alternates by 3-12 spirals, the 3rd stamen pre- ceding the 4th by 1–3 spirals, held above tepals in a circle around the pistil, sometimes partially inclined over and obscuring it; anthers white tinged with maroon, 0.6-0.8 mm long, 0.5-0.8 mm wide; thecae oblong, 0.3-0.4 mm wide, not at all or somewhat divaricate; pollen purplish violet fading to yellowish cream, sometimes faintly cinnamon- scented. /nfructescence sessile with spathe per- sisting; to 41 cm long; spadix 20-25(44) cm long, 2.0-4.5 cm diam., with the berries scattered throughout; berries violet-purple, obovoid, truncate at apex, 4-6 mm long, 3-4 mm diam.; mesocarp mealy; seeds (rehydrated) oblong, ca. 5 mm long, 2 mm diam., adhering very closely to carpel wall throughout adaxial length Anthurium caucavallense is known only from the Cauca River Valley (hence the name) or its vicinity in the departments of Antioquia and Valle in Colombia, from 900 to 1,920 m. Terrestrial or epiphytic, it occurs in steep, dry, more or less exposed areas or in the understory of disturbed forest in the premontane wet forest life zone. This species is distinguished by its hooked cata- phylls, а more or less D-shaped petiole that is usually 3-5-ribbed abaxially, its primary lateral veins which often spread at a broad angle, by its violet-purple, cylindroid and weakly tapered spa- dix, and its violet-purple berries. Anthurium caucavallense is most similar to А. glaucospadix in its leaf and petiole shape and habitat. The latter species, however, has straight cataphylls and bluish green, glaucous spadix. Other distinctive differences of А. glaucospadix are its obtuse to emarginate leaf apices and its primary lateral veins which always arise at a more acute angle. Anthurium caucavallense may also be confused with A. fendleri, a species widespread in Colombia, but that species differs in its promptly withering spathe, thinner leaves, and early-emergent fruits. COLOMBIA. Bogotá-San Agustin, Bogner s.n. (culti- vated) (M). ANTIOQUIA: Mpo. Venecia, 4.2 km E of Bo- lobolo, road to Venecia, Hacienda La Plata, 920 m, 6?1'N, 5° Zarucchi & Echeverry 4656 (В, M, МО); n etween Jerico and Tamesis, ca. 5 mi. of Tamesis, 05?42'N, 75°44’W, 1,920 m, Croat 70035 (MO); along road from Salgar to El Dauro near border with Choco, 05°59'N, 76°07'W, 1,830 m, Croat 69896 (MO). vaLLE: Calcedonia (bought; said to have been col- lected locally), Croat 51984 (COL, DUKE, F, MO, NY, MA, TEX); Cali- Buenaventura, Km 24, 1,750 m, Es- cobar 2770 (HUA); between Loboguerrero and Cisneros, along Quebrada de Los Indios, 346' 30"М, 76%42'W, 500 m, Croat 62804 (K, M, МО, NY, 05); Cartago-San José el Palmar, 10-15 km W of Ansermanuevo, 900 m, 4?40'N, 76%02'W, ш 56729 (CAS, COL, JAUM, К, MO, NY, QCA, U Anthurium colonicum K. Krause, Bot. Jahrb. Syst. 54(Beibl. 118): 123. 1916. TYPE: Pan- ama. Colón: forest around Portobelo, 5-2 m, Maxon 5801 (holotype, US). Figures 88, 9 Epiphytic; stem short, 5-7 cm diam.; leaf scars obscured by root mass and cataphylls, 1.5-1.8 cm high, 3.5-4 cm wide; roots moderately numerous, descending, green, smooth, slightly cata- phylls coriaceous, erect, lanceolate, 20-36 cm long, acute to obtuse at apex with a subapical ЕШШ which appears hooked, green, drying tan (B & yellow-red 9/10), persisting intact eventually as linear fibers or deciduous. Leaves spreading; pet- ioles 26-90 cm long, subterete, narrowly to broad- ly and bluntly sulcate adaxially, the margins blunt, not raised, rounded abaxially, the surface pale- speckled; geniculum 2-6 cm long; blades coria- ceous, ovate-triangular to oblong-ovate, short-acu- minate at apex (the acumen to 30 mm long), broad- ly to deeply subcordate at base, 39-83 cm long, Volume 78, Number 3 1991 Croat 625 Anthurium sect. Pachyneurium (9)13-44 cm wide, broadest below the middle, the margins undulate; anterior lobe 38-76 cm long, the posterior lobes -17 cm long, about as broad as long; sinus arcuate, arcuate with blade decurrent on petiole or hippocrepiform in large blades; upper surface weakly glossy to semiglossy, dark green, lower surface semiglossy to glossy, paler; midrib flat to slightly angular-raised at base, gradually becoming acutely to obtusely angular toward the apex above, prominently and convexly raised below; basal veins 2-4 pairs, usually free to base, sometimes coalesced for 1-2.5 cm; posterior rib naked, sharply turned upward on outer margin; primary lateral veins 11-17 per side, departing midrib at 50-65” angle, straight to weakly arcuate, convexly raised near the midrib, becoming sunken toward the margin above, prominently and nar- rowly raised and paler than surface below; inter- primary veins few, weakly sunken to raised above, raised below; tertiary veins obscure or weakly sunk- en above, weakly raised, darker than surface below; collective vein arising from one of the lowermost primary lateral veins or lst basal vein, sunken above, raised and darker than surface below, 3-5 mm from margin. /nflorescences spreading to pen- dent; peduncle 15-48 cm long, 6-13 mm diam., 0.5 as long as petiole, terete; spathe strongly re- flexed and recurled, markedly undulate, coria- ceous, dark green heavily tinged with violet-purple, lanceolate to broadly lanceolate, 10-18.5 cm long, 2.5-6 cm wide, broadest just above the base, in- serted at 60—65? angle on peduncle, acuminate at apex (the acumen cuspidate), acute to obtuse at base; stipe to 1 cm long; spadix green tinged with violet-purple, sessile or sometimes stipitate, ta- pered, (5)15.5-28 cm long, 1.2-3 cm diam., near base, 0.7—1.7 cm diam. near apex; flowers 4-lobed, 2.3-3 mm long, 2.8-3 mm wide, the sides sigmoid; 6-11 flowers visible in principal spiral; tepals matte, punctate; lateral tepals 1.4-1.9 mm wide, the inner margins broadly convex, turned up against pistil; pistils emergent to 1-3 mm, pale green, longer than the stamens at anthesis; stigma linear, ca. 0.5 mm long, droplets appearing ca. 1 week before stamens emerge; stamens emerging from the base, lateral stamens quickly followed by alternates in rapid succession, held above tepals and against pistil; filaments fleshy, transparent, 2-2.5 mm long; anthers 0.9-1.2 mm long; thecae ovoid-ellipsoid, slightly divaricate; pollen white. /nfructescence pendent; berries orange (B & K yellow-red 8/7.5), narrowly ovoid-ellipsoid, acute at apex, ca. 10 mm long; seeds 2 per berry, pale yellow-green, brownish at apex and base, 4–4.8 mm long, 2.3-2.5 mm diam., enveloped by sticky, amber substance. Anthurium colonicum is endemic to Panama and ranges from Veraguas to Сосје, Panamá, and Colón provinces (also in San Blas) from near sea level to 1,150 m in tropical wet and premontane rain forest. This species is distinguished by its relatively elongate, generally subcordate leaf blades and its frequently stubby spadix with exserted stamens, but especially by its sharply pointed, early-emer- gent pistils, its wavy-margined spathe, and its elon- gate, sharply pointed, orange berries. nthurium colonicum is most closely related to А. nervatum, which has similarly veiny leaf blades, a reflexed purplish spathe, and moderately acute orange berries. The latter differs in generally hav- ing more numerous (12-30 vs. 7-12) primary lateral veins, but especially by its straight to merely twisted (but not markedly undulate marginally) spathe and its blunt pistils, emerging with the an- thers, and moderately flat tepals at anthesis. In addition, the stamens of А. colonicum are white to yellowish when fresh, whereas those of 4. ner- vatum are orange at anthesis. PANAMA. COCLÉ: La Pintada-Coclecito, 600 m, 8?45'N, 80°30'W, Hamilton & Davidse 2854 (МО); Llano Gran- de-Coclecito: S of Cascajal, Continental Divide, 800-900 m, 8°45'М, 80°25'W, Knapp 1946 (MO); road to Co- clecito, 12 mi. from Llano Grande, 200 m, 8?47'N, 80287, Churchill et al. 4169 (MO); El Copé Region, Alto Calvario, El Potroso, Folsom & Collins 1541 (MO); Alto Calvario, 900 m, Folsom 2493 (MO); N of Alto Calvario, N of El Соре, Limón, 800-1,000 m, Folsom 5820 (MO); El Valle Region, Cerro Gaital, N of El Valle, La Mesa, 800-900 m, 8*40'N, 80*7"W, Knapp & Dress- ler 4912 (MO); El Valle de Antón, N of Cerro Gaital, La (МО); Portobelo- Nombre de DN hr Indio, 50 m, Croat = 02 (В, ВМ, Е, М, А, 05), y Sedo 33640 MO). PANAMA: Cerro ampo Tr 00 m, t Jefe, Ca 27096 (MO); > 0 Сагі a 6.8-12 km Am Hwy., 450 m, Croat 2500964 (В, Е, К, MO, NY, РМА, 5, Кеа ТЕХ, US, VEN), 25174, 33760, 49114 (CAS, MO), 67344 (MO), Mori & Kallunki 2285 (MO), Pos et al. 8757 (MO), Sytsma 1724 (MO, РМА); 10 m Interamerican Hwy., 350 m, Croat 33822 "i PMA): д Le ed 1 km fro Једини | Hwy. 350 m, W, de Nevers 4976 (MO , РМА), VERAGUAS: gic үз "Region Cerro Tute, near Е еј Agricola Alto Piedra, 1,070 m, Antonio 1984 (MO); Rio Santa Maria, N of school, а 5388 (MO); 15 km past school, Rio waters, 500 m, Sytsma & Andersson 4762 (MO); Rio 626 Annals of the Missouri Botanical Garden Tercero Brazo, 11 km beyond Santa Fe, 650 m, Croat 25613 (MO); Rio Dos Bocas, 11 km from school, 450 m, Croat 27544 (F, M Anthurium concolor K. Krause, Notizbl. Bot. Gart. Berlin-Dahlem 11: 606. 1932. TYPE: Panama. Colón: Rio Indio de Gatun, 0-300 m, Pittier 2798 (holotype, US). Figures 90, 94-96 Epiphytic or epilithic; stem to ca. 25 cm long, ca. 2.5 cm diam.; roots dense, spreading to erect- ascending, pale greenish white, ca. 4 cm long, 3- mm diam.; cataphylls narrowly coriaceous to subcoriaceous to nearly membranous, lanceolate, weakly sometimes ooke in the upper part, 1-ribbed near apex, 11-18 cm long, acuminate at apex with a subapical apiculum ca. 6 mm long at apex, drying dark brown to reddish brown (B & K yellow-red 3/10), persisting = intact, ultimately deciduous. Leaves erect; petioles (2)8- 16 cm long, 10-15 mm diam., erect, D-shaped to + quadran- gular, shallowly sulcate adaxially, with thin, erect margins, prominently and unevenly 3-8-ribbed abaxially; geniculum somewhat thicker and slightly paler than petiole, 1-2 cm long; blades coriaceous to subcoriaceous, usually oblanceolate to + elliptic, sometimes narrowly obovate, acute at apex, nar- rowly acute to narrowly rounded (less often round- ed-subtruncate) at base, (30)40-110 cm long, (13)18-36 cm wide, broadest above the middle, the margins moderately undulate; both surfaces semiglossy, upper surface medium green (B & K yellow-green 6/7.5), drying greenish to grayish brown; lower surface moderately to conspicuously paler greenish to reddish brown (B & K yellow- green 7/10); midrib flat with medial rib at base, becoming broadly acute toward the apex below, drying reddish brown; primary lateral veins 10- 17 per side, departing midrib at 50-65? angle, ascending = straight or slightly arcuate to near the margin, prominently and convexly raised above, less so below; tertiary veins scarcely visible above, slightly darker than surface below, prominulous and raised on both surfaces when dried; reticulate veins not visible below when fresh, weakly visible when dried; collective vein arising from near the apex, less prominent than primary lateral veins, 2-10 mm from margin. /nflorescences arching- erect to spreading; peduncle 30-75 cm long, 5- 9 mm diam., 2-4.4(11)X as long as petiole, green, sometimes purple-speckled, usually subterete, sometimes l-ribbed; spathe reflexed, subcoria- ceous, green tinged with violet-purple, narrowly oblong-ovate or oblong-elliptic 5.5-14 cm long, 1.4-3.5 cm wide, inserted at 10-70? angle on peduncle, gradually to abruptly acuminate at apex, acute to obtuse at base; spadix dark purple-violet (B & K red-purple 2/5), sometimes green tinged purple-violet, sessile or subsessile, subcylindroid or weakly tapered, 5.5-16 cm long, 7-9(12) mm diam. near base, 5-7(9) mm diam. near apex; flowers rhombic, 2-3.3 mm long, 2.4-3.5 mm wide, the sides + straight parallel to spiral, sigmoid perpendicular to spiral; 5-9 flowers visible in prin- cipal spiral, (3)6–10 in alternate spiral; tepals matte to semiglossy weakly to conspicuously and densely punctate, forming large nectar droplets on surface at anthesis; lateral tepals 2-2.3 mm wide, the inner margins straight; pistils weakly emergent, green tinged with violet-purple; stigma broadly ellipsoid, 0.5-0.6 mm long, brushlike, droplets appearing 5-8 days before stamens emerge; stamens emerg- ing rapidly in a scattered pattern throughout, some- times those at apex emerging first, held barely above level of the tepals, sometimes erect when dried; anthers purplish brown to purple-violet, 0.8— 1 mm long, 0.8-1.1 mm wide, in a tight cluster obscuring pistil; thecae oblong-ellipsoid, = divari- cate; pollen purplish (B & K purple 6/10). /n- fructescence pendent; spathe persisting at least in early fruit; spadix to 30 cm berries red, obovoid, 7-8 mm long, ca. 5 mm diam.; pericarp with numerous raphide cells; me- socarp transparent, gelatinous; seeds 2 per berry, m long, 2.5 cm diam.; white, oblong-ellipsoid, ca. 4 mm long, ca. 1.3 mm diam Anthurium concolor ranges from central Pan- ama on the Atlantic slope to northern Colombia 100 m in tropical wet forest and wetter parts of tropical moist forest. It is to be expected in Costa Rica in Limón Province. This species is distinguished by its short, cylin- drical, violet-purple spadix held erect at anthesis, by its tendency to form globular droplets on the (Chocó) from sea level to tepals, by its purplish pollen, its bright red berries, and by the petioles, which are broadly sulcate adax- ially and usually prominently several-ribbed abax- ially. Anthurium concolor is apparently not closely related to any other species, but has been confused with A. salviniae due to similarities in leaves and habit. Anthurium salviniae can be readily distin- guished by its long-tapered, pale lavender spadix. Because of its similar leaf and petiole shape, А. concolor might also be confused with A. upalaense, but that species has a long-tapered, green spadix, a longer, recurled spathe, and orange rather than purple anthers. Volume 78, Number 3 1991 Croat 627 Anthurium sect. Pachyneurium COLOMBIA. CHOCO: hill behind be pus 0-100 m D'Arcy 14217 (MO). PANAMA. BOCAS DEL TORO: uplata 377 (US). CANAL AREA: Fort Randolph, Standley 28723 (US); Fort Sherman, Standley 31018 (US); Matachin- Las Cascadas, Cowell 324 (NY). CHIRIQUÍ: Barro Colorado РС Croat 8154 (MO); Rio Provindencia, S of Colón, Tyson & Blum 3936 (MO), 3999 (SCZ). COLON: Garote (Cultivated: originally collected Ai D'Arcy), Croat 50108 (BM CAS, CM, MO, PMA, SEL, S); Portobelo Region, uv Croat 33637 (B, F, K, MO, PMA, RSA, U, 05), Rio Indio de Gatün, 0-300 m, Pittier 2798 (US). Anthurium consobrinum Schott, Oesterr. Bot. Wochenbl. 5: 66. 1855. TYPE: Nicaragua. Rio San Juan: along Rio San Juan, Friedrichsthal s.n. (destroyed; Schott Aroideae #353 serves as the type). Figures 99-101. т“ о inum var. cuneatissimum Engl., Pflan . IV. 23B(Heft 21): 176. 1905. boron 5 енен (Engl.) Croat, Ann. Missouri Bot. Gard. 70: 285. 1983. TYPE: Costa Rica. Limón: Llanuras de Santa Clara, 250 m, Donnell Smith 6811 (holotype, B; isotypes, K, US). Epiphytic; stem short, 1-2 cm diam.; leaf scars conspicuous, 0.7 cm high, 0.9 cm wide; roots nu- merous, dense, descending to spreading or as- cending, white to pale green, smooth, moderately elongate and slender, blunt, to 11 cm long, 2-4 mm diam.; cataphylls subcoriaceous, lanceolate, 5.5-15 ст long, acute to caudate-acuminate at apex, light green, drying reddish brown (B & K yellow 6/2.5), persisting intact, eventually decid- uous. Leaves erect to spreading; petioles 2.5-9 cm long, 6-10 mm diam., D-shaped, broadly sul- cate adaxially, margins blunt to sharply raised, rounded abaxially, the surface pale-speckled; gen- iculum thicker and paler than petiole, becoming fissured transversely with age, 0.5-2 cm long; blades subcoriaceous, oblanceolate to broadly oblanceo- late, long-acuminate at apex, attenuate to acute to narrowly rounded at base, 19-84 cm long, 3.5- 2] cm wide, broadest above the middle, the mar- gins broadly undulate; upper surface matte to semi- glossy, medium green, lower surface semiglossy to glossy, slightly paler, drying matte, brown, occa- sionally greenish; midrib flat at base, becoming sharply acute-raised toward the apex above, higher than broad to round-raised at base below, becoming convexly raised toward the apex; primary lateral veins 12-18 per side, departing midrib at 40-60° angle, straight to near the margin, then arcuate and joining the margin, convexly raised above and below (more so below); tertiary veins obscure above, weakly visible below; collective vein arising from about the middle of the blade or absent, weakly sunken above, raised below, 5-12 mm from mar- gin. Inflorescences erect to spreading, shorter than leaves; peduncle 20-43 cm long, 5-6 mm diam., 4-7 х as long as petiole, medium green, terete; spathe spreading to reflexed, subcoriaceous, green, sometimes tinged with maroon, linear-lanceolate to oblong-ovate, 4.5-7.5 cm long, 0.8-2.5 cm wide, inserted at 30-90? angle on peduncle, abruptly acuminate at apex (the acumen inrolled), obtuse to rounded at base; spadix green to white, becoming pinkish to orange to tan to reddish violet, subcyl- indroid, prominently curved, 5.2-9 cm long, 0.8- 1.5 cm diam.; flowers 4-lobed, 2-3.4 mm in both directions, the sides weakly to jaggedly sigmoid; 9-15 flowers visible in principal spiral, 12-16 in alternate spiral; tepals matte, greenish white turn- ing pinkish tan; lateral tepals 1.1-1.3 mm wide, the inner margins broadly convex, turned up against pistil, the outer margins 3-4-sided; pistils early emergent, exserted to 2.2 mm, sharply pointed, pale green to white; stigma ellipsoid to circular, 0.5-0.7 mm long, becoming brushlike; stamens emerging slowly from the base, held above then retracting to surface of the tepals; filaments fleshy, colorless, 0.5-1 mm long, 0.6 mm wide; anthers white, 0.7-1 mm long, 0.7-1 mm wide; thecae narrowly ovoid, slightly divaricate; pollen white. Infructescence pendent; spadix to 13 cm long; berries greenish white, reddish to purplish at apex, ovoid, narrowly acute to prominently beaked at apex, ca. 10 mm long, 5 mm diam.; mesocarp clear, juicy; seeds 2 per berry, white to greenish white, oblong-ellipsoid, 3-3.4 mm long, 2 mm diam., 1.5 mm thick, enveloped by gelatinous, transparent substance extending ca. 1 mm beyond seed on both ends. Anthurium consobrinum ranges from Nicara- gua to Panama, from sea level to 850 m (mostly less than 450 m). It is a common species from tropical wet forest and premontane wet basal belt transition life zones on the Caribbean slope. This species is distinguished by its oblanceolate blades; short, sulcate, and conspicuously sheathed petioles; whitish roots that are florescences with upturned nontapered spadix; the markedly protruding styles; and the berries that are greenish white at the base and reddish at the apex. Anthurium consobrinum is probably most easily confused with А. fatoense of Costa Rica and Pan- ama, differing from that species in its abaxially rounded petioles and nearly oblong spadix, and by its berry coloration. Anthurium fatoense has a 628 Annals of the Missouri Botanical Garden quadrangular petiole, a cylindroid-tapered spadix, and pale yellow or orange berries. This species is also similar to, and may be most closely related to, А. acutifolium, which differs in being terrestrial or epilithic and in having a long- tapered spadix and greenish yellow berries. Costa Rica. WITHOUT LOCALITY: Taylor 11593 (NY). ALAJUELA: W o ina, Molina et al. 17329 (NY, US); between Naranjo and Aguas Zarcas, NE of Quesada, Croat 46949 (MO); Zapote, Taylor 18163 (МУ); NW of Zarcero, Finca Los Ensayos, Croat 43615 (MO); 22 km NE of Quesada by air, 4 km W of Muelle San Carlos, 10928'N, 84°30'W, Liesner 14173, 14174 (MO); 4 km SE of Fortuna, then 2.5 km SW on jeep road, 400-500 m, 20°29'N, 84?43'W, Liesner et al. 15203 (MO); San Carlos, along Río Penas Blancas, Haber & Bello 1754 (MO); Cañas-Upala Road, 2-4 km М of Bijagua, 400- 600 m, Almeda & Nakai 4043 (MO), Burger & Baker 9857 (DUKE, F, MO, NY), Croat 36264 (MO), sisi 20094 (DUKE); 13.8 km N of Bijagua, 100-150 m, Croat 36423 (МО); Bijagua, Utley & Utley 3907 (DUKE, МО); Rio Zapote, km S of Rio Canalete, 100 m, Croat 36383 (МО); Miravalles, above Bijagua, 850 Gomez et al. 19065 (MO). GUANACASTE: Rio Las Ee Quebrada S ЙЕ -Quebrada Sanguijue- la, Hacienda Montezuma, 450 m, Grayum et al. 4878 (CM, CR, MO, NY); Santa Maria National Park, 600 m, 10°37'N, 85°17'W, Liesner 5072 (MO). HEREDIA: E of Rio Sarapiquí, Tirimbina, 150-250 m, 10°24'N, 84°7'W, Burger & Burger 8090 (F, MO), Proctor 32168 (LL, MO); N of Quebrada Tigre, NE of Finca El Plástico, 8 km SW of Las Horquetas, 450-550 m, 10?18'N, 84*02'W, Grayum & Sleeper 6527 (MO); Río Peje- Rio Sardinalito, Atlantic slope of Volcán Barva, 700-750 m 10°17.5'№, 84*04.5'W, Grayum & Jermy 6786 (MO); Finca La Selva, OTS Field Station on Rio Puerto Viejo, just E of its junction with Rio Sarapiqui, 100 m, Almeda et al. 5101 (CAS), Croat 44304 (MO), Hammel 11627 ( . à (DUKE, MO), Sperry 622 (MO), Wilbur 37240, 37624, Wilbur & Jacobs 34205, 34963 457 1 (NY); 4.5 km SE of bridge at Puerto Viejo, 50 m, 10926'N, 83°58’W, Stevens 13494 (MO); San José- Puerto Viejo, vic. Chilamente, 11.6 mi. N of Cariblanco, 10°27'N, 84%05'W, 100 m, Croat 68360 (MO). LIMÓN: Cahuita- Limón, 0-10 m, 9?44'N, 83?2'W, Baker & Burger 151 (BM, CAS, CM, F, MO, US); Finca Castilla, m, Dodge & Goerger s.n. (MO); 29 air km f Tortuguero, Hacienda Tapezco and Hacienda La Suerte, 40 m, 10%30'N, 83%47'W, Davidson et al. 6723, 6768 (MO), Davidson & Donahue 8413, 8461, 8733 (МО); Parque Tortugero, 40 m, 10?10'N, 83?34'W, Robles 1159, 1233 (МО); 10-80 m, 10?31'N, 83°01’ м 1518 (MO); between Barra del Colorado and ocean beach, 0- 2 m, 10947'М, 83*35'W, fee 24138A (MO): hills 2 air km SSE of Islas Buena Vista in Rio Colorado, 14 air N of Lag 83°38'W, Stevens 23654 (MO); Rio Colorado, 2 km upstream from downstream branch of Cano Bravo, 5 m, 10%43'N, 83°40'W, Stevens 24032 (MO). NICARAGUA. CHONTALES: 4 km of Santo Domingo, 280 m, 12°17'N, 85°6'N, و ا‎ А al. 3811 (MO). GRANADA: N of Mom- bacho, San José, 700-800 m, 11?49'N, 85?58'W, San- dino 2508 (MOL. MATAGALPA: NE side of m oen Blancas, Matagalpa- Waslala, Grijalva & Mor 61 (MO). Rio SAN JUAN: 20 km NE of El Castillo, Río et watershed, 200 m, Neill 3421 (MO); 3 km from conflu- ence of Rio San Juan and Rio Sabalo, Sabalo, 40-50 m, Araquistain 3229 (MO); 2 km N 11?2'N, 84°29'W, Moreno 23281 (MO); Santa Eduviges, Rio Sabalós, 70-80 m, 11°3’N, 84°29'W Moreno 23017, 23037 (MO); Cano Chontaleno, 20 km NE of El Castillo, 200 m, Neill 3332, 3376, 3420 (MO), Neill & Vincelli 3488 (MO, TEX), 3622 (МО); Cerro El Gigante, Castillito-Caño de Oro, 10 m, 10%43-44'N, 84%54'30"W, Martínez 2167 (MEXU); Rio Santa Cruz at confluence with Río San Juan, 42 m, 11?2'N, 84?24'W, Moreno 25532 (MO). zELAYA: 6.3 km S of bridge at Colonia Yolaina, SW of Colonia Naciones Unidas, 200- 300 m, 11?36'N, 84°22’W, Miller & Sandino 1100 (МО); Stevens 4819, 4826, 6411, 6412 (MO); 1.3 km SE of intersection with road Nueva Guinea- Colonia Ver- dún, road to Colonia Yolaina, Colonia La Esperanza, 180- 200 m, 11?40'N, 84?26'W, Stevens 6308 (МО); road to Colonia Yolaina and Colonia La Esperanza, 180-200 m, Ипсеш 166, 168 (MO); 1.5 km SE of Estación Experimental “El Recreo," Soza et al. 448 (MO); Colonia Yolaina, Nueva Guinea, 300 m, 11*38'N, 84?21'W, (ba quistain 3126 (MO); 1 km N of El Zapote, 6 km Colonia Verdun, 250-300 m, 11934'М, 84?24'W, ps « Vega 27907 (MO); Bluefields, N of Base Camp, Proctor et al. 26913, 27010 (LL, MICH, NY, 05); Caño Costa Riquita, 1.8 km SW of Colonia Naciones Unidas, 150- 180 m, 11?43'N, 84918", Stevens 4966, 5058 (МО); Cano Monte Cristo, 10 m, 11?33'N, 87?48'W, Moreno & Sandino 14704 (MO); Cerro Las Nubes, 10-60 m, 11937'М, 83°52'W, Moreno 14975 (MO); Las Faldas, 40-60 m, 11?36'N, °51'W, Moreno 14811 (МО); German Pomares- Las Benitas, 100 m, 11°36'N, 83*51'W, Moreno & Sandino чш 14923 (MO); 1 km before German Pomares, 10 m, 11?35'N, 83?51'W, Moreno 14840 (MO); Cano Monte id and Cano El Consuelo confluence, 10 m, 11?35'N, 83*51'W, Moreno 15033 (MO) Quebrada La Talolinga, 170 m, 11%51-52N, 84*26-27'W, Miller & Sandino 1176, 1178 (MO); Rio жо Gorda, Atlanta, “La Richard," 20-30 m, 11?32'N, °5'W, Moreno & Sandino 12967, 13008, 13114 MOS Atlanta, “La Richard," Loma San Jorge, 150-160 m, 11?31'N, 84*4'W, Moreno & Sandino 13041 (MO); Rio Rama, Cano Zamora, 10 m, 11?57'N, 84?16'W, Stevens 8861 (MO); Salto La Oropendola, 15-25 m, 1195 7'N, 84°17'W, Stevens 8954 (MO). Доре, COLÓN: between Portobelo and Madre de Dios, 1.2 mi. beyond junction to Isla Grande, 9%40'N, 79°35'W, Croat 49812 (MO). vERAGUAs: Santa Fe Region, Cerro Tute, vic. Es- cuela Agricola Alto Piedra, 600-1,400 m, 8*30-32'N, *07"W, Knapp € Kress 4357 (B, MO), Knapp & Dressler 5435 (B, MO), McPherson 10722 (MO). Anthurium coriaceum С. Don in Sweet, Hort. Brit. ed. 3. 633. 1839. Figures 97, 98, 102. PIE coriacea Graham, Edinb. Phil. J. 14: 353. 1826, on Salisb. (1796). TYPE: Brazil. Rio de Janeiro: Harris s.n. (cult. Kew). Volume 78, Number 3 1991 Croat 629 Anthurium sect. Pachyneurium Anthurium coriaceum eri icher, Pos A 240. 1837, 3.1] G. Don in t у Anthurium glaucum Schott, Wiener Z. Kunst 1829(3): 828. 1829, nomen nudum Anthurium glaucescens Kunth, Enum. Pl. 3: 73. 1841. TYPE: Origin unknown (ype m not е papi. invalidly p | synonym Pothos Maced а 1. Conc., Fl. ‘Flom: Archivos do Mus. Nac. 390 Mes TYPE: Brazil. exact location unknown. Plate 122 in Flora Flumiensis 1825 [1829] саң m Engl., Bot. Jahrb. Syst. 25: 399. 8 Brazil io de Janeiro: Gavea, Glaziou 71 (cult " Berlin) (holotype, B). Anthurium subcaulescens (Vell. Conc.) Stellfeld, Arq. Mus. Paran. 8: 177. 1950. Terrestrial or epilithic, sometimes epiphytic; stem less than 20 cm long, 3.5 cm diam.; roots mostly descending, to 5 mm diam.; cataphylls co- riaceous, unribbed, broadly lanceolate, 12-14 cm long, acute to obtuse at apex, drying straw-colored, becoming brown toward base, persisting semi-in- tact. Leaves erect-spreading; petioles (3)1 9-60 ст long, (4)10-15 mm diam., terete, sometimes ob- tusely flattened adaxially, the surface sometimes pale-speckled; geniculum moderately thicker and paler than petiole, 1-2.5 cm long; sheath 4—5.5 cm long; blades coriaceous, broadly elliptic to some- what oblong, sometimes somewhat lanceolate or oblanceolate, acute to rounded and minutely apic- ulate at apex, acute to obtuse or rounded (rarely shallowly cordate) at base, (35)60-110 cm long, (6)15-28(39) cm wide, broadest usually above the middle, the margins moderately undulate; upper surface matte to weakly glossy, medium green, lower surface semiglossy, slightly paler; midrib above flat to broadly convex at base, becoming convexly to acutely raised toward apex, prominently convex elow; primary lateral veins numerous, departing midrib at (40)50—70(80)° angle, straight to slightly arcuate to the collective vein, weakly sunken to weakly raised and obscure above, similar but more visible below, raised above and below when dried; interprimary veins almost as conspicuous as pri- mary lateral veins; tertiary veins prominulous above and below on drying; collective vein arising near the base, equally as prominent as primary lateral veins, 5-20(30) mm from margin. Inflorescences erect; peduncle 11-44 cm long, (3)5-9 mm diam., (0.2)0.5-1.7(3.6)X as long as petiole, terete; spathe erect, coriaceous, green, broadly lanceolate-ellip- tic, (7.5)10-24 cm long, (1.7)2.5-4.3 cm wide, broadest near the base, acute to acuminate at apex, acute at base; spadix pale purplish to maroon, sessile, tapered, erect, straight to slightly curved, 10-31 cm long, 8-12 mm diam. midway, 4—5 mm diam. near apex; flowers rhombic, 2.5-3.1 , 1.8-2 mm wide, the sides straight to weakly sigmoid; 13-23 flowers visible in principal spiral, 8-14 in alternate spiral; tepals matte, smooth; lateral tepals 1.6-2 mm wide, the outer margins 2-sided, the inner margins rounded; stigma oblong- ellipsoid, 0.5 mm long, depressed medially; anthers 0.5-0.7 mm long, 0.6–0.7 mm wide, inclined over and obscuring pistil; thecae ovoid-oblong, slightly or not divaricate; pollen pale yellow on drying. Infructescence erect; spathe persisting; spadix to 18 cm long, 2 cm diam., with the berries scattered throughout; berries dirty white to pale lavender, greenish at apex, obovoid-oblong, truncate at apex, mm lon 7 mm long, 5 mm diam., the pericarp with raphide cells; seeds 2 per berry, ovoid-ellipsoid, 4.5-5 mm long, 3-3.5 mm wide, 1.7-2 mm thick, flattened on one side, oblique and with a short, hooked ap- pendage at apex. Anthurium coriaceum is endemic to coastal southeastern Brazil, ranging from central Espirito Santo to Guanabara with a single outlying collection on Dos Corais Islands off the coast of Paraná State. It is expected in some of the few remaining natural areas along the coast between Rio de Janeiro and Paraná states. The species occurs terrestrially, or on rocks, less often epiphytically, in dry areas on windward slopes below 500 m This species is readily distinguished by its large, coriaceous, green-drying, more or less matte leaf lades with numerous primary lateral veins that are as conspicuous as the interprimary veins, and a collective vein which arises from near the base and is remote from the margin. The peduncle may be shorter or slightly longer than the petioles. Anthurium coriaceum is not closely related to or confused with any other species. In some re- spects (especially venation), it is like members of sect. Urospadix, which are especially common in this part of Brazil, but has involute vernation and thus shows its affinity with sect. Pachyneurium. BRAZIL. WITHOUT LOCALITY: Glaziou 7 1 (B (cult. Berlin)), 15575, 16519 (C), s.n. (F, US), Warming s.n. (C), Wid Ж 1081 (S), Lassen s.n. (С). ESPIRITO SANTO: along . BR-259 between Colatina and көбү. at km 38.5 ы junction of BR-259 and BR-101, 14.2 km E of н 19933'5, 40*36'W, 160 m, p 61907 (MO, R); Domingos Martins, Fazenda Kautsky, 550-860 m, Croat 61815 (R). GUANABARA: Barra da Tijuca, Pabst & Klein 5228 (В); Gavea, Ule 4872 (HBG). PARANÁ: n; dos Corais, Hatschbach 32538 (MO, NY). RIO DE JANEIRO without locality, Glaziou 17335 (K); Pào de Кой, аташа 429 (GUA), Casari & Vilaca гго до Геше, Кеппеду 838 (Е); Мро. Niteroi, Alto Moinas, Praia de Itaipucu, 100-250 630 Annals of the Missouri Botanical Garden m, 22%53'S, 43°07'W, Croat 53757 (CM, K, MO, NY, RSA), 53781 (MO); Itaipu, Morro das Andorinhas, Ar- aujo & Vilaca 3847 (GUA); Represa do Camorim, Ar- ajo 3751 (GUA); road to Represa do Camorim, Estrada da Boca do Mato, Vergem Grande, Rocha 151 (GUA); Cabo Frio, Arraial do Cabo, Pontal beach, Segadas-Vian na et al. 665 (NY). SAO PAULO: cultivated at Floricultura Campina Hda., Hutchison 8850 (MO) Anthurium cotobrusii Croat & R. A. Baker, Brenesia 16(Suppl. 1): 43. 1979. TYPE: Pan- ama. Chiriqui: above San Félix, along mining road, 18-27 mi. off Pan-Am Hwy., 1,200- 1,500 m, Croat 33058 (holotype, MO 2381190-91; isotypes, AAU, B, CAS, CM, COL, CR, DUKE, F, GH, К, L, M, MBM, MEXU, NY, P, PMA, RSA, SEL, US, W). Figures 107, 108. Terrestrial or often epilithic; stem usually less than 15 cm long, 1-3 cm diam.; roots descending, greenish or grayish brown, smooth to weakly pu- bescent, moderately thick and elongate, to 6 mm cataphylls subcoriaceous, lanceolate, 4.5- 11 ст long, acute at apex, green, sometimes tinged with purplish, drying light brown, persisting semi- intact, eventually as fine linear fibers. Leaves erect to spreading; petioles 15-65 cm long, 4-7 mm iam., = terete, weakly sulcate to flattened adax- ially, rounded abaxially; geniculum thicker than petiole, (1)3-5 cm long; blades subcoriaceous, nar- rowly ovate to triangular, long-acuminate at apex diam.; (the acumen downturned, weakly apiculate), shal- lowly and broadly cordate at base, 16-65 cm long, 9.5-33 cm wide, broadest at base, the margins weakly undulate; anterior lobe (13)18-30(37) cm long, the posterior lobes (5)7-13(15) cm long, di- rected upward and slightly inward, narrowly round- ed at apex; sinus arcuate to broadly parabolic when flattened, hippocrepiform when not, (4)7-9 cm deep; upper surface semiglossy, dark to medium green, lower surface glossy to semiglossy, conspic- uously paler, both drying matte to weakly glossy, yellowish green; midrib obtusely raised at base, becoming more sharply raised and then sunken at apex above, obtusely to convexly raised and more prominent below; basal veins 3—5 pairs, all free to base or 3rd—4th coalesced 4-6 cm, weakly raised to flat above, weakly raised and darker than surface mary lateral veins 4—8 per side, departing midrib at 50-60? angle, broadly arcuate, sharply to weak- ly raised or weakly sunken above, weakly raised to flat and darker than surface below; interprimary veins obscure above, flat and darker than surface below; tertiary veins obscure above, weakly visible below; reticulate veins obscure above and below; collective vein arising from one of the uppermost basal veins or one of the lowermost primary lateral veins, sunken above, slightly raised and darker than surface below, almost as prominent as primary lateral veins, 3-8 mm from margin. Inflorescences erect-spreading; peduncle (4)20-60 cm long, 4- (0.7))-1.7X as long as petiole, terete; spathe reflexed, subcoriaceous, green to pale green or green with purplish veins or heavily suffused with purple on inner surface, sometimes tinged with purple only at base, ovate to narrowly ovate, weakly boat-shaped to flat, (2.5)5-15 cm long, (1.7)2.7-4.5 cm wide, inserted at 60-90° angle on peduncle, acuminate at apex (the acumen inrolled), acute to rounded or shallowly cordate at base; stipe 1-2 cm long in front, 2-6 mm long in back; spadix green to brownish green to dark pur- ple-violet, cylindroid to weakly tapered, erect to 10 mm diam., curved, (2.5)5-25 cm long, 7-15 mm diam. near base, 3-7 mm diam. near apex; flowers + 4-lobed, 3.5-5 mm long, 3.5-4 mm wide; 3-6 flowers visible in principal spiral, 7—10 in alternate spiral; tepals sparsely punctate, more densely so near the margins, semiglossy to matte, the inner margins broadly convex; pistils emergent, matte, green; stigma linear, purple-violet or green, ca. 0.6 mm long, with brushlike papillae; droplets appearing 1— 3 weeks before stamens emerge, sometimes per- sisting when lst and 2nd stamens are opening; stamens weakly exserted, in a tight cluster above pistil; filaments whitish, soon retracting, 0.4 mm long, 1 mm wide; anthers orange to yellowish, 0.6— 0.9 mm long, 1 mm wide; pollen pale orange fading to white, yeasty-scented. /nfructescence erect spreading; spathe persisting; spadix 12-18 cm long, 1.5-2 cm diam.; berries pale orange (B & K yel- low-red 8/7.5), obovoid, darker and rounded at apex with radial ridges, 12-15 mm long, 8-11 mm diam.; mesocarp juicy, orange; seeds usually 2 per berry, oblong-ellipsoid, 6.5-7.5 mm long, 2.5-4 mm diam., encased in flattened, translucent envelope that projects up to 3 mm beyond the apex of the seed and up to 1 mm beyond the lateral margins. Anthurium cotobusii is found in Costa Rica and Panama, in premontane wet forest at elevations of 600 to 1,850 m. This species is most easly distinguished by its subcoriaceous leaves which have more or less ob- scure reticulate veins when dried, and by its prom- inently tapered, violet-purple spadix with stamens in tight clusters. In the dried state the flowers have Volume 78, Number 3 1991 Croat 631 Anthurium sect. Pachyneurium their tepals turned somewhat upwards, which, in conjunction with the clustered stamens, gives the spadix a rough appearance. Other diagnostic fea- tures are the primary lateral veins often free to the margin in the basal half of the blade, the thick, grayish roots that are densely long-pubescent on drying and the early emergent berries that are quadrangular with prominent ridges radiating from the center to the corners. Anthurium cotobrusii is an atypical member of sect. Pachyneurium, probably related to other cor- date and subcordate, orange-fruited species such as А. colonicum, A. nervatum, A. ranchoanum, and А. watermaliense. Specimens of A. cotobrusii with longer posterior lobes can be confused with А. watermaliense, but that species is distinguished by its early emergent pistils held above the tepals before anthesis, and by its often persistently long-exserted stamens. The species was placed in sect. Cardiolonchium by Croat & Baker (1979) and later in sect. Be- lolonchium (Croat, 1983), but was confirmed as a member of sect. Pachyneurium by its involute vernation. Costa RICA. PUNTARENAS: Las Cruces-Neily, Fila de Cal, Gómez 19649 (CR, МО); Rio Coto Brus, 23 km N of La Unión, Panama border, Croat 26674 (MO); San Vito de Coto Brus- Villa Neily, Fila de Cal, Cuesta Fila de Cal, 300-600 m, 8°41'N, 82°56.5'W, Hammel 14160 (MO); Cordillera de Talamanca, Tres Colinas, 1,800-1,850 m, 9°7'N, 83°4'W, Davidse et al. 25652 (МО); Cerro Fran- tizius- Cerro Pittier, Rio Canasta, 9.5 air km NW of Agua Caliente, 1,500-1,600 m, 9?2'N, 82°59'W, Davidse et al. 28366 (MO); Cerro Pan- do, 1,000-1,800 m, 8°55'М, 82?45'W, Barringer & Gómez 1617 (F); Refugio de Fauna Silvestre, Penas Blancas de Esparza, 1,000-1,400 m, 10°7’ 50"N, 84?40'25"W, Herrera et al 295 (DUKE, F, MO, TEX, US); cultivated at Las Cruces Bo- tanical Garden, 1,300 m, Croat 44452 (МО); orig- inally from Las Alturas, Croat 44387 (MO). Pan- AMA. CHIRIQUÍ: above San Félix along mining road 18-27 mi. off Pan-Am Hwy., above Chami or turnoff to Escopeta, 1,200-1,500 m, Croat 33058 (AAU, B, CAS, CM, COL, CR, DUKE, F, GH, K, Г, M, MBM, MEXU, MO, NY, P, PMA, RSA, SEL, US, W), 331474, 33152 (MO); Cerro Col- orado, 7.6 km from main road, 1,450-1,750 m, Folsom et al. 4840 (MO); 17 km NE of San Félix, 13-14 km by road NE of bridge over Rio San Félix, 1,000 m, /Vee 10699 (MO); Cerro Hornito, beyond Gualaca, 2-3 km E of Finca Linares, vic Planes de Hornito, 1,400-1,900 m, Croat 48850 (CM, COL, CR, ENCB, LE, M, MO, U); 1,750- 1,900 m, 8°41'N, 82°10'W, Croat 67983 (MO). Anthurium cowanii Croat, sp. nov. TYPE: British Guiana. Kaieteur Plateau, forest along Potaro River, ca. 1 mi. above Kaieteur Falls, 470 m, Cowan & Soderstrom 2230 (holotype, BH; isotypes, K, NY, US). Figure 103. Planta epiphytica; internodiis brevibus, caule 1-1.5 hyllum per —€— semi-intactum; petiolus , D-formatus; lamina oblanceolata, 50-90 cm е 13. 5- 30 cm lata; pedun- culus 49-63.5 cm longus, 3-7 mm diam.; spatha 9-16 cm longa, 1.5-2.5 cm lata; spadix stipitatis, (9)15- 18(21) cm longis, 6-7(10) mm diam., paa ureus-rubescens ad brunneo-rubescens. Baccae ignota Description based on dried material only. Epi- phytic; stem ca. 1-1.5 cm diam; roots dene, numerous grayish brown ‚са. 2-3 mm diam.; ceous, at least 8 cm long, acute at apex, brown (B & K yellow 4/2.5), persisting semi-intact, even- tually as fine linear fibers; petioles 8.5-13 cm long, 7-10 mm diam., D-shaped, apparently sulcate adaxially, with the margins prominently raised, probably rounded abaxially; geniculum slightly thicker than petiole, 0.8- m long; sheath (2.5)5.7–6.5 cm long; Hades coriaceous, oblan- ceolate, acute to shortly acuminate at apex, nar- rowly rounded to obtuse at base, 50–90 cm long, 13.5-30 cm wide, broadest above the middle, the margins apparently undulate; both surfaces semi- glossy to weakly glossy, grayish green to brownish green; midrib convexly raised above, slightly paler than surface, prominently convex to higher than broad below, somewhat darker than surface; pri- mary lateral veins 9-16 per side, departing midrib at 25—50° angle, straight, becoming somewhat ar- cuate toward the apex, prominently convex above and below, more so below; interprimary veins not cataphylls apparently набори present; tertiary veins prominulous, raised above and below; reticulate veins weakly raised; collective vein arising from near the apex, raised above and below, 3-11 mm from margin. /nflorescences with peduncle 49-63.5 cm long, 3-7 mm diam., 6.1– 7.5X as long as petiole, probably terete; spathe apparently erect to spreading, coriaceous, green, oblong-lanceolate, 9-16 cm long, 1.5-2.5 cm wide, broadest near the base, shortly acuminate at apex (the acumen inrolled, 7-15 mm long), acute at base; stipe 1.5-4(7) cm long in front, 0.5-1.5(2.5) cm long in back; spadix dark purple-red to brown- red, long-tapered, somewhat curved, held at 150- 165° angle from peduncle, (9)15-18(21) cm long, 632 Annals of the Missouri Botanical Garden 6–7(10) mm diam. near base, 4-6 mm diam. near apex; flowers rhombic, 4-4.6 mm long, 2.5-2.8 mm wide, the sides straight; 7-8 flowers visible in principal spiral, 5-6 in alternate spiral; tepals mi- nutely papillate, semiglossy on drying; lateral tepals 2.3-2 m wide, the inner margins broadly rounded, the outer margins 2-sided; pistils minutely papillate; stigma slitlike, 0.4–0.5 mm long; anthers 4-0.5 mm obscuring pistil; thecae oblong, not at all or slightly long, 0.5-0.6 mm wide, contiguous, divaricate; pollen fading to white. Fruit unknown. Anthurium cowanii is known only from the Kaieteur Plateau in gallery forest along the Potaro River in Guyana at 470 m, where it was reported as an epiphyte in a premontane moist forest life zone (though the area is mostly savanna). This species is characterized by its large, oblan- ceolate and short-petiolate leaf blades which dry greenish gray above and more brown and semig- lossy below. Also, the dark purple-red, tapered spadix on a peduncle much longer than the petioles is an important distinctive feature. Anthurium cow- anii is not likely to be easily confused with any other species, but А. loretense and А. vaupesian- um, from lowland western Amazonia, may bear a resemblance in the color of the leaf blades on drying. Anthurium cowanii has been cultivated at the New York Botanical Garden. An herbarium vouch- er at NY, prepared in 1948, was made from a living plant brought back as Maguire & Fanshawe 26173, from the Potaro River Gorge, but no field voucher has been located. The cultivated specimen resembles field-prepared specimens of 4. cowanii quite well, but leaf blade shape is more obovate- oblanceolate (vs. oblanceolate). he species is named in honor of Richard 5. Cowan, who collected in the Guiana region in the early 1950s and early 1960s and who, with T. R. Soderstrom, collected most of the known material of А. cowanii. GUYANA. Kaieteur Plateau, Potaro River, са. 1 mi. above Kaieteur Falls, 470 m, Cowan & Soderstrom 2230 (BH, K, (BH, US); (cultivated), Maguire & Fanshawe 26173 (NY). Anthurium crassinervium (Jacq.) Schott, Wei- ner Zeitschr. 1825: 828. 1825. Pothos cras- sinervia Jacq., Icon. Pl. Rar. 3 (1793), t. 609, Coll. 4. 122. TYPE: t. 609 serves as the type. Figures 104, 105, 109, 110. Anthurium ellipticum K. Koch & Bouché, Ind. Sem Berol. App. 6. 1853. TYPE: Venezuela. Ca- racas, Gollmer s.n. (lectotype, B; photo seen (Field Mus. Neg. #F-011915)) Anthurium rugosum Schott, Oesterr. Bot. Z. 8: 387. 1858. TYPE: Venezuela: Caracas, Gollmer s.n. (B, lectotype; photo seen (Field Mus. Neg. #F-012065)). Anthurium egregium Schott, Prodr. 475. 1860. TYPE: Schott о 486 (microfiche #13 A 10) serves as the lec ). Anthurium ‚о Schott, Bonplandia 10: 347. 1862. TYPE: Locality m a cultivated source; Fi Anthurium preussii Engl., Pflanzenr., IV. 23B( 68. 1905. TYPE: Venezuela. Carabobo: bello, Preuss 1552 (holotype, B; isotype, BM). Epiphytic, epilithic or terrestrial, sometimes co- 2.5-4 cm diam.; ascending, smooth; cataphylls subcoriaceous, 7— lonial; stem short, roots dense, 13 cm long, acute at apex, drying medium brown (B & K yellow 4/7.5), persisting + intact, weath- ering into reticulum of fibers. Leaves erect to spreading; petioles 3-33.5 cm long, 4-11 diam., D-shaped to quadrangular or thicker than broad, broadly and shallowly sulcate to V-sulcate, mm rarely narrowly sulcate adaxially, the margins acute, rarely inclined inward, flattened to rounded or 3- 5-ribbed abaxially; geniculum somewhat thicker and paler than petiole, 0.8-2 cm long; blades mod- erately coriaceous, oblanceolate, gradually acu- minate or rarely rounded at apex, acute to some- what obtuse at base, 25-142 cm long, 11-52 ст wide, broadest usually above the middle, the mar- gins undulate; both surfaces matte to semiglossy, slightly paler below, often pale-pustulate and/or with pale or dark punctations; midrib flat with a conspicuous medial rib near the base above, be- coming acutely raised near the middle, sharply raised below, sometimes 3-ribbed at base; primary lateral veins 6-14 per side, departing midrib at 40—70° angle, + straight or arcuate to the margin, prominently raised near the midrib above, then sunken and merging with margin, raised below, ibn raised and paler on both surfaces; inter- primary veins not apparent; tertiary veins flat and slightly visible above, visible and darker than sur- face below; collective vein arising in the upper Y or in the upper М of the blade, prominulous when dried. Inflorescences erect to spreading; peduncle (13)20-98 cm long, 5-12 mm diam., 1-3(5)X as long as petiole, terete, sometimes ribbed near the base of spathe; spathe spreading to reflexed, mod- erately thin, green, sometimes tinged with purple, lanceolate, 8-12.5 cm long, 1.5-2 cm wide, broad- est near the base, often decurrent at base; spadix dark purple or violet to green or green tinged with purple, tapered, sessile or stipitate to 12 mm, (6)1 2- Volume 78, Number 3 1991 Croat 633 Anthurium sect. Pachyneurium 33 cm long, 7-10 mm diam. near base, 4-7 mm diam. near apex; flowers rhombic to 4-lobed, 2.4— 3.1 mm long, 1.6-2.8 mm wide, the sides jaggedly sigmoid; 7-10 flowers visible in principal spiral, 5-10 in alternate spiral; tepals matte; lateral tepals 1.2-1.5 mm wide, the inner margins rounded, tinged with violet-purple; pistils emergent, violet- purple; stigma linear, 0.5 mm long; stamens етегр- ing from the base of the spadix, laterals emerging to midway followed by alternates in rapid succes- sion, arranged in a circle around the pistil just above the tepals; anthers pale orange (B & K yellow-red 8/7.5) 0.7 mm long, 1 mm wide; thecae ovoid, slightly divaricate; pollen pale orange, white when dried. /nfructescence pendent; berries red, ovoid, basally attached to spadix by 4 tepalar fibers, 7.6-8.2 mm long, 3-3.5 mm diam. on rehydra- tion; pericarp studded with raphide cells; seeds 1- 2 per berry, pale yellowish, with raphide cells, ellipsoid, 3.6-4.4 mm long, 1.6-2.2 mm diam., 0.8-1.6 mm thick, with a mucilaginous apical ap- pendage. Anthurium crassinervium ranges throughout northern Venezuela in the states of Aragua, Cara- bobo, Falcón, Lara, Mérida, Miranda, Portuguesa, Táchira, Trujillo, Yaracuy, Zulia, and the Distrito Federal, from sea level to 1,800 m. It has also been found in Colombia in the departments of Ces- ar, Guajira, Magdalena, and Santander from 400 to 1,600 m, and on Curagao in the Netherlands Antilles at 350 m. The species is ecologically quite variable, occurring in tropical moist, tropical dry, premontane wet, premontane wet (transitional to warm), premontane moist, premontane dry, sub- tropical dry, and subtropical thorn forest transi- tional to subtropical dry forest. Although quite variable in size, А. crassiner- vium can be recognized by its lanceolate, usually acute at the base blades with undulate margins and often with pustular or pale to dark punctations on the abaxial surface, by its lanceolate and sometimes early-deciduous spathe, tapered spadix, and red berries with the seeds attached apically to the car- per wall Bey a mycilaginous appendage: is that should be men- tiohed are tentatively ей here. Two, Bunting & Holmquist 4327 and 4328, from Bolivar State in Venezuela (the only collections known from Bo- livar), are quite typical for the species except for the unusually large leaf blade size and the unusually high number of flowers in the primary spiral (11 flowers per spiral). Another collection of note is Bunting 4397, from Táchira, with 13 flowers in the primary spiral, and a spathe 5 cm broad. The overall size of the latter plant is reported to be unusually large. Several collections from lower el- evations north of the Cordillera de la Costa, es- pecially those from near Cata in Aragua, Venezuela (Bunting 2197, 2198, 2199, 2200, 4353, 13501, 13508, 13509, 13510, 13511), are unusual in having the petioles narrowly sulcate and sometimes with the margins convergent adaxially. While this material is interpreted by Bunting (pers. comm.) as А. ellipticum, the original description of А. ellipticum makes no mention of petiole shape, and subsequent illustrations by Schott (1984) and En- gler (Engler's Araceae Illustrations No. 249) show the petiole as shallowly and broadly sulcate with inconspicuous margins adaxially. Petiole shape in A. crassinervium is highly variable, and the Ara- gua specimens exhibit just one of the extremes. Curiously, in his 1905 revision, Engler illustrated Anthurium ellipticum (fig. 22) as having a prom- inently sulcate petiole with the margins curved inward. There is no indication that А. ellipticum is out of the range of variation in А. crassinervium; however, the illustration of А. ellipticum in Das Pflanzenreich represents a plant like those men- tioned above from near Cata. These may represent a distinct taxon, perhaps a subspecies of А. cras- sinervium. In addition to the petiole cross-section differences, these plants also have longer, more tapered spadices, and are always epilithic or ter- restrial (often in sand). This taxon, if recognized, would probably include А. preussi, described from a similar area near Puerto Cabello, and perhaps also A. fontanesii, but not A. ellipticum. Petiole shape in А. crassinervium is highly variable, and this represents just one of the extremes. Anthurium crassinervium is closely related to А. wagenerianum, but the latter generally has a short, stout spadix and a short, broadly lanceolate spathe. Its leaf blade margins are minutely undulate in dried specimens, instead of wavy, and are api- cally cuspidate, instead of acute. Furthermore, the lower leaf epidermis in dried specimens of А. wag- enerianum is not highly light-reflective at low mag- nifications, and the concentric epidermal cell pat- tern observed in 4. cras The fruits of 4. wagenerianum are distinctive as well. See that species for further details. Engler distributed several specimens that he determined as А. wagenerianum, but which have a markedly tapered spadix like that of А. crassinervium ап therefore have been included there. The range of А. overlaps that of another species which possibly could be confused with it, namely А. fendleri, but the latter can be distinguished by its usually thinner blades that are sinervium is not apparent. 7 C ooLlULCI ULULTIL Annals of the Missouri Botanical Garden usually narrowly rounded to subcordate at the base, and by its thin, caducous spathe. Anthurium bonplandii subsp. виауапит might also be confused with А. crassinervium, as both have similarly shaped leaves that are pustular or punctate abaxially. The petiole of the former is C- or D-shaped and rounded abaxially, while the latter usually has quadrangular petioles which are usually ribbed abaxially. The berries of А. bon- plandii subsp. guayanum are rather dry upon rehydration, and its seeds are attached by a band of fibers. Anthurium crassinervium, in contrast, has the more typical seed for sect. Pachyneurium, attached to the berry by a mucilaginous appendage. Pothos crassinervia Jacq. was depicted in Cur- tis's Botanical Magazine (t. 2987) by J. D. Hooker in 1830 and was said to have come from Demerara (Guyana); however, this was a living collection sent by James Fraser, Esq., and was possibly not locally collected. To date there are no indications the species occurs in the Guianas. Lectotypification has been necessary in the cases of A. crassinervium, A. ellipticum, A. rugosum, A. egregium, and А. fontanesii. In the first case, Jacquin clearly described his new species, Pothos crassinervia, in 1790, while his dilatation, here lectotypified, was not published until 1793. In the second case, Koch mentioned no specimen in his original description of A. ellipticum, but did say that it occurred outside of Caracas, and that it was "brought to Europe by the famous Moritzi from Caracasan regions” (translated from Latin by Dan Nicolson, pers. comm.). It is unclear whether he was referring to a Moritzi collection, but it is likely that he had seen the May 1852 Gollmer collection that is here lectotypified, perhaps after its having been brought to his attention by Moritzi. In the third case, Schott described А. rugosum in 1858, citing no specimen. However, in 1860, in his Prod- romus, he mentioned both a Gollmer and a Wag- ener collection; the Gollmer collection has here been accepted as the lectotype. In the fourth case Schott described А. egregium in 1860, again citing no specimen. His plate of drawing 482 deposited at W and represented by microfiche number 13A-7 in his [cones Aroideae et Reliquiae, serves as the lectotype of this species. Lastly, Schott cited no specimens in his original description of 4. fonta- nesii, citing only that it was cultivated at Schon- brun. His illustration 496 serves as the type. NETHERLANDS ANTILLES. CURAGAO: 350 m, Wilde 6A (WAG). COLOMBIA. CÉSAR: Sierra Nevada de Santa Marta, SE slopes, Donahui, along trail to Chimencumena, 1,300 m, Plowman & Davis 3699 (COL, СН); road from La Paz to Manaure, 460 m, Cuadros & Gentry 3468 (MO), 3465 (MO), Guajira, Serrania La Macuira, SE edge of Cerro Huararech region, 550 m, Sugden 63 (K); Cerro Manzano, 500-760 m, Saravia 2424 (COL, US); Gua- saira, 600-750 m, Saravia & Saravia 3561 (COL, US); Palua, 650 m, Sugden 216 (COL, FHO, K). MAGDALENA: Sierra de Pireja, E of Manaure, Hacienda Nuevo Hori- zonte, El Podrido, 1,550-1,600 m, Cuatrecasas & Кот- его C. 25375 (US); Mpo. Santa Marta, Constante- Pue- blito, 100-400 m, Romero C. 8023 (COL). SANTANDER: Rio Lebrija, NW of Bucaramanga, 400-700 m, Killip & Smith 16298 (US). VENEZUELA. Cult. Berlin, Engler 193 (GH). ARAGUA: i ‚ Cata, Bunting 493 (МО); 50 m, Bunting 4355 (МУ Cata, abov reps de Cata, АЫ 2197 (MY), 2198 (NY), 2199, O (MY); Dtto. Girardot, Cata-Catica, 10-100 m, de 4904 (MY); Henri Pittier National Park, « 50 m, Carnevali et al. 2329 (VEN); Ocumare- Turiamo, Bunting & Holmquist 1978, 1979 (NY); Магасау-Ос- umare, 0.5-2 km NW of Rancho Grande, hillsides of Cordillera de la Costa, 1,200 m, Bunting 2015 (МУ); near Rancho Grande, 1,200 m, Bunting 1966 (NY); Cerro La Mesa, 1,500-1,900 m, Bunting 4721 (NY); La Cumbre de Rancho Grande trail, 1,200-1,300 m, 10°21'N, 67°39'W, Bunting 3266 (MY), Davidse et al. 16732 (VEN); Cata-Cuyagua, 8-9 km E of Cata, 400 m, 10?29'N, 68*42'W, Bunting et al. 13501, 13508, 13509, 13510, 13511 (MO); Colonia Tovar-El Limón, Bunting 2146 (NY). BOLIVAR: Altiplanicie de Nuria, 500 m, Bunting & Holmquist 4327, 4328 (MY). CARABOBO: Puerto Cabello, Pre B, BM), Bunting 4353 (NY); Quizandal, Trujillo 9581 (MY). DISTRITO FEDERAL: Caracas Higuerote, Bunting 3275 (МУ); Caracas-La Guaira, Rose & Rose 21 753 (US); 25 km E of La Guaira (MO), Bunting 2144-1 (МО); 2 km E Bunting & Steyermark 2144-2, 2144-3 (NY); Maique- tia, 40 m, Andre 289 (K); Silla de Caracas, 1,300 m Elias 475 (F); Caracas Botanical Garden, 870-980 m, Berry s.n. (MO), Braun 9 (VEN), Croat 38338 (MO, PMA, SEL, US), 54412, 54415 (MO), 54416 (CM, MO), 54418 (CM, F, МО); Cerro El Avila, Quebrada Chacaito, 1,600 m, Manara s.n. (VEN); Rio Macarao, Montes 74, 279 (VEN); Rio San Julian, just above Carrabelleda, 10- 300 m, Bunting 2050 (NY), 2050 N, 2050 J (NY). FALCÓN: Cerro Socopo, 1,200-1,560 m, 10?29'N, Nacional Quebrada de la Cueva El Toro, 600-900 m, 10950'N, 69?7'W, Liesner et al. 7864 VEN); Sierra de San Luis, Curimagua-San Luis, 1,300- m, Steyermark 99149 INT. US, VEN); Sierra de San Luis, Coro-Churumagua, m, Bunting 2844, 2846 (NY); Dtto. Bolivar, El сен 340 m, Ruiz et al. 481 (МЕМ); Dtto. Silva, Cerro Chishiritisbe: above Soledad, 200 m, 10%51'N, 68?20'W, Steyermark & Manara 110812 (MO, VEN); Dtto. Zamora, Cerro Mam- ostal, 400 m, 11?27'N, 69*17'W, González 1068 (MO, VEN). LARA: Sanare- Yacambü, road through Parque Na- cional Yacam 69°36' 70°01'W, Croat 6062 1 (CM, К, MO, RSA); Dtto. Jiménez, Alto del Viento- Cerro Pando, 1,000-1,700 m, 9°39-42'N, 69*34-36' W, Davidse & González 21186 n dirt road 17 km E of Duaca, 1,200 m, 10?17'N, 2'W, Croat 60617 (AAU, MO, VEN); Dtto, Torres, iin Zulia, Agua Linda, 5 km E of El Venado, 1,300 m, Bunting & Fucci Volume 78, Number 3 1991 Croat 635 Anthurium sect. Pachyneurium 13475 (NY); 1,100-1,200 m, Bunting & Stoddart 9753 (NY). MERIDA: Mérida-La Azulita, La Chorrera (waterfalls N of hwy.), between Las Cruces and La Calera, 8%35'N, 71°16’ w. Croat 54794 (МО); 28 km W of Mérida- El Vigia intersection, 1,120 m, Bunting 2602, 2602J (MY, NY); 9 km above Plaza Bolívar in La Azulita, SE of La Azulita, 1,590 m, 8°44'N, 71°26'W, Croat 54856 (MO); Mérida-Lagunilla, ca. 2 km S of Mérida, near intersection with road to Га Azulita, Bunting 2740 (MY, МУ); Mé- rida-El Vigia, 26 km above El Vigia, 850 m, Bunting B (MY, NY); Cano Zancudo-La Azulita, border of Cano Blanco (Puente Hierro), Bunting 4382 (MY); 6 km above Cano Zancudo, near bridge across Cano Ron (sec- ond bridge), Bunting 2798 (MY); El Vigia- Tovar, Km 5, 3 km S of junction to Mesa Bolivar, 5 km N of Santa Cruz de Mora, 450 m, 8?26'N, 71?44'W, Croat 54877 (B, MO); 5 km М of El Morro, 1,800 m, 8*25'N, 71*10'W, Hahn & Grifo 3320 (MO); Montana Salinas, Mesa Bo- livar, 1,300-1,400 m, Bernardi 633 (NY); Tovar de rida-Panamerican Hwy. m above Zea, 975 m, qe 2574 (NY). MIRANDA: e Caracas- Higuerote, of Osma, Bunting 3298 (CM, MO, NY); Morros de Эр Guairita 880-1,000 m, Berry 1829, 1897 (VEN); Quebrada Garita headwaters, N of Mérida, Vista Linda, 1,000 m, 10°26'N, 66°49'W, Steyermark & Berry 111961 (VEN). PORTUGUESA: 15 km E of Chabasquen, 67 km NNW of Guanare, 1,450-1,520 m, 9?26'N, 69*54'W, Steyermark et al. 126795 (MO). TÁCHIRA: Lobatera- Urena, 13 km NE of Lobatera, near La Vic- toria, 3,650-3,750 ft., Bunting 2352 (MY, NY); 1,216- 1,250 m, Bunting 2351, 2351 J (MY, NY); UA San Antonio, 11 km W o 1 ; of Las Delicias, Bunting 4397 (MY); 6 km e Las Delicias, 5,500 m, Bunting 2313B (MY), mus 23 13A (NY); San Cristóbal-Santo Domingo del Táchi асыга, Agua m, 7?28'N, 72°09'W, Liesner & González 10865 (VEN); Pregonero-La Fundación, 19 km S of Delicias, 45 km SW of San Cristóbal, 19 km N 1,300 m, 9?42'N, 72*25'W, Croat 55031 (МО); San Cristóbal-Chorro del Indio-Caño Seco-La Florida, 1,100- 1,200 m, Bunting 11645 (NY); 1,100 m, Вее 13403 (NY); San Cristóbal-Cordero, San Rafael-Cordero, 900- 1,000 m, Bunting 4814 (NY); Maracay-Choroní, 19 km from Maracay, along S slope of northern cordillera, 1,280-1,300 m, 10?18'N, 67°24'W, a 54499(CAS, F, MO); La Fria, at Las Parés: along Caño Aguas Calien tes, Iv bar m, Bunting 13390 (NY); La Grita- m H . 7-8 km N of La Grita, 1,600-1,7 Bunting. 11668 (NY); Dtto. Ayacucho, Buntin ^ Cha. cón 4983 (MO, NY); Dtto. Junin, Villa Paez-Betania, 2,050-2,350 m, Bunting 4958 (NY); Dtto. Cárdenas, above Palmira, 1,800 m, Bunting 4847 (МО); Dtto. Cárdenas, on the outskirts of Cordero, 1140 m, Bunting 4864 (NY); above Palmira on the outskirts of San Cris- tóbal, 1,800 m, Bunting 13451, 13452 (МУ); Dtto. Lobatera, La Cazadora, 1,600 m, 7°55’М, 72?18'W, van des Ortiz 5614 (MO); 2,000 m, van der Werff & Ortiz 5 km NE "y N of Salom, NE of Nirgua, W of Valencia, 1,200 m, 10°11’N, 68°30'W, Croat 54652 (MO, RSA); Cerro La Chapa, N , Bunting 2 (NY). zULIA: Perijá, 1,175 m, yeh 1926 6 (US); da d de Perijá, Río Omira-kuna (Tumuriasa), near Colombian border, SE of Pishikakao and Iria, 1,470-1,56 nek et al. 91156 (MO, VEN); i de trail, 1,200- 3266 (MY) 1,300 m, 10°21'N, 67°39'W, Bunting Anthurium crenatum (L.) Kunth, Enum. PI. 3: 75. 1841. Pothos crenata L., Sp. Pl. ed. 2: 1373. 1763. TYPE: Virgin Islands: St. Thom- as, C. Plumier illustration published by J. Bur- man, Plant. Amer. Fasc. t. 39. 1756 serves as the type. Figures 106, 111, 115 ide dnd acaule var. portoricensis Kuntze, O. Ktze. . 2:738. 1891. TYPE: Puerto Rico, Kuntze s.n. (holotype, MO; isotype, K). Anthurium acaule var. s Engl., Bot. Jahrb. Syst. 25: 362. 1898. TYPE: Puerto Rico, Sintenis 1582 (holotype, B; чё К). Epiphytic ог epilithic, sometimes forming dense clumps; stem to 9 cm long; roots moderately dense, spreading to descending, pale green pubescent, elongate, blunt, 3-4 mm diam.; cataphylls sub- coriaceous, narrowly lanceolate, weakly 1-ribbed near apex, 9.5-13 cm long, narrowly acute at apex with a subapical apiculum, light green, drying reddish brown (B & K yellow 5/2.5), persisting semi-intact, eventually deciduous. Leaves erect to spreading; petioles 2.5-29.5 cm long, 4-17 mm diam., sharply to bluntly D-shaped, shallowly and broadly sulcate adaxially, the margins sharp but not raised, rounded to 3-ribbed abaxially, the sur- face weakly pale-speckled; geniculum slightly thicker and paler than petiole, 0.6–1.3 cm long; blades moderately coriaceous, broadly to narrowly oblanceolate or elliptic, sometimes + oblong, acute to acuminate at apex (the acumen apiculate), acute (rarely weakly subcordate) at base, (18)23-100 cm long, 4-28 cm wide broadest near or above the middle, the margins sinuate to undulate; upper surface semiglossy, medium to dark green, lower surface weakly glossy to semiglossy, paler; midrib acutely angled at base, gradually becoming weakly sunken near the apex above, prominently higher than broad at base, becoming prominently convex toward the apex below, paler than surface; primary lateral veins 8-18 per side, departing midrib at 636 Annals of the Missouri Botanical Garden (30)40-75° angle, + straight to arcuate-ascending to the margin, raised above, convexly raised below; tertiary veins weakly raised above and below, darker than surface, drying prominently raised; reticulate veins weakly visible above, prominulous and slight- ly darker than surface below, conspicuously raised when dried; collective vein arising in the upper уз of the blade or absent, 5-12 mm from margin. Inflorescences erect to spreading or pendent; pe- duncle 19-72 cm long, 3-7 mm diam., 1.6-14.4 х as long as petiole, green, terete to ellipsoid in cross section; spathe erect to spreading, subcoriaceous, sometimes flushed pink or maroon, usually light green (B & K yellow-green 7/10), linear-oblong, with the margins curled under, 4-14.5 cm long, 0.4-1.5 ст wide, acuminate at apex, obtuse to rounded at base; spadix reddish or purplish violet to maroon (dull, dark bluish purple fide S. Mayo 1982), long-tapered, somewhat curved, 6.5-25 cm long, 3-8 mm diam. near base, 2.5-4 mm diam. near apex, broadest at the base; flowers scented like rotting fruit prior to anthesis, rhombic to square, 1.5-3 mm long, 1.6-2.2 mm wide, the sides prom- inently and smoothly sigmoid; 5-8 flowers visible in principal spiral, 8-12 in alternate spiral; tepals matte to weakly glossy, densely, minutely papillate; lateral tepals 0.6-1.4 i straight to rounded, the outer margins 2-sided; pistils emergent 2-3.5 mm, glossy, green; stigma 0.2-0.4 mm long, brushlike; stamens emerging in a slow, regular sequence from the base, the laterals preceding the alternates by 10 spirals; filaments translucent, broader than anthers, ca. 0.5 mm long; anthers 0.3-0.6 mm long, 0.5-0.8 mm wide; thecae oblong-ellipsoid to oblong-ovoid, weak- ly divaricate; pollen yellow-orange, fading to white (B 4 K yelow 8/5). Infructescence spreading- pendent; spathe withered, usually deciduous; spa- dix to 2.5 cm diam.; berries bright red K red 4/5), obovoid, glossy, 8-10 mm long, 5-6 mm diam.; mesocarp mealy, white, sweet, becom- mm wide, the inner margins linear, ing acrid; seeds tan to pale yellow, oblong-elliptic, 3.8-6 mm long, 2-2.8 mm wide, 1.5-1.8 mm thick, with a short gelatinous appendage at apex. Anthurium crenatum is endemic to the north- western part of the West Indies, where it is known from the Dominican Republic, Puerto Rico, and the small islands to the east of Puerto Rico, namely Vieques, Culebra, St. Thomas, St. John, Tortola, and St. Croix (all but Vieques are part of the U.S. Virgin Islands). It occurs principally in humid for- ests from near sea level to 900 m. “Dry evergreen woodland" is noted for Mori & Woodbury 16999, from St. John. This species is characterized by its usually nar- rowly oblanceolate, short-petiolate blades which have conspicuously raised tertiary and reticulate venation on drying and sometimes even on live plants, and by its long-pedunculate inflorescence, which is usually pendent, with a long-tapered, dark purplish violet to maroon spadix and a promptly withering, linear-oblong spathe. nthurium crenatum is not easily confused with any other species and is the only member of sect. Pachyneurium in the areas where it occurs. It is perhaps most closely allied to А. crassinervium of northern South America and Сигасао. Anthurium crenatum has been confused since Schott's time with А. acaule (Jacq.) Schott. See the discussion of the history of this taxonomic con- fusion by Mayo (1982), whose paper dealing with the West Indian bird's nest Anthurium species demonstrates that Æ. acaule is actually a com- pletely different species endemic to Martinique and is probably a member of sect. Urospadix, with most of its relatives in Brazil. Howard (1979) er- roneously associates the name А. асаше with А. hookeri. BRITISH VIRGIN ISLANDS. TORTOLA. Dotys, Fishlock 368 (GH, К); High Bush, 375 m, Britton & Shafer 811 (МУ, US). Dominic 2 Пау выс. Parque Nacional Los Haitises, Monte Plata 0 m, Garcia et al. 664 (NY); Гай, 100- 500 m, а 2601 (US); Miches-El Seibo, at crest of 5. Oriental, 600 m, Gastony et al. 694 (СН, NY, S); Río Comatillo, Bayaguana, 150 m, Liogier 22465 (BSD): Río San Juan, T 1125 (05); барана del Маг, 100 m, НИ das Y, US). LA ALTAGRACIA: 21 km W o , 200- a m, Pimentel 83 (JBSD, NY); “El be here old road parallels Rio Dueye), N of Boca de Yuma, 18°24'N, 68°37'W, Zanoni et al. 10685 JBSD); La С olorada, near headwaters of Rio Llano, Km 22 on road from Santana to La Colorada, NW of hwy., 18°46'N, 68°44'W, Mejia € Ramirez 11133 (JBSD, ~ 68°50'W, 100 m, Mejia & Ramirez 11229 (JBSD); 21 km W of Higuey, road to El Seibo, Loma del Penon, 200-350 m, Zanoni & Mejia 17332 (JBSD); 0.5 km from Otrabanda on road from Cruce de Los Islenos and Nisbon at Arroyo Yaoya, 18?39'N, 69397, Zanoni et al. 18967XX (JBSD). SAMANA: 9 km NE of Samana, rd. to La Laguna, and Loma Pan de Azucar, 50 m, 19?15'N, 69°16'W, Gentry & Zanoni 50600 (MO); Loma Pan de Azucar, E of La Laguna, NE of Santa Barbara de Samaná, 19?15.5'N, 69*18'W, 100 m, Zanoni & Gentry 34116 (JBSD); Peninsula de Samaná, Loma El Fronton, 19°17'N, 69°10'W, 100-256 m, Zanoni et al. 29346 (JBSD, MO, NY, USD); Sanchez, Rose et al. 4396 (NY, US), Wright et al. 536 (GH, US); Las Canitas Mts., 330 m, Taylor 7 (GH, 05); 3.5 km E of Las Terrenas, 5 km 5 of Callejon, 1991 18'N, 69930 77 , Zanoni & Mejia 17634 (JBSD). SAN CRISTÓBAL: 8.5 km from Comatillo on road to Cruce de Pilancon and Pilancon, 18%52'N, 69°37'W, Zanoni & Mejia 16411 (JBSD); Río Capita (de Trinidad), 15 km from Comatillo on road to Sierra de Agua and Trinidad Mejia, 18*50'N, 69°34'W, Zanoni & Mejia 16318(JBSD, O, USD). siEBo: Cordillera Oriental, along rock road Volume 78, Number 3 1991 Croat 637 Anthurium sect. Pachyneurium 9.2 km SE of Miches, along Río Yeguada Arroyo San- tiago, 100 m, 18%55'N, 69%9"W, Croat 68550 (МО); Arroyo Las Cabirmas, 5 km S of Miches-Nisibon Hwy., о Zanoni et al. 15946 (JBSD); El Salado- Otrabanda, Hig- uey, pia 12243 (СН, K, 05); Hato Mayor del Rey, 400-500 m, Allard 13632 (05); Hato Mayor-Sabana del m Guamira, Jimenez et al. 4129 (NY); Hato Mayor del Rey, 18?49.5'N, 'W, Zanoni et al. 11966X (JBSD); Higuey, Howard & Howard 9790 (BM, GH, NY, US); 11.7 km W of El Valle, trail to Trepadora Alta, 18*58'N, 69?28'W, Zanoni et al. 21283 (JBSD). Ha- WAII. Oahu, cultiv ied at Foster Gardens, Wahiawa, Croat 45013 (МО), 45016 (B, MO). Puerto Rico. Adjuntas, Las Cruces, Sintenis 4205 (MO, NY, P); Arecibo-Utu- & Cowell 345 (NY, US), Eggers 1142 (P), oo & Мт вв 846 (МУ, 05); Вауатоп, Агзепе .n. NY), Stevenson 384 (US), Underwood & Griggs ae NY. US); tli Finca Sanchez, Hioram i m, Acevedo & Chinea , Webster & Miller 8607 (US); Candelaria, oe et al. "279 (NY, US); W of Candelaria, Km 2 on Route 2, 80 m, 18?24'N, 66°14' I то 5762 (МО, SEL); Cayey-Guayama, Underw р 13 (NY); Guayama Road, Quebrada Arriba, Goll et al. a Juana Diaz, Penas de las Cuevas, Britton Я Marble 2296 (NY, US); Laguna Tortuguero, Wagner 286 (U); 2 mi. N of Mameyes, 50 m, Hartley 13373 (U, US); Maricao, ves 479 (B yagüez, Heller 628. Wagner 203 (GH); San Juan, Hioram (NY, US), Hornbeck s.n. (W), is bs s s.n. (K, NY), Riedel s.n. (P к ened Britton Britton 7270 (NY); ا‎ & Heller 1267 (NY, US); Utuado, Britton 5224 (NY); Yauco, Garber ~ ~ 104 (CH, K, NY); Dorado Beach Forest, Dorado, Dorado Beach Hotel, 10 m, Croat 60851 (B, CM, MO, RSA); El Dorado forest, N coast, 18*28'N, 60*16'W, Luteyn & Luteyn 11466 (NY); Río Abajo State Forest, Hwy. 621, near end of asphalt road, 360-390 8?19'N, 66°40'W, Croat 60860 (MO, US); Rio Piedras, ond 151, 901 (NY); Sierra de Luquillo, Mt. Jimenes, Sintenis 1582 (К); Sierra de Naguabo, 210-675 m, Shafer 3392 (NY). CULEBRA: Britton & Wheeler 114 (МУ, US). MANATÍ: 2 mi. NE of Manati, Hansen et al. 9068 (MO, USF). VIEQUES: Eggers s.n. , Grosgourdy 13 (P); Isabel Segunda-Capo Cielo, Shafer 2377 (NY, US). U.S. ViRGIN ISLANDS ST. JOHN: Bethania- Rosenberg, Britton & Shafer 225 (NY, US); Bordeaux, 330 m, Britton & Shafer 545 S). sr. THOMAS: Bourne Resolution, Britton & is 428 (NY); above Cinnamon Bay, road to Herman m, 60-120 m, Mori & Woodbury 16999 (MO); Pearl- те Beschition, Britton & Marble 1320 (NY, US), Eggers 307 (GH, NY), Friedrichsthal s. n. (K, W), Krebs s.n. (K, W), s.n. E (US), 814 E Eggers s.n. (Engl. Arac. no. 191) (CH, , US) , Anthurium cubense Engl., Bot. Jahrb. Syst. 25: 364. 1898. TYPE: Cuba. Monte Toro, 300 m, Eggers 5402 (holotype, B; isotype, К). Fig- ures 1, 112, 113, 116, 117. Epiphytic or terrestrial; stem short, to 3 cm diam.; roots dense, ascending, green when young, becoming gray, pubescent at base, becoming smooth, tapered or ending abruptly, ca. 5-10 ст long, 3-5(8) mm diam.; cataphylls subcoriaceous, lanceolate, (4)10-22 cm long, narrowly acuminate at apex, drying reddish brown (B & K yellow-red 4/10), persisting + intact, weathering into coarse linear fibers and persisting around stem. Leaves erect; petioles (4)6-16(36) cm long, (2)5-12 mm diam., C-shaped to D-shaped, broadly sulcate adax- ially, vedad to obscurely flattened abaxially; ge- niculum slightly thicker than petiole, 0.5-1.5 cm long; sheath (1)3-8 cm long; blades subcoriaceous, usually oblanceolate-elliptic, sometimes + elliptic, short to long-acuminate at apex, usually obtuse at base, sometimes acute or weakly rounded, (20)33- 95 cm long, (5.5)13-38 cm wide, broadest at or near the middle, the margins conspicuously un- dulate; upper surface semiglossy, medium green (B K green 5/2.5), lower surface matte, slightly paler, both drying grayish; midrib flat to convexly raised at base, becoming obtusely raised toward the apex above, broadly angular to 3-ribbed below; primary lateral veins 6-11 per side, departing mid- rib at 30-50 angle, + straight to convexly raised above and below, slightly paler than surface; tertiary veins obscure, weakly raised to the margin, prominulous when dried; reticulate veins visible, darker than surface, obscure when dried; collective vein arising in the upper Y to upper М of the blade or absent, equally as prominent as primary lateral veins when present. Inflorescences erect, rarely spreading, usually several at different stages of development per plant; peduncle (5.5)8-33 cm long, (2)5-10 mm diam., 0.4-1.2x as long as petiole, green tinged violet, terete; spathe reflexed at anthesis, rarely recurled, subcoriaceous, green tinged with violet-purple, ovate to lanceolate, (2.5)5-16 cm long, (0.6)1–5 cm wide, broadest near the base, acute to acuminate at apex, rounded at base; stipe to 2 cm long in front, to 1 cm long in back; spadix purplish (B & K purple 4/5) to pale violet-purple (B & K purple 6/7.5), cylin- droid, slightly tapered toward both ends, (2.5)8— 19 cm long, (4)6-10 mm diam. near base, (3)5- 9 mm diam. near apex; flowers rhombic to 4-lobed, (1.3)2.1-2.6 mm long, (1.7)2-2.6 mm wide, the sides weakly sigmoid, sometimes straight; 9-12 flowers visible in principal spiral, (9)14—16 in al- ternate spiral; tepals matte, E punctate, slightly roughened; lateral tepals 0. mm wide, the inner margins straight to we saldo convex, the outer margins 2-sided; pistils weakly raised, greenish, darker than tepals; stigma linear, 0.4 mm long; stamens emerging in a scattered pattern through- 638 Annals of the Missouri Botanical Garden out the spadix, laterals emerging almost to apex before alternate emerge, held just above tepals and against pistil, obscuring the latter at anthesis; an- thers white tinged with red-violet; thecae oblong, not divaricate; pollen white. Infructescence spread- ing or erect; spathe persisting, brown; spadix ca. 14-27 cm long, 2-2.5 cm diam. (not including exserted berries), with the berries scattered throughout; berries orange-red (B & K yellow-red 5/2.5), принц a subtruncate and subapiculate at 4.5-6 mm diam.; mesocarp juicy; usually d CUM 2 seeds per berry, yellow (B & k ds 9/5), oblong to ellipsoid, 5— 7 mm long, 3-3.4 mm diam., with a gelatinous, sticky, transparent appendage. apex, 9-14 m Anthurium cubense is found in Cuba and drier parts of eastern Guatemala and Yucatan on the Atlantic slope, and in Nicaragua, Costa Rica, and Panama on the Pacific slope. In South America, it extends along the Caribbean coast from western Venezuela to Colombia, with one outlying collection known from the Río Patia in southern Colombia. It occurs in tropical dry forest, drier parts of trop- ical moist forest, and premontane wet forest from sea level to 700 m. It is most common in Cuba and Nicaragua. This species is recognized by its short petioles, short inflorescences (much shorter than the leaves), its stubby, purplish to pale violet-purple spadix that is tapered toward the apex (or sometimes toward both ends), and by its bright red berries. It has no apparent close relatives and is not easily confused with any other species. Engler (1905) also recognized material of this species under the name Anthurium recusatum Schott from Cuba. That name, however, has been misapplied and is probably synonymous with 4. fendleri from South America. Anthurium fendleri is not closely related to A. cubense, though it has a similar rosulate habit. It differs chiefly in having a long-pedunculate inflorescence, narrowly tapered spadix, promptly withering spathe, and purple ber- ries. COLOMBIA. Isla de Providencia, Smooth Water Road, Torres-Romero 295 (COL). ATLÁNTICO: Barranquilla, Elis 1295 (F, US), 1404 (Е). BoLívAR: Turbaco, Torrecilla, Killip & Smith 14656 (GH, NY, US); Mo o. San Juan Nepomuceno, Santuario Nacional de los Colorados, 230- 250 m, 9*58'N, 75°10'W, Cuadros & Keating 60679 MO). MAGDALENA: Dibulla, Seifriz 263 (US); Santa Mar- ta, Smith 2763 (GH, K, NY); Quebrada El Cedro, 350- 400 m, Moreno 415 (COL). Costa RICA. GUANACASTE: km S of Penas Blancas, 15 km М of La Cruz, 1.5 km of main road, 0-300 m, 11°11'N, a id _ 4652 (МО). PUNTARENAS: Cabo Blanco Nature Res Burger & Liesner 6627 (Е, MO). ME deut звера, Зарта ѕ.п. (Р); d: epe Hioram 3937 (US); Ma- NY, US); Matanzas, Meisner s.n. 0 m, Eggers 5402 (B, K), Wright 3208 (BM, СН, К); Orientali, Wright 601 (GH, К); San Diego de los Banos, Britton et al. 6854 (NY), León 4233 (NY, P). LA HABANA: La Guácima, Abarca & O'Donovan 4688 (GH, NY); San Antonio, Hitchcock s.n. (F), Van Hermann 822 (F, NYy Vento, Curtiss 647 (BM, F, GH, , А , Р). Las VILLAS: Hanabanilla Falls, Britton et al. 4838 (NY); Santa Clara, Rowe 8643 (GH), 8644 ۵ У). MATANZAS: Mat tanzas, Britton & Shafer 377 (NY). RIENTE: Bayate, 6202 (NY); Firmeza-Gran Pie- dra, Shafer 8936 (GH, 1573 (МУ); San Luis, А 2324 (МУ upper part, 300 m, Taylor 195 (NY). PINAR DEL RÍO: El Guama, Palmer & Riley 317 (05); Rio Mestanza, Britton et al. 10152 (US, NY); Sierra de Anafe, Britton et al. 9608 (NY), Wilson 11492 (NY, USy Sierra Guayaba, Shafer 13851 (A, NY); Mpo. Vinales, Sierra del Infierno, across Hwy. from Granja “Moncada,” 18°14'N, 68397, 150 m, Proctor et al. 39314 (JBSD); Pasco Real, Van Hermann 988 (GH). aa El Cobre, Pollard & Palmer 389 (GH, MO, NY, US, F). GUATEMALA. ALTA VERAPAZ: Cubilgiiitz, аа 8606 (US). GUYANA: Demerara, cultivated at Glasgow Bot. Gard., Hooker s.n. (К). NICARAGUA. BOACO: Boaquito-Sta. Lucia, 4 km E of Boaca, Hacienda Рећаз Blancas, 400-600 m, Grijaiva & Soza 4197 (MO, SAR). CHONTALES: Ипсеш 86, 87 (MO); Acoyapa-Rio Oyate, 0.4 km SE of bridge over Quebrada Niscale, 50 m, 11?47'N, 85%01'W, Henrich & Stevens 161 (MO); 1.9 km S of La Libertad, 530 m, 12°12'N, 85?*10'W, Stevens & Montiel 17513 (МО); 5.3 km W of Puente Lovago, Km 160.1, 150-170 m, 12%00'N, 85?12'W, Stevens & Henrich 20524 (МО); 1.5 km E of San Pedro de Lovago, 260-280 m, 12°07'N, 85°07'W, Moreno 16047 (MO); 2 km S of Acoyapa, Nichols 1723, 1725 (МО); 2 km N of Santo Tomás, rd. to San Domingo, 340-380 m, 12%04'N, 85067, Mo- reno 16208 (MO); 2.8 km N of Cuapa, 400-500 m 12°17'N, 85?23'W, Pipoly 1571 (CAS); М of Сиара, Stevens 3662, 6107 (МО); Rio El Bizcocho, Juigalpa- 350-400 m Acoyapa-Morrito, 50 m, 21631 (MO). esTELí: Kukamonga, km 167, Belén, 660- 800 m, 13?15'N, 86?21'W, Moreno 21822 (MO). GRA- NADA: road to Cutirre, 10-13 km NE of Granada, 350- 400 m, 11%50'N, 86°56-57'\/, Moreno 1471, 1516 о а 85?45'45"W, Grijalva 1 э (мо; neta M Granada, Isla El Carraco, 35 m z & Rivi 543 (MO, à Volcán um Neill 2730 (cult т Kew) (B, K, M , MO, PMA, US); NE side, Atwood & Neill AN82 (МО); 10 km S of Granada, finca “Las Delicias,” 340-360 m, 11°51°N, ] 8447 (MO) Finca Сиште, 300-400 m, 5955 5, Guzmán & Castro 2152 (МО); S of Hacienda Cutirre, 300-400 m, 11%49'N, 85°55'W, Moreno 16442 dies de Las Delicias- Hda. Cutirre, 400-500 m, 11?50'N, 6'W, Moreno 16478 (МО); las Fincas Maria Aux- rd El Cacao, 400-500 m, 11°51'N, 85%56'W, Mo- reno 442 (MO). MANAGUA: Managua, Maxon et al. 7538 (US); Managua, cultivated, Araquistain 398 (MO); Cen- со National Park. pp 4444 (B, MO). RÍO SAN JUAN: San Volume 78, Number 3 1991 Croat 639 Anthurium sect. Pachyneurium Carlos, Atwood 2028 (MO); W of San Carlos, 30-35 m 11?08'N, 84°47'W, Sandino 1773 (MO). RIVAS: а де Ometepe, Volcán Concepción, Altagracia “Та Sabana, 200-500 m, pe М, 85?35'W, Robleto 133 (МО); М side of Volcán Concepción, 3 m, 11?33'N, 85°37'W, Moreno 19827 (МО); NE B. Volcán Concep- ción, 140-160 m, 11?33'N, у , Sandino 4240 (MO); Volcàn Maderas, Mérida, ka Abejas,” 500- 700 m, 11°26'N, 85°32'W, Robleto 293 (MO); NE side of Volcán Maderas, 600-900 m, 11?27'N, 85°29'W, Moreno 19580, 19600B (MO); Volcán Masearas, Santa Cruz, Stevens 6626 (MO). ZELAYA: Caño Majagua, Ste- vens 6894 (MO). PANAMA: CHIRIQUÍ: 1 km E of Remedios, 8?14'N, 81*50'W, Nee 10104 (MO, PMA); Río Esti, Gualaca-Fortuna, 0.6 km E of Gualaca, 700 m, Croat 48793 (FTG, MO). VENEZUELA: FALCÓN: Dtto. Silva, Gol- fete de Guare, “Та Penita," S of Chichiriviche, 10-100 m, 10%54'N, 68°16-17'W, Steyermark & Manara 110434 (MO). ZULIA: е Perayra, SE of Ma- chiques, Steyermark 99847 (MY); Serrania de Perija, Río Palmar, Chitty & Benkowski 3061 (VEN); SE of Rio Catatumbo, 20-100 m, 9?06'N, 72?142'W, Liesner & Gonzalez 13274 (МО); Dtto. Colón, Laguna de Congo, EI 55 (MY); Dtto. Mara, El Palmar, Medina 2 (MY) Cano Indio, Hacienda Cano Azul-Base of E Yolanda, Steyermark et al. 122640 (VEN); Cerro Negro, SE of Rio Guasare, Steyermark et al. 122687, ~ Cosira, SW of Los Angeles de Tucuco, Davidse et а! 18531 (MO, VEN). Anthurium curtispadix Croat, sp. nov. TYPE: cuador. Pastaza: Puyo; originally collected by Mickey Carmichael, Croat 7 (holo- type, MO 3104658; isotypes, AAU, B, K, M, NY, ЕЗА, U, US; cult. at MO). Figures 114, 118, 123. Caulis ad 21 cm longus, 3 cm diam.; cataphyllum persistens semi-intactum; petiolus 3.5-14 cm longus, 4- diam., D-formatus ad C-formatus, adaxialiter sul- catus; fanda elliptica ad oblanceolata, 17-57 cm longa, .5 m lata; nervis primariis lateralis 5-7 utroque. Pedunculus 5-10 cm longus, 4 mm diam., effusus; spatha pallida viridis, oblongo-lanceolata, 3-3.5 cm longa, 1.5- 2 cm lata; spadix erectus, cylindricus, 3.5-4.7 cm longus, 6-7 mm diam., pallide viridis. Baccae ignotae. Stem to 21 cm long, ca. 3 cm diam.; leaf scars conspicuous or obscured by root mass, 0.5- high, 0.6-0.8 cm wide; roots ascending to oa ing to descending, green to light brown (B & K yellow-red 9/7.5), smooth to sometimes weakly pubescent, tapered, elongate, 4-7 mm diam.; cata- phylls subcoriaceous, weakly 1 -ribbed near the apex, 5-8 cm long, narrowly acute and apiculate at apex, drying tan to reddish brown, persisting semi-intact, eventually as a reticulum of fibers with the apex remaining intact. Leaves erect to spreading; peti- oles 3.5-14 cm long, 4-9 mm diam., D-shaped to C-shaped, slightly sulcate adaxially, with the mar- gins sharply to obtusely raised, rounded to some- times obscurely 1-2-ribbed abaxially, the surface obscurely short-lineate; geniculum paler and thick- er than petiole, becoming minutely and transverse- ly fissured and scurfy abaxially, (0.7)1-2.2 cm long; blades subcoriaceous, elliptic to oblanceolate, gradually acuminate at apex (the acumen apicu- late), narrowly acute to obtuse to weakly rounded at base, 17-57 cm long, 7.5-20 cm wide, broadest at or above the middle, the margins weakly un- dulate; upper surface semiglossy, medium green, lower surface matte to weakly glossy, paler, drying greenish to yellowish green; midrib flat at base with blunt medial rib, narrowing and acute toward the apex above, convexly raised and paler than surface below; primary lateral veins 5—7 per side, departing midrib at 30—45° angle, + straight, raised and weakly paler than surface above, convex below; interprimary veins absent; tertiary veins obscure above, visible below; reticulate veins obscure above, darker than surface below; collective vein arising from near the apex, weakly sunken above, prom- inulous below, 2-9 mm from margin. /nflores- cences spreading, usually several present at once; peduncle 5-10 cm long, 4 mm diam., equaling to usually ca. 0.5 х as long as petioles, green (B € K yellow-green 6/7.5), terete; spathe reflexed- spreading, coriaceous, pale green, sometimes weak- ly tinged with red (B & K yellow-green 8/10), oblong-lanceolate, 3-3.5 cm long, 1.5-2 cm wide, broadest in the lower Уз, inserted at 75? angle on peduncle, gradually to abruptly acuminate at apex (the acumen inrolled, hooked), acute to decurrent at base, the margins strongly downturned, meeting at 60—80? angle; spadix pale, dull green, somewhat tinged with brown, sessile, cylindroid to sometimes weakly clavate, erect, held at ca. 90? angle from peduncle, 3.5-4.7 cm long, 6-7 mm diam. near base, 3-4 mm diam. near apex, broadest at the base; flowers rhombic to 4-lobed, 1.7-2.3(2.6) mm long, 1.3-2.3 mm wide, the sides straight to jag- gedly sigmoid; 13-15 flowers visible in principal spiral, 9 in alternate spiral; tepals matte, minutely papillate; lateral tepals 1.5-1.8 mm wide, the inner margins broadly rounded, the outer margins irreg- ularly 3-sided; pistils weakly emergent, purplish violet; stigma linear, slitlike, 0.3 mm long; stamens emerging in a prompt, erratic sequence, the laterals extending to the apex in a scattered pattern before the 3rd stamen emerges, laterals followed by al- ternates in a rapid succession but very irregularly 640 Annals of the Missouri Botanical Garden on the spadix (sometimes the 4th stamen appearing first); anthers white, 0.6 mm long, 0.7 mm wide, obscuring pistil; thecae ellipsoid, not divaricate; pollen drying white. /nfructescence not seen. Anthurium curtispadix is known only from a living collection made near partment, Ecuador, by Mickey Carmichael. It was probably collected in an area of premontane wet forest at ca. 1,000 m This species is distinguished by its short, spread- ing inflorescence with an erect, short, stubby spadix (hence the name) with many (9-15) flowers per spiral. It is probably most closely related to ernestii, which differs in having an erect inflores- cence and cataphylls persisting as an intact network of fine, pale fibers. uyo in Pastaza De- ECUADOR. PASTAZA: Puyo, cult. at MO and originally collected by Mickey Carmichael, ca. 1, m, Croa 55207 (AAU, B, К, M, MO, NY, RSA, U, US) Anthurium dombeyanum Brongn. ex Schott, Prodr. 477-478. 1860. TYPE: Peru, without exact locality, Dombey s.n. (holotype, P). Fig- ures 11, 13, 119-122, 127. Anthurium e K. Koch & Sello in K. Koch, 1221 xi. 276. 1868. TYPE: ES locality un- k nown, of cultivated sain said to ua, vic. dum rigidissimum Engl., Pflan nzenr. M Ней 2. 1905. ТУРЕ: Peru. Junin: Tarma, above Huacapjistana, 1,900-2,000 m, Weberbauer 2007 (holotype, B). Anthurium б aE um var. mutatum Engl., Pflanzenr. f E 21) 292. 1905. TYPE: Peru. Junin: ve Huacapistana, 1,900-2,000 m, We- ur 2158 (holotype, B). Terrestrial or Vig rarely epiphytic; stem 20-30 cm long, 2-6 c roots dense (the root mass to 20 cm der рини to pale reddish, grayish when dried, sometimes with raphide cells on the surface, elongate, blunt at apex, to 40 cm long, 4-10 mm diam.; cataphylls subcoriaceous, lanceolate, 5-25 cm long, narrowly rounded to bluntly acute at apex, pale green, drying brown, persisting semi-intact at upper nodes, otherwise as coarse linear fibers, sometimes with the apex re- maining intact. Leaves erect to spreading; petioles (3)10-30 cm long, 5-18 mm diam., erect to ат.; spreading, D-shaped, slightly thicker than broad in juveniles and lacking the medial rib adaxially, sul- cate in age with an obtuse to acute medial rib, rounded (rarely 1-2-ribbed) abaxially, the surface pale-speckled; geniculum slightly thicker and paler than petiole, 0.5-2.5 cm long; sheath 3-11 cm long, extending halfway or throughout the petiole in smaller leaves; blades subcoriaceous to thickly coriaceous, oblong-lanceolate to oblanceolate or narrowly obovate, often elliptic in smaller leaves, bluntly acute or acuminate at apex (the acumen minutely apiculate), long-attenuate to obtusely rounded or subcordate at base, (10)40-190 cm long, (2)10-35 cm wide, broadest above or near the middle, the margins flat in smaller leaves or markedly undulate especially in larger leaves; up- per surface glossy to semiglossy, dark to medium green, occasionally developing a bluish, glaucous covering with age, lower surface usually semi- glossy, rarely matte, concolorous or paler than upper surface; midrib above acutely raised, below obtusely ribbed at base, becoming acutely angled in upper Уз, and then convexly rounded toward the apex, paler than surface or concolorous with it; basal veins usually present and aggregated in subcordate leaves, arcuate-ascending, free to base; primary lateral veins (7)10-20(24) per side, de- parting midrib at (18)40-60(70Y angle, usually gently arcuate-ascending to within 1 cm of the margin, then abruptly ascending to the margin, convexly raised above; tertiary veins scarcely vis- ible above, moderately to distinctly visible below, sometimes darker than surface, very weakly raised below or flat, drying raised and conspicuously vis- ible; collective vein arising from near the base to near the apex, less prominent than primary lateral veins, sunken to weakly raised above, raised below, 2-10 mm from margin. Inflorescences erect- spreading to spreading-arching, shorter than or almost equaling leaves; peduncle (15)30-65(90) cm long, 2-11 mm diam., (2)3-7(10)X as long as petiole, pale green, sometimes tinged with red- violet, green to dark brown when dried, subterete, sometimes l-ribbed, flexible; spathe spreading to reflexed at anthesis, recurle riaceous, green, sometimes tinged with red-violet, speckled with raphide cells, linear-lanceolate, 7— 20 cm long, 1-3.5 cm wide, broadest near base, acute at apex, decurrent at base; spadix olive-green to grayish, becoming dark pink or maroon (B & К red-purple 2/2.5) to purplish at anthesis (В & K blue-purple 4/10), weakly to moderately ta- pered, rarely cylindroid, sessile to stipitate to 2.5 cm, + erect, rigid, held at 130-180? angle from peduncle, 4-28 cm long, 4-17 mm diam. near , subcoriaceous to co- Моште 78, Митбег 3 1991 Croat 641 Anthurium sect. Pachyneurium base, 3-11 mm diam. near apex, broadest near the base; flowers squarish to 4-lobed, to 3 mm long when fresh, 1.6-2.6 mm long when dried, to 2.5 mm wide when fresh, 1.4-2.2 mm wide when dried, the sides = straight to smoothly sigmoid; 4-14 flowers visible in principal spiral, 3-8 in alternate spiral; tepals matte; lateral tepals 1-1.6 mm wide, the inner margins tinged with pink, becoming pur- plish, broadly convex, the outer margins 2—4-sided; pistils scarcely emergent, green, with raphide cells; stigma oblong, slitlike, 0.4–0.601.0) mm long; sta- mens emerging irregularly from the base of the spadix in a scattered pattern, slightly exserted, lateral stamens emerging to midway, the laterals preceding the alternates by 5-18 spirals, inclined over and obscuring pistil; filaments tan, with raph- ide cells, exserted ca. 0.5 mm, 1-2 mm long, 0.6– 0.9 mm wide; anthers pinkish, 0.6-1.1 mm long, 0.5-1 mm wide, inclined over the pistil; thecae obovoid, scarcely or not divaricate; pollen pale orange to pale yellow fading to cream, yeasty scent- ed. Infructescence semi-erect; spathe persisting; spadix 5-27 cm long, 1.5-2.5 berries scattered throughout; berries violet-purple to reddish violet (B & K purple 3/10), obovoid, bluntly rounded at apex, 6-8 mm long, 5-6 mm diam.; pericarp with raphide cells; mesocarp pulpy, white; seeds 1–2 per berry, green, oblong, 3.5-8 mm long, 2.5 mm diam., with a gelatinous ap- pendage at apex. cm diam., with Anthurium dombeyanum is an Andean species ranging from central Ecuador to southern Peru at 950 to 2,760 m. Collections from Ecuador are from the provinces of Tungurahua and Loja (a somewhat aberrant collection is known from Mo- rona-Santiago; see below), while Peruvian collec- tions are represented from Amazonas and Caja- marca to San Martin, Huánuco, Pasco, Junin, and Cuzco. Specimens from Ecuador ostensibly occur in lower montane moist forest to premontane moist forest life zones. In Peru, the species is present in lower montane and montane moist forest and pre- montane dry forest. The species occurs in season- ally dry habitats and often inhabits rocky or ex- tremely precipitous sites, with low nighttime temperatures. This species is recognized by its short-petiolate, coriaceous, usually markedly undulate blades, its long-pedunculate inflorescence with a moderately short-tapered, usually purple spadix, thick, green, lanceolate reflexed spathe and minutely papillate tepals with a more or less erose inner margin. Another feature common to most plants, despite their markedly variable blades (ranging from nar- rowly obovate to oblong-oblanceolate) is the pri- mary lateral veins, which frequently extend nearly to the margin in an almost straight line before turning markedly upward and extending along the margin and gradually merging with it. The only species which might be confused with A. dombeyanum is A. leonianum, from northern Ecuador, but only two collections of А. dombey- anum are known from anywhere near this area and fall short of reaching the same province (Im- babura). Anthurium leonianum is most easily dis- tinguished by its petioles, which are 5-8-ribbed, rather than rounded abaxially. See that species for further discussion. Anthurium cymatoph yllum is doubtfully placed here in synonymy. Typical material is no longer existent, and the species is represented by a tracing of the type done by Masters and deposited at Kew and by a photograph taken before the type was destroyed in World War II. The species was de- scribed from a plant of cultivated but unknown origin. There was speculation it originated in Brazil, but it is almost certainly not of Brazilian origin. It most clearly fits into 4. dombeyanum, although the three-ribbed abaxial petiole surfaces are rare in that species. The leaf of Anthurium dombeyanum demon- strates great morphological plasticity in size and shape. А common feature of many collections is an obtuse to rounded leaf base with an aggregation of many veins in the lower few centimeters of the base. Yet plants in the same population, or even on the same individual, may commonly exhibit acute leaf bases lacking such an aggregation of veins. Comparison of Croat 58323 and Croat 58324 demonstrates the range of leaf bases found within a population. Inspection of the various sheets of Croat 57707 or Croat 58323 demonstrates this tendency within one individual. The species is also tremendously variable in size, ranging from rather tiny plants collected in exposed areas (e.g., Vargas 3988) to huge plants growing in more mesic hab- itats (e.g., Croat 58323). Several collections are worthy of special men- tion. Two individual collections, Croat 58366 and Ellenberg 3538, are from unusually low elevations, i.e., 630 m and 370 m, respectively. The Croat collection, from Cajamarca, Peru, is perhaps a distinct species, having blades that dry yellowish green rather than the more typical brown color, and a weakly glaucescent aspect to its mature leaves. The Ellenberg specimen, from Amazonas, Peru, on the road between Bagua and Nazareth, is not unusual in any manner other than its aberrant low elevational occurrence. Sparre 19272 was also Annals of the Missouri Botanical Garden collected at a rather low elevation (700-800 m) and is the only collection known from Morona- Santiago Province in Ecuador; its leaf blade is also unusually broad. Vargas 6209, from Cadena in uzco, Peru, is aberrant in having the peduncle only 25 the length of the petiole (vs. (2)7-10x longer in typical А. dombeyanum), although the spadix is rather immature. Two collections from near Quillabamba in Cuzco, Peru (Croat 50962, 5091 9) differ in having the petioles 3-ribbed abax- ially, whereas most, if not all, other material has в rounded abaxially. Barbour 2583, from Amazonas in Peru, is remarkable in having a spathe 4 cm broad near the base and a petiole recorded as terete. The spathe of Croat 58192, from San Martin, approaches the dimensions of the Barbour collection. ECUADOR: LOJA: Cerro Campana, Vilcabamba- Yan- gana, Km 7, 1,900 m, 79?16'W, 4°17'5, Lojtnant & Molau 15002 (AAU, GB); Loja-San ch 2,100 m Asplund 18056 ($); Km 25-32, 2,200 m, Dodson & Thien 597 (MO), 659 (MO, US, T 19 km N of Loja, 2,070 m, 4°10’S, 79710", pin 50847 (AAU, B, CM, СВ, K, M, MBM, MO, ОСА, SEL, US); Mollococha, 10 km W of Vilcabamba, 1,600 m, vorn ndersson 21769 (GB). MORONA- о Е aquiza, Misión Bom boiza, Misión On ig ts m, Sparre 19272 (5). TUNGURAHUA: 2 iro B); , Sodiro s.n. (B; photo, MO). PERU: Without Meg Ta be ey s.n. (P). AZONAS: Pron, agua, 1 m E of main plaza in Bagua, 5 km above La Peca, 1,050-1,160 m, 5?33'S, 78°21'W, Croat 58358 (MO, USM); 12 km E of La Peca (by trail), 1,700 m, Barbour 2583 (MO); Bagua rande, Longa Grande, Buenos Aires (Calpon), 1,420 i ) О); Bagua- Nazareth, 370 1,500-1,600 m, 5%55'S, 77*53'W, Knapp & Alcorn 7538 (MO); Río Utcubamba, SE of Naranjitos, Gentry et al. 61375 (MO); Suyobamba, 3 km NE of Pedro Ruiz Gallo, 1,400 m, 5%55'S, 77*58'W, Gentry et al. 61290 (MO); Chiclayo-Río Utcubamba, trail above hwy., NW of Pedro Ruiz Gallo, 1,300-1,400 m, Young & Eisenberg 289 (MO); Moyobamba- Bagua, NW of Pedro Ruiz Gallo, 1,820 m, 5°52’S, 77*56'W, Croat 58314 (МО); at jet. of road to Chachapoyas, NW of Pedro Ruiz Gallo, 1,440 m, 5%56'S, 77%56'W, Croat 58321, 58324 (MO, USM, VBD), 58322 (CAS, CM, MO, NY, RSA, USM), 58323 (K, MO, RSA, SEL, US, USM); Lago Pomacocha, Mo- yobamba-Chachapoyas, Km 339-34 2,200-2,250 m, 5%47'S, 77*53-54'W, Croat 58254 (MO, NY, USM), 58281 мана Prov. Chachapoyas, На- cienda Shani, Chachapoyas, Cevasco s.n. AYACUCHO: Valle de San Miguel, Media Naranja, 2,000 m, 13?13'S, 73°95'W, Guerrera 2034 (Е). CAJAMARCA: Prov. Cajamarca, Huambos, 2,200 m, Ferreyra 8409 (NY, USM); Prov. Jaén, Rio Chamaya, Bagua—Olmos, 35 km E of Pucara, 70 km of Rio Maranon bridge, 630 m, 6?0'S, 78%52'W, Croat 58366 (AAU, C, CAS, CM, GH, K, KYO, MO, NY, SEL, U, US, USM). cuzco: Prov. Calca, Vileabamba, 2,700 m, Vargas 3988 (CUZ); Prov. Convención, Choquellowanca, 1,450 m, Vargas ~ 12947 (CUZ); TES pre Cocalpampa, 150 km NW from Cuzco, 1,2 m, Nuñez et al. 6816 (MO); Rosario pu 0 E 2 о кы Chávez 3334, 3336 (МО), Vargas 20674 (US); Quillabamba-Cuzco, just S of Qui- llabamba, 1,050- 1, 200 m, 1220'S, 72°44-45'W, Croat 50919 (CM, MO), 50962 (COL, M, MO, MY, USM); Santa Teresa-Chaullay, at Quellomayo, 139 km from Cuzco, 1,200-2,600 т, 13°08'S, 72°36'W, Nunez & Motocanchi 8752 (MO); Urusayhua, 1,400 m, Chavez MO); na Mopillo, Yupanqui-Rio Apurimac, Po- ener) 000 m, Davis et al. 1261 (F, SEL); Rio Urubamba, Quillabamba, 1,100 m, Solomon 3111 (MO, F); Prov. Quispicanchis, Cadena, 1,060 m, Vargas 6209 (CUZ) Masca Pata, Hda. Itio, 2,000 m, Vargas 3053 (CUZ); Prov. Urubamba, Km 88-92, 2,500 m, Vargas 3396 (CUZ); km 110 Machupicchu-Cuzco, 2,000 m, 13909'10"5, 72°31'W, Nuñez 8679 (MO) 2,900 m, Nunez 8423 (МО); Machupicchu, Jackson & Jackson P-1 (F); 2,000 m, EE 904 (К); 2,300 m, Vargas 21642 (CUZ, MO); 2,300 m, Zahner s.n. (М); railway, Km 96, 2,300 m, Palmer 181 (К); Laderas, 2,100 m Vargas 2982 (CUZ); dee 4 stops on salou above Machupicchu, 2,500 m, 13°11'S, 72°30'W, Croat 50969 (MO); Tankarpata, 2,650 m, Vargas 13576 (CUZ, US). HUÁNUCO: Tingo Maria Region, Huamincha, 1,800 m, Woytkowski 34220 (F, G, MO, UC); Prov. Huánuco, Кош 2,300 m, Woytkowski 34331 (Е, MO, UC); Río Tulca, Huánuco-Tingo Maria, km 443, 6 km N of Acomayo, 2,450 m, 9°04’S, 76?04'W, Cdi 57842 (CAS, СМ, МО, NY, RSA, SEL, US, USM). Junin: Prov. Acobamba, Acobamba-Oxapampa, Carretera 20B, 2: km NE of Acobamba, 2,170 m, Jones 9138 (LAM, МО); Prov. Tarma, Tarma-Oxapampa, 38 km NE of Tarma, on Carretera 20B, 2,226 m, Jones 9105 (LAM); Tarma- San Ramón, 1,000-3,000 m, 11?10'S, 75?20-45'W, Sullivan et al. 1067 (F, MO, SEL); e m, 11?10'S, 75°33'W, Smith & Canne 5931 (B, МО); Carpapata, 2,400-2,500 m, Cerrate 2810 (USM); (pum 2,760 1,800-2,400 m, Killip & Smith 24328 (NY, US); Ma tichacra, near Huacapistana, 2,100-2,200 m, Ferreyra 11180 (NY); 28-34 km NE of Tarma, 2,400- Dillon & Turner 1354 (Е, BM, K, MO, ВЗА, U, M SR Ge (МО). La LIBERTAD: Río Amazonas , Vencerem 7140" W, uo P al. 45343 s 'Moyobamba- -Cha- chapoyas, Km 400-404, 1,150-1,280 m, 5°%45'S, 77°29'W, Croat 58204 (MO, USM), 58192, 58207 (MO). Anthurium ernestii Engl., Pflanzenr., IV. 23B(Heft 21): 80. 1895 a. Anthurium ernestii var. ernestii. TYPE Peru. San Martin: Pongo de Cainarachi, Ule 6325 (lectotype, B). Figures 124, 128-130 Epiphytic; stem 6-20 cm long, 1-3 cm diam.; roots dense, green, whitish when dried, velutinous, 2—5 mm diam.; cataphylls subcoriaceous, 2-ribbed, Volume 78, Number 3 1991 Croat 643 Anthurium sect. Pachyneurium 3-12 cm long, acuminate at apex, drying reddish brown to light brown (B & K yellow-red 4/10), persisting intact, eventually as reticulum of fine, straw-colored fibers with the apex remaining intact. Leaves erect-spreading; petioles 4-30 cm long, 3- 13 mm diam., D-shaped, sulcate and with a medial rib adaxially, the margins acute, rounded to 1–5- ribbed abaxially; geniculum thicker and paler than petiole, 0.5-2.3 cm long; sheath 1.5-6 cm long; blades subcoriaceous, elliptic to oblanceolate to ob- long-oblanceolate, acute to acuminate at apex (the acumen flat), attenuate to obtuse, rarely rounded at base, (25)40-60(131) cm long, (6)7-20(36) cm wide, broadest at or above the middle, the margins broadly undulate, frequently concave in the lower part of the blade; both surfaces matte to semiglossy, generally drying green; midrib broadly and acutely raised at base, becoming flat toward the apex above, acutely raised below with a prominent central ridge at base, becoming slightly convex at apex; primary lateral veins 6-15 per side, departing midrib at 30—70° angle, straight to arcuate-ascending to the margin, prominently raised near the midrib, flat to sunken near the margin above, prominently raised below, drying orangish; tertiary veins prominulous, raised on both surfaces; collective vein arising in the upper Уз of the blade, 5-13 mm from margin. Inflorescences erect; peduncle 3-25(45) cm long, 2-6 mm diam., 0.5-3(5.7)X as long as petiole, ale green, terete; spathe erect to reflexed or re- curled, subcoriaceous, green to green tinged with red at base (B & K yellow-green 6/7.5), lanceolate to broadly lanceolate, 4-12 cm long, 1-3(4) cm wide, broadest near base or sometimes at or near middle, inserted at 45-80? angle on peduncle, abruptly acuminate to acute at apex, acute at base; spadix pinkish to magenta to purplish or green at anthesis, sessile, rarely stipitate to 0.5-1 ст, cyl- indroid, slightly tapered, erect, slightly curved, 3- 11(20) ст long, 3-14 mm diam.; flowers 4-lobed, 1-1.9 mm long, 1.4-1.7 mm uds. the sides sig- moid; 5-8(17) flowers visible in principal spiral, 7—11(16) in alternate spiral; tepals matte, densely papillate; lateral tepals 0.4-1 mm wide, the inner margins straight to convex, the outer margins 2- 3-sided; pistils raised, the exposed portion rect- angular to squarish, often reddish violet, darker than tepals; stigmas slitlike to ellipsoid, 0.2-0.6 mm long; stamens emerging in a regular sequence, the laterals preceding the alternates by up to 27 spirals, the 3rd stamen preceding the 4th by up to 25 spirals; filaments ca. 0.6 mm long, 0.8 mm wide; anthers purplish or pinkish to pinkish white, 0.2-0.7 mm long, 0.4-0.6 mm wide, inclined over and obscuring the pistil; thecae ellipsoid, not di- varicate; pollen pale yellow fading to white. /n- fructescence spreading; spathe withered; berries red-violet to violet (B & K purple 3/10-4/10), subglobose to obovoid, rounded at apex, 4-10 mm long, 3-6 mm diam.; seeds 1-2 per berry, obovoid, 2—4 mm long, 1.5-2.5 mm diam., creamy brown, with sticky appendage at apex; mesocarp grayish, juicy, somewhat mealy Anthurium ernestii var. ernestii ranges throughout western Amazonia, from southern Co- lombia (Putumayo) to Peru (Amazonas, Loreto, San Martin, Huánuco, Pasco, and Madre de Dios), and western Brazil (Acre and Amazonas as far east as the Rio Madeira), mostly below 500 m (rarely to 800 m) in tropical moist, premontane wet, and tropical wet forest life zones. This species is recognized by its fine, thin, pale brown to pale tan network of cataphyll fibers with the uppermost cataphylls often skeletally intact or nearly so and by its oblanceolate leaf blades which generally dry green or greenish brown with the primary lateral veins more or less orange (es- pecially on the lower surface). Also characteristic is the generally short peduncle and especially the cylindroid spadix which is about 5-8 x (rarely to 13 x) longer than broad and often has a minutely prickly appearance owing to the tight clusters of stamens and the upturned tepals on dried collec- tions (or more rarely the exserted styles). Like most of the widespread species of sect. Pachyneurium, A. ernestii var. ernestii is quite variable overall and is not easily understood when only specimens representing a portion of the geo- graphical and/or morphological extremes are at hand. Among the most variable characters for the species are spadix and fruit color. Spadix color has been reported as greenish yellow, pale green, pale olive-green, orange-buff, pink, pinkish, yellowish pink, grayish pink, light purplish pink, reddish, reddish brown, light pink-brown, purplish white, light violet, violet, pale purple, purple, reddish cream, or cream turning magenta. It is generally not a dark color (at least at anthesis), nor a color easily qualified. As in many species of Anthurium, the spadix often changes color during the course of anthesis. Fruits are variously described as violet, dark violet, bright purple, purple, or red. The morphological plasticity of Anthurium er- nestii is rather remarkable and is geographically clinal in nature, with plants increasing in size from north to south. In Ecuador, most notably in Napo, plants tend to be relatively small with short petioles and small inflorescences, and leaves mostly drying distinctly green with the characteristic or- Annals of the Missouri Botanical Garden angish primary lateral veins. Additionally, leaf blades are proportionately narrower in Ecuador (averag- ing 4.6X longer than broad) than in Peru (aver- aging 3.4X longer than broad). Peru is believed to be the center of distribution for the broadly interpreted taxon comprising Ап- thurium ernestii, and the departments of Loreto (including newly created Ucuyali Department), San Martin and Madre de Dios in particular harbor the greatest amount of morphological diversity. This is particularly true of the Rio Huallaga and Rio Ucayali basins (especially the region between the east banks of the latter and Acre, in Brazil). In Loreto, and to a lesser extent in Huánuco and Pasco, the specimens seen exhibit an overall tendency toward a significant increase in size and show a greater departure from the “typical” con- dition of leaves drying green or greenish. Many dry brown or brownish green, but still largely ex- hibit the orangish primary lateral veins. Anthurium ernestii typically has а peduncle about twice as long as the spadix, but in San Martin Department a few collections from Distrito Tocache Nuevo are very large in overall size and have peduncles 30-42 cm long and about 3x longer than the spadix. The number of flowers per spiral for the spadices of these plants represents the high- er range limit for the taxon. Peduncle length in relation to petiole length is greater in the northern part of the range (a col- lection from southern Colombia has a peduncle 5.7 X longer than the petiole, the greatest disparity observed). In Peru, the peduncles are normally about equally as long as the petioles. Overall, the peduncle is commonly no more than 3x longer than the spadix. Anthurium ernestii appears to be most closely related to Anthurium galactospadix. It differs at once from А. ernestii in having the spathe longer than the stouter spadix, and in having more (13- 28) flowers per spiral. Anthurium ernestii is apparently also closely allied to А. uleanum, which is largely sympatric. In general, 4. uleanum may be distinguished from A. ernestii by its more slender, glaucous-looking spadix, longer peduncle and petioles, and leaf blades which mostly dry brown. In addition, 4. uleanum lacks the characteristic pale, skeletally intact cata- phylls and orangish drying primary lateral veins of A. ernestii. Occasionally, spadices of dried collec- tions of what are believed to be А. ernestii var. ernestii may have a thin, waxy layer when they are very young, but this is usually soon sloughed off, at least before anthesis. Harling 1118 is an example. Anthurium ernestii comprises two varieties, with var. ernestii encompassing the range of the species. Variety oellgaardii, which occurs in central Ec- uador in Pastaza, differs from the typical variety in having a proportionately more slender, tapered spadix, a spreading, rather than erect, peduncle, and leaves which dry bright green. In Ecuador, peduncle length for var. ernestii ranges up to the lower limit for var. oellgaardii, but rarely exceeds it and is often much shorter. A source of possible confusion is a Field Museum photograph of the type of А. ernestii. It depicts two different herbarium specimens. The one on the left (Ule 6325) is the lectotype of A. ernestii. The specimen on the right (apparently now lost) is la- beled Ule 5776 and closely resembles the speci- mens discussed under 4. galactospadix (Cid & Nelson 2563, Prance et al. 7471). The latter specimen consists of part of a leaf blade, a petiole, and an inflorescence. This specimen apparently represents a mixed collection since a note on the specimen indicates that Ule 5776 is actually a Gesneriaceae and that the specimen in the pho- tograph is possibly Ule 5788. However, Ule 5788, represented by a plant at Geneva, is a species with a much shorter petiole than the one depicted in the photograph. BRAZIL. ACRE: Mpo. Cruzeiro do Sul, Mpo. уна ul, 62492 (MO); Rio Moa, Boiador, 150 m, a Rosas А. зи пење A MO), 62408 А, МО). AMAZO о Acre, Prance et al. 2338 (INPA); Rio e Ule 5788 (С); Mpo. Humayta, Rio Madeira, Tres Casas, Krukoff 6097 (NY), 4 6496, 6548 (СН, NY), Krukoff 6547 (С COLOMBIA: PUTUMAYO: El Wiskey, 13 ln S of Umbria vic. Finca Santa María, 300 m, dp 2066-А (F). ECUADOR: MORONA-SANTIAGO: Patuca, Harling 1118 (Sy; Rio Paute, Méndez, 700 m, Harling 1090 (5); Mendez-Paute, 600 m, 2?44'S, 78*19'W, Lojtnant & Molau 14549 (AAU); piod 6 m, 2°40'S, 78°14'W, Harling 984 (S). м Communa San Isla, 260 m, 0°29'S, 7692 et al. 39799 (AAU); Puerto ею Siona, ЕН jet. with Rio Napo, 3.5 km E of Puerto Маро, 1?02'S, TPAT'N, Un. 58885 (MO); 2-5 km SSW to WSW of San Pablo de las Secoyas, path to Shushufindi, 300 m, 0°15'5, 76°21'W, Brandbyge et al. 32548, 32796, 33359, 33412, 36222 (AAU); trail to Shushufindi, E of San Pablo de Las Secoyas, Jaramillo & Coello 2685 (AAU, ОСА); N of San Pablo de Las Secoyas, Jaramillo p Coello 2804 (AAU, QCA); о Sector Huashito, О km N of Coca, 250 m, 0?20'S, 77*05'W, Gudino РИ 1 (МО); San Pueblo, 35 ni E of Do Agrio, Pii didi 165 (K); Lago Agrio-Puerto El Carmen de Pu vic. Tara m E of Lago Agrio, 240 m 76923", Croat 58632 (MO); Rio Aguarico, Cuyabeno, ~ Volume 78, Number 3 1991 Croat Anthurium sect. Pachyneurium S of the river, 200 m, 0?17'S, 75°53'W, Holm-Nielsen et al. 21543 (AAU, MO); Dureno, Pinkley 245 (ECON), Jaramillo & Coello 3029 (ОСА); Lago ps 7.2 km of Rio uana 270 m, 4 N, 76°51 (AAU), Nielsen 76092 (AAU); Reserva Faunistica de Cuyabeno, N of La aguna 76°11'W, 0*01'N, 8 et al. an (AAU); Rio Shushufindi, near confluence with Rio Agu rico, 300 m, 0?18'S, 78°20'W, Balslev 4831 (NY); Tena Puyo, 5 km S of bridge over Rio , 510 m, 1905'5, 2. Croat 58921 (МО); Tena. Asplund 9423 5 m W of Tena, 500 m, 00%01'S, 77°51'W, Cro p (MO, QCA); Rio Cotapino, 50 о E of Tena, Cotapino, 500 m, Harling et al. 7023 (GB); Río Arajuno, Hacienda Aguinda, 450 m, 1%07'S, 77°36'\/, Marles EE 13A (F); Rio Due, Baeza-Lago Agrio, Km 65, trail to Chuscuyacu, “Zona Reserva Ecológica Cayambe- Coca," Jaramillo & Coello 3061 (AAU); Rio Napo, Ahuano, 1°04'S, 77*31'W, Lugo 245 (СВ, MO); Parque Nacional Yasuní, Anangu, 260-350 m, 0%31-32'S, 76?23'W, Lawesson et al. 39473 (AAU), Luteyn & Mori 8530 (NY), Luteyn et al. 8700 (MOy; Pozo petrolero Daimi 2, 200 m, 0%55'S, 76°11'W, Cerón & Hurtado 4088, 4238 (MO); near confluence with Rio rp Ballesteros (mil- itary post), Harling et al. 7368 (GB, МО); confluence 1 1 li T 250 m, Ceron et al. 2425 (MO, QCNE); Coca, Rio Payamino, 350 m, Harling & Andersson 11928 (GB); Payamino, Reserva Floristica “Е Chuncho,” 0930'5, 77°1'W, 250 m, Ceron et al. 2275 (MO, ОСМЕ), 2286 (МО, ОСМЕ), 2516 (МО, ОСМЕ); E of Puerto Misahuallí, 00 m, Besse et al. 1194 (SEL); 8 km below Puerto Misahualli, 1.5 km S of Río Маро, 450 m, 1%04'S, 77°36'W, Palacios et al. 448 (MO, NY, Vom QCNE); Santa Rosa, 32 km E of Tena, 0*01'N, 77*30'W, Lugo 2065 (GB); Rio Napo- Tena, 8 km SE a Tena, Shingui- pino, 480 m, Grubb et al. 1530 (K, NY); Jatún Sacha T Reserve, 8 km ESE of Puerto Misahualli, 400 1°04’S, 77°37'W, Miller et al. 2198, 2200 (МО); 400 m, Cerón 1330 (MO, QCNE), 1742 (MO, QCNE), 1694 (MO, QCNE), 2578 (MO, QCNE), 2671 (MO, CNE); Rio Pacuno, С 300 т, 0?40'S, 77°20'W, Whitmore 746 (К); Rio Putumayo at Colombian frontier, Baeza-Puerto El Carmen de Putumayo, 43.7 km SE of m e у W of Gua- rumo, 240 m, 0°7'N, 76°39'W, Croat 58591 (MO, ОСА); T Wai si aya, northern tributary of Río Aguarico, 300 ‚ 0°15'S, 76°21'W, Brandbyge et al. 32652, 32653, 36195. 36199 Sar bii Nacional Parque, 200 m, 0955'S, 76° n & Hurtado 3866 (MO), 230 75°35'W, Lawesson et al. 43314, О (AAU); ne Garza Cocha, 200 m, Lawesson et 4 43500 (AAU); 80 km upriver from Nuevo Rocafuerte, 225 m, Foster 3671, 3703B СЕ сз 47 (ОСА); епо, ‚ 0°2'5, 76942", ón & s 3131 (MO); Cantón rellana, sector Hua: к 20 km N of Coca, 250 m 0°20'S, 77*05'W, Rubio 259 (MO); Cantón Tena; Rio Blanco ус at headwaters of Río Huambuno, 6 k f Ahu 440 m, 1?00'S, 77°40'W, Kohn 1136 (МО). p "Puyo- Dies de Agosto, 8.2 km NE of Diez de Agosto, 970 m, 1?27'S, 77*51'W, Croat 59051 (MO, ОСА); Rio Bobonaza, oil exploration camp Chicirota, 300 m, 2222'5, 76940 W үс apod et al. 35301 (AAU); Rio Capihuari, repr i of Rio Pastaza, 285 m, 2°30- 31'S, 76*50-55'W, Oellgaard et al. 35102 (AAU); Río Curaray, S dec of river, mouth of Ri QAME), Palacios & Neill 719 (К, MO, QAME); N side of river, las lagunas Potoamo, 230 m, 1°30’S, 76*30' W, Neill & Palacios 6818 (MO, NY, QAME); Lorocachi, 200 m, 1°36-38'$, 75°58-59'W, Jaramillo et al. 30918, 31150, 31568, 31584 (AAU), 30965B (MO), 31188 AAU, MO); Centro-Oriente, Tzapino, Jaramillo & Coel- lo 3602 (QCA), 3644 (QCA); Tiwaeno, EM & Coello 3207 (QCA). PERU. AMAZONAS: Pro — Quebrada de Apigkagentsa, 720 m, Кауар 602 (MO); e Etseketai, 240 m, Kayap 841 (МО); Huam- 5 km Е of Chávez Valdivia, 200-250 m, 4°30'S, 78:30" W, Апсиазћ 1095, 1260, 1336 (MO), Berlin 2046 (MO), 2080 (MO, PMA), Kujikat 72, 161, 237, 302 (MO); S of Huampami, S of river, 266-283 m Berlin 1689 (MO); S of Huampami, trail to Sasa, 250 m, Berlin 1707 (MO); above Quebrada Tuhusik, 5 min. down river from Chávez err 230-270 m, Berlin 560 (МО); Rio Cenepa-Quebrada Kayamas, 270-300 m, Berlin 514 (MO); Rio E 250-280 m, Ka- yap 1251, 1339 (MO); Río Магапоп, above Pongo de Manseriche, 200 m, Wurdack 2456 (US), Mexia 6132a (UC); Rio qu E of the river, Galilea, 180 m, Hu iini 71 po мен 2-3 km behind unity | bur Za, of Pinglo, 200 m, 3°50/5, 1 1940' W, Би ак 1793, 2025, 2086, 2275 (МО), e 395, 690, 712, 840 (MO); 1-2 km below La Poza, m, ону 3, 264, 306, Рећа 84 (МО); 400 m ied La Poza, 180 m — Barbour 4429 (MO); E of H 200-600 m, Berlin кожу Ам trail E from La Реса into Serrania de Bag 000-1,400 m, Gentry et al. 23090 (MO); ныг КО 310 = Кауар 936 MO); Mesones-Muro Hwy., 8 of Montenegro, at km 286, 8 km E of Montenegro, 650 dnt S, 78°20 МА. Hutchison & Wright 3764 (UC). н со: Tingo Mar Region, Bella Durmiente, 700 m 8 (USM): Tingo Maria~Monson, Cuevas de Guchar “Parque Na- cional Tingo Maria," Rio Patay, 650 m, 992 'S, 76712", Croat 57930 (MO); Prov. Leoncio Prado, Dtto. Кира Rupa, Rio Huallaga, 700--800 m, Croat 21023 (MO); 750-800 m, Plowman & Ramirez 7568 (F, MO); Tingo Maria Region, 675 m, Croat 50978 (MO, RSA), Solomon 3388 (MO); W of Tingo Maria, 700-800 m, Schunke 10516 (MO); Prov. Pachitea, W of Codo de Pozuzo, 500-1,000 m, 9%40'S, 75?28'W, Foster 9217 (МО); Bosque Nacional de Iparia, Rio Pachitea, 1 km above Tournavista, near Miel de Abeja, 300-400 m, Schunke 1827 (Е, NY, US), 2240 (Е). LORETO: Río Ampiyacu, Pucaurquillo, Davis et al. 819 (F, SEL); Prov. Alto Ama- zonas, Balsapuerto, 150-350 m, Killip & Smith 28695 F, NY, US); Rio Huallaga, Lagunas, Croat 17838 (MO, USM); Yurimaguas, 135-200 m, Ferreyra 4949 (USM), Killip & Smith 28074 (NY, USy Puerto Arturo, below Yurimaguas, 135 m, Killip & Smith 27733 (NY, 05); Río Paranapara, Yurimaguas, less ii 5 17931 (MO), Rio Pastaza, Andoa 76°25'W, Croat 51255 (CM, МО), Vásquez 4435 (MO). m Ferreyra 13808 ~ 646 Annals of the Missouri Botanical Garden Vásquez & Jaramillo 848 (MO, RSA), Vásquez et al. 2985 (MO); Prov. Loreto, San Jose de Parinari, 150 m, 4°32'S, 74°30'W, Vásq uez & Jaramillo 3326 (MO); Pucayacu (native community Shimaco-Urarina), Vás- quez 6036 (MO); Rio Mararion, 1 hr. above Saramuro, 4%40'S, 75?00'W, Diaz & Ruiz 872 (MO, SEL); Rio Samaria, Atuncocha, Vásquez et al. 4362 (MO); Río Santiago, 200 m, Mexia 6142a (MC, UC); Campamento 2-Flor de Yarina, 140-160 m, 5%02'S, 74°30'W, eile et al. 38068 (МО); Prov. Maynas, Sacarita de Yanayac 90 m, McDaniel & Rimachi 25633 (MO), 25668, 25684 (IBE, МО); Santa Maria de Nanay, Quebrada Yarina, 150 m, 3°55'$, 73°40'W, Vásquez et al. 12200 (MO, US); Iquitos Vigne Rio Maniti, Recreo, NE of Iquitos, 2*50'W, Vásquez & Jaramillo 1139 ui a 120 m, 4°10’S, 73920", Vásquez et al. 315 (MO); Quebrada Tahuayo, above Tamishiyaco, Croat з 19744 (МО); Rio Amazonas, village Yanamono, near mouth of Rio Маро, 150 m, 3?25'S, 72°50'W, Croat 50125 (from a live plant col- lected by A. Gentry) (MO); Yanamono, Explorama Tourist Camp, Croat 61762 (AMAZ, MO, WIS), Gentry et al. 31496, 36637, 61833 (MO), Contry 299124 (МО); Mariscal Castilla, 106 m, 3°55'$, 70*30'W, Vásquez & Jaramillo 9344 (MO); Rio Gueppi, tributary of Rio Pu- tumayo, 8 km from mouth of river, Puerto Peru-Rio Napo trail, 200 m, Gentry et al. 21963 (F); Explorama Inn, ca. 2 km W of Indiana, 130 m, Gentry et al. 55988 (MO) Rio Napo, Croat 20188 (MO); Rio d trib. of Rio Ampiyacu, Brillo Nuevo, 2°40'S, 72°00'W, Balick et al. 1040 (GH); Prov. Maynas, Du poe i i io Мотоп, Quebrada Momoncillo, 200 DE DIOS: ser Shintuya, Pantiacolla, 480-840 m, Gentry et al. 27359 (MO); confluence of Rio Tambopata and Rio La Torre, of Puerto Maldonado, 12%50'S, 69°20'W, Smith et al. 118, 610 (US), 638 (NY, US), Smith 230 (US), Smith 307 (MO, US); Мапа National Park, Cocha Cashu Uplands, 400 m, 11%45'S, и E: 5540 (CUZ, MO, USM); Rio Manu, 350-4 ‚ Davidson 117 (MO), Foster 5811 (F), Foster A и 6482 of Puerto Rio Palotoa (Rio Panticolla), үш of Alto Madre де ae NW of EE 500 m, Foster & Terborgh 6749 xapampa, Iscozazin- Villa America, 350- “4501 m, 10°11'S, 75°15'W, Smith 2854 (MO, US); Palcazü Valley, Iscozacin, less than 500 m, Salick 7265 (МО); Rio Pichis, 10 km downriver, E of island, Puerto udez, Reserv i “Nue 14.1 km beyond bridge over Rio Pulcache, 600 m, 7?41'S, 76°40'W, Croat 58029, 580314 (MO), 58030 (CM, MO, USM); Pongo de Cainarachi (Shanusi), Ule 6325 (В); Prov. Mariscal Cáceres, Madre Mia, 760-880 m, Boeke & Ramírez 1281 (NY); Prov. Mariscal Cáceres, tag Tocache Nuevo, Tocache Nuevo, Е of bridge, 500 ‚ Schunke 7790 (MO); Quebrada de Canuto, 500 m, 12107 (IBE); Colegio Agropecuario de Tocache, 400 m, 8°13'5, 76°32'W, Croat 57994 (CAS, Е, DUKE, GH, M, MO, RSA, SEL, TEX, US, USM), 57995 (B, K, MO, USM); 5 km S of Cachaco, 42 km N of Tocache Nuevo, 330 m, 7%58'S, 76°38’ “ Croat 58048 (MO, МУ, USM); Fundo de Las Flores, 550-600 m, 8?10'S, 76*33'W, Croat 50991 (CM, MO), 50992 (F, MO, SEL), 50993 (MO, NY), 50994 (B, CM, IBE, К, M, MO, RSA, USM); Fundo Geoglifico del Sr. Luis Luden, Quebrada de Ishi- chimi, 400 m, Schunke 8107 (MO); Granja Santa Ysabel, old rd. to Limón, 400 m, Schunke 10326 (MO); Palo Blanco, W of bridge, 600-700 m, Schunke 5624 (F, NY, U, US); Palo Blanco, E of bridge, road to Shunte, 500-800 m, Schunke 7391 (CM, MO); Quebrada Pu- cayacu, chacra del Sr. Alfredo Sinarahua, 400 m, Schunke 12001 (IBE); Rio Huallaga, W of Escuela de Balsa Pro- bana, 400 m, Schunke 3637 (COL, F, NY, 05); 1-2 km downriver from Tocache Nuevo, Miramar, 500 m, 8°10'5, 76?28'W, Plowman et al. 11417 (Е, USM); Quebrada de Tanta, Schunke 4045 (F); Rio Tocache, Fundo Melodia, road to Shunte, 800 m, Schunke 7496 (CAS, MO); Rio de La Plata, 550-650 m, 8?10'S, 76°25'W, Croat 51000, 51029, 51033, 51035, 51041 (MO), 51002 (MO, USM); 5 km МЕ of bridge over Rio Huallaga, Fundo del Sr. Manuel Gatica, less than 100 m, 8?08'S, 76?23'W, Croat 57986 (MBM, MO, USM), 57987 (CM, MO, RSA, USM), 57988 (F, MO, P, US, USM); Fundo del Sr. Manuel Gatica, 660-700 m, Schunke 12136, 12138, 12186 (MO); NE, of Tocache Nuevo, 400-600 m, Schunke 8354, 8372, 8373, 8375 (Е, MO), 8376 (Е, MO, NY). UCAYALI: Prov. Ucayali, Rio Pisqui, 7°45'$, 75°01'W, Tessman 3249 (NY); Prov. Colonel Portillo, Km 99 on Carretera Federico Basadre, Arbo- retum Von Humboldt, 350 m, 8?45'S, 75?05'W, Diaz et al. 680 (К, MO); Arboretum Von Humboldt-Puerto Inca, 5 km E of Pucallpa-Tingo Maria road, Km less than 500 m, 8?17'S, 74*58'W, Croat 51051 (MO); Bosque Nacional de Iparia, NW of Rio Ucayali, 1 km below Iparia, 250 m, 9?12'S, 74°30'W, Schunke 2783 , NY). b. Anthurium ernestii var. oellgaardii Croat, var. nov. TYPE: Ecuador. Pastaza: between oil exploration camp Chichirota & Destacamento Cabo Pozo, 300 m, 2%25'S, 76%39'W, Oell- gaard et al. 34881 (holotype, MO 3039089; isotypes, AAU, QCA). Figure 131. A | РЦ | E + RH o r ffert a var. typi p graciliore elongatoque. Description based on dried material only. Epi- phytic; stem 1–1.5 cm thick; roots dense, whitish or brownish, velutinous, moderately short, 1 -3 mm diam.; cataphylls subcoriaceous, lanceolate, un- ribbed, 5-8 cm long, acuminate at apex, tan, per- sisting as reticulum of fine and dilacerated fibers. Leaves erect-spreading; petioles 7-15 cm long, 3- 6 mm diam., D-shaped, sulcate with a prominent medial rib and erect margins adaxially, 3—5-ribbed abaxially; geniculum thicker than petiole, 0.5-1 cm long; sheath 3-5.5 cm long; blades subcoria- ceous, obovate to oblanceolate-elliptic, short-acu- minate at apex, gradually attenuate toward the base, acute to obtuse at base, 54-87(100) cm long, broadest above the middle, the margins broadly undulate; midrib acutely raised 8-34 cm wide, Volume 78, Number 3 1991 Croat 647 Anthurium sect. Pachyneurium above, obtusely raised and faintly reddish below; primary lateral veins 10-16 per side, departing midrib at 30-50? angle, straight to arcuate-as- cending to the margin, prominently raised above, less so below; tertiary veins raised, prominulous; collective vein arising in the upper Уз of the blade, raised above and below, prominulous, 2-10 from margin. /nflorescences spreading-pendent; Ко 23-54 cm long, 3-9 mm .9(5.9)х as long as petiole, terete; spathe re- i: subcoriaceous, greenish, sometimes tinged with violet-purple, oblong-oblanceolate, 5-30 cm long, 0.9-3.5 cm wide, inserted at 40—70° angle on peduncle, acuminate at apex, acute at base; stipe 5-20 mm long in front, 1-10 mm long in back; spadix purple, cylindroid, slightly tapered, curved, erect, 7-8 cm long, 4 mm diam. near base, З mm diam. near apex; flowers 4-lobed, 1.2- 1.5 mm long, 1.2-1.5 mm wide, the sides sigmoid; 5-6 flowers visible in principal spiral, 10-12 in alternate spiral; tepals minutely papilate, pale- punctate; lateral tepals 0.8 mm wide, the inner margins rounded, weakly scarious, minutely erose, the outer margins straight to rounded, 2-sided, paler; stigma oblong-ellipsoid, 0.3-0.7 0. m wide; anthers 0.5 mm long, 0.5 mm wide; thecae oblong-ellipsoid, slightly divaricate. Infructescence pendent; spathe persisting; spadix 7-60 cm long, 0.7-3 cm diam.; berries dark violet- purple in upper Уз becoming whitish toward the I subglobose to cylindroid, 5-10 mm long, 3- mm diam.; pericarp with white raphide cells; 1o 2.8-4 mm long, 1.5-1.9 mm wide, 1.0-1.2 mm thick, with a gelatinous, transparent append- age at both ends. mm iam., 2.2- mm long, Anthurium ernestii var. oellgaardii is known at present only from east central Ecuador in the province of Pastaza in the vicinity of the type locality on the Rio Bobonaza and Rio Bufeo at about 300 m in a tropical moist forest life zone. Some of the collections are reported from peri- odically flooded forest dominated by Mauritia palms. This variety is distinguished by its large, broadly oblanceolate blades, which dry green with reddish brown or yellowish major veins and are attenuate at the base, and especially by its long-pedunculate inflorescence which is spreading-pendent at anthe- sis and produces a stout purple infructescence to 60 cm long and 3 cm diam. in fruit. It is noteworthy that only the type specimen has an infructescence of such large dimensions; all others are substantially more slender. This variety shares most features of the typical variety, except the slender, tapered spadix. The latter differs in its proportionally shorter, cylin- droid, spadix, generally shorter peduncles (in re- lation to petiole length) which are erect, and blades which usually dry brownish or, if greenish, very rarely clear green. Anthurium ernestii var. oellgaardii can also be confused with 4. uleanum, which has similar leaves, a long peduncle, and a similar, tapered inflorescence. That species differs in having tepals which bear a waxy, powderlike, grayish white cov- ering on drying, and leaves which dry brownish. In contrast, the tepals of 4. ernestii var. oell- gaardii are minutely granulose and glossy on dry- ing, and the leaves dry green. Most of the specimens examined are of fruiting material, and the upper limits of flowering spadix length are not known. This variety is named in honor of Benjamin Oellgaard, who participated in several expeditions to Ecuador and who made all known collections of var. oellgaardii. ECUADOR: PASTAZA: Río Bobonaza, Cachitama-outlet of Rio Bufeo, 300 m, 2*20'S, 76?40'W, Oellgaard et al. 34733, 34735, 34737 (AAU); Destacamento Cabo Pozo- oil ARAM e camp Chichirota, 300 m, 2%25'S, 76°39'W, Oellgaard et al. 34881 (AAU, MO), 34882 (AAU); kcal tb Cabo Pozo-La Boca, near outlet to Rio Pastaza, 275 m, 2°30-35'$, 76°38'Ұ, Oellgaard et al. 34933, 34937 (AAU) Rio Bufeo, N tributary of Rio Bobonaza, 300 m, 2°20'5, 76?40'W, Oellgaard et al. 34764, 34801 (AAU), 34769 (AAU, MO). Anthurium eximium Engl., Bot. Jahrb. Syst. 25: 412. 1898. TYPE: Costa Rica. Puntarenas: in forest near Punta Mala, Tonduz 67668 (lec- totype, BR; isolectotype, CR). Figures 125, 132. Epiphytic; stem elongate, 10—100 cm long, ca. 1 cm diam.; roots ascending to descending, pale green, pubescent, moderately elongate, somewhat tapered, 3-5 mm diam.; cataphylls subcoriaceous, narrowly lanceolate, 3.7-10 cm long, acute at apex with a subapical apiculum, yellowish green, dryin brown, persisting intact, eventually deciduous. Leaves spreading; petioles 7-12 cm long, 6–7 mm diam., D-shaped to subterete, broadly to narrowly sulcate adaxially, the margins blunt or sharp, not raised, abaxially rounded to 3-ribbed; geniculum thicker and paler than petiole, 0.5-1 cm long; blades subcoriaceous, narrowly oblanceolate, long- acuminate at apex (the acumen apiculate), nar- rowly acute to acute to obtuse or narrowly rounded at base, 23-80 cm long, 3.3-11.5 cm wide, broad- est above the middle, the margins weakly undulate; 648 Annals of the Missouri Botanical Garden upper surface weakly glossy to matte, medium green, lower surface weakly glossy to semiglossy, paler, drying matte, greenish; midrib obscurely and ob- tusely angular at base, becoming acutely raised toward the apex above, higher than broad at base, becoming round-raised toward the apex, conspic- uously paler than surface below; primary lateral veins 10-12 per side, departing midrib at 40—45? angle, weakly arcuate to the collective vein, sunken above, raised below; interprimary veins almost as conspicuous as primary lateral veins; tertiary veins obscure above, slightly darker than surface below; collective vein arising from near the base, sunken above, raised below, equally as prominent as pri- mary lateral veins, 1-4 mm from margin. /nflo- rescences spreading, shorter than leaves; peduncle 29-32 cm long, 3-5X as long as petioles, green tinged reddish, terete; spathe spreading to reflexed, green, lanceolate-ovate to ovate, 5-7.7 cm long, 2-2.6 cm wide, clasping at base; spadix creamy white to yellow to grayish or brown, rd ds to clavate, held at 90? angle to peduncle, 4.5-6 c long, 8- 10 mm diam. near base, 9 mm diam. near apex; flowers rhombic, 1.2-1.6 mm long, 1.3-1.8 mm wide; 10-12 flowers visible in principal spiral; tepals matte, densely papillate; lateral tepals 0.8- 1.1 mm wide, the inner margins straight, the outer margins 2-sided; stigma ellipsoid, 0.3-0.4 mm long; stamens emerging somewhat irregularly in a com- plete sequence, the laterals preceding the alternates by 1-2 spirals; anthers pale white, 0.6-0.7 mm long, 0.5-0.6 mm wide; thecae oblong, not divar- icate; pollen yellow, faintly aromatic at anthesis. Infructescence to 13 cm long, 1.5-1.7 cm diam.; berries pinkish to bright red, obovoid, ca. 4.6-7 mm long, 4—5 mm diam.; seeds 2 per berry, yellow, oblong, slightly flattened, 2.5-3.1 mm diam., 0.9-1. mm long, 1.2- mm thick. Anthurium eximium is endemic to Costa Rica and is found strictly on the Pacific slope, in tropical wet forest formation from sea level to 800 m. This species can be distinguished by the oblan- ceolate, epunctate leaves, persistent cataphylls, clasping, lanceolate spathe, cylindrical to clavate spadix, and bright red berries. Anthurium eximium most closely resembles 4. bradeanum and A. spathiphyllum (which are both restricted to the Atlantic slope), but differs from both by having entire cataphylls. 4nthurium bra- deanum also differs in having typically more elliptic leaf blades and 16-20 flowers per principal spiral (10-12 in A. eximium). Anthurium spathiphyl- lum also differs from 4. eximium їп having leaf blades with 20-30 primary lateral veins per side (vs. fewer than 13 in .4. eximium). In his monograph, Engler (1905) cited Ап- thurium eximium as a synonym of А. consobri- num. The leaf shapes in both species are very similar, but A. eximium has the collective vein arising from near the base, while 4. consobrinum has the collective vein arising from about the mid- dle. In addition, Schott’s illustration of A. conso- brinum clearly shows a long, slightly tapered spadix with pistils and berries greatly exceeding the tepals. Anthurium consobrinum is a common species of the Atlantic lowlands, and numerous collections are consistent with Schott's drawing. The berries of 4. consobrinum are white with purple, pink, or orange apices as opposed to the solid bright red berries of А. eximium. Anthurium eximium is a member of sect. Pachyneurium despite certain features it shares with sect. Calomystrium, including the thick, рег- sistent, intact cataphylls, and the presence of linear raphide cells on both blade surfaces. Costa RICA. PUNTARENAS: 2.5 mi. SE of Golfito, along hwy. from Rio Claro to Golfito, 60 m, 8°36'N, 8394", Croat 6; 7622 (МО). Punta Mala, Tonduz 6768 (BR, CR); Palmar Norte, 100-200 m, Croat 35121 (MO); Santo Domingo de Golfo Dulce, Tonduz 7260 (BR, CR, US); 100 ft., Allen 5211 (US); Osa Peninsula, Corcovado National Park, 0-1 km W of park headquarters at Sirena, -200 m, Gentry & Acevedo 48690 (B, CM, MO), Knapp 2192 (MO, PMA), Liesner 2846 (MO), 2996 (MO); Parque Nacional Corcovado, 0-150 m, 8°27-30'N, 83°33-38'W, Kernan 70 (CR, Е, K, MO); Rincon de Osa, 20-300 m, Liesner 1934 (MO); 5 km W of Rincon de Osa, NW of airfield, 50-200 m, 8?42'N, 83*31'W, Burger & Liesner 7286 (MO, NY); E base of Fila =P rigones, 1 km 5 and 2 km W of Canasas, 12 km 5 of Rincon de Osa, 60 m, 8?34'N, 83?25'W, Croat & Gra yum 59823 (МО); Las Cruces pre Garden, Croat 57240 (MO). SAN JOSÉ: Santa Maria de Dota, Croat 55204 (MO); Cerro Pelon, N base, A E of main road, 0.5 km S of San Martin de Puriscal, 800 m, 9?44'N, 84*23'W, Стауит & Baker 4694 (MO); Montañas Ja- maica, 2.5 E of Bijagual de Turrubares, Carara Reserve, 460-520 m, 9?45'N, 84?33.5'W, Grayum et al. 5478 (MO); Rio Negro, ca. 1.5 km E of Santa Rosa de Puriscal, 320 m, 9?42'N, 84?23'W, Grayum et а! 8339 (CR, ). Anthurium fasciale Sodiro, Anales Univ. Centr. Ecuador 19: 288. 1905. ТУРЕ: Ecuador. Тип- gurahua: Rio Pastaza, Sodiro s.n. (holotype, B; isotype, €). Figures 126, 133 Epiphytic or terrestrial; stem climbing, to 50 cm long, 1-3 cm roots moderately few descending, drying 2 mm diam.; cataphylls sub- coriaceous, lanceolate, 14-24 cm long, acute at diam.; apex, drying whitish, straw-colored, persisting as a reticulum of . Leaves mostly spreading; petioles 11-40(60) ‹ cm long, 5-11 mm diam., Volume 78, Number 3 1991 Croat 649 Anthurium sect. Pachyneurium D-shaped, flattened or convex with an acute medial rib adaxially, the margins sharply raised, rounded abaxially; geniculum moderately thicker than pet- iole, 1-2 cm long; sheath 2-7 cm long; blades coriaceous, oblong-linear to broadly oblong, nar- rowly acute at apex, obtuse to shortly attenuate or narrowly acute at base, 47-120 cm long, 5- 20 cm wide, broadest at or near the middle, the margins moderately to conspicuously undulate; up- per surface weakly glossy to semiglossy, slightly quilted, dark green, lower surface semiglossy to glossy, slightly to moderately paler; both surfaces rying green to yellow-brown; midrib convexly raised, or sometimes acutely raised toward the apex above, paler than surface, acutely raised below; primary lateral veins 17-33 per side, departing midrib at 60—90° angle, slightly arcuate to the collective vein, convexly raised above, flat and darker than surface below, drying raised on both surfaces; interprimary veins almost as conspicuous as primary lateral veins; tertiary veins moderately obscure above, darker than surface below, raised above and below when dried; reticulate veins ob- scure; collective vein arising from the base, equally as prominent as interprimary veins, (1)2-7 mm from margin. /nflorescences erect-spreading; pe- duncle 8-40(62) cm long, 2-8 mm diam., 0.6- 2.3 as long as petioles, terete; spathe reflexed, subcoriaceous, green, ovate to broadly oblong, ca. 6.5-12 cm long, 1.5-3 cm wide, broadest in the lower half, the acumen inrolled, 4-8 mm long, acute to narrowly acute at base; spadix dark red, cylindroid to slightly tapered, sessile or stipitate to o mm, erect, straight or slightly curved, 5.5-12 cm long, 5-10 mm diam .; flowers (dried) rhombic, 1.6-1.7 mm long, 1.2-1.5 mm wide, the sides jaggedly sigmoid to straight; 8-11 flowers visible in principal spiral, 5-8 in alternate spiral; tepals semiglossy, minutely papillate, lateral tepals 0.7— wide, the inner margins erose, the outer margins 2—3-sided; pistils yellow-brown on drying, smooth, lacking raphide cells, rounded at apex, not at all protruding above the tepals; stigma linear, ca. 0.3 mm long; stamens emerging from near the middle, inclined over and obscuring pistil at an- thesis, held against the pistil post anthesis; anthers 0.4–0.6 mm long, 0.4–0.5 mm wide; thecae ob- ovoid-oblong, weakly divaricate. /nfructescence erect to pendent; spathe persisting, green; spadix са. (7.5)12-16 cm long, 1.5-2 cm diam.; berries red, obovoid, 7-8 mm long, 5 mm 1-2 per berry, broadly oblong to rounded, flattened on one side, 4.4-5 mm long, 3-3.5 mm diam., 1.5-2 mm thick, with a transparent, gelatinous appendage at base and dark gelatinous appendage at apex. diam.; seeds A member of series Multinervia, Anthurium fasciale occurs in Morona-Santiago, Napo, Pas- taza, and Tungurahua provinces in Ecuador, at 970-1,600 m in premontane moist and very wet premontane forest life zones. This species is distinguished by its rather long cataphylls which quickly weather to pale fibers, its abaxially rounded and adaxially flattened or me- dially ribbed petioles with sharply raised margins, and its oblong leaf blades drying green or yellowish with the primary lateral veins rather numerous, but unequal in prominence and ascending at a broad angle from the midrib. Another distinctive feature is the dark red spadix with exserted sta- mens. A collection which may belong here (Croat 50586), made from a cultivated plant reportedly collected south of Puyo on the road to Macas, differs in having stamens not exserted and a rel- atively shorter peduncle (0.4 х as long as the pet- ioles), but agrees well with A. fasciale in other characters, including pistil color (on drying). The type specimen and another Sodiro collection closely resembling it are unusual in having pedun- cles 16-25 cm longer than any other specimens. Anthurium fasciale is likely to be confused only with А. penningtonii, also from high elevations on the eastern slopes of the Andes and also a member of series Multinervia. Anthurium penningtonii differs in having more elliptic leaf blades with more numerous, sharply raised primary lateral veins, and a more slender spadix. This species is also similar to 4. soukupii, rang- ing from southern Ecuador to Bolivia. It shares similar leaf blades and a similar weakly tapered spadix. Anthurium soukupii occurs at a higher elevation (2,200 to 2,400 m) and also differs in having a more or less triangular petiole and fewer flowers per spiral (7 or less). In addition, the dark- drying pistils of that species are densely marked with raphide cells and have prominently exserted styles. In contrast, the pistils of А. fasciale dry yellow-brown, with no raphide cells visible and are broadly rounded at the apex with the stigmas not at all protruding. ECUADOR: MORONA-SANTIAGO: boy de Calagras, 11 km E of San Juan Bosco, Descan 1,600 m, 3%02'S, 78°21'W, Vid et al. 104954. (SEL). NAPO: Holin- d Km 3 of Guamani, 1,200 m, 0°43’S, 7938", m et al 8624 (MO). PASTAZA: cultivated 5 of Puyo, road to Macas, Croat 50586 ( Agosto and Arajuno, 8.2 E of Diez de Agosto, 970 m, 1?37'S, 77%51'W, Croat 59017 (AAU, B, BM, CM, ‚ К, M, MO, QCA, (MO, QCA); 2 km NE of Mera, Hacienda San Antonio 650 Annals of th Missouri Botanical Garden de Baron von Humboldt, 1,100 m, 1°27'S, 78?06'W, Palacios et al. 58 (MO); Mera, Lugo 25 (S); Rio Pastaza, Mera, Lugo 1117 (GB) El Topo, 1,250 m, 1°25’S, 78°10'W, Harling 3403 (S). TUNGURAHUA: Río Pastaza (Río Chambo), Sodiro s.n., 8/ 1904 (C), Sodiro s.n., 12/ 1906 (B) Anthurium fatoense K. Krause, Notizbl. Bot. Gart. Berlin-Dahlem 11: 611. 1932. TYPE: Panama. Colón: Dos Bocas, Rio Fato Valley, 40-80 m, Pittier 4227 (holotype, US). Fig- ures 134-136. Epiphytic; stem 10-60 cm long, 1-2 cm diam.; leaf scars obscured by root mass; roots dense, as- cending to descending, numerous, pale green to white, smooth to very pubescent, short to elongate, usually tapered, 1-4 mm diam.; cataphylls sub- coriaceous to coriaceous, oblong-lanceolate, 5—14.5 cm long, acute at apex with a subapical apiculum, pale yellow-green, sometimes tinged with red, dry- ing medium brown, persisting semi-intact, soon de- ciduous. Leaves erect to spreading; petioles 2.5— 9.5 cm long, 5-8 mm diam., sharply quadrangular to angularly D-shaped to occasionally subterete, flattened to broadly sulcate adaxially, sharply 1- 3-ribbed or sometimes rounded abaxially, the sur- face minutely pale-speckled; geniculum paler and thicker than petiole, 0.7—1(2) cm long; blades sub- coriaceous, oblong-oblanceolate to oblanceolate to oblong-ovate, long-acuminate at apex, acute to ob- tuse to narrowly rounded at base, 24—83 cm long, 6-26 cm wide, broadest at or above the middle, the margins weakly and broadly undulate; upper surface matte to glossy, medium green, lower sur- face semiglossy to glossy, paler, drying greenish to greenish brown; midrib flat at base, becoming sharply and narrowly raised toward the apex above, raised and sharply 2-ribbed or prominently convex- raised at base below, becoming convex to flat to- ward the apex; primary lateral veins 8-16 per side, departing midrib at 40—50° angle, arcuate- ascending, narrowly raised above, less so and darker than surface below; interprimary veins less con- spicuous than primary lateral veins; tertiary veins obscure above, darker than surface below; collec- tive vein arising from near the apex or absent, 2- 9 mm from margin. /nflorescence erect-spreading to arching-pendent, shorter than or equaling leaves; peduncle 14-48 cm long, 2-6(9)X as long as petioles, green to reddish, weakly striate, terete to subterete, faintly ribbed; spathe spreading to re- flexed, subcoriaceous, light green, sometimes tinged with purple, narrowly lanceolate, 5.5-9(13.5) cm long, 0.7-1.3(2) cm wide, broadest near base, in- serted at 45—90° angle on peduncle, long-acumi- nate at apex (the acumen inrolled), obtuse to sub- cordate at base; spadix green to green tinged with red to pale reddish to reddish violet or purplish, oblong-cylindroid to long-tapered, 5-15 cm long, 4-10 mm diam. near base, 2.5-4 mm diam. near apex; flowers rhombic to weakly 4-lobed, 1.4-2.8 mm long, 1.8-3 mm wide, the sides smooth to jaggedly sigmoid; 5-9 flowers visible in principal spiral, 5-12 in alternate spiral; tepals matte to weakly glossy, minutely and densely papillate and mm wide, the inner margins straight to broadly convex, some- pale punctate; lateral tepals 1.2-1. what erose, the outer margins 2—4-sided; pistils prominently to weakly raised, weakly glossy, pa- pillate, green to white; stigma 0.4–0.5 mm long, brushlike; stamens emerging in slow progression from the base in a complete sequence, the laterals preceding the alternates Бу 6-34 spirals, the 3rd stamen preceding the 4th by 1-12 spirals, ar- ranged in a circle around pistil; filaments trans- lucent, 0.5 mm long, 0.4 mm wide; anthers yellow to white, 0.5-0.6 mm long, 0.5-0.8 mm wide; thecae broadly to oblong-ovoid, not divaricate; pol- len yellow fading to white, sweet-scented. /nfruc- tescence spreading; spadix to 26 cm long, with berries scattered throughout; berries cream to greenish yellow or pale orange, oblong-ovoid to ovoid-ellipsoid, acute at apex, 4-7.8 mm long, 3.3-4. diam.; pericarp thickened; seeds 2 per berry, pale yellow with numerous pale, granular raphide cells on the surface, oblong-ovoid, е or not at all flattened, 2.6-3 mm long, -1.8 mm diam., 1-1.3 mm thick, with а (чирий. gelatinous, sticky appendage. Anthurium fatoense is known from southern Costa Rica and Panama, occurring principally on the Caribbean slope in wetter parts of tropical moist forest, as well as in tropical wet, premontane wet, and lower montane wet forest, from sea level to nearly 1,200 m. It also occurs on the Pacific slope in the Fortuna area of Chiriquí Province in Pan- ama. Anthurium fatoense is considerably variable in size of the leaves and inflorescences as well as in berry color. It is distinguished by its slender, erect, pointed roots, short, conspicuously sheathed peti- oles, green-drying blades, and its tapered, greenish spadix with the stamens closely clustered over the pistil, the latter being not at all apparent. This species is most easly confused with А. con- sobrinum, but that species has prominently emer- gent pistils that emerge well ahead of the stamens and extend several millimeters above the stamens. Моште 78, Митбег 3 1991 Croat 651 Anthurium sect. Pachyneurium Costa RICA: ALAJUELA: il ae Aguas Zarcas, 8.5 km NE of Villa Quesada Hwy , 600 m, Croat 46976 (MO); 2 km N of Santa а 15 km N of Boca Arenal on Quesada- Muelle San Carlos-Los | oad, 100 m 10?38'N, 84°31'W, Liesner et al. 15 24(MO). CARTAGO: cultivated, Madison 6957 (MO, ЗЕТ) LIMÓN: 7 km of Bribri, 100-250 m, Barrer 200 m, 9?40.5'N, 83°02'W, Grayum & Hammel 5787 p Rio Sizaola, between Limón and Shiroles, 6.5 mi. of Bribri, 6 mi. SW of Bambú, 50 m, Croat 43295 МА PUNTARENAS: pee de Coto Brus, m W of San Vito de Java, 1,200 m, 8*49'N, 82*58'W, cultivated ч Газ сев MU si Botanical pere Croat 57229 (CR, MO). P OCAS DEL TORO: km S of 00 7 m, Antonio 3147 BN Cerro 6 (MO); Portobe Road, и. 1.2-8 km N of D тај Divide, 310-1,000 m, Chur- ampshire & Whitefoord 484 (BM); along road from Fortuna Dam toward Chiriqui Grande, 10 mi. from Continental Divide, 120 m, 8955'М, 82*10'W, MePherson 8587 (МО); Rio Guarumo, Chi- riquicito to 5 mi. S along Río Guarumo, Lewis et al. 2051 (MO); Chiriqui Lagoon, Fish Creek Mts., von Wedel 2350 (MO). CHIRIQUÍ: Fortuna Road, E of Fortuna, 1,200- 1,600 m, Folsom et al. 5293 (MO); between Gualaca and Fortuna Dam site, 5.9-11.8 mi. N of Los Planes de Hornito, 1,100-1,470 m, Antonio 5175 (MO), Croat 48705 (MO, F, DUKE, NY, RSA, S, US), 48716 (FTG, MO), 50095 (F, MO, TEX, W), 49893 (FTG, MISSA, MO), Knapp & Vodicka 5517 (MO); vic. Fortuna Dam, N of the lake, 1,100 m, McPherson 10407 (MO); along trail from Forestry Nursery to Rio Hornito, 1,100-1,200 m, 8°45'N, 82°15'W, Thompson 5029 (CM). состЕ: Ce- rro cm pue 2055 (MO); Llano Grande-Coclecito, tal Divide, 400-600 m, Antonio 1358, 3598 bo pen 49220 (MO), D'Arcy & Sytsma 14690 (MO), Hammel 1619 (MO), 3483 (MO, SEL), Sytsma 3880 (MO, РМА); S of Cascajal, Continental Divide, 80 m, 8°45'N 5'W, Knapp 1961 (MO); 9- 12 mi. from Llano Grande, 170-330 3994, 4165 (MO), Hammel 1808, 1908 (MO); El Соре Region, Alto bs iie 600-800 m, Folsom 3183 (MO); Е Copé W of sawmill, Hammel 2 2441 (MO); Con- пбн Б Лр sawmill above El Соре, 750-800 m, 8°40'N, 80°36'W, Knapp & Dressler 3501 (B, MO); El Valle Region, El Valle de Antón, 2 km vic. La Mesa, 860-9 Valle de pec N of Cer Hammel 3902 (MO). ph San Lucas, 100 m, Н 4549 (МО); pee Bruja, S slope, Rio Escandaloso, 200 m, Hammel 3197 (MO); Portobelo Region, Río Escan- daloso, т E Manganese mine, m, Hamme о = А о ч ct о = 2 > сл © © о ~ = > A 0 m 79°40'W, McPherson 8513 (MO); Río Boqueron head- waters, fork with Rio Nombre эў Diosito, 170 m, Наттеі 47 s, 40-80 m, Pittier 4227 о = < 9 w © 5 3 3 ® ә. л ~ o № = = un р d Lm d & =] Ф za Ф gion, Cerro Tute, beyond Escuela Agrícola Alto Piedra, чы of Santa Fe, 900-1,250 m, Antonio 2910, 4039 MO), Croat 48940 (MO), Sytsm ma & Antonio 3049 un 3068 (К, MO), Sytsma & Andersson 4658 (МО); etween Escuela Agricola Alto Piedra and Rio Dos Bocas, 350-770 m, Croat 25941 (MO), 27365 (MO), 27583 (MO), 27749 (MO); 5.9 mi. N of Escuela Circlo Alto de Piedra, between Santa Fe and Río San Luis, 480 m, Croat 66940 (MO); Rio Primero Brazo, 5 mi. NW of Santa Fe, 700-1,200 m, Croat 23115 (MO); 3.9-5 mi. N of Santa Fe, 500-1,000 m, Folsom 2949 (MO, PMA), Gentry 2953 (MO); 1-3 km N of Escuela Agricola Alto de Piedra, N of Santa Fe, 900-1,250 m, Mori & Kallunki 2560 (MO); 6-7 km W of Santa Fe, 970 m, Nee 9807 (MO), Nee 11220 (MO, РМА); between Santa Fe and Rio Ca- lovébora, 735 m, Croat & Folsom 34124 (B, CAS, COL, CR, DUKE, F, MBM, MEXU, MO, NY, PMA, RSA, US), Knapp & Sytsma 2623 (MO), Hamilton & Krager 3965 (MO). Anthurium fendleri Schott, Prodr. 468. 1860. TYPE: Venezuela. Colonia Tovar, Fendler 1343 (lectotype, К; isolectotype, GH). Figures 137- 139 Anthurium recusatum Schott, Prodr. 467. 1860. T tura, vic. Las Juntas near Dagua, 200-6 mann 5332 ме а B; isotype, Anthurium canasas Croat, Aroideana (1): E: Panama. КМ Dist. Са above Cariasas on road to Los Valles, 37051 (holotype, MO; isotypes, B, COL, F, PMA, S, SCZ, US). Epilithic, epiphytic or terrestrial, on steep rocky banks; stem to 15 cm long, 1-3 cm diam.; roots ding, greenish to brownish, densely pubescent, moderately elongate, tapered, 2-5 mm diam.; cataphylls moderately thin, broadly lanceolate, 5-10 cm long, acute with sub- apical apiculum at apex, yellowish red, heavily pale-speckled, drying brown, persisting semi-intact to fibrous, eventually deciduous. Leaves erect- spreading; petioles (2)4—47 cm long, 5-8 mm diam., D-shaped to subtriangular, narrowly to broadly sul- cate adaxially, the margins sharply to bluntly raised, rounded to 1—5-ribbed abaxially, the surface me- dium green, pale-speckled; geniculum thicker and paler than petiole, becoming fissured transversely with age, (0.5)1–2.5 cm long; blades chartaceous to subcoriaceous, oblanceolate to broadly elliptic, gradually acuminate at apex, usually rounded or shallowly cordate, sometimes acute to obtuse at moderately numerous, descen 652 Annals of the Missouri Botanical Garden base, (15)30-81 cm long, (6)10-34 cm wide, broadest just above the middle, the margins broadly undulate; upper surface semiglossy, dark to me- dium green, lower surface weakly glossy to semi- glossy, slightly paler; midrib flat at base, becoming convexly raised toward the apex above, slightly paler than surface, prominently higher than broad and acutely angular at base below, becoming con- vexly raised toward the apex; primary lateral veins 8-12 per side, arcuate-ascending, the lowermost free to the margin, loop-connecting to the collective vein in apical third, convexly raised, sometimes sunken in grooves above, prominently convex be- low; tertiary veins sunken above, raised and darker than surface below; reticulate veins obscure above, weakly visible below; collective vein arising from about the middle of the blade, sunken above, raised below, slightly less prominent than primary lateral 4-12 spreading or stiffly erect; peduncle 30-79.5 cm long, 1.6 to prominently ribbed abaxially, sometimes also adaxially; spathe spreading and recurving or loose- veins, mm from margin. Inflorescences -3.8 x as long as petiole, terete or weakly ly coiled, caducous, thin, soon drying and weath- ering, green or green tinged with violet-purple (B & K purple 5/10), linear-lanceolate to narrowly triangular, 5.5-11.5 cm long, 0.9-2 cm wide, broadest near the base, inserted at 60-80? angle on peduncle, narrowly acuminate at apex, acute at base, the margins meeting at 40—50? angle; stipe green, 6-18 mm long; spadix maroon (B & K re purple 2/10), long-tapered, 7.5-21 cm long, 4- 7 mm diam. near base, 2.5-4 mm diam. near apex; flowers square to rhombic, 2.5-3 mm in both directions, the sides smoothly to jaggedly о 4-8 flowers visible in principal spiral, 5-10 alternate spiral; tepals matte, weakly Din e densely and minutely papillate; Mer tepals 1.3- 1.8 mm wide, the inner margins broadly convex, the outer margins 2—3-sided; pistils early emergent, 0.1-0.2 mm, green becoming violet-purple, some- times green with violet-purple splotches; stigma ellipsoid, light purple-violet, 0.6 mm long, dry and full of slightly exserted bristles at anthesis, becom- ing dark and caviform, droplets conspicuous, ap- pearing 2-4 days before stamens emerge; stamens emerging in a slow progression from the base, the laterals preceding the alternates by 3-5 spirals, the 3rd stamen preceding the 4th by 1-3 spirals, held just above tepals; anthers creamy to yellow, 0.4—0.5 mm long, 0.8-0.9 not divaricate; pollen bright to pale yellow, soon mm wide; thecae ovoid, fading to white. /nfructescence pendent; berries violet-purple, obovoid, rounded at apex, 6.7-7 mm long; mesocarp thick, gelatinous, white; seeds l- 2 per berry, purplish, 4.5-5 mm long, 2-3.5 mm diam., 0.7-2 mm thick Anthurium fendleri ranges from sea level to usually less than 1,000 m, less frequently to 2,430 m, generally in partially shaded areas, in tropical dry, tropical moist, premontane wet, and premon- tane dry (presumably in premontane moist as well) forest life zones in Venezuela and Colombia. In Panama, it is rare and occurs in tropical moist forest. Anthurium fendleri was previously known only from the Cordillera de La Costa of Venezuela west of Caracas, but is now known to range throughout most of western Colombia into the Serrania de Perijá in northeastern Colombia and western Ven- ezuela, along the Rio Meta in the state of Apure and southward along the eastern edge of the Andes to the Departments of Meta and Vichada, in south- western Colombia. Disjunct populations include those in the Chocó, Colombia, in the Serrania del Darién, Panama, and in central Panama on the Pacific slope near Cariasas in Veraguas Province. Apart from 4. jenmanii, A. fendleri appears to be the only species of sect. Pachyneurium that ranges from northern South America (well outside the Amazon Basin) to lowland Amazonia. While А. jenmanii achieves this by wrapping around the Guianas and into Amapá, Brazil, А. fendleri ranges south through Colombia, as outlined above. This species differs from most members of the section by virtue of its thinner leaves and dark purple berries, and sometimes markedly rounded to subcordate leaf bases. It is further distinguished by its rich green, oblanceolate-elliptic to broadly elliptic leaves, which dry more or less olive-green and bicolorous, by its sunken tertiary veins, its ribbed petioles, thin, caducous spathe, slender red- dish brown spadix, and violet-purple berries. Anthurium fendleri is apparently most closely related to А. bonplandii, and some Colombian material may be confused with that species. Ап- thurium bonplandii differs in having leaves that dry brown and are usually conspicuously dark- glandular punctate and/or pustular (especially be- low) and acute at the base. This species is variable in leaf size and shape and the base may vary from nearly acute to sub- cordate. Specimens from Apure in Venezuela seem to have somewhat more coriaceous leaf blades. Croat 56578, collected on rocky cliffs near the Río Claro in Antioquia, Colombia, has unusually coriaceous leaf blades. Another aberrant collection is Stergios et al. 4568, from Portuguesa, Vene- zuela, in Municipio de Ospino; it has leaf blades Volume 78, Number 3 1991 Croat 653 Anthurium sect. Pachyneurium which are more coriaceous and less concolorous than usual, and obtuse-cuneate at the base. Other collections from the same area are from lower elevations, and have leaf blades that are more typically thinly coriaceous and bicolorous on dry- ing, and conspicuously shallowly cordate at the base Though А. recusatum Schott was described from a cultivated plant of unknown origin, the Schott illustrations show it to be remarkably similar to 4. fendleri, and it is here considered synonymous. COLOMBIA. Rio Magdalena, Von Bayern 44 (BOG). SEL, US, VEN); N sec ien of саћоп of Rio Claro, 325 m, 9°93’N, 74°39'W, Cogollo 1342 (COL, MO); central section of canyon, 350-475 m, Cogollo 615 (COL, MO), 1302 (COL, MO); Dabeiba, 1,350 m, Backley & Gut- tierrez 1822 (COL). ATLÁNTICO: Barranquilla, Elias 1477 (F); Sabana Larga, SSW of Barranquila, Elias 1296 (COL, Е, GH, US). BOLÍVAR: Juan Arias-San Pedro, Castenada 9855 (COL). BOYACÁ: Rio Casanare, Esmeralda, 130 m, Cuatrecasas 3818 (US); СЕЗАВ: Río Magdalena, above Canasas on road to Los Valles, Rincón Hondo, Allen 285 (MO). сносб: Serrania del Darien, W of Unguia, Rio Tigre Base Camp, 300 m, Gentry & Aguirre 15232 COL); Acandi, Bahia Zardi, 0-5 m, 8?24'N, 77°07'W, Betancur et al. 1249 (MO). CUNDINAMARCA: Cascajal- Tabla Grande, Nocaima Region, 1,000 m, Plow- man et al. 5246 (COL). GUAJIRA: Sierra de San Antonio, Cuchilla de La Nueva, 300 m, 11%05'N, 73?30'W, Cua- dros & be «d 2962 (MO). MAGDALENA: Rio Frio-San Andrés de La Sierra, El Platanal, i 1697 (05); Santa Marta, 1,500 2 Smith et al. 2305 (K), 2309 (BM, F, GH, MO, ‚ US). META: 310 m, 2°96’М, 73?10'W, Madison p? (GH); Villavicencio- FM ada, N of Granada, 520 m, Croat 55519 (MO); central moun- onda, 460 Plowman et al 4210 ( > US). SAN ANTONIO: Sierra de nta Marta anbury-Tracy 449 (K). м Rio Surata Valley. о 400-600 5). ЗАМТАМОЕВ SUR: Висагатапва- Дара! irón, 660 m, 7°03'N, 73°12'W, Croat 56545 (AAU, B, CAS, CM, COL, DUKE, K, M, n = ‚ ЗАВ, SEL, Es US); Rio Carare- Rio M , Puerto Berrio, 100- dalen 700 m, Haught 1710 (US) Mu Flor ida v Verse Los Cauchos, S of Bucaramanga, 1, I'N, 73°05'W, Croat 56388 u COL, K, L, MBM. "MO; : Buenaventura, Las Juntas and Dagua, 200-600 m, po n 5332 (B, F); Cali- Buenaventura, Quebrada La Guinea, m E of Cis neros, 220 3?A7'N, 76°46'W, Croat 62821 (ЕТ, IBE, M gua-Pepita, 302 m, André 373 (K); Cordillera Occidental, La Cumbre, 1,700-2,100 m, Pen nell & Killip 5756 (N Roc rritorio Faunistico del Tauparro, “incelli 1201 (MO). PANAMA: VERAGUAS: ve Canasas on road to Los abov Valles, Croat 37051 (B, CM, COL, DUKE, ENCB, F, K, KYO, MEXU, MO, NY, PMA, QCA, RSA, S, SAR, SCZ, TEX, VEN, US). VENEZUELA: Cultivated at Jardin Botánico de Ciudad Univ., Braun 2 (VEN). APURE: Dis- trito Achaguas- Rómulo Gallegos, Río Capanaparo, Hato San Joaquin, 100 m, 6*48-52'N, 68*45-48'W, Gragson & Gragson 92 (US). ARAGUA: Colonia Tovar, Fendler 1341, 1343, 1344 (GH); Corgua, El (MY) Por M es 3833 (MY); {жиек Trujillo 4889 (MY), Го, 2022C, 2022A (NY); Erie Na- Puig-Ross 14271 (МО); Barinas- Mérida, of peus. along Hwy. 1, 450 m, Croat 54754 (MO, VEN) between Barinitas- Altamira road and Calderas, S m, е на 2263 (MY, NY). cojeDEs: Ditto. El Lim of Tinaco, 100 m, Delascio 4187 (VEN). FALCÓN: 440-1,200 m, Liesner et al. 8283 (VEN). FALCÓN-LARA: Cerro Socopo, 500-1,000 m, Liesner et al. 8432 (MO, NY, VEN), 8441 (MO, VEN). GUÁRICO: N of San Juan де los Morros at La Puerta, 380 m, Bunting о LARA: Рио. Jiménez, Paso de Angosturo, 500 m, Steyermark & Espinoza 107627 (MO, VEN); Dtto. Talay гепсіпо, le de el Altar, 250 m, Steyermark et al. 109917 (F, VEN). MERIDA: Beguilla- Mucuchachi, 1,065- 2,430 m, Steyermark 56316 (F); El Vigia- Tovar, N of Santa Cruz de Mora, 4 00-200 12554 (МО); Dtto, Guanare, UNELLEZ property, 9°04'N, 69?49' W, A d 1306 (M ción, 800- 1,20 A et al. 4568 Me TACHIRÁ: La Fría-La Gri m, Е 2487, 2508(M 1 1,250 m, Bunting 2350 (MY); 13 km of Lobatera, Bunting 2349 (NY); vic. San Cristóbal, 1,000 m, Bun- ting 4. 369 (MO, NY); 925 m, Bunting 4837 (NY); San 5 m пи 13020 (NY); San Cristóbal-Santo Domingo, 250- 0 m, Bunting 23844, 2398 (MY, NY); 250 m, Вип- |. 2399 (NY); 300-350 m, Bunting 2384B (NY); E of Ayari, 250 m, Steyermark & Rabe 96659 (MO, VEN), = et al. 119464 (MO, VEN); Rio Frio, Junction Rio Quinimari- Río Frío, 480-500 m, Steyermark & a 119115(M / 450 т, Liesner & Guarigila 11835 (МО, VEN); Sierra el Casadero, Las Dantas-Las Adjuntas, 850 m, Steyermark et al. 120173 (B, СМ, K, MO, VEN). TRUJILLO: Sabana de Mendoza- Valera, Bunting 2825 (MY ); Valera, е 10793 (05); 20 km NE of Agua Viva, E of Valer 50 m, Bunting 2817 (NY). YaRACUY: E of Salóm, W s Valencia, NE of Nirgua, Kn m, Croat 54612 (BM, K, MO, VEN). ZULIA: pm a de Perijá, Rio Palmar, 450 m, Chitty & anto "3090 (VEN); Dtto. Bolivar, Piedras Blancas- El Rio Chiquio, W of Embalse, 6.5 km N of Piedras Blancas, 75-125 m, Bunting & Alfonso 8029 (NY); 20-25 km W of La Villa, 225-350 m, Bunting et al. 654 Annals of the Missouri Botanical Garden 8799 (МУ); Len Colón, Casigua el Cubo, 150 m, Bun- ting et al (NY); on Еў outskirts of Casigua El Cubo, sector p 2 4 km N of Casigua, Bun- ting et al 7 (NY) Dita. Lagunillas, Serrania de БРАН 13 km N of Embalse, 550-600 m, Bunting et al. 11279 (МУ); Dtto. Mara, 400-530 m, Steyermark et al. 122687 А (MO). Anthurium galactospadix Croat, sp. nov. TYPE: razil. Acre: N bank of Rio Juruá, opposite Cruzeiro do Sul, Prance et al. 2924 (holotype, INPA; isotypes, NY, US). Figures 140, 143, 193, 343, 344. Planta epiphytica; internodia brevia, 2-3 cm diam.; cataphyllum lanceolatum, persistens semi-intactum; pe- aie 11.5-28 cm longus, 5-10 mm diam., adaxile uni- nervis, abaxile с vel trinervis; о Јаје ођ- коок. -elliptica, 51-92 cm longa, 18-34 ст lata; nervis primariis lateralis 7-14 utroque. Pedunculus (4)6-11 ст longus, 3.5-5 mm diam., 0.4-0.7-plo oe quam petiolo; spatha late lanceolato-elliptica, (4.5)5.5-12 cm longa, 1.5-3.5 cm lata, a spadix leviter contractus, 7.5-11 cm longus, 8-14 mm diam., eburnius. Bacca ignotae. Epiphytic; stem 2-3 cm diam.; roots dense, short, 3-5 mm diam.; cataphylls coriaceous, lan- ceolate, 7—9 cm long, acute to emarginate at apex, drying reddish brown, persisting semi-intact as a reticulum of fibers, splitting at apex. Leaves erect- spreading; petioles 11.5-28 cm long, 5-10 mm diam., D-shaped, with a prominent medial rib and el raised margins adaxially, rounded to 3-ribbed abaxially; geniculum shaped like and slightly thicker than petiole, 0.5—2 cm long; sheath 2.5-6.5 cm long; blades coriaceous to subcoria- ceous, broadly oblanceolate-elliptic, acute to acu- minate at apex, often attenuate with concave mar- gins to long-acute to obtuse at base, 51-92 cm long, 18-34 cm wide, broadest above the middle, the margins moderately to strongly undulate; mid- rib sharply raised above, 3-ribbed near the base, ecoming obtuse to acute below; primary lateral veins 7-14 per side, departing midrib at 40—50° angle, den straight or arcuate to the margin, acutely raised above, convexly raised below; in- terprimary veins absent except toward base and apex; tertiary veins obscure above, flat and darker than surface below, prominulous on both surfaces on drying; reticulate veins obscure; collective vein absent. Inflorescences erect; peduncle (4)6-11 cm long, 3.5-5 mm diam., 0.4-0.7 X as long as pet- iole, terete; spathe usually erect to erect-spreading, sometimes reflexed, coriaceous, dark green to greenish white, broadly lanceolate-elliptic, flattened to navicular, (4.5)5.5-12 cm long, 1.5-3.5 ст wide, broadest at or near the middle, obtuse at apex (the acumen apiculate), acute at base, some- times decurrent for up to ca. 1 cm; spadix pale green to creamy white, Pera scarcely tapered, erect, slightly curved, (4)7.5-11 cm long, 8-14 iam. midway, 7-8 mm diam. near apex; flowers square, 1.1-1.9 mm long, 1.1-1.8 mm wide, the sides mostly straight parallel to spirals, jaggedly sigmoid perpendicular to spirals; 13-25 flowers visible in principal spiral, 15-28 in alter- nate spiral; tepals matte, smooth or covered with a waxy layer or minutely papillate; lateral tepals 0.5-0.8 mm, the inner margins rounded, weakly turned up against the pistil, the outer margins 2-sided; pistils emergent before stamens appear, square, ca. 1 mm in both directions, pale purplish lavender when fresh, colored + like the tepals when dried; stigma 0.5 mm long, slitlike when young, becoming brushlike; stamens emerging somewhat sporadically from near the base, the laterals emerg- ing to the apex before the first alternates appear, held erect on drying; anthers 0.4-0.6 mm long, 0.4-0.5 mm wide; thecae oblong-ellipsoid, not di- varicate. Infructescence erect; spathe persisting; spadix (4)6-13 cm long, 0.8-1.7 cm diam.; ma- ture fruits unknown. mm Anthurium galactospadix is restricted to the western Amazon basin in “Trapecio Amazonico" in extreme southeastern Colombia, east-central Lo- reto near the Brazilian border in Peru, and in Acre and southern Amazonas in Brazil. It comprises epi- phytes in tropical wet and tropical moist forest life zones (at least in Peru and Colombia), at 300 m or less. This species is readily distinguished by its thick, relatively short, whitish spadix with the spathe usu- ally longer than the spadix, its short peduncle, and its broadly elliptic-oblanceolate blades which are often attenuate with conspicuously concave mar- gins toward the base. It is most closely related to A. ernestii, which differs in having usually thin, straw-colored cataphylls that perisist intact as fine reticulate fibers (vs. semi-intact with more coarse fibers), a shorter and narrower spathe and much less stout spadix which is variously colored but never whitish, and in usually having somewhat orange primary lateral veins upon drying (especial. ly on the lower surface). In addition, the spadix of A. ernestii always surpasses the spathe, while in the present species it is shorter than the spathe. The specific epiphet for this species comes fro *galacto-" (Greek), meaning milky white, alluding to the color of the spadix. Two collections from Acre in Brazil (Cid & Nelson 2563 and Prance et al. 7471) are ten- Volume 78, Number 3 1991 Croat 655 Anthurium sect. Pachyneurium tatively placed here. They differ in having a stem only 1-1.5 cm diam., leaf blades 43-76 cm long, and 11-20 cm wide, a peduncle 0.2-0.5 x the length of the petiole, a spathe 4.5-6.5 cm long, 1.5-1.7 cm wide, and a spadix 3.1-5 cm long and 4-7 mm diam. In addition, the Prance et al. col- lection reports the spadix color as brown, and in the Cid & Nelson collection the peduncle equals the petiole in length. Another aberrant specimen is Tessmann 5463, from Yarina Cocha in Loreto, Peru, which has a more slender spadix and a spathe more than twice as long as the spadix. All of the above are probably aberrant specimens rather than representatives of distinct taxa. BRAZIL: ACRE: п Ё e I Branco, 44 km E of Cruzeiro do Sul, Faz ona Cabuca, 150 m 1*38'S, 72°35'W, ipe 62329 АРА, МО); Зепа Mad- ureira- Rio Branco, Km 5.5 , 9*04'S, 68°39'W, Croat & Rosas 62721 TO T 'K, MO, NY, US); Rio Acre, Seringal de São Francisco, Ule 9206 (K, MG); Rio Moa, Cruzeiro do Sul, m d „ T9388. o. Sena Ma- dureira, Sena Madureira, Cid & Nelson 2563 (INPA); Sena Madureira-Rio Branco, Km 7, Prance et al. 7659 (NY, INPA); Ж» Tarauaca, Tarauaca, 8°08'S, 70°50'W, Ргапсе et al. 7471 (NY, INPA, М id = е Schultes & Black 8527 (COL, US). Реко. 0: Yarina Cocha, Pucallpa, Tessmann 5463 bs Prov. fond Portillo Rio Abujao, Caserio de J.C. M iategui- Mejico, one hour by boat from Caserio de ae: 300 m, 8?15'S, 73°45'W, Diaz et al. 753 (MO). T Anthurium glaucospadix Croat, sp. nov. TYPE: Colombia. Cundinamarca: along road between Tena and La Mesa, 6 km NE of La Mesa, 1,300 m, 4?39'N, 74?26'W, Croat 55257 (holotype, MO 3033409; isotypes, B, CAS, COL, F, GH, K, MBM, NY, ЕЗА, SEL, US; live at MO). Figures 141, 142, 144, 145, Planta terrestris aut epiphytica; internodia brevia, 3- aut late lanceo- formatus vel C-formatus aut plus minusve subquadran- cm longa, eolata, apica ple- rumque rotundato; nervis primariis lateralibus (7)9-18 utroque; pedunculus 30-60(90) cm longus, 3-10 mm A por oblongo-lanceolata, purpureo-violacea aut viridis, (6)10-32 cm longa, (1)2-3.5 cm lata; spadix glaucus, caeruleo-viridis, longus et graduatim contractus, (7.5)14-33 cm longus, 5-12 mm diam. Baccae rubrae, oblongae, apice quadrato. Terrestrial or epiphytic; stem to ca. 30 cm long, 3-5(6) cm di white to pale green, soft, fuzzy, heavily woolly- pubescent, short, tapered, less than 10 cm long, 3-6 mm diam.; cataphylls subcoriaceous, lanceo- late to broadly lanceolate, 3-12 cm long, acute to prominently emarginate at apex, green tinged with red, drying tan to reddish brown, persisting as a reticulum of fibers. Leaves semi-erect to spreading, occasionally pendent; petioles 4-26(63) cm long, 4—13 mm diam., sharply D-shaped or C-shaped to = quadrangular, flattened to sulcate, often with a medial rib adaxially, the margins acutely raised, rounded to 3-5(8)-ribbed abaxially, the surface short-lineate, sometimes pinkish at base; geniculum slightly paler and thicker than petiole, becoming minutely fissured and scurfy adaxially, 1-3.5 cm long; sheath 3-7 cm long; blades moderately co- riaceous, mostly oblanceolate, sometimes elliptic or oblong-oblanceolate, mostly rounded or rounded- emarginate on larger leaves, acute to slightly acu- minate on smaller leaves at apex (the acumen diam.; roots dense, ascending, gra -'3h inrolled, apiculate, sometimes inequilateral, down- turned), mostly attenuate or acute (rarely rounded to obtuse) at base, 30-126 cm long, (8)15-47 cm wide, broadest in upper third, the margins slightly to moderately undulate; upper surface glossy to semiglossy or weakly glossy, medium green or yel- ow-green, lower surface semiglossy or occasionally matte, slightly paler, sometimes weakly to con- spicuously whitish glaucescent; midrib flat at base, — becoming obtusely to acutely raised toward the apex above, prominently obtuse-raised to 3-6- ribbed at base, becoming prominently convex to- ward the apex below, paler than surface above and below; primary lateral veins (7)9-18 per side, de- parting midrib at (30)50-70(80Y angle, weakly arcuate-ascending to the margin, sometimes ap- pearing sinuate, convexly raised above, promi- nently convexly raised below, paler than surface; interprimary veins few, scarcely visible, sunken above, weakly raised below; tertiary veins weakly sunken or weakly raised above, prominently raised below; collective vein arising from near the apex, occasionally in the lower third (especially in smaller plants), sunken above, raised below, 3-9 mm from margin. Inflorescences semi-erect to spreading or pendent, shorter than, or almost equaling leaves; peduncle 30-60(98) cm long, 3-10 mm diam., 2-7 X as long as petioles, green weakly tinged with red, pale short-lineate, terete to subterete, occa- sionally ribbed abaxially; spathe spreading to re- 656 Annals of th Missouri а Garden flexed, sometimes twisted, occasionally with prom- inently revolute margins, subcoriaceous, matte above and below, green to green heavily tinged with red or purple-violet (especially on adaxial sur- face), oblong-lanceolate to oblong-triangular, (6)10— 32(55) cm long, (1)2-3.5(4.0) ст wide, broadest near the base, inserted at 20—35° angle on pe- duncle, long-acuminate to abruptly short-acumi- nate at apex (the acumen often inrolled, 15-20 mm long), acute or decurrent to 6.5 cm at base; stipe rarely to 14 mm; spadix glaucous, bluish reen, sometimes weakly tinged with purple or erect to pendent and cm long, 5-12 mm yellow, very long-tapered, + curved upward, (7.5)14-33 diam. near base, 3—4 mm diam. near apex, broad- est at the base; flowers rhombic, (1.6)1.8-3 mm long, (1)1.4-2.6 mm wide, the sides straight to jaggedly sigmoid; (5)6-9 flowers visible in principal spiral, 8--13 in alternate spiral; tepals glaucous, matte, minutely papillate, with a few scattered droplets, lateral tepals 1-1.6 mm wide, the inner margins broadly convex to straight, the outer mar- gins + straight or bluntly angled to obtusely 3-sided; pistils emergent, papillate, matte, squarish, pale green; stigma ellipsoid becoming circular, 0.4-0.6 mm long; stamens emerging rapidly in a regular and complete sequence over а 5—6-day period, the laterals preceding the alternates by (3)10-14 spi- rals, the 3rd stamen preceding the 4th by 1-3 spirals, inclined over and obscuring pistil; anthers pale yellow, 0.2—0.8 (dried) to 1 (fresh) mm long, 0.4–0.9 mm wide when fresh; thecae oblong, not divaricate; pollen pale yellow fading to cream, sweetly scented like rotting fruit. /nfructescence with spathe persisting; spadix 12-53 cm long, 1.5- 3.5 cm diam.; berries reddish purple, obovoid to oblong, 7-12 mm long, 4-5.5 mm diam.; pericarp thin, with raphide cells; mesocarp gelatinous; seeds 1 -2 per berry, oblong, 4-5 mm long, 1.6-2.2 mm 1-1.2 pendage. diam., mm thick, with a mucilaginous ap- Anthurium glaucospadix is endemic to Colom- bia in the departments of Risaralda, Santander, Santander Sur, Cundinamarca, Valle and Huila, from 500 to 2,100 m. The distribution is unusual in that it included both slopes of the Cordillera Occidental (but only an unusually dry region of the Pacific slope). Anthurium glaucospadix is terres- trial or epiphytic on dry, rocky, steep slopes, ex- posed areas and thickets or as an understory herb in disturbed forest in premontane moist and pre- montane wet forest life zones. his species is recognized by из long-tapered, bluish green, glaucous spadix (hence the name), its distinctive, usually emarginate or apically rounded blade, and by its reddish purple fruits. This species is most closely related to 4. cau- cavallense, also from Colombia, which differs by having a nonglaucous spadix, hooked cataphylls, more acute blade apices and more broadly spread- ing primary lateral veins. Worthy of mention is Triana 683/ 1702, a mixed collection, two sheets of which are mounted with the leaves of a dicotyledonous plant, apparently of the Lecythidaceae. Although this number was re- corded as being from Cundimarca, Triana 682 is from Pasto. су ы CUNDINAMARCA: Fusagasuga-Girardot, km , 7 km $ of Fusagasuga, 1,520 m, 4?25'N, 74°25'W, uh 52000 (AAU, B, CAS, CM, "COL. G, JAUM, K, M, MEXU, MO, NY, QCA, SAR, US), 55428 (K, MO, NY, US); Fusagasuga- Melgar, 4 km of Fusagasuga, 1,525 m, Barclay et al. 3645 (US); Mosquera-La Mesa, km 32, Laguna Pedro Palo above Finca San Јове, 2, 000- 2,250 m, Gentry & Fallen 17159 (С La Mesa, 6 km NE of La Mesa, 1.300 14?26'W, eie 55257 (B, CAS, COL, CM, DUKE, ENCB, F, СН, К, MBM, MO, NY, OOM, QCA, RSA, ЗАК, SEL, TEX, VBD, US, 2); Anolaima, 1,800 m, Daniel & Augusto 4567 (COL); 6 km from La Mesa, 1,470 m, Forero & Garzón 307 (COL); usar Blanca, Choachi-La Calera, 1,900 m, Acosta- Arteaga 2 MO); Quebrada Cabana, Tocaima-Pubenza, 380- о т, Killip et al. 3837 1 (COL, US); Rio Bogota, above Apulo (Rafael Reyes), 2,100 m, Plowman & Davis 3760 (COL); Santandercito, bridge over river, 1,600 m, Silva Mojica 404 (COL); Anapoima, WSW of Bogota, 2,000 m, Triana 683/1702 (BM, COL, P, US); Mpo. Tena, Lago Pedro Palo, 3 km N of Tena, 2,080 m, Fernández & Mora 1445 (COL). HUILA: Cordillera Oriental, E of Neiva, 1,300- 1,800 m, Rusby & Pennell 998 (NY); 0.5 km above Vegalarge, 1, 300 m, Croat 55274 (MO); San Augustin, 1,500 m, Bogner 846 Lise (K). META: Cordillera Oriental, Dryander 2154 (US). Pasto: 2,200 m, Triana 682 (US). RISARALDA: along road between Pueblo Rico and Istmina, 14 km NW of Pueblo Rico, 5°03’N, 76900" W, 780 m, Croat 70848 (MO). SANTANDER: 0.5 km S of Berveo, 15 km NW of Velez, 2,000-2,500 m, Fassett 25107 (MO, US). SANTANDER SUR: Mpo. Florida, culti- vated at Jardin Botánico "Eloy Venezuela," 880 m, 7%05'N, 73°07'W, Croat 56546 (MO). VALLE: Buga- Buenaventura: near Lago Calima, 1,430 m, 3?54'N, 76°33'W, Croat 56747 (COL, MO); Km 44, Vereda El Trapiche, 1,140 m, 3%48'N, 76°37'W, Croat 56753 (CM, COL, JAUM, МО, QCA, U, 05); Cali- Buenaven- tura, NW of Loboguerrero at km 58.8, m, 3?42'N, s W, Croat 575 18А (MO), 57519(COL, F, JAUM P, ОСА, US, WIS); Cisneros, 300- ч m, Killip moms (COL, F, US); Dagua, Killip 5441 (GH, NY, US), 5442 (NY); 1, 200 m, He. oe & joie m (MO); Mun. р, along ns and Restrepo, 1,0 , Ram E Reyna As (CUVC, F, MO). Anthurium guanchezii Bunting, Phytologia 60(5): 294-295, fig 3. 1986. ТУРЕ: Vene- Volume 78, Number 3 1991 Croat 657 Anthurium sect. Pachyneurium zuela, Amazonas: Dpt. Atures, Cano Cabeza de Monteco, affluent of Rio Autana, 110-120 m, 4°52'N, 67?27'W, Guanchez & Melguei- ro 3451 (holotype, VEN; isotype, VZM). Terrestrial; internodes short, drying 2.5 cm diam.; cataphylls persisting semi-intact; petioles 30 cm long, sheathing 1.7 cm at the base, cross- sectional shape not known; geniculum 0.8-1 cm long, remote from the blade base by 4.5 cm; blades pergamentaceous to subcoriaceous, broadly ovate, gradually acuminate at apex (the acumen 25 mm long), weakly cordate at base, 33.5 cm long, 22.5 m wide; sinus 1.8 cm deep; the lower surface brown-punctate; the lower naked portion of midrib subterete and narrowly sulcate; posterior rib naked 1.5-2 cm; primary lateral veins ca. 8 per side, slender, departing midrib at ca. 40? angle, the lower 2-3 pairs at broader angle, arcuate-ascending to the margin; collective vein lacking or arising only in the upper fourth of the blade. /nflorescences slightly longer than leaves; peduncle 58 cm long; spathe reflexed-recurved, narrowly ovate, 9 cm long, 2.5 cm 1.2 cm at base; stipe 1.7 cm long in front, 4 mm long in back; spadix weakly tapered, 8.2 cm long (apex missing), 7 mm diam. at anthesis, to 1.1 cm wide, acuminate at apex, decurrent diam. in pre-fruiting condition. /nfructescence not known. Anthurium guanchezii is endemic to Venezuela and known only from the type locality at 110 to 120 m in a region which lies near the boundary between the tropical moist forest and premontane wet forest life zones. The description given here is based only on the original type description and the species is therefore imperfectly known. Still, there is little doubt that . guanchezii is a member of sect. Pachyneurium and is probably most closely related to А. irami- rezae, which is also glandular-punctate, but differs in lacking a remote geniculum and in having a cartilaginous rather than a subcoriceous blade with basal veins not naked at the sinus. In addition, 4. iramirezae has a narrower spathe, a sessile spadix and occurs at 1,350 m on granite outcrops. Anthurium guanchezii is also related to А. bon- plandii because of its glandular-punctations. It is easily distinguished from any of the varieties of that species by its ovate blades with a remote geniculum. While some specimens of А. bonplan- dii subsp. bonplandii have a remote geniculum, none have such ovate blades. VENEZUELA: AMAZONAS: Dept. Atures, Cano Cabeza de Manteco, affluent of Rio Autana, *'Manteco" rapids, 110- 120 m, 4%52'N, 67°27'W, Guanchez & Melgueiro 3451 (VEN, VZM). Anthurium halmoorei Croat, Ann. Missouri Bot. Gard. 70(2): 301. 1983. туРЕ: Mexico. Na- yarit: along Hwy. 28 between Tepic and Jal- cocotan at Km 15.5, ca. 1,000 m, Croat 45337 (holotype, MO 2690192-193; iso- types, CAS, K, MEXU, MICH, SEL). Figures 146, 147, 152. Epiphytic or epilithic; stem 1.5-5 cm diam.; roots dense, spreading to descending, tan to green- ish, smooth, short, and stubby, 7-10 mm diam.; cataphylls coriaceous, lanceolate, (5)17-23 cm long, acute and apiculate at apex, green, sometimes reddish, drying tan to brown, persisting with apex remaining intact, dilacerating at base. Leaves erect to spreading; petioles 15-26 cm long, 7-18 mm diam., D-shaped to thicker than broad, broadly sulcate adaxially with the margins sharply raised to obtuse, sharply l-ribbed to obscurely 3-ribbed or rounded abaxially, the surface pale-speckled; geniculum paler and thicker than petiole, 1–2.5 cm long; blades coriaceous, oblanceolate to oblan- ceolate— elligtio; short-acuminate at apex (the acu- men shortly pie) acute to attenuate at base, 2-108 cm long, cm wide, broadest above the middle, the margins prominently undulate; up- per surface semiglossy, medium green, lower sur- face semiglossy to matte, much paler; midrib flat to obtusely angular at base, becoming narrowly acute and then flat toward the apex above, prom- inently higher than broad апа 1-ribbed at base, becoming convexly raised toward the apex; pri- mary lateral veins 5-14 per side, departing midrib at 40—55° angle, straight then arcuate-ascending to the margin, convexly raised, much paler than surface above, prominently raised and paler than surface below; interprimary veins less conspicu- ously raised above, visible and flat below; tertiary veins obscure above, darker than surface below; collective vein arising from near the apex or absent, flat to weakly sunken above, weakly raised and darker than surface below, 6-11 mm from margin. Inflorescences spreading, shorter than leaves; pe- duncle 32-43 cm long, 9-10 mm diam., 2-2.5x as long as petiole, subterete, obscurely angled; spathe spreading, subcoriaceous, green, sometimes tinged with purple (B & K yellow-green 6/10), ovate to broadly ovate, 13-16 cm long, 6.2-9.7 cm wide, broadest near the base, inserted at 70? angle on peduncle, acuminate at apex, acute at base; spadix green, purple or green heavily tinged with purple (B & K red-purple 2/5), cylindroid, 658 Annals of the Missouri Botanical Garden 9-21 cm long, 11-27 mm diam. near base, 5-7 mm diam. near apex; flowers 4-lobed, 2.4-2.8 mm long, 2.7-3.2 mm wide, the sides sigmoid; 8-14 flowers visible in principal spiral, 12-20 in alter- nate spiral; tepals matte, weakly punctate, minutely papillate; lateral tepals 1.5-1.8 mm wide, the inner margins convex; pistils slightly raised, green, tinged with purple in the area of stigma; stigma linear, 0.4-0.6 mm long, droplets appearing 4—5 days before stamens emerge; stamens emerging + rap- idly from the base, inclined over the pistil; flaments exserted, ca. 0.3 mm long, 0.8 mm wide; anthers orange; thecae ellipsoid, scarcely divaricate; pollen orange, fading to white (B & K yellow 7/2.5). Infructescence 20-23 cm long, 3 cm diam., pen- dent; spathe usually persisting, green; berries pale yellow, greenish at tip, irregularly obovoid, 12-15 mm long, 7-9 mm diam.; mesocarp fleshy, with numerous raphide cells; seeds 2 per berry, pale tan, subglobose, weakly flattened, 5-6.5 mm long, 3 m diam., ca. 2.5 mm thick, with a sticky appendage at both ends. Anthurium halmoorei is restricted to western Mexico in the states of Nayarit, Jalisco, and Mi- choacán in seasonally very dry forests from 450 m to 1,000 m. It usually occurs on rocks on very steep slopes in forested areas. This species is characterized by its generally oblanceolate or oblanceolate-elliptic blades with usually free-ending primary lateral veins; by its petiole which is acute abaxially; and especially by its ovate to ovate-elliptic spathe, green to purplish spadix, and greenish yellow mature berries. Anthurium halmoorei is most easily confused with А. schlechtendalii subsp. jimenezii, which differs in having the petiole subquadrangular and usually flat abaxially, and in having a lanceolate spathe and red berries at maturity. The latter taxon occurs only in Guerrero and southern Oaxaca. MEXICO: MICHOACÁN: Aguililla, McVaugh 24741 (MICH); Coalcomán de Matamoros, Rzedowski ied (ENCB, MICH). NAYARIT: Tepic-Jalcocotán, at Km 1 Hwy. 28, 1,000 m, Croat 45337 (CAS, F, K, MEX Tepic, Norris utlan Navarro, McVaugh 10224 (MICH), Wilbur 2289 (MICH); El Tui- to-Puerto Vallarta, Delgado 359 (MICH), Hernandez 2608 (MEXU); S of San Pedro, Koeppen & Iltis 593 (BH); Talpa de Allende, McVaugh 23408 (MICH). Anthurium hammelii Croat, Monogr. Syst. Bot. Missouri Bot. Gard. 14: 109. 1986. TYPE: Panama. Bocas del Toro: 5 km ENE of Cerro Pate Macho, near Finca Serrano, NE of Bo- quete, 1,675 m, Hammel 6160 (holotype, MO 2802252). Figures 148, 153. Description based on dried material only. Epi- phytic; stem 1 cm diam.; cataphylls subcoriaceous, 6 cm long, brown, persisting semi-intact; petioles 33-34 cm long, 4 mm diam., apparently subterete; geniculum | cm long; blades subcoriaceous, oblong to oblong-elliptic, acuminate at apex (the acumen flat, 15-20 mm long), obtuse at base, 30-31 cm long, 5.5-7 cm wide, broadest = or near middle, the margins weakly undulate; xly raised above, more prominently so below; primary lateral veins 9-11 per side, departing midrib at 60° angle, arcuate to the collective vein, raised above and below; interprimary veins few, almost as conspic- uous as primary lateral veins; tertiary veins pro- minulous above and below; collective vein arising from the base, less prominent than primary lateral veins, 2-4 mm from margin. Inflorescences erect, shorter than leaves; peduncle 28 cm long, 3 mm diam., slightly shorter than petiole, terete; spathe subcoriaceous, dark reddish purple, ovate-rounded, 8.5 cm long, 7 cm wide, broadest at or near middle, the apex round and minutely apiculate, the base cordate; stipe 1 cm long in front and back; spadix cream, cylindroid-ellipsoid, 2.1 cm long, 8 mm diam. midway, broadest at the middle; flowers rhombic, 1.5-2.4 mm long, 2-2.5 mm wide; 5- 7 flowers visible in principal spiral; lateral tepals 1-1.5 mm wide, the inner margins rounded; pistils not emergent; stamens emerging well above tepals, then retracting to level of tepals, held in tight cluster above pistil; anthers 0.6 mm long, 0.6 mm wide; theca ovoid, scarcely divaricate. /nfructes- cence not seen. Anthurium hammelii is known only from the type collection made in Bocas del Toro, Panama, t 1,675 m in lower montane rainforest This species is distinguished by its leaves with the petiole longer than the blade, its oblong-elliptic blades, its large, ovate-rounded spathe, and cyl- indroid-ellipsoid, cream-colored spadix. Anthurium hammelii is similar in overall ap- pearance to А. protensum subsp. arcuatum, which differs in having proportionately shorter petioles, a much narrower spathe, and a more elongate, tapered spadix. Volume 78, Number 3 1991 Croat 659 Anthurium sect. Pachyneurium PANAMA. BOCAS DEL TORO: ENE of Cerro Pate Macho, near Finca Serrano, NE of Boquete, 1,675 m, Hammel 6160 (MO). Anthurium harlingianum Croat, sp. nov. TYPE: Ecuador. Napo: between Lago Agrio and Baeza at Rio Oyocachi, 1,620 m, 0?20'S, 77%55'W, Croat 50290 (holotype, MO 2819231.32; isotypes, K, NY, QCA, RSA, USM). Figures 7, 16, 149, 150, 154, 155. Planta оа d brevia, 3-5(6) cm diam.; cataphyllum lan rmie vel cucculato, per- sistens semi- m petiolus 4.5-13 cm longus, 7-15 mm diam., adaxile uninervis, abaxile trinervis-quinque- nervis; lamina (32)50- 90(150) ст longa, 18-40(60) cm lata, oblongo-oblanceata; nervis primariis lateralibus 12- 10-25(47) cm longus, 4-7 mm diam. Baccae albae usque ad purpureae, oblongae, 7-8 mm longae. Epiphytic; stem to 30 cm long; internodes short, 2.5-5(6) cm diam.; roots dense, ascending to spreading, whitish to green-gray, velutinous, to ca. 13 cm long, 3-6 mm diam.; cataphylls lanceolate and hook-shaped or cucullate, less often straight, subcoriaceous, short, to 7 cm long, persisting semi- intact. Leaves erect-spreading; petioles 4.5-13 ст long, 7-15 mm diam., D-shaped, obtusely angled with a medial rib adaxially, rounded to 3-5-ribbed abaxially; geniculum paler and moderately thicker than petiole, 1–2.5 cm long; sheath 2-10 cm long; blades coriaceous to subcoriaceous, broadly oblong- oblanceolate to = elliptic, acute at apex, shallowly cordate to truncate, rarely acute to narrowly acute at base (usually in younger plants), (32)50-90(150) cm long, 18-40(60) cm wide, broadest above the middle, the margins broadly undulate; upper sur- face weakly glossy to matte, medium green, lower surface matte, slightly paler; midrib above obtusely raised at base, becoming acutely raised toward the apex, rounded to obtusely angled below; primary lateral veins 12-23 per side, departing midrib at 30-60? angle, ascending straight to the margin, becoming arcuate near the margin, convexly to acutely raised above, convexly raised below, slight- ly paler than surface above and below; tertiary veins prominulous and darker than surface below; collective vein arising from about the middle of the blade (more often absent), prominulous, 3-10 mm from margin. /nflorescences erect-spreading, shorter than leaves; peduncle 21-50(100) cm long, 3-6 mm diam., (2.6)4-6(11)X as long as petiole, green, sometimes tinged red-violet, terete; spathe erect-spreading to recurled, sometimes twisted, subcoriaceous to coriaceous, green to violet-purple to maroon, lanceolate, 6.5-22(40) cm long, 1.5- 4.5 cm wide, inserted at 50—70° angle on peduncle, acuminate at apex, acute at base; spadix dull salm- on-pink to pale violet-purple (B & K purple 5/5), rarely greenish, sessile, slightly to moderately ta- pered, slightly to рани ian curved, 10–25(47) cm long, ca. 9-10 mm diam. near base, 4-9 mm diam. midway, 2-4 mm diam. near apex, broadest at the base; flowers square to rhombic or 4-lobed, 1.6-2.5 mm long, 1.5-2.4 mm wide, the sides straight to sigmoid; 6-13 flowers visible in principal spiral, (5)8-16 in alternate spiral; tepals matte to weakly glossy, minutely papillate; lateral tepals (0.4)0.8-1.3 mm wide, the outer margins 2-4- sided, the inner margins straight to slightly round- ed; pistils scarcely emergent to prominently raised, gray-green to purplish; stigma oblong-ellipsoid to slitlike, 0.3-0.6 mm long, reddish to violet-purple, darker than surrounding pistil; stamens emergin from the base of spadix, in a complete sequ. the laterals preceding the alternates by ca. spirals, the 3rd stamen preceding the 4th by ca. 2—9 spirals; anthers yellow-red (B & K yellow-red 8/2.5), 0.4-0.6 mm long, 0.3-0.9 mm wide, in- clined over the pistil; thecae broadly ovoid to ovoid- triangular, slightly or not divaricate; pollen creamy white, white when dried, the scent yeasty or like rotting fruit. /nfructescence pendent; spathe per- sisting; spadix 16-28 cm long, 1.5-2.5 cm diam.; berries white to violet-red, 7-8 mm long; seeds 1- 2 per fruit, oblong, 1.7-3 mm long, 1.3-2 mm wide, 0.8-1.4 mm thick, with a gelatinous ap- pendage at both ends. Anthurium harlingianum is endemic to the low- er eastern slopes of the Colombian and Ecuadorian Andes at 440 to 1,800 m (mostly above 750 m), in areas of tropical wet, premontane wet, and wetter parts of tropical moist forest, where it is almost strictly epiphytic. It occurs in a broad arc from the Serrania de Macarena in Meta Dept. in south- ern Colombia, to the Province of Zamora—Chin- chipe in Ecuador. Probably also belonging to the same species are several Peruvian collections from the vicinity of Tingo Maria in Huánuco, and one each from Junin and Pasco. However, there is room for doubt, since these areas are well out of the normal range, and no collections are known from a large intervening area. Anthurium harlingianum is distinguished by its generally large size, hook-shaped to cucullate cata- phylls, erect-spreading inflorescences, and by its leaves which usually dry dark brown. Anthurium harlingianum is closely related to 660 Annals of the Missouri Botanical Garden and perhaps only subspecifically distinct from 4. loretense and А. cataniapoense, both of which usually occur at elevations below 300 m and oc- cupy positions at opposite ends of the arc in the lowland Amazonian region. Anthurium loretense occurs at the south end of the arc in eastern Loreto Dept. in Peru, and A. cataniapoense occurs at the north end in the watersheds of the Rio Negro and Rio Orinoco in Venezuela. Both of these lowland species are distinguished from 4. harlingianum by their pendent inflorescences and more glossy leaf blades. The blades of A. harlingianum are typically matte with less prominent tertiary veins than those of the other two species. Of these two species, A. loretense is closest to A. harlingianum geographically and taxonomically. It differs in hav- ing an inflorescence with both the spathe and spadix averaging about 6 cm longer than in 4. harlin- gianum, and the spadix is typically more long- tapered, averaging two-thirds the length of the peduncle. In А. harlingianum, the inflorescence is typically erect at anthesis, but may become pen- dent, and the spadix is usually more stubby. It averages only about half as long as the peduncle and has a proportionally somewhat shorter spathe (ca. 3/5 as long). All three of the above taxa share a lanceolate, usually hook-shaped cataphyll and similarly short- petiolate, coriaceous, oblanceolate leaf blades which frequently have closely set primary lateral veins at the base. In addition, all three have similarly D-shaped petiole cross sections with raised margins and an obtuse medial rib, as well as purplish violet spadices (pale purple and also ranging to other colors in the case of A. harlingianum). Aside from cultivated material, the spadix of Harling & Andersson 24496, from Могопа-Ѕап- tiago, represents the lower end of spadix-length range (6.5 cm long) and is unusual in being stipitate by 5 mm. Croat 58162, from San Martin Dept. in Peru at 700 m, also has a stipitate spadix (6 mm in back); with its larger spadix it appears to link the last-mentioned specimen with typical ma- terial from Ecuador and with collections from fur- ther south in Peru. However, in view of its ter- restrial habit on limestone rocks and leaves which dry pale green, it may represent a distinct taxon. More collections of middle-elevation Pachyneu- rium from the Department of Amazonas in Peru are needed to help solve this and similar problems in the section. Two collections from the same lo- cality in Junin, Peru (Killip & Smith 25528, 25604), are unusual in being quite small for typical A. harlingianum, with leaf blades up to 10 cm wide, less than 50 cm long, and spadices са. 4—7 cm long. Anthurium harlingianum is named in honor of the Swedish botanist Gunnar Harling, who coau- thored the Flora of Ecuador project with Benkt Sparre, and who has led several collecting expe- ditions to Ecuador. COLOMBIA: META: Sierra de la pem northern pla- teau, 1,000 m, Philipson 2342 (BM, COL); Rio Güejar- Rio Sansa, 500- E т о 479 с ы Rio Güejar-Morro Ballo | Tablazo, 500-1,2 ‚ Cha parro et al. 74 (COL), ША еї al 440 ( ona PUTUMAYO: El Whiskey, S of Umbria, vic. Finca Santa Marta, Plow- man 2066 (F); Río Mocoa, m E of Mocoa, 800 m, Plowman 2027 (COL, Е, GH); 700 m, 1?10'N, 76°33'W, Croat е (MO). ECUADOR: MORONA-SANTIAGO: 6 km món, rd. to La Union, 1, m, Harling yr vet 24496 (GB); а е Bomboiza, Misión Salesiana, 85 km NNE of , 700-800 3°26'S, 78°32'W, Sparre 19106 (Sx Саша, Za. mora, 5 km 5 of El Pangui, 800 m, 4?21'S, 78°50' Croat 50816 (MO, QCA, RSA), 50817 (MO); 31 ма N of Yangzatza, 1,000 m, 4?11'S, 78°49'W, Croat 50786 (CM, MO); 20 km S of San Juan Bosco, 1,300 m, Mad- ison & Besse 7559 (MO, SEL). маро: Lago Agrio- Baeza, 3 ($); Rio MUS cuim 23 km NNE of бар, N of El Chaco, 1,500-1,620 m, 0°18'S, 77°48'W, Balslev & Madsen 10512 (AAU), Croat 50290 (K, MO, NY, QCA, RSA, USM); 43 km E of Lago Agrio, 560 m, 0°07'N, 77°15'W, Croat 50437 (MO); 67 km W of Lago Agrio, 760 m, Croat 49506 (MO, QCA); 67.5 km W of Lago Agrio, 1,180 m, 0%01'N, 77°19'W, Croat 50477 (MO, QCA); 72.5 km W of Lago Agrio, 1,166 m, Croat 49518 (MO, NY, ОСА); 74.5 km W of Lago Agrio, 1,220 m, 0°00'N, 77°20'W, Croat 50455 (МО); Lago Agrio-Coco, 26 km N of Coco, 450 m, 0°29'S, 76°55'W, O 5042 1 (BM, MO, US); Lago Agrio- Rio San Miguel, .5 km N of Lago Agrio, 450 m, 0%'N, 76?50'W, Croat 50319 (MO, QCA); 2 km 5 of Rio San Miguel, 2 km N P Lago Agrio, 470 m, 0%08'N, 76°50'W, Croat 50350 (CM, МО, ОСА, RSA), 50352 (M, MO), 50355 (AAU, CAS, CM, MO, QCA); Tena, Asplund 10182 ($); Tena-Puyo, 1.7 km S of Rio Маро, 440 m, 1904'5, 77*47'W, Croat 58899 (MO); Tena-Puyo, 5 km 5 of bridge over Rio Napo, 510 m, 1905'S, 77°47'W, Croat 58922 (MO, QCA, UC); Tena-Puyo, 58-61.5 km N of Puyo, 500-600 m, Croat 49629 (MO, QCA), 49663 (МО); Río Aguarico, road to Сере Ferry Crossing, E of Lago Agrio, 450 m, 0%2'N, 76°50'W, Croat 50429 (МО); Rio Tena, 2 km W of Muyuna, 5.7 km W of Tena, 500 Cantón Archidona, Challuayacu, Carretera Hollin- Loreto, Km 25-31, 1,230 m, 0943'S, 77?40'W, Hurtado 1161 (MO); Carretera Hollin-Loreto-Coca, between Rio Pu cuno and Rio Guamani, 0?46'S, 77°26’W, 1,100-1,200 m, Céron 2901 (MO, QCNE); Cantón Loreto, Huaticocha, Carretera Hollin- Loreto, 575 m, Hurtado et a (MO); Reserva Biologica Jatan Sacha, Rio E of Misahuallí, 1°4’S P (MO, QCNE), Palacios 2465 (MO). PASTAZA: Mera, 1,160 m, Harling 3689, 3815 (5); Rio Pastaza, Mera, Lugo 1116 (СВ); Мега-Вапоз, 1 km E of Topo, 1,300 m, Volume 78, Number 3 1991 Croat 661 Anthurium sect. Pachyneurium 1°27'5, 78°10'W, Palacios et al. 190 (MO, NY, QAME, QCNE); 3 km from turnoff from Puyo-Mera road, Tar abita, 1,000 m, Croat 49688 (MO, QCA Puyo- Daños, 5.7 km W of Shell, 1,070 m, 1?26'S, 78*09'W, Croat egro, 2 km W of El Topo, 1,160 m Croat 59110 (F, MO, QCA); Puyo-Macas, 19 km S of uyo, 1,200 m, 1%37'S, 77*53'W, Croat 50558 (MO, QCA), 50565 (МО, QCA), 50574 (CM, MO, QCA), 50576 (MO); Puyo- an 33 km S of Puyo, 16 km S of Escuela Fiscal VAR 900 m, 1?38'S, 77*52'W, Croat eda (MO, QCA); 3.8 km from Puyo-Macas road, road t Canelos, 850 m, 1?37'S, 77*51'W, Croat 58984 4 (MO. . ZAMORA-CHINCHIPE: 2 km М of Zamora, Madison 2460 (SEL); carretera ч PETRO- шош under construction, via Аиса, 11 m S of Coca, S of Río Tiguino, 320 km, 1?15'S, aa Zak & Rubio 43 18A (MO); Loja- Zamora, 15 km above Zamora, 1,800 m, Madison 2503 (SEL). PERU: HUANUCO: Tingo Maria Region, Prov. Huánuco, Huánuco-Pucallpa, Tingo Mar- ia, Naranjillo, 600-700 m, Ferreyra 2210 (USM, NY); л Leoncio Prado, Rio Huallaga, 650-800 m, Croat 57955 (F, MO, RSA, USM), Gentry et al. 15925 (MO), 15970 (F, MO), Plowman 7 yon mírez ias (F, MO, US), Solomon 3387 (MO); 2 pe 1,000 m, Winter s.n. (USDA er a at Un of Hawaii d 398) (MO); Prov. Leoncio Prado. Dtto. Rupa Rupa, Cerro Quemado, E of Tingo María, 800-850 m, 5 hunke 10498 (MO); Tingo María- Pu- pun La Divisora, 1,650 m, 9%07'S, 75*48' W, Plowman n. (Selby Garden 78-1918) (MO). JUNÍN: 2 trail, Yapas, 1,350-1,600 m, Killip & Smith 25528, 25601 (NY, US), 25604 (US). pasco: Prov. Oxapampa, 20 p from Pte. Paucartambo, rd. to Oxapampa, 1,100 10?41'S, 75?24'W, Smith 7570 (MO). SAN MARTIN: Moy. obamba- Chachapoyas, 13.5 km W of Naranjos, km marker 415, 700 m, 5?38'S, 77*25'W, Croat 58162 (MO, USM). Anthurium holmnielsenii Croat, sp. nov. TYPE: Ecuador. Pichincha: between Pacto and Nue- vo Azuay, 15.3 km N of Pacto, 1,320 m, O°11'N, 78°04’W, Croat 61640 (holotype, MO 3422019; isotypes, B, K). Figures 156, 157, 159-161. Planta epiphytica vel terrestris; internodia brevia, 1.5— cm diam.; cataphyllum lanceolatum, persistens fibris linearis; petiolus (18)24-47 cm longus, 7-12 mm diam., acute D-formatus; lamina coriacea, oblongo- oblanceolata -98.5 ст longa, (11)13-27 cm lata; nervo F app e basi. Pedunculus 24-64 cm longus, 3-6 mm iam., 1.4-2-plo longiorus quam ape pa 6-17.5 cm long i 5-2.7 cm lata; spadix 0.5 cm longus, viridis, dm contractus; baccae mide d 3.5-4.7 mm longae, 2.3-2.7 mm Epiphytic, occasionally terrestrial; stem 1.5-3.5 cm diam.; leaf scars 1.5-1.7 cm high, 2.2-2.5 cm wide; roots dense, descending, pale, moderately elongate; cataphylls coriaceous, lanceolate, 17-25 cm long, narrowly acute at apex, drying light brown to tan, persisting as linear fibers at the apex of the stem. Leaves spreading; petioles (18)24-47 cm long, 7-12 mm diam., sharply D-shaped and slight- ly thicker than broad, broadly convex to weakly ribbed medially adaxially, the margins sharply raised to winged, rounded abaxially; geniculum thicker and paler than petiole, becoming fissured trans- versely with age, 1–3 cm long; blades coriaceous, oblong-oblanceolate to oblanceolate-elliptic, long- acuminate at apex, narrowly acute at base, (50)76- 98.5 cm long, (11)13-27 cm wide, the margins broadly undulate; upper surface matte to semi- glossy, dark green, lower surface semiglossy to glossy, slightly paler; both surfaces drying con- colorous, green to yellowish; midrib above promi- nently convex and paler than surface, below thicker than broad at base of the blade, becoming prom- inently convex toward the apex and paler than surface; primary lateral veins numerous, to more than 20 per side, departing midrib at 45—60? angle, + straight to the collective vein, moderately ob- scure, weakly sunken to raised above, darker than surface and weakly raised below; interprimary veins almost as conspicuous as primary lateral veins, numerous; tertiary veins obscure, raised on drying; collective vein arising from near the base, weakly sunken above, equally as prominent as primary lateral veins on drying and when fresh, 5-25 mm from margin. Inflorescences spreading; 24—64 cm ong, 3-6 mm diam., 1.4-2(3.4)x as long as pet- iole, green tinged with purple, terete; spathe erect to spreading, subcoriaceous, green, sometimes heavily tinged purplish or red, broadly lanceolate to oblong-lanceolate, 6–17.5 cm long, 1.5-2.7 ст wide, broadest near the base, acuminate at apex (the acumen inrolled), acute to obtuse at base; stipe 1-2.2 cm long in front, 1-2 mm long in back; spadix green (B & K yellow-green 5/5), also re- ported as brownish, long-tapered, subsessile, straight to prominently curved, 9-20. ong, 6-9 mm diam. near base, 3—4 mm diam. near apex; flowers square, 2.3-2.7 mm in both directions, the sides smoothly to jaggedly sigmoid; 7-8 flowers visible in principal spiral, 4—5 in alternate spiral; tepals matte, te деше and minutely papillate; lateral tepals .4 mm wide, the inner margins straight, the outer margins 2—4-sided; pistils slightly raised, mi- nutely papillate, green; stigma linear, 0.4-0.5 mm long, the laterals preceding the alternates by 8-9 spirals, the 3rd stamen preceding the 4th by 1-3 spirals; filaments fleshy, translucent, somewhat flat- tened, 0.7-0.8 mm wide; anthers dark yellow, 0.6– 0.8 mm long, 0.6–0.9 mm wide; thecae oblong- ovoid, not divaricate; pollen orange. Infructescence with the spathe persisting; spadix 16.8-31 cm long, 662 Annals of the Missouri Botanical Garden 1.4-2.3 cm diam.; berries orange, oblong-ovoid, slightly beaked, 6.5-7.4 mm long, 4-4.5 mm diam.; pericarp somewhat thickened, with linear raphide cells; mesocarp transparent, gelatinous; seeds l- 2 per berry, yellowish brown when dried, oblong, sometimes ovoid, 3.5-4.7 mm long, 2.3-2.7 mm diam., 1.6-2 mm thick, with a gelatinous, sticky appendage at both ends. A member of series Multinervia, Anthurium holmnielsenii is known only from Ecuador in the provinces of Los Ríos, Imbabura, Pichincha, and Cotopaxi at 250 to 1,500 m in tropical moist and premontane wet forest, where it grows either ter- restrially or as an epiphyte. This species is distinguished by its concolorous, green-drying leaf blades with numerous primary lateral veins and a collective vein arising from near the base, sharply D-shaped petioles, long-tapered, green spadix, and orange berries. Anthurium holmnielsenii is closely allied to А. acutissimum, and differs in having petioles with sharply raised margins and spreading-pendent to more or less erect leaves with broader blades (3.3- 6.5(7) vs. (6.7)7.5—10(1 1) times longer than broad). This species is named in honor of Lauritz B. Holm-Nielsen, who has coordinated much of the botanical activity in Ecuador by the University of Aarhus, Denmark. ECUADOR. COTOPAXI: Rio Guapara, 20 km NW of El elis 250 m, Sparre 17113 bis Td IMBABURA: Col- , 840 m, " Solis ы 2889 (Е). Los Rios: aN Blanco, У ingo- Esmeraldas, 3 km S of Km 24, 250 m, 0*5'S, 79°15' W, rae 50689 MO, QCA). PICHINCHA: El Paraiso- ка al, 3 km from El Paraiso, 1,500 m, ai pk о. а $n 37773 (AAU); Pacto- Nuev у, та of La Esperanza, 1,300 т, Holm- pond s "iA 24542 Pu МО); 15.3 km М of Pacto, oad, 2.3 km М of Paraiso, 1,320 m, 0?11'N, 78%04'W, Croat 61640 (B, K, MO). Anthurium iramirezae Bunting, Ann. Missouri Bot. Gard. 76: 917-918, 1989. TYPE: Ven- ezuela. Amazonas: Dpt. Rio Negro, Serrania de Tapirapeco, outcrop of granitic rocks, 1,350 m, 1?20'N, 64*55'W, I. Ramirez & Las- kowsky 310 (holotype, VEN). Figure 353. Epilithic; stem thick, reclining horizontally; in- ternodes short; cataphylls persisting. Leaves few; petioles 19-40 cm long, subterete, sulcate adaxi- ally, the margins rounded; geniculum 1.7- long, 1-1.2 cm thick; blades held erect on petioles, ovate, obtuse to short-acuminate at apex (the acu- men with a cusp to 1.3 cm long), cordate at base, 30-37 cm long (equaling or twice as long as pet- ioles), 19-24 cm wide (1.5-1.7 x longer than wide); .3 cm sinus arcuate; upper surface glossy, lower surface slightly paler and brown glandular-punctate; major veins prominently raised on upper surface; basal veins 2 pairs, free to base (not naked at the sinus); primary lateral veins 7(8) per side, departing midrib at 45-55? ascending to the margin; tertiary veins prominent on both surfaces. Inflorescences longer than leaves; peduncle 45-71 cm long; spathe recurved or re- flexed, persistent, green, narrowly oblong-ovate, 11 cm long, 0.8 cm wide, decurrent at base for 8 mm; spadix dark purple, becoming green with age, dd tapered, sessile, 13.3-15.3 cm long, 7.5 m diam.; 6-8 flowers visible in principal spiral. иле to 1.5 ст diam.; berries unknown. angle, almost straight or weakly arcuate- Anthurium iramirezae is endemic to Venezuela, known only from the type Dr in an area of premontane rainforest, а! 1, The description given here is cd only on the original type description and the species is therefore imperfectly known. This species is distinguished by its broadly ovate, subcordate, glandular-punctate blades and narrow- ly oblong-ovate spathe. It is most closely related to А. guanchezii, but that species differs in having the geniculum remote from the base of the leafy portion of the blade, as well as by having naked basal veins and a narrowly ovate spathe which is 3.6 х longer than broad. Anthurium iramirezae is perhaps most easily confused with А. wurdackii, but that species has proportionally narrower (about 1.9x longer than broad) blades with angular rather than acute sinus and epunctate lower surface, and its spathe de- current for 3. 7.5 ст VENEZUELA. AMAZONAS: Rio Negro, Serrania de Tapi- rapeco, outcrops of granitic rocks, 1,350 m, 1?20'N, 64*55'W, Ramirez & Laskowsky 310, 318 (VEN). Anthurium jenmanii Engl, Pflanzenr. IV. 23B(Heft 21): 72. 1905. TYPE: Guyana. Es- sequibo, Jenman 5760 (lectotype, К; isolec- totypes, BRG, NY). Anthurium oe Bunting, Acta Bot. Venez. 10: 270. Venezuela. Monagas: between Guachardo & oa del Guácharo, 1 km above El Guacharo, Bunting 2653 (holotype, MY). Anthurium trinitatis Engl., Pflanzenr. ТУ. 23B(Heft 21): 73. TYPE: Trinidad, Fendler 741 (holotype, K; isotype, NY). Epiphytic, epilithic or terrestrial; stem short, (1)1.5-3 cm diam.; roots dense, spreading to as- cending, pale green to brownish, thick and blunt, ca. 5-10 cm long, 3-6 mm diam.; cataphylls Volume 78, Number 3 1991 Croat 663 Anthurium sect. Pachyneurium broadly lanceolate, subcoriaceous to coriaceous, (3)5-7 cm long, acute at apex, drying dark brown, persisting semi-intact or as a reticulum of fibers. Leaves erect to erect-spreading; petioles (3)7- 20(38) cm long, 6-15 mm diam., bluntly to sharply D-shaped, flattened to weakly convex adaxially, rounded abaxially, the surface pale-speckled; ge- niculum slightly paler and moderately thicker than petiole, 0.5-2.5 cm long; sheath 1-6 cm long; blades coriaceous, broadly oblanceolate to elliptic (rarely ovate-elliptic), shortly and abruptly acute (rarely long-acuminate or obtuse-rounded) at apex, narrowly acute to obtuse to broadly truncate to sometimes weakly subcordate at base, (22)40-103 cm long, (8)11–52 cm wide, broadest near the middle or in the upper fourth, the margins mod- erately and broadly undulate; upper surface semi- glossy to glossy, yellowish green, lower surface matte to weakly glossy, moderately paler, drying yellow-green, matte; midrib flat at base, soon be- coming bluntly acute above and conspicuously pal- er than surface, prominently convex to higher than broad at base, becoming obtuse-rounded toward the apex below, slightly paler than surface; primary lateral veins 5-13 per side, departing midrib at 30-60” angle, broadly arcuate-ascending to the margin, the uppermost merging into a collective vein, raised near midrib, sunken at margin, mod- erately paler than surface above and below; inter- primary veins absent except near the base, less conspicuous than primary lateral veins; tertiary veins sunken above, raised below, raised when dried; reticulate veins prominulous above and below when dried; collective vein arising in the upper third of blade or absent, equally as prominent as primary lateral veins, 6-15 mm from margin. /nflores- cences spreading-erect to spreading-pendent, equaling or longer than leaves; peduncle 26-86 cm long, 5-9(15) mm diam., 3.4-6X as long as petiole, pale-speckled, colored like petiole, some- times ringed with purple, terete to elliptic; spathe reflexed, usually soon withering and becoming twisted, subcoriaceous, pale green tinged with pur- ple, becoming purple at anthesis, oblong to lan- ceolate, 11-50 cm long, 1.5—4.5 cm wide, broad- est near the base, inserted at 30-60? angle on peduncle, acuminate at apex (the acumen 8 mm long), narrowly acute to obtuse, sometimes decur- rent for 3-4 cm at base; stipe absent or to 4-5 cm long in front, to 1 cm long in back; spadix dark maroon to dark purple, sometimes bluish pre-an- thesis, tapered to cylindroid, usually sessile to weakly stipitate, nearly straight to slightly curved, 10.4- 47 cm long, 6-9(12) mm diam. near base, ca. 4 mm diam. near apex, broadest at the base; flowers rhombic, 1.8-2.7 mm long, (1.3)1.7-2.2 mm wide, the sides straight to slightly sigmoid; 6-14 flowers visible in principal spiral, 6-10 in alternate spiral; tepals matte to semiglossy, minutely papillate; lat- eral tepals 1-1.2 mm wide, slightly convex, erose, the outer margins 2-sided; pistils not emergent, whitish; stigma slitlike, colored like tepals, 0.5 mm long; stamens emerging in a slow, regular sequence from the base, the laterals preceding the alternates by 7-12 spirals, the 3rd stamen preceding the 4th by 1-8 spirals, held against the pistil and obscuring it; anthers pale yellow to orange, 0.4—0.5 mm long, —0.7 mm wide; thecae ovoid, divaricate; pollen yellow, fading to whitish, moderately sweet-scent- ed. Infructescence pendent; berries pale reddish purple to purple or white tinged with violet apically, obovoid, rounded at apex, to 10 mm long, 7 mm diam.; seeds 2 per berry, oblong to obovoid, 8 mm long, 3 mm diam. Anthurium jenmanii ranges from Trinidad and Tobago and Venezuela (Sucre and Monagas; one collection from northeastern Bolivar) to Guyana, Surinam, French Guiana, and Amapá in Brazil. It is the only species, besides 4. fendleri, which is extra-Amazonian in northern South America and which also occurs in lowland Amazonia. Instead of achieving this via the Guianas, A. fendleri ranges south to the departments of southern Colombia. Anthurium jenmanii occurs mostly below 500 m in moist forest as well as in dry open woodlands and granite outcrops, where it may be terrestrial, epiphytic, or epilithic. is species is recognized by its large, coriaceous blades with impressed tertiary veins on the upper surface (when fresh), which usually dry yellow- green. Also characteristic are its purple spadix and spathe, the latter usually withering early, and by its reddish purple berries. The leaf bases are highly variable, ranging from acute to somewhat subcor- date (see below). Staminal progression is markedly slow, the spadix requiring over one month under office conditions (room temperature) to complete anthesis throughout its length of Specimens from throughout the range of А. јеп- manii may have basally obtuse-rounded or even subcordate leaf blades; Sheffer 265, originally col- lected in Trinidad and cultivated at the University of Hawaii, and Sastre 1665, from the Tumuc- Humac Mountains in Brazil, are examples of the latter end of the spectrum. One collection, Geyskes 16, from the Tumuc-Humac Mountains along the Brazilian border in Surinam, is noteworthy in hav- ing leaf blades conspicuously pustular on the lower surface. This character is typical for Anthurium 664 Annals of the Missouri Botanical Garden bonplandii subsp. guayanum, but that subspecies typically has often dark, plate-shaped glands in addition and occurs at higher elevations; no col- lections of the latter are known from the Tumuc- umac Mountains. Sastre 1665 and Granville 1179, also from these mountains, are more typical of А. jenmanii and have no pustules on the leaf blades. This region is in general poorly known, and further collections would greatly help to understand the geographical distribution and variation of 4. bonplandii and А. jenmanii, and also А. cata- niapoense. Another noteworthy collection is Cre- mers 7405, from north-central French Guiana, which is aberrant in having a leaf blade so attenuate at the base that the geniculum appears remote from the blade by up to 4.5 cm In 1975, Bunting described 4. englerianum, based on material collected in Sucre in the Coastal Cordillera of Venezuela, 1 km above Guácharo. Examination of the type and material collected at the same site (Croat 54389) revealed no characters that can be used to separate А. englerianum from A. jenmanii, despite the fact that the type locality of A. englerianum is at 970 m elevation. It is here synonymized for the first time. WITHOUT LOCALITY. Potter 5299 (NY); cultivated, ен 45053 (MO). BAHAMAS. Cultivated, Nassau, Pro ). BRAZIL: Serra и 0 m, Sastre 1 665 (САУ, 05). (IAN, NY); Cachoeira Utussansain, island in Cachoeira, , US); Rio Araguari, downriver from orto Platon, Pires et al. 1152 (NYy Rio Jari, , 3°33'N, 51937 & 823 (US); Cachoeira Mu. Roche, Froes 26695 (IA META Luetzelburg 20265 (M); AMAPÁ- PARÁ: erra Тач umaque, ма Rio Cumina, Затра dir a u ba Granville et al. 9779 (CAY, P, US); Bassin du Haut- Marouini, 230 m, Granville et al. 9908 (CAY), 200 m, 9291 у не Bassin де l'Oyapock, 10 m, Cremers 9936 (CAY); Camopi River above its mouth at Tamouri River Granville 2097 (CAY, US); Grand Canori, Oldeman 1979 (CAY) Montagne Maripa, Granville 2870 (CAY, Р); Pedra Alice, Irwin et al. 47566 (NY); opposite Pedra Alice, Irwin et al. 47604 (МУ ); Savanes- Roches, Fleuve Oyapock, Oldeman 2575 (MO); St. Elie, Prevost 740 (CAY); NW of Sinnamary, W of Orstom “Ecrex” Project, St. Elie Track, 250 m, Croat 53848 (MO); Brazilian border, Trois Sauts, 300-350 m, Grenand 1091 (CAY, P); Massif T Emerillons, 450 m, ORA 6594 (CAY, K) 300-350 m, Cremers 6639 (CAY); Montagne des Trois А A second peak, Oldeman 2609 (CAY, US); Monts de la Trinite, Inselberg NW of Monts de la Trinité, Cremers 7405 (K); Mt. St. Marcel, Haut Oyapock, 300- 450 m, Sastre 4412 (PX Haut Tampoc, S of Pier Kourou, Cremers 4517 (CAY); Tumuc-Humuc, Granville 1179 CAY, Py Brazil and Surinam border, summit of Palou- louimeenpeu, 707 m, Granville 1103 (CAY, K, MO). GUYANA: Cayenne, Fevillet 1126 (U); Tumatumari, Glea- son 172 (NY); Rockstone, Gleason 617 (GH, NY); Ma- zaruni station, Forestry Dept. 2939 (K); Quebrada Ca- monnie, Jenman 2026 (K); Demerara River, Christianburg, Bartlett s.n. (BRG); Río Essequibo, Morabilli Creek, near Bartica, 0 m, Sandwith 427 (K); Mazaruni River, upper part, Leng 355 (МУ); Potaro-Siparumi Region. Chena- pou, Amerindian village (Patumona), 50 km upstream fron Kaieteur Falls, 450 m, 5%00'N, 59°34'W, Kvist et al. 302 (05). ESSEQUIBO: Jenman 5760 (ВЕС, К, NY); ~ 44872 (MO, SEL). SURINAM: Cultivated, originally сој- lected by Dodson, Paramaribo, Madison 1927 (SEL) Nassau Mts., Marowijne River, 550 m, Cowan & Linde- man 39100 (F, NY, US, U); Paloemeu River, Palaimee Tumuc-Humac Mts., Te- NY); Wosia Falls, 300 m, Daniels & Jonker 833 (NY, U). saRAMACCA: Wilhelmina Gebergte, WSW of Juli An- atrop, Werkhoven 16712 (BBS, SEL). Товасо. Great Dog River, Eggers 5802 (US); Little Tobago, Broadway 4909 (US), Dinsmoore 38 (WIS), Purseglove 6325 (K, US); Mason Hall, Broa d 4832 (US). TRINIDAD: Broad- way 4106 (BM), Fendler n. (BM), Fendler 741 (BM), atu Croat 53915 bows Mor ruga, Broadway 7621 Мај 76 ( an Eos m, Philcox et al. 8075 (TRIN); en route to Huevos ulf o 17050 (BM, ie г NEZUELA. pd AR: E x Nuria, E of Mia 500-600 m, Steyermar (NY, US); Rio Toro ( (Rio Grande) Ri o La Reforma- uid Rico, N of El Palmar, 200-250 m, Steyermark 88110 Е, NY, US, VEN). monacas: El Guácharo-La Cueva Guácharo, Bunting 2653 (MY, NY), 4351 (NY) km above Guacharo, 1.5 km below Cuevas del а 970 m, 10?10'N, 63931" М, Croat 54389 (B, MO, VEN). SUCRE: El Pilar-Guariguen, 4-10 km 5 of El Pilar, 10°31'N, 63%06'W, Croat 54382 (Е, IBE, К, M, MO, SEL, US); Quebrada del Purgatorio, E of Grapa, Lasser is Vareschi 3851 (VEN); Peninsula de Paria, Cerro Pa- ao, 300 m, 10°41'N, 62°3'W, Milliken et al. 38 (МО); сећи Colón, Broadway 170 (СН, NY, US); Cuman- acoa-Caripe, La Cuestra, 6 km NW of Las Piedras, 760 m, Bunting 2697, 26975 (NY); Lago de Guanoco, Lasser & Vareschi 3906 (VEN); Dtto. Cajigal, Santa Isabel, Fernandez 3349 (MY). ~ Volume 78, Number 3 1991 Croat 665 Anthurium sect. Pachyneurium Anthurium johnsoniae Croat, sp. nov. TYPE: Originally collected in Colombia. Magdalena: between Santa María and El Campano, vic. of Minca, road above Santa Maria via Bonda, cultivated by Marilyn and Al Johnson, Croat 62835 (holotype, MO 3656446). Planta terrestris; internodiis brevibus, caule 1-3 cm m persistens intactum mox deciduum; quadrangularis, complanatus ialiter, 2-3 costatus abaxialiter; lamina late oblan- ceolata, (24)50- 70 cm longa, (11)17-30 cm lata; nervis pureus in apice, ochraceus ad basim. Baccae ignotae Stem short, + slender, 15-50 cm long, 1-3 cm diam.; roots moderately numerous, spreading to descending, grayish white to whitish green, gla- brous, minutely but conspicuously warty, elongate, weakly tapered, 3-5 mm diam.; cataphylls thinly subcoriaceous, narrowly straight-lanceolate, prom- inently 1-ribbed throughout, 7-18 cm long, acute to obtuse and with subapical apiculum, pale green (B & K yellow-green 6/5), drying reddish brown (B & K yellow-red 4/7.5), persisting intact, even- tually deciduous. Leaves erect to spreading; peti- oles 12-29 cm long, 4-10 mm diam.; D-shaped to quadrangular, flattened to shallowly sulcate adaxially, sometimes with the margins weakly raised, prominently 2—3-ribed abaxially; geniculum thick- er than petiole, becoming fissured transversely with age, 1.3-2 ст long; blades subcoriaceous, broadly oblanceolate, abruptly and shortly acuminate at apex (the acumen apiculate), obtuse to acute to narrowly acute at base, = ст long, (11)1 7-30 ст wide, broadest above middle, the margins broadly undulate; upper surface weakly glossy, dark to medium green (B & K yellow-green / 10), lower surface semiglossy to weakly glossy, moderately paler; midrib flat at base, becoming obtusely and narrowly angular toward the apex above, paler than surface, prominently higher than broad with sharp ridge at base, becoming convexly raised and flat toward the apex below; primary lateral veins 6-16 per side, departing midrib at 50-90? angle, straight to weakly arcuate, convexly raised, much paler than surface above, slightly raised below; tertiary veins obscure above and darker than surface below; collective vein arising from below the middle, weakly sunken above, slightly darker than surface below, 3-7 mm from margin. /nfloresences with peduncle 24 cm long, 3-5 mm diam., 0.9X as long as petioles, green subterete; spathe reflexed, subcoriaceous, green, lanceolate, 9.8 cm long, 1.4 cm wide, broadest near the base, acute at apex (the acumen inrolled, 3 mm long), acute and somewhat d spadix maroon in the upper 4, brownish yellow- ochre at base (B & K yellow 5/5), long-tapered, sessile to subsessile (stipe to 2 mm long), slightly curved, held at 160? angle from peduncle, (7)16- 22 cm long, 5-8 mm diam. near base, 2-3 mm diam. near apex, broadest at the base; flowers rhombic to square, 1.5-2 mm long, 1.2-1.5 mm wide, the sides straight, becoming smoothly sig- moid; 7-8 flowers visible in principal spiral, 10- 12 in alternate spiral; tepals densely and minutely papillate; lateral tepals 0.8-1 mm wide, the inner margins broadly rounded, the outer margins 2-3- sided; pistils not seen; stamens barely emerging above level of tepals and persisting, the laterals preceding the alternates by 12-14 spirals, the 3rd stamen preceding the 4th by 7—8 spirals, held in a tight cluster above the pistil; anthers yellow, 0.4— 0.5 mm long, 0.5 mm wide, obscuring pistil; thecae oblong-ovoid, slightly divaricate; pollen bright yel- low. Infructescence with the spathe persisting; spa- dix 23 cm long, 2.8 cm diam., bearing berries in t at Dase; 9 the basal portion only; berries orange, broadly el- lipsoid, apparently truncate and with radial ridges at apex, 10-12 mm long, 4.8-5 mm diam.; seeds 2 per berry, dark yellowish with numerous pale raphides on the surface, oblong, flattened, 4.7-6 mm long, 2-2.5 mm diam., 1.2-1.7 mm thick. Anthurium johnsoniae is known only from a live collection made by Marilyn and Al Johnson near Minca in Magdalena Province in Colombia, apparently i in a premontane wet forest life zone. This species is distinguished by its glabrous, mi- nutely warty roots, leaf blades with primary lateral veins spreading at a broad angle and with a col- lective vein mostly arising from near the base, and by its ellipsoid, orange berries. Because of the broadly spreading primary lateral veins and basal collective vein origin, A. johnson- iae resembles most closely А. caucavallense, also from Colombia, which has smooth roots and purple berries. The species is named in honor of one of the collectors, Marilyn Johnson, of Miami, Florida. COLOMBIA. MAGDALENA: between Santa Marta and El Campano, vic. of Minca, road above Santa María, via Bonda, originally collected by Marilyn & Al Johnson, Croat 62835 (B, COL, K, MO, NY, US). Anthurium knappiae Croat, sp. nov. TYPE: Peru. San Martin: Lumas, on old trail from Yum- batos to San Antonio de Cumbasa, S of Sha- pajilla, lower slopes of Cerro Isco, 400-500 666 Annals of the Missouri Botanical Garden Kna et al. 8544 (holotype, MO 3483655: isotypes, B, US, USM). Planta epiphytica; caulis i vim cataphyllum аа 7m in fibris pallidus; petiol 2 cm longus, 6- diam., sulcatus; lamina late elliptica, 57-61 cm bon 18-19.5 cm lata, basi attenuata; pedunculus 74-93 cm E: ceolata, viridis, 10-12 1 3 ст lata; зрафх cremeus ia viridis, cylindroidius, 6-11 cm longus, ca. 4 mm m dia Description based on dried material only. Epi- phytic on rotten logs; stem unknown; cataphylls ca. 7 cm long, apparently persisting as weathered pale fibers. Leaves erect-spreading; petioles 34.5- 42 cm long, 6-7 mm diam., sulcate with the mar- gins apparently acute, apparently acutely 1-ribbed abaxially; geniculum ca. 1 cm long; sheath 4.5 cm long; blades subcoriaceous, broadly elliptic, acu- minate at apex, long-attenuate at base, 57-61 cm long, 18-19.5 ст wide, broadest at or near the middle, the margins flat, both surfaces yellow to olive-green; midrib convexly raised above, promi- nently raised and apparently acutely angled toward the base below; primary lateral veins 12-16 per side, departing midrib at 50—80° angle, slightly arcuate to the collective vein, weakly raised above and below; interprimary veins much less to almost as conspicuous as primary lateral veins; reticulate veins weakly raised; collective vein arising from the base, equally as prominent as primary lateral veins, 10-16 margin. Inflorescences erect; peduncle 74-93 cm long, 4-5 mm diam., 1.8-2.4X as long as petiole, green, terete; spathe spreading to reflexed, subcoriaceous, green, lan- ceolate, 10–12 cm long, 1–1.3 cm wide, broadest near the base, inserted at ca. 20? angle on pedun- cle, acuminate at apex (the acumen with apiculum mm from 4 mm long, inrolled and recurled), prominently decurrent for 2-3 cm at base; spadix 99 “creamy, probably somewhat greenish (post anthesis), cyl- indroid, sessile, 6-11 cm long, 4 mm diam.; flowers square, ca. 4 mm in both directions; 5-6 flowers visible in principal spiral, 4—5 in alternate spiral; tepals with conspicuous pale raphide cells; lateral tepals 1-2 mm wide, the inner margins Mna rounded, the outer margins 2-sided; anthers ] mm long, 1.4-1.8 mm wide, inclined over ilis pistil; thecae ovoid, slightly divaricate. /nfructes- cence (immature) erect-sreading; spathe persisting, green; spadix 8 cm long, 1 cm diam.; berries green, ellipsoid Anthurium knappiae is endemic to Peru, known only from the type locality, at 400 to 500 m, probably in an area of premontane wet forest. The type specimen was collected in the Tarapoto region, about 20 km north of Tarapoto. It is one of six or more Pachyneurium species from this region. This species is distinguished by its broadly ellip- tic, long-petiolate leaves which dry green and have the collective vein arising from near the base, and by its long-pedunculate inflorescence with a green- ish spadix. This species is most closely related to А. oxy- carpum, differing in its petiole being 2-3 х longer, its attenuate leaf base (only rarely attenuate for A. oxycarpum), its much longer peduncle (74-93 cm long vs. 18-55 cm long for 4. oxycarpum), and its pale, raphide-cell-covered, nonglaucous tepals (vs. conspicuously glaucous and apparently free of raphide cells in 4. oxycarpum). The new species is named in honor of Sandra Knapp, who made the type collection. PERU. SAN MARTÍN: Cerro Isco, lower slopes, old trail from Yumbatos to San Antonio de Cumbasa, Lumas, 5 of Shapajilla, 400-500 m, 6?20'S, 76?24'W, Knapp et al. 8544 (B, MO, US, USM). Anthurium krukovii Croat, sp. nov. TYPE: Bra- zil. Amazonas: Maraa, Rio Japurá, Lago Ma- raa, vic. of Maraa, 1%51'S, 65?36'W, Plow- man et al. 12211 (holotype, INPA; isotype, MO 3117449). Planta terrestris aut epiphytica; internodia brevis, 5- mm diam.; cataphyllum lanceolatum, persistens in- tactum; petiolus 19.5-38 cm longus, 3-7 mm diam., pura anguste sulcatus; nins ovato-elliptica, 26-40 cm lon | редипс и oblongo- cde. 8 cm longa, 7 m 3.5 cm longus; spadix atrorubens ad brunneus, 12-15 cm longus, 4-5 mm diam.; baccae violaceae. — = Description based on dried material only. Ter- restrial or epiphytic; stem 5-10 mm diam.; roots nse, numerous, spreading, pubescent, relatively short, bluntly pointed, 1-5 mm diam.; cataphylls lanceolate, thin, 4–8 cm long, acute at apex, brown, persisting intact, eventually deciduous. Leaves with petioles 1 m long, ca. 3-7 mm diam., subterete, narrowly sulcate with rounded margins adaxially, rounded abaxially; geniculum darker than petiole, 0.6-1.5 cm ovate-elliptic, long-acuminate at apex, acute to semi- rounded at base, then narrowing abruptly to the geniculum (making geniculum seem remote from base by 4-11 cm), 26-40 cm long, 11-18 cm wide, broadest below or near the middle, the mar- gins concave toward the base, apparently shallowly undulate; иш semiglossy, green to yellowish green; long; blades subcoriaceous, ly convex above, sharply acute below; primary lateral veins 7-10 per side, de- Volume 78, Number 3 1991 Croat 667 Anthurium sect. Pachyneurium parting midrib at 40—65? angle, + straight, raised above and below, more sharply so below; tertiary veins weakly raised; reticulate veins obscure above, weakly raised below; collective vein arising from near the base to about the middle of the blade, equally as prominent as primary lateral veins, 5— 15 mm from margin. /nflorescences with peduncle 31-57.5 cm long, ca. 3-4 mm diam., equaling or to 1.8x longer than petiole, light to olive-green, terete; spathe recurled, membranous, pale reddish green, oblong-lanceolate, to 8 cm long, ca. 0.7 cm wide, broadest near the base, narrowly acute at apex, acute and decurrent at base; stipe 1.5-3. cm long in front, (0)8-17 mm long in back; spadix dark red to brown, long-tapered, usually long-stip- itate, 12-15 cm long, 4-5 mm ca. 2 mm diam. near apex; flowers + square or rhombic when dried, 1.2-1.8 mm in both direc- tions, the sides straight to smoothly sigmoid; 5-7 flowers visible in principal spiral, 8-9 in alternate spiral; lateral tepals 0.9-1.2 mm wide, the inner margins straight to broadly rounded, the outer mar- ins 2-sided; pistils emergent, not raised; stigma ellipsoid, 0.5-0.7 mm long; anthers 0.2-0.3 mm long, 0.4-0.5 mm wide, inclined over pistil; thecae ovoid, slightly divaricate. Infructescence with spathe deciduous; spadix ca. 1 cm diam.; berries violet, (rehydrated) globose to oblong-ovoid, rounded at apex, 3.5-3.7 mm long, 2.6-3.5 mm diam.; peri- carp somewhat thickened with moderately numer- ous pale raphide cells; seeds 1 or 2 per berry, brown, oblong-ovoid, somewhat flattened laterally, 2.5-2.8 mm long, 1.8-2 mm diam., 1.2-1.4 mm thick. iam. near base, Anthurium krukovii is known from a few col- lections made in the central, northern and south- western parts of Amazonas, Brazil, below 150 m, probably in a tropical moist forest life zone. The species is char 1 mainly by the shape of its leaf blades, which are rounded and abruptly attenuate at the base and narrowed even further to the geniculum in a manner that makes the latter seem remote from the base of the blade by 4-11 cm. Anthurium krukovii is mostly closely allied to A. atropurpureum var. arenicola, which differs in having more coriaceous and usually narrower leaf blades which are long-attenuate (with more or less parallel margins) toward the base, and in having mostly shorter petioles. The same general differ- ences apply to the typical variety of А. atropur- purem, which is smaller in overall size. The occurrence of clear pustules on the under- surface of the dried leaf blades in the present species is of interest. The same feature occurs, albeit rarely, in А. atropurpureum, the presumed closest relative of А. krukovii. The two collections of A. krukovii which show this feature are Byron et al. 621 and Duarte 6958. The new name Anthurium krukovii honors Bo- ris Alexander Krukoff, one of the first major Amer- ican collectors to explore Amazonian Brazil, and whose philanthropy funds many important botan- ical functions to this day. In addition, Krukoff was the first known collector of the present species. BRAZIL. AMAZONAS: Barcelos, Duarte 6958 (RB); Rio Embira (tributary of Rio Tarauaca), 7%30' hé 70915", Krukoff 4927 (NY); Mpo. Maraa, Rio Japurá, vic. Maraa Lago Maraa, dene et al. 12211 (MO, ПАРА); Sitio Fortaleza, 7 km NW of Maraa, Plowman et al. 1227 (INPA, NY); mouth of Igarape Maraa and Lago Maraa, Plowman et al. 12380 (NY, INPA); Mpo. Tefe, Lago Tefe, Vila Nogueira, Porto Camanaus, Byron et al. 621 (MO, INPA) Lago Tefe, NW shore, Plowman et al. 12442 (MO, INPA). Anthurium lanjouwii Jonk. & Jonk., Acta Bot. Neerl. 15: 133. 1966. TYPE: Surinam. Emma Mountains, 5 slope, 850 m, Daniels & Jonker 1101 (holotype, U). Figure 348. Terrestrial; stem to 15 cm long; internodes short, 1-1.5 cm diam.; roots moderately dense, spread- ing, to 25 cm long, the younger roots light brown, villous; cataphylls narrowly triangular, 3.4-5 cm long, persisting semi-intact toward apex, as fibers toward base. Leaves with petioles 7-23 cm long, drying 3-7 mm diam., subterete, sulcate adaxially, drying with sharp margins, rounded abaxially; ge- niculum sharply sulcate, 0.5-1 cm long; blades subcoriaceous, elliptic to ovate-elliptic, acuminate and apiculate at apex, obtuse to rounded or some- times cuneate at base, 23-36 cm long, 12-19 cm wide, drying yellow-green to yellow-brown; midrib prominently riased on both surfaces, drying angular and paler than surface below; primary lateral veins 6—9 per side, departing midrib at 40—60° angle, moderately straight to the margin, then markedly ascending and merging with the margin or loop- connecting with the next higher vein, raised on both surfaces; tertiary veins drying prominulous below; collective vein lacking or arising below the middle of the blade. /nflorescences erect, shorter than leaves; peduncle 17-39 cm long; spathe promptly withering and soon deciduous, green tinged with purple, lanceolate, 2.5-4.5 cm long, 0.7-0. cm wide, inserted at ca. 45? angle on peduncle, mucronate at apex, acute at base; spadix purplish brown, weakly tapered, short-stipitate, 5-14 cm long, drying 4-6 mm diam.; flowers rhombic, 2.5— 668 Annals of the Missouri Botanical Garden 3.3 mm long, 1.7-2.3 mm wide, the sides smoothly sigmoid parallel to spiral, jaggedly sigmoid perpen- dicular to spiral; 5-7 flowers visible in principal spiral; tepals matte and brown on drying; lateral tepals 0.8 mm wide, the inner margins rounded, the outer margins bluntly triangular; pistils not at all emergent; stigma linear, slitlike, to 0.8 mm long, stamens emerging in a regular sequence, held just above the tepals in a tight cluster above pistil, inclined over and obscuring the pistil; anthers 4 mm long, 5.5 mm wide; thecae ovoid, moderately divaricate. /nfructescence unknown. Anthurium lanjouwii is endemic to Surinam, known only from steep southern slopes of the Emma Mountains, on sandstone and dolomite, at 850 m. This species is most closely related to Anthurium bonplandii and may prove to be only subspecific with that taxon. It differs from that species in being isolated geographically, having more or less elliptic blades and lacking plate-glands or pustules. It is geographically closest to 4. bonplandii subssp. uayanum, which has much larger, oblanceolate to obovate blades with conspicuous dark glandular- punctations. FRENCH GUIANA. Sommet Tabulaire, 40 km SE of Saul, 650-700 m, Cremers 6469 (CAY). SuRINAM. Emma Mountains, south, sandstone slope, 850 m, Daniels & Jonker 1101, 1241, 1264 (U). Anthurium latissimum Engl., Pflanzenr. IV. 23B(Heft 21): 292. 1905. ТУРЕ: Peru. Junin: Prov. Tarma, La Merced, Rio Chanchamayo, 1,000 m, Weberbauer 1939 (holotype, B; isotype, МО-3122266). Figures 167-170. + Terrestrial or epiphytic; stem to 30 ст long, 2-3.5 cm diam.; roots dense, descending, white, fuzzy, short, ca. 3 mm diam.; cataphylls subco- riaceous, lanceolate to hook-shaped, 2-ribbed, 5- 11.5 cm long, acute to acuminate at apex, drying reddish brown, persisting semi-intact, eventually as a reticulum of fibers. Leaves erect-spreading; pet- ioles 8-48 cm long, 4-12 mm diam., sharply D-shaped to C-shaped, flattened to sulcate and with a medial rib adaxially, the margins sharply to blunt- ly raised or bluntly acute, rounded to 3-5-ribbed abaxially, the surface pale-short-lineate, sometimes tinged with purple; geniculum paler and thicker than petiole, becoming calloused and minutely transverse-fissured with age, 1—2 cm long; sheath 3-10 cm long; blades subcoriaceous, broadly ob- lanceolate to obovate-elliptic to broadly elliptic, acute to acuminate at apex, acute to attenuate at base, (26)50-125 cm long, 10.5-36.5 cm wide, broadest at or above the middle, the margins un- dulate, concave toward base; upper surface semi- glossy, dark to medium green, lower surface semi- glossy to weakly glossy, moderately paler; midrib above flat to bluntly raised at base, becoming sharp- ly acute toward the apex and paler than surface, below bluntly acute at base, becoming broadly con- vex to acutely raised toward the apex; primary lateral veins 6-14 per side, departing midrib at angle, retrorse to 130? toward the base, + straight to arcuate ascending to the margin, raised above, bluntly convex below; interprimary veins not visible; tertiary veins weakly raised above, flat and darker than surface below; collective vein arising from near the base or near the apex, prom- inulous, 2-10 mm from margin. /nflorescences erect to spreading; peduncle (10)30-98 cm long, 2-9 mm diam., X as long as petiole, green to reddish green, terete; spathe reflexed-spreading, sometimes recurled, subcoriaceous, sometimes red- dish, usually green tinged with red or purple (B & K yellow 5/7.5), broadly lanceolate, (3.5)8–12(18) cm long, 1.3-2.5 ст wide, broadest near the base, inserted at 40—60° angle on peduncle, acute to acuminate at apex (the acumen apiculate), acute to rounded at base; spadix olive-green to gray- green to brownish to violet-purple (B & K red- purple 2/2.5), sometimes weakly glaucous, slightly tapered, sessile or stipitate to 2 cm, erect, straight to curved, held at ca. 180? angle from peduncle, 5-13(28) cm long, 4-7 mm diam. midway, 2-4 mm diam. near apex; flowers rhombic to 4-lobed, 2-2.3 mm long, 2 mm wide, the sides smoothly sigmoid to jagged; 7—10 flowers visible in principal spiral, 5-7 in alternate spiral; tepals minutely pa- рате, matte, white-pustulate; lateral tepals 0.5- 0.6 mm wide, the outer margins 3-4-sided, the inner margins weakly curved; pistils papillate, mi- t to maroon; stig- ma linear, 0.4 mm long; stamens emerging in à regular sequence, the laterals preceding the alter- nates by 4-14 spirals, the 3rd stamen preceding nutely granulose, raised, red-violet the 4th by 5-7 spirals, arranged in a circle around the pistil; anthers pinkish to reddish brown, 0.5- 0.7 mm long, 0.5-0.7 mm wide, thecae divaricate; pollen yellow fading to white, yeasty-scented. /n- fructescence with spathe persisting; spadix 9-17 cm long; berries maroon to deep red-violet (B & K red-purple 2/5), obovoid, flattened to depressed at apex, 6.5-7.5 mm long, 5-6 mm diam.; me- socarp transparent, gelatinous; seeds 2 per berry, white tinged with red-violet at base, 4.5-6 mm long, 2-3.5 mm diam Anthurium latissimum is endemic to Peru, where it ranges from San Martin to Huánuco, Junin, and Volume 78, Number 3 1991 Croat 669 Anthurium sect. Pachyneurium Ayacucho at 500 to 1,800 m in tropical moist, premontane moist, or tropical dry forest life zones. his species is characterized by its usually long- petiolate blades which are rather abruptly attenuate in the lower third with concave margins, and es- pecially by having the primary lateral veins in the lower part of the blade departing the midrib at a right angle to the midrib, or even markedly retorse toward the base, before arching upwards toward the margins. Anthurium latissimum is perhaps most closely related to А. ernestii, which has generally smaller leaves that lack the retrorse or perpendicular veins, and usually occurs below 500 m. Younger plants, represented by Killip & Smith 24612 for example, may have small blades that are scarcely attenuate at the base. They also differ in having a stubbier spadix with only five to six flowers per spiral. Also noteworthy is Croat 5117 1, which dried greenish in contrast to most other, brown-drying material. It differs in no other re- spect. While the tepals of most dried collections are at least sparsely papillate, a collection of cultivated origin believed to be this species (Croat 522304) has densely granulose tepals which appear to be paler than those of most other collections. PERU. Without locality, cultivated at MO: Croat 522304 (B, CAS, CM, Е, К, M, MO, NY, P, RSA, S, M). AYACUCHO: Huanta-Rio Apurimac, Auina, 750- 1,000 m, Killip & Smith 22836 (МУ, US). Schunke 10498 (MO). JUNÍN: San Ramón ps apampa, of turnoff to Satiapo, 700 m, 10953"М, 75?18'W, Croat 57727 (AAU, BM, MO, RSA, USM); Prov. Tarma, Tarma-San Ramón, Schunke Hacienda, va Merced, 1,300 m, Macbride 5678 (Е); Tarma-San Ramón, Schunke Hacienda, 1,400-1,700 m, Killip & Smith 24589, 24612 (NY, US); Rio Chan- ко La Merced, 1,000 m, Weberbauer 1939 (B, MO); йөре re of Rio Chan nchamayo, ca. 1,000 m, Soukup 44 US). SAN n Río luallaga, Tingo Maria-Huánuc d km fro m, Allard 21977 (US) Prov La yobamba, 10 km NW o 76°43'W, Croat 51171 м US); Pr ceres, Dtto. Uchiza, Tingo María- dice Nuevo, behind Ramal de Aspusana, 25.8 km N of turnoff to Tocache Nuevo from Tingo Maria-Pucalpa er 500 m, 9%07'S, 76°03'W, Croat 57967 (MO). Anthurium lennartii Croat, sp. nov. TYPE: Ec- uador. Loja: Celica—Zapotillo Road, ca. 3 km below Pozul, 1,400 m, Harling & Andersson 18071 (holotype, MO 2908943; isotype, GB 897-61). Figure 171. Planta epiphytica; internodia ca. 1.5 cm diam.; cata- phyllum persistens in fibris tenuibus; petiolus 7-11.5 cm longus, 5-7 mm diam., D-formatus; lamina oblongo-ellip- tica, 55-98 cm longa, 10-17.5 cm lata, apice obtusa ad rotundata; d 20.5-37.5 cm longus, 3-4 m diam.; spatha 7.5-9 cm longa, 1.5-2 cm lata; spadix lentiter contractus, 7 cm longus, 5-6 cm diam., viridis ad lilicinus; baccae rubentes ad atropurpureae. Description based on dried material only. Epi- hytic; stem ca. 1.5 cm roots moderately numerous, pale grayish, shortly pubescent, mod- erately elongate, ca. 2-3 mm diam.; cataphylls subcoriaceous, ca. 10 cm long, light brown, per- sisting as linear fibers, soon deciduous. Leaves with petioles 7–11.5 cm long, 5-7 mm diam., D-shaped, with the margins sharply raised adaxially, probably 2-3-ribbed abaxially; geniculum slightly thicker diam.; than petiole, sometimes fissured transversely, 0.8— 1.5 ст long; sheath 3.5-5.5 cm long; blades sub- coriaceous, oblong-elliptic, obtuse to semi-rounde at apex (the acumen shortly apiculate), obtuse to narrowly rounded at base, 55- m long, 10- 17.5 cm wide, broadest at or near the middle, both surfaces matte, green to yellowish; midrib convexly raised above, somewhat acute at base, becoming prominently convex toward the apex below; pri- mary lateral veins 16-23 per side, departing midrib at (40)55—85? angle, arcuate to the collective vein, convexly raised above and below and paler than surface; interprimary veins moderately numerous, raised on both surfaces; tertiary veins weakly raised above and below; collective vein arising from the base or near the base, less prominent than primary lateral veins, raised above and below, 2-7 mm from margin. Inflorescences with peduncle 20.5- 37.5 ст long, са. 3-4 mm diam., 2.9-3.7X as long as petiole, greenish to yellowish green, prob- ably terete; spathe apparently spreading, subco- riaceous, green, broadly linear to linear-lanceolate, 1.5-9 cm long, 1.5-2 cm wide, broadest near the base, abruptly acuminate at apex (the acumen in- rolled), acute at base; stipe 13 mm long in front, 3 mm long in back; spadix green, weakly tapered, 7 cm long, 5-6 mm diam. near base, 4 mm diam. near apex, broadest at the base; flowers rhombic to 4-lobed, 1.6-1.9 mm long, 1.5-1.8 mm wide, the sides usually jaggedly sigmoid, sometimes smoothly sigmoid; 6-8 flowers visible in principal spiral, 5-6 in alternate spiral; tepals roughened; lateral tepals 1-1.3 mm erose, broadly convex, the outer margins 2-sided; pistils exposed, yellowish; stigma slitlike, dark, ca. 0.4 mm long; anthers 0.5-0.7 mm long, 0.7–0.9 mm wide; thecae ovoid to oblong-ovoid, 0.3-0.4 mm wide, not divaricate. /nfructescence with spathe wide, the inner margins 670 Annals of the Missouri Botanical Garden persisting; spadix 15.5-26 cm long, 2.2-2.5 cm diam., with tepals becoming enlarged and conspic- uous, often purplish; berries red to dark purple, after rehydration ovoid to ovoid-ellipsoid, acute to + rounded at apex, 6.3-8.5 mm long, 3.7-4.3 mm diam.; pericarp translucent with pale, linear raphide cells; mesocarp gelatinous, translucent, reddish; seeds 1-2 per berry, yellowish, ovoid to oblong-ovoid, flattened, 3-3.5 mm long, 2-2.7 mm 1-1.5 mm thick, enveloped by gelatinous substance. diam., Anthurium lennartii is endemic to southern Ec- uador, where it occurs in Loja Province, in lower montane moist to premontane dry forest life zones, at 1,400 to 2,700 m. It is expected in adjacent Peru. This species is characterized by its greenish- drying, oblong-elliptic blades that are obtuse to narrowly rounded at the apex, its basally originat- ing collective vein running fairly close to the mar- gin, its moderately numerous primary lateral veins, and by its weakly tapered spadix. Unique features of Anthurium lennartii are its conspicuously ac- crescent, purplish tepals and red to purple berries. Anthurium lennartii appears to be superficially similar to 4. manabianum. The two species share in common general blade color on drying and spa- dix and spathe shape, but differ in blade shape and stipe length, which is up to 3 cm in 4. тапа- bianum. An important difference between these species is their distribution, with A. lennartii oc- curring in southern Апдеап Ecuador and А. man- abianum in the coastal part of the country, in Manabi Province at 450 m. Anthurium lennartii is also similar in its ap- pearance to 4. sodiroanum Engl., sect. Xialophyllium, which vaguely resembles oth- er species of sect. Pachyneurium. Anthurium len- nartii differs from Anthurium sodiroanum by hav- ing shorter nodes, collective vein running closer to the margin, more prominent primary lateral veins, a slightly broader spathe, and an elevational range to 2,700 m The new species is named in honor of Lennart Andersson of the University of Goteborg, who col- lected, with Gunnar Harling, all known species of A. lennartii. a member of ECUADOR. ТОЈА: 8 km W of Celica on road to Alamor, 2,000 m, Harling & Andersson 22159 (GB); Celica- Gauachanama, Km 8, 2,700 m, Harling & Andersson 22302 (GB); Celica-Zapotillo, 3 km below Pozul, 1,400 m, Harling & Andersson 18071 (GB, MO). Anthurium leonianum Sodiro, Anales Univ. Centr. Ecuador 17(123): 256. Jan. 1903. TYPE: Ecuador. Imbabura: W slopes of Volcan Cotatachi, 0°22'N, 78?20'W, Sodiro s.n. (ho- lotype, B; isotype, QPLS). Figures 172-174. Terrestrial on rocky slopes; stem to 3 cm diam.; roots spreading-descending, whitish, appearing pu- escent when fresh, thick, rather short, promi- nently tapered; cataphylls subcoriaceous, lanceo- late, longer than petioles, acuminate at apex, drying brown, persisting as fine linear fibers. Leaves erect- spreading to spreading; petioles 5-22 cm long, 10- 15 mm diam., D-shaped, flattened and with a me- dial rib to broadly and sharply sulcate adaxially with the margins prominently raised, 5-8-ribbed abaxially, the surface pale-speckled; geniculum thicker and paler than petiole, 1-1.5 cm long; sheath 6-8 cm long; blades coriaceous, obtuse to acute at apex, rounded to acute at base, (23)43- 101 cm long, 8-29 cm wide, broadest at or near the middle, the margins prominently undulate; up- per surface matte to semiglossy, dark green, lower surface matte to weakly glossy, paler, both surfaces drying yellowish green; midrib flat at base, becom- ing acutely raised toward the apex and conspicu- ously paler than surface above, prominently higher than broad at base, becoming acutely raised toward the apex below; primary lateral veins 10-15 per side, departing midrib at 45—60° angle, ascending + straight to near margin, then arcuate-ascending and merging with margin, prominently convexly raised above and below; tertiary veins prominulous when dried above and below; collective vein arising from near the apex, raised above and below when dried, usually less than 5 mm from margin. /nflo- rescences + erect, equaling or longer than leaves; peduncle 46-91 cm long, 10-17 mm diam., 4- 5x as long as petiole, subterete to 1 -ribbed; spathe reflexed, coriaceous, green, long-lanceolate, 29- 35 cm long, 2-5 cm wide, broadest near base; spadix dark purple (pre-anthesis), becoming green- ish brown, long-tapered, 23.5-42 cm long, 10- 12 mm diam. near base, 5-7 mm diam. near apex; flowers 4-lobed, 2.2 mm long, 1.5 mm wide; 14– 15 flowers visible in principal spiral, 7-9 in alter- nate spiral; tepals weakly and minutely papillate; pistils weakly emergent; anthers ca. mm long. Infructescence, 2.5-3 cm diam.; berries dark pur- ple at apex, obovoid. Anthurium leonianum is endemic to Ecuador in the province of Imbabura from 900 to 1,500 m, in lower montane dry or premontane moist forest life zones. Моште 78, Митбег 3 1991 roat 671 Anthurium sect. Pachyneurium This species is recognized by its thick, erect to spreading leaves which dry yellow-green, its peti- oles which are sharply sulcate adaxially and 5-8- ribbed abaxially, and by its long-tapered, deep purple spadix which becomes greenish brown at anthesis. It is also reported to have deep purple berries. Anthurium leonianum might be confused with A. dombeyanum, which, in Ecuador, occurs only in the provinces of Loja and Tungurahua (only two collections are known from the latter province). Anthurium dombeyanum differs mainly in having the petioles rounded abaxially. In addition, the leaf blades of А. leonianum are conspicuously yellow- green on drying, while in 4. dombeyanum they are usually brownish or dull grayish green. The spadix of А. dombeyanum is generally shorter than that of А. leonianum. Anthurium leonianum also bears some resem- blance to two species from central Colombia, 4. glaucospadix and А. caucavallense. Both have, at least at times, yellowish green leaf blades on drying, though the color more closely approaches that of some specimens of 4. dombeyanum. Al- though the cataphylls of the type specimen of 4. leonianum at Berlin appear to be hooked and re- folded, they are straight and lanceolate on living collections from the same area. It is here believed that the cataphylls of the type specimen are so shaped because of the way it was prepared. This character, along with the lower angle of the primary lateral veins (45-60 vs. 50-90?) and peduncle- petiole ratio (peduncles 4-5 vs. 1-2 times longer than the petioles in A. leonianum), serve to dis- tinguish this species from А. caucavallense. An- thurium glaucospadix also differs in having the primary lateral veins depart the midrib at a broad angle, as well as in having a bluish green, glaucous spadix and red berries. The name Anthurium leonianum was used in the Flora of Rio Palenque (Dodson & Gentry, 1978), based on Dodson & Tan 5389, but that collection represents a new and unrelated species, А. spar- reorum, described in the present treatment. The latter species has a thinner blade with the collective veins arising at the base and running close to the margin, whereas А. leonianum has a coriaceous blade with the collective vein arising from one of the primary lateral veins in the upper М of the blade. A character that may be used to distinguish the two immediately is the number of primary lateral veins per side (10-1 (15)20-30 reorum is known from lower elevations, generally around 200-300 m in premontane wet forest. in A. leonianum vs. in A. sparreorum). Anthurium spar- ECUADOR. IMBABURA: Ibarra-Lita, 1,500 m, Cobb 21 Е (МО, QCA); Salinas-Lita, 900-1,000 m, 0%45'N, 78°15'W, eer 2262 (К, MO); Tercer Paso, on San Lorenzo RR, 1,100 m, Madison et al. 4960 (MO, SEL); Volcán ul W slopes, Sodiro s.n. (B, QPLS); Río Meta, Ibarra-Lita, E of La Carolina, border of Carchi Prov., 1,090 m, Croat 38986 (MO) Anthurium lindmanianum Engl., Bot. Jahrb. КАР 25: 367. 1898. TYPE: Brazil. Mato Gros- о: Cupim near Palmeiras, Lindman 2455 1 y (lectotype, S). Figures 175-177. Anthurium douradense deg Rev. Goiana Med. 16: 31- 33. 0. TYPE: Brazil. Goias: Serra Dourada, Rizzo 4532 (holotype, UFC, isotype, RB). Terrestrial or epilithic, rarely epiphytic; stem 1–5 cm diam.; roots moderately dense, descending, fuzzy, drying 2-4 mm diam.; cataphylls subcoriaceous, lanceo- late, 2–9(15) cm long, acute to acuminate at apex, frequently creeping over ground, green, drying reddish brown, persisting + intact or as coarse linear fibers. Leaves erect-spreading; petioles (2)11-57(65) cm long, (2)8-18 mm diam., bluntly to sharply D-shaped, flattened to slightly sulcate adaxially, rounded abaxially, the surface pale-speckled; geniculum slightly paler and thicker than petiole, (0.2)0.8—2 cm long; blades coriaceous to subcoriaceous, oblanceolate to broadly elliptic, acute to obtuse to short-acuminate at apex, cu- neate-attenuate to obtuse to rounded, rarely shal- lowly cordate at base, (14)20-84 cm long, (4.2)12- 29(34) cm wide, broadest usually above the middle, the margins usually flat, sometimes broadly un- dulate; upper surface glossy to semiglossy, medium green (B & K green 2/2.5), lower surface matte, paler; major veins sometimes paler on both sur- faces; midrib convexly raised above, prominently and acutely raised below (about as high as broad); primary lateral veins (4)7—10(14) per side, de- parting midrib at (30)40—60(80)° angle, arcuate- ascending to the margin or to the collective vein, raised, sometimes becoming sunken toward margin above, raised below; tertiary veins weakly etched, and sometimes concolorous above, prominulous and darker below, conspicuously raised on both surfaces when dried; collective vein arising from about the middle to near the apex of the blade or absent, sunken above, weakly raised below, 5—13 mm from margin. /nflorescences erect to spreading, equaling or longer than leaves; peduncle (15)35-134 cm long, (1)2-12 mm diam., 1.3-2.3(14.6)X as long as petiole, terete; spathe spreading to reflexed, sometimes recurled, subcoriaceous, withering and/ or deciduous, yellow-green (В & К yellow-green 672 Annals of the Missouri Botanical Garden 8/10), lanceolate, (3)5-17 cm long, (0.5)1-2 cm wide, broadest near the base, inserted at 20—40? angle on peduncle, long-acuminate at apex (the acumen ca. 3 mm long), acute to obtuse at base; spadix pinkish to dull brownish olive-green to dark red or purplish violet, oblong, weakly tapered, ses- sile or stipitate to 4 cm, erect, (1.5)4-12(20) cm long, 3-7 mm diam. midway, 1.5-4 mm diam. near apex, moderately tapered; flowers 4-lobed, 1.8-3 mm long, 1.2-2.7 mm wide; (2)4-9 flowers visible in principal spiral, 4—10 in alternate spiral; tepals matte, minutely papillate; lateral tepals 0.9— 2.2 mm wide, the inner margins straight to broadly rounded, sometimes minutely and unevenly erose, the outer margins 2—5-sided; pistils not emergent before anthesis, slightly raised at anthesis, caviform with erect, reddish margins; stigma linear becoming circular, 0.2-0.3 mm long; lateral stamens pre- ceding the alternates by 2-16 spirals, the 3rd stamen preceding the 4th by 7—10 spirals; anthers 6 mm long, 0.4-0.7 mm wide, inclined over d obscuring the pistil; thecae ovoid-ellipsoid, not divaricate. /nfructescence with spathe deciduous; spadix normally more than 15 cm long, bearing berries in the basal portion only; berries dark purple to dark red toward the apex, white below, obovoid, "hollow" when rehydrated, 7-9 mm long, 4-6 mm diam.; seeds 1—2 per berry, 5.5-6.8 mm long, 3-3.5 mm diam., 2-2.3 mm thick, weakly apic- ulate at both ends, pale-punctate. Anthurium lindmanianum is known from Brazil in the states of Mato Grosso, Rondónia, Goiás and in the southern and central eastern parts of Pará at 90 to 1,000 m. It grows terrestrially in sandy soil, frequently along stream banks, among or on sandstone rocks, in gallery forest and scrub This species is recognized by its broadly elliptic or sometimes oblanceolate blades, which are round- ed to abruptly acuminate at the apex, by its long- pedunculate inflorescence, and by its berries, which are purple at the apex and whitish at the base. It is also distinguished by its stem creeping over the surface of the ground and by its leaves being erect from the apex of the stem. The stem may even be subterranean. In cultivation, the species grows best when given direct access to water, e.g., when grow- ing on a large brick placed in a pool of water. The appressed stem adheres closely to the brick (above the water line) in this situation. Anthurium lindmanianum is most easily con- fused with А. bonplandii subsp. bonplandii, which occurs in the same state of Brazil (Pará). Typical specimens of А. bonplandii from Рага, all col- lected north of the Rio Amazonas, mostly differ in having conspicuous, dark brown plate glands on the undersurface of the leaf blades, and promi- nently raised tertiary venation on both leaf blade surfaces on drying. In contrast, the tertiary ve- nation of А. lindmanianum is inconspicuous on the upper surface and generally less conspicuous on the lower surface on drying, and dark brown (or colorless) glandular punctations are normally not present and are not conspicuous where ob- served. In Рага, А. lindmanianum occurs in the area of the Serra dos Carajas in the central eastern part of the state, and in the Serra do Cachimbo (including the immediate lowlands to the north, i.e., the Rio Cururú) in the southwest. Thus, the species appears to prefer upland areas, a trend which con- tinues toward the south, where it is more prevalent. few collections from Рага deserve mention, being rather markedly different from typical ma- terial and possibly meriting subspecific recognition once the taxon is better represented in herbaria by collections from this area. These collections, Berg & Henderson 493, Bockerman 248, Sperling et al. 8, and Secco et al. 136, are all smaller than average for the species in all aspects and have a prominently stipitate spadix. In addition, the pe- duncle may be up to 14.6 х larger than the petiole (vs. 1.3-2.3 times for typical material), and the spadix of two of these collections is reported as "black ZIL. Cultivated at Munich Bot. Gard., Bogner 586 мој cultiva Burle-Marx, Croat 57174 (МО); i iego, California, Croat iranhas, m, Irwin et al. 17703 Caiapo, 12 km S of ПА near small creek, 720 m, шс ion 8505 (МО); 48 К 1,000 m, 17°12'5, 51 (МО, SEL), Prance & Silva 59634 (K, NY); Serra Dou- with road to Монахан on GO 070 from Goiania to Goias, 750-800 m 1997 S, 50%02'W, Th 5778 (NY). MATO "GROSSO (U); 5 km E of Chapada dos Guimaraes, road to EM BRATAL, 720 m, Prance et al. 19375 (NY); 270 km М of Xavantina, 12%54'S, 51°22'W, Gifford 146 (NY); Cerrado-Pantanal, Matas de а 600 m, Marti- nelli 366 (K); Serra Itapirapuan, Lindman 2407 1/2 (cited by Engler (1898) as 2455 1/2) (B, S); Araguaia, diregao Rondopolis, Hutchison 8548 (UEC); Fazenda Cachimbo, Cordeiro 1087 (MG, US); base camp, 12?49'S, 51°46'W, Harley et al. 10631 (К); Xavantina-Sáo Felix, 12%54'S, 51%52'W, Ratter et al. 960, 966 (К); Xavan- tina, 12954'5, 51?22'W, Gifford 145, 146 (К); Ponte de Pedra, Hatschbach & Koczicki 33205 (K); Cupim, near disch 1623 (K); баи до Копсадог, 84- 85 km N of Volume 78, Number 3 1991 Croat 673 Anthurium sect. Pachyneurium Xavantina, 550 m, Hunt 5800 (K), Hunt & Ramos 5740 (K, NY), Irwin et al. 16446 (GH, IAN, MO, TEX); о BR158, 11?17'S, 51?45'W, Thomas et al. 4440 (MO); : Mpo. Rondenepolis, Serra da Petrolina, Hatschbach 34125 (K). PARÁ: Mara- ba, Alto da Serra, Secco et al. 136 (MO); Rio и |. canal SE of Missão Cururú, Alto Tapajós, 140 m, 7°35’S, 57°31'W, Anderson 10602 (NY); 2 hours downstream from Missào Cururü, Mouro, Alto Tapajós, 100-200 m, 7915'5, 57°55'\/, Anderson 11080 (COL, К, МО, NY, U); Tapajós, Rosa & Santos 1906 (MG, MO, NY); Serra do Cachimbo, Cachimbo, 500-600 m, 9?20'S, 54*53'W, Bockermann 248 (UB), Pereira 1821 (RB); Serra dos aoe Serra Norte, A (МС, MO); 20 km NW of Serra Norte mining camp, less than 500 m, 5%55'S, 50%26'W, pes et al. 1697 (INPA, ); Serra Norte, Maraba, Clareira N-1, Cava dante & Silva 2631 (MG), Silva et al. 1631, 1866 (MG); Mpo. Itaituba, Serra do Cachimbo, 5 km from Cachimbo airport, along Rio Formiga, 500-600 m, 9°23’S, 54°55'W, Silva et al. 135 (NY, INPA). RONDONIA: ca. 35 km WSW of Ariquemes, Mineraceo Taboca a sangana, 10%02'S, 63°20'W, Zarucchi et al. 2650 (Е, INPA, MG, MO, NY, RB, US); 4 km from Ariquemes, |» 364, 200-500 m, 9%55'S, 63?06'W, Vieira et al. 8 (MG, MO, NY). Anthurium linguifolium Engl., Pflanzenr. IV. 23B(Heft 21): 162. 1905. ТУРЕ: Ecuador. Manabi: between El Recreo and Agua Amar- ga, Eggers 15530 (lectotype, B; isolectotypes, F, K). Figures 178, 180. Terrestrial; stem to 19 cm long, са. 1 cm diam roots dense, ascending, green to whitish, de ре. to smooth, short, 1–1.5 cm long, 5-6 mm diam.; cataphylls subcoriaceous, linear-lanceolate, 3-10.5 cm long, obtuse to acute or acuminate at apex, drying reddish brown (B & K yellow-red 4/10), persisting — intact, splitting at base. Leaves erect- spreading; petioles 4-7 cm long, 5-7 mm diam., erect-spreading, subtriangular, convexly raised to obtusely ribbed, sometimes broadly sulcate adaxi- ally, the margins sharply raised, somewhat rounded and sharply 1-ribbed abaxially, conspicuously swol- len to 14 mm diam. at base; geniculum slightly paler and conspicuously thicker than petiole, 0.5-2 cm long; blades subcoriaceous, oblong-linear, acute, sometimes apiculate at apex, acute to rounded at base, 25-85 cm long, 2.5-6.5 cm wide, broadest at or near the middle, the margins broadly and shallowly undulate; upper surface matte to semig- lossy, medium green, lower surface matte, con- spicuously paler; midrib above obtusely raised at ase, becoming sharply and acutely raised and higher than broad toward the apex, pale-speckled, slightly paler than surface, below acute throughout, (-70)° angle, not well distinguished from interprimary veins, slightly arcuate to the margin, weakly sunken to weakly raised in shallow grooves, rather obscure above; interprimary veins almost as conspicuous as primary lateral veins; tertiary veins obscure above, conspicuously darker than surface below, prominulous on both surfaces when dried; collective vein usually arising from near the apex, e base, obscure above, equally as prominent as interprimary veins below, sometimes from near t 1—7 mm from margin. /nflorescences erect-spread- ing to spreading, shorter than or equaling leaves; peduncle 38-57 cm long, 4 mm diam., 7.6- 19(38)x as long as petiole, green, terete; spathe spreading to reflexed-spreading, coriaceous, yel- low-green, sometimes tinged with purple abaxially (B & K yellow-green 5/7.5), narrowly lanceolate, 7-8.5 cm long, 0.8-1.5 cm wide, broadest near the base, inserted at 45? angle on peduncle, acu- minate at apex (the acumen inrolled), obtusely rounded at base; spadix glaucous, dull lavender- purple (B & K purple 5/2.5), sessile, slightly ta- pered, cylindroid, erect, 5-8.2 cm long, 7-9 mm diam. near base, 4-6 mm diam. near apex, broad- est at the base; flowers 4-lobed, 2.1-2.6 mm long, (1.5)2.4-3 mm wide, the sides weakly sigmoid; 4)7—9 flowers visible in principal spiral, (3)5- 7 in alternate spiral; tepals matte; lateral tepals 0.8— ~ 1.4 mm wide, the inner margins weakly concave to slightly rounded, the outer margins usually 2-, sometimes 3-—4-sided; pistils emergent, bright green В & К yellow 5/2.5); stigma ellipsoid, 0.3-0.4 mm long; stamens emerging in a regular sequence from the base, the laterals preceding the alternates by 9 spirals, the 3rd stamen preceding the 4th by 3 spirals, borne at edge of tepals in a circle around the pistil; anthers straw-colored (B & K yellow-red 9/10), 0.6 mm long, 0.8 mm wide; thecae ovoid, slightly or not divaricate; pollen pale yellow (B & K yellow 9/2.5), sweetly scented at anthesis. /n- fructescence with persistent spathe; fruits not seen. Anthurium linguifolium is known from only two collections made in coastal Ecuador in Manabi province, between Bahia de Caraquez and Canoa, near sea level in a very dry tropical forest life zone. This species is characterized by its linear-oblong, weakly undulate blades, short petioles, rather elon- gate stem with short, more or less erect roots throughout its length (above ground) and by its glaucous, dull lavender-purple spadix. 674 Annals of the Missouri Botanical Garden The closest ally of А. linguifolium is A. bar- clayanum. The latter differs in greenish (rarely violet-purple) spadix, much larger size, longer pet- ioles and oblanceolate blades. Anthurium barclay- anum is sympatric with the present species at one locality at least, but they occur in very different habitats. The first is epiphytic, the second terres- trial in open areas. In addition, the ratio of peduncle to petiole length is much greater in A. linguifolium (7.6–19(38) vs. 3.5-6(9)). ЕСЏАРОК. MANABÍ: 6 km N of Bahia de Caraquez, 3- 4 km E of ocean, 50 m, 0?30'S, 80?22'W, Croat 50697 (К, MO, NY, ОСА, US); El Recreo-Agria Amarga, 0?29'S, 80°27'W, Eggers 15530 (B, F, К, NY). Anthurium llewelynii Croat, sp. nov. TYPE: Peru. San Martin: Prov. Tarapoto, Rio Mayo, near Cumumbigue, 6 km 5 of Тагарою-Моуо- bamba road at Km 15, 350 m, 6%23'S, 76°39'W, Croat 51092 (holotype, MO 2819899; isotypes, B, К, M, NY, RSA, SEL, US, USM). Figures 181, 183-185. Planta terrestris; internodia brevia, 2.5-4 cm diam.; cataphyllum lanceolatum, persistens in reticulum fibra- m diam., adaxile undatus; lamina oblanceolata ad anguste oblanceata aut elliptica, (29)50-110 cm longa, (5)7-15(20) cm lata, base plerumque abrupte rotundata vel subcordata; nervis primariis lateralibus 4-12 utroque; 5 mm diam.; spatha cm pri n spadix eviter contractus, 3.6-9 cm lon- gus, iam., olivaceus r Bugs baccae atropurpu- reae, пен obovoidae, 3-5 m Usually terrestrial, rarely epiphytic; stem creep- ing, to 20 cm long; internodes short, 2.5-4 cm diam.; roots numerous, dense, descending, whitish, velutinous, 4-5 mm diam.; cataphylls subcoria- ceous, lanceolate, 3-7 cm long, acute to acuminate at apex, drying brown (B & K yellow 3/7), per- sisting as a reticulum of fibers. Leaves erect- spreading to spreading; petioles (5)1 2-40 cm long, 3-8 mm diam., erect, sharply D-shaped, flattened to slightly convex adaxially, rarely with a medial rib, the margins sharply raised, rounded abaxially; geniculum paler than petiole, becoming calloused with transverse fissures, conspicuously thicker than petiole, 0.5-2 cm long, sometimes extending be- yond leaf base up to % its length; sheath 2-6 cm long; blades subcoriaceous, oblanceolate to nar- rowly oblanceolate or elliptic, acute to acuminate at apex, narrowing toward the base, ending abrupt- ly, obtuse to rounded or shallowly cordate at base, (29)50-110 cm long, (5)7-15(20) cm wide, broad- est at or above the middle, the margins moderately to broadly undulate; upper surface glossy to semi- glossy, occasionally matte, dark to medium green B green 3/7.5), lower surface matte to semiglossy, considerably paler; midrib pale-speck- led, conspicuously to scarcely paler than surface, broadly rounded-raised at base, becoming acutely angled toward the apex above, broadly rounded- raised to bluntly angled and slightly paler than surface below; primary lateral veins 4—12 per side, departing midrib at 35-90" angle (rarely retrorse to 1109), arcuate-ascending to the margin, соп- vexly raised and paler than surface above, less prominently raised and darker than surface below; interprimary veins almost as conspicuous as pri- mary lateral veins; tertiary veins obscure; collective vein arising from near the base or in the upper third of the blade, weakly raised above and below, 3-19 mm from margin. /nflorescences erect to spreading; peduncle 23.5-75 cm long, 3-5 mm iam., equaling to 5 х as long as petiole, green to brownish green, pale-speckled, terete to subterete; spathe spreading to reflexed, coriaceous to sub- coriaceous, green, sometimes tinged with red at margins, lanceolate to ovate-lanceolate, 5-9 cm long, 1-2.5 cm wide, broadest near the base, in- serted at 45—60° angle on peduncle, abruptly acu- minate to acute at apex (the acumen inrolled), acute at base; spadix reddish olive-green, sessile, cylindroid, ray tapered, erect or slightly curved, 3 iam. midway, 3-4 mm diam. near мны flowers rhombic to 4-lobed, strongly scented like rotting fruit, 1.5-3 mm long, ‚ 5 mm -2.4 mm wide, the sides jaggedly sigmoid to straight; 4—8 flowers visible in principal spiral, 5- 7 in alternate spiral; tepals matte, sparsely white- punctate, densely papillate; lateral tepals 0.6–0.8 mm wide, the inner margins straight to broadly convex, the outer margins 2—4-sided; pistils weakly raised, olive-green to maroon; stigma linear-oblong, caviform, 5 mm long; stamens emerging in a scat- tered pattern from the base, the laterals preceding the alternates by up to 17 spirals, the 3rd stamen preceding the 4th by 13 spirals, arranged in a circle around the pistil; anthers pinkish to yellow tinged with pink, 0.3-0.6 mm long, 0.5-0.9 mm wide; thecae oblong-ovoid, slightly divaricate; pol- len yellow fading to white. Infructescence spread- ing-pendent; spadix 11-18 cm long, 1-1.3 cm diam.; berries dark purple (В & К red-purple 2/2.5), globose-obovoid to subglobose, truncate at apex, 5 mm long, 3-5 mm diam.; pericarp mod- , with raphide cells; seeds 1—2 per berry, brownish puple; oblong- ellipsoid, 4-5 mm long, 2.1-2.5 mm diam., 1.5- erately thickened; mesocarp gelati Volume 78, Number 3 1991 Croat 675 Anthurium sect. Pachyneurium 2 mm thick, with a gelatinous appendage at both ends. Anthurium llewelynii is endemic to Peru, known principally from an area of tropical dry and tropical moist forest in the vicinity of Tarapoto, at 130 m, and also in the vicinity of Tingo Maria in the Department of Huánuco to 750 m. The species is terrestrial, usually growing in sandy soil along streams, on exposed roadbanks, in shrub forests, or on rocky cliffs above riverbanks. This species is characterized by its creeping stem, its comparatively long, sharply D-shaped, erect petioles, the usually markedly arching-spreading, narrowly oblanceolate blades (2.2-3.5, averaging 2.8x longer than the petioles) which are usually abruptly and narrowly rounded to shallowly cordate at the base, as well as by the more or less green, moderately short, bluntly tapered spadix and dark purple fruits. Anthurium llewelynii is not easily confused with any other species growing in the same area and habitats in which it occurs. Only 4. plowmanii shares similar dry forest habitats in this region, but that species differs in having a peduncle equal to or shorter than the spadix, and a deeply sulcate petiole. Anthurium tarapotense also occurs in the re- gion, but occurs at slightly higher elevations (360– 530 m) in a premontane moist forest life zone. It differs in being less robust and having a generally broader, short-petiolate blade which is attenuate at the base. It is of interest that individuals of А. llewelynii brought into cultivation and grown in more mesic and crowded conditions may develop blades that are narrowly acute at the base, rather than the more characteristic narrowly rounded to subcor- date condition. Compare, for example, vouchers prepared in the field and vouchers made from greenhouse plants, respectively, of Croat 51092. Anthurium llewelynii is named in honor of Llewelyn Williams, who collected extensively in the Tarapoto area for the Field Museum in December 1929 CULTIVATED. poene hes Selby 81-76-1, Croat 57212 (B, MO). PER : Prov. Leoncio Prado, Tingo María Region, Rio Helat 750 m, Croat 21 086 US). SAN MARTÎN: Juanjui-Tarapoto, 35 km S of Тага- poto, 130 m, 6?46'S, 76?21'W, Croat 58082 (AAU, CAS, СМ, GH, К, MO, NY, USM); El Abra, 29 km S of Tarapoto, 450-540 m, 6?40'S, 76?20'W, Gentry & Smith 45014, 45091A (MO); Morales, W of Tarapoto, 360-900 m, ај 5721 (Е); Tarapoto, 350 m, Wil- liams 5737, 6625 (Е); Tarapoto- Yurimaguas, Km 5, Kennedy Ау (ex cult.; = Plowman 13300) (Е); та ауо, 6 km S of Tarapoto- Moyobamba road at Km 1 Cunumbigue, 350 m, 6?23'S, 76?39'W, Croat т (AAU, B, CM, DUKE, ENCB, IBE, JBGP, К, KYO, L. LE, M, МО, NY, RSA, SAR, SEL, TEX, US, USM); Puente Colombia-Shapaja, 280 m, Plowman 6018 (СН); Rio Shilcayo trail to Boca Toma del Shilcayo, N of Ta- rapoto, 400 m, 6°30’S, 76°22'W, Knapp & Alcorn 7336 (K, МО); Dtto. Tarapoto, Tarapoto- Yurimaguas, km 13, 750-1,000 m, Rimachi 5782 (MO). Anthurium loretense Croat, sp. nov. TYPE: Peru. Loreto: Prov. Maynas, Dist. Iquitos, Rio Mo- món and Quebrada Momoncillo, 1 km from the Rio Nanay, 200 m, Croat 51226 (holo- type, MO 2813874-80; isotypes, B, CAS, DUKE, F, K, NY, US, USM, VEN). Figures 17, 182, 186. Epiphytic; stem 15-50 cm long, (1)2.5-5 cm diam.; roots dense, spreading-ascending to spread- ing-descending, green-gray to whitish, puberulent to smooth, slender, ca. 1.5-7 cm long, 5-6 mm diam.; cataphylls subcoriaceous, sometimes coria- ceous, 4—6 cm long, hook-shaped, sometimes cu- cullate at apex, green, drying brown (B & K yellow 4/5), persisting with apex remaining intact and as a reticulum of fibers at base. Leaves erect to spreading; petioles 2.5—10(16) cm long, 6-20 mm -shaped to quadrangular, flattened to sulcate with a medial rib and the margins sharply raised adaxially, 3-8-ribbed abaxially; geniculum slightly thicker than petiole, prominently fissured transversely with age, 0.7-3 cm long; sheath 2- 8 cm long; blades coriaceous to subcoriaceous, obovate to oblanceolate to elliptic or sometimes lam., + spathulate, acute to obtuse at apex (the acumen apiculate), usually abruptly rounded to shallowly cordate, sometimes truncate or acute to attenuate at base, (36)56-120(168) cm long, 30-50(62) cm wide, broadest usually above the middle, the mar- gins broadly undulate; upper surface semiglossy, dark to medium green (B & K yellow-green 4/7.5), lower surface semiglossy to sometimes matte, slightly paler; midrib flat to obtusely raised at base, be- coming sharply acute toward the apex above, slightly paler than surface, acutely raised at base, becoming convexly raised toward the apex below; primary lateral veins 10-21 per side, departing midrib at (20)30-60° angle, + straight to the margin, prom- inently raised above, slightly paler than surface, prominently convexly raised near the midrib below, becoming acutely raised toward the margin; inter- primary veins obscure on both surfaces, promi- 676 Annals of the Missouri Botanical Garden nulous when dried; reticulate veins obscure above, darker than surface below; collective vein arising in the upper third of the blade or absent, flat to weakly sunken above, 4–10(15 Inflorescences spreading to spreading-pendent, shorter than leaves; peduncle 25-80 cm long, (2)4— 7 mm diam., 4-12(20)X as long as petiole, green to green tinged with maroon or purple-violet, te- rete; spathe erect to spreading-reflexed to recurled, coriaceous, occasionally subcoriaceous, green heavily tinged with maroon (B & K red-purple 2/7.5), lanceolate to oblong-linear, 5.5-24(35) cm long, 1.2-3m wide, broadest near the base, inserted at 75-80? angle on peduncle, obtuse to long-acu- minate at apex, acute to weakly decurrent at base; stipe 3-11 mm long in front, ca. 1 mm long in back; spadix maroon to reddish violet (B & K red- purple 3/10), cylindroid, scarcely to long-tapered, weakly curved, 7.5-28 cm lo near base, 2-6 mm diam. near apex; flowers square or rhombic to 4-lobed, 1.3-2.1 mm long, 1.3-2.4 mm wide, the sides smoothly to jaggedly sigmoid; 7-13 flowers visible in principal spiral, 5-9 flowers visible in alternate spiral; tepals matte, densely and mm from margin. ng, 6-11 mm diam. minutely papillate, with few droplets; lateral tepals 0.6-1 mm wide, the inner margins straight to broadly convex, the outer margins 2—4-sided; pis- tils emergent, not raised, purple to reddish; stigma linear, purplish violet, 0.3-0.7 mm long; stamens emerging in a regular sequence from the base, the laterals preceding the alternates by ca. 18 spirals; anthers pinkish to orange-purple, 0.3-0.5 mm long, 0.2-0.7 mm wide, inclined over and obscuring the pistil; thecae oblong-ellipsoid to obovoid, slightly or not divaricate; pollen yellow, fading to white, faintly yeasty-scented. Infructescence pendent; spathe de- ciduous or persisting; spadix 24-46 cm long, 1.5- 2 cm diam.; berries purple-violet, oblong; mesocarp juicy; seeds white. Anthurium loretense is currently known mainly from Peru in Loreto (hence the name) and Ucayali Departments at less than 350 m in a tropical moist forest life zone. A few collections are also known from Napo and from Morona-Santiago in Ecuador, Amazonas, Pasco, and San Martin in Peru, near Leticia in southernmost Colombia, and southwest- ern Amazonas in Brazil. This species is characterized by its hook-shaped cataphylls, D-shaped petioles, large, broadly ellip- tic-oblanceolate blades drying greenish brown to brown, and by its long-pedunculate, mostly pendent inflorescence with the spathe usually 25 or up to 34 as long as the long-tapered, maroon to purplish violet spadix and often held parallel to it. Although the base of the leaf blade is normally shallowly cordate with several congested primary lateral veins per side (appearing basal), leaf blades of younger plants are frequently acute to attenuate at the base. Anthurium loretense is probably most closely related to (and perhaps only subspecifically distinct from) A. cataniapoense from southern Venezuela. That taxon shares many features, including similar petioles, cataphylls and inflorescences. Anthurium cataniopoense differs, however, in having a gen- erally much smaller and more slender spadix and a more slender peduncle. Anthurium loretense is also similar to 4. harlingianum, of Ecuador and Colombia, which differs in occurring at 450 to 1,800 m and in having a mostly erect inflorescence averaging about 6 cm shorter than that o loretense. See under А. harlingianum for details. An interesting species also apparently closely related to А. loretense is А. vaupesianum, known only from Amazonian Colombia. It differs most markedly from A. loretense by its small size, more obovate leaf blades, and proportionally more slen- der spadix. One collection from near Iquitos (Croat 19342) is noteworthy. Though closest to A. loretense among the species in the area, it appears closer still to А. harlingianum because of its erect inflorescence, moderately stubby spadix and leaves drying a very similar brown color. However, it differs from both of the above species in having the spathe longer than the spadix. It perhaps represents a new spe- cies; Williams 2465 from near Leticia, Colombia, appears to represent the same taxon. wo further collections from Loreto (Vasquez et al. 4882 and Rimachi 5569) possibly belong here. The Vásquez collection has a leaf blade only 54-62 cm long and petioles 17-21 cm long, and the Rimachi collection has a spadix (at anthesis) reported as green. In addition, both collections have blades long-attenuate toward the base (ending abruptly acute at the base) and shortly stipitate spadices. Another collection (Smith 2881) from Oxapampa in Pasco Department, Peru, appears to belong here as well, although it is very far removed from the known range of А. loretense. Gentry et al. 29912 (perhaps a mixed collection) from near Iquitos, has leaf blades that are very long-attenuate in the lower third and narrowly acute at the base. It otherwise appears to belong here. Mention should be made of several collections from the rather isolated valley of the Rio Santiago and the adjacent Rio Cenepa valley in Amazonas Department, Peru, the only ones known from this department. These differ in having more flowers per spiral (10-18 vs. 5-13) and relatively shorter Volume 78, Number 3 91 Croat 677 Anthurium sect. Pachyneurium peduncles (1.5-2 vs. 4-12(20)x longer than the petioles). Local names for these plants include "uyayu ugkunamu" and “kagkur пика" (Huambisa tribe); “tukum” and “тип tukui” (Rio Cenepa). BRAZIL. AMAZONAS: Rio Jutai, 5%12'S, 69*00'W, Traill С Cano Guacaya, 700 ft., 0930'S, 70940' W, s & Cabrera 16239 (GH). Еси JADOR. MORONA- SANTIAGO: E of Montalvo, 260 m, 1?49'S, 76°42'W, Zak & Espinoza 4590, 4680 (MO). NaPo: Lago Agrio-Coca, weed road 7.2 km S of Rio Aguarico, 76° 270 m, 0°02’N, 51'W, Croat 58643 (MO, ОСА); Lago Agrio- Rio San Miguel, 5 km N of Lago Agrio, 280 2-1 76*55'W, Croat 58689, 58691 (МО, QCA); 3 km N of Lago Agrio, Croat 50344 (MO) Lago P San Miguel, 3 km from San Miguel, 350 m, Besse et al. 1549 (SEL); Lago Agrio- Puerto El Carmen 13 (MO). маро: Puerto Napo- Misahualli, junction. n Rio Misahualli and Rio Маро, Ve- т Venecia, 3. of Misahualli, 370 m, 1902'5, 42'W, Croat 58895 (МО, ОСА); Rio а 1,5 b. upstream from Puerto livar, 300 m, 0906'5, 76°10'W, Brandbyge et al. 33700 (AAU, M N of Laguna Grande, 265 m, Poulsen 797 10 (AAU); Parque Nacional Yasuní, 230 m, 0%52'S, 76°05'W, Cerón 3354 (MEXU, MO, QCNE); 200 m, Cerón & Gallo 4937 (MO, QCNE), 53396 в, МО, ОСМЕ); Anangu, in NW "22- 23'W, 0*32'S, Korning Huampami m Eo 430'S, 78°30'W, e 398 (MO); Rio Santiago, Que- 76?28'W, Gentry et al. 29672 (MO): Prov. Loreto, Tigre, Vista Alegre, 240 m, 2%40'S, 75°35'W, Lewis et al. 12841 (MO); Prov. Maynas, Alpahuayo (Estación IIAP), Vásquez et al. 5981 (MO); Iquitos Region, Río ev (IBE); Rio ioa Recreo, q 72°50'W, Vásquez & Jaramillo 1128 (MO); Rio те rear of Fundo TBohallo, l hr. upriver from mouth, 130 m, Diaz & Jaramillo 78 (CM, K, M, MO); 1 km з the Rio Nanay, Quebrada МатопсШо, 220 m, 3°43’S, 73°20'W, Croat 51225 (CAS, DUKE, К, ES US, USM), 51226 (AAU, B, CM, K, MO, RSA USM); below Balcon, ca. 95 m, McDaniel & ا‎ 26283 (IBE); Rio Amazonas, S of Iquitos, Croat 19342 (MO, F); NE of Leticia DE Caballococha, 3°54’W, 70%32'W, Williams 2465 (F); Quebrada de Sinchiqui below Santa Maria de па | McDaniel & Rimachi 23873 (IBE); Оно. Indiana, Explorama Inn, 220 m, 3° 23'S, 73°02’W, Croat 61648 (AMA 7, MO); Rio Amazonas, Quebrada del caserio de San 90 m, Rimachi 5055 P try et al. 29912 (CM, ^ MO), 61651 (CM, MO), Vásquez & Jaramillo 6321 (MO); 3°24'$, 72*49'N, 150-180 m Croat 61766, 61783 (AMAZ, MO), Gentry et al. 36577 (MO); across Rio Amazonas from Isla °55' be ا‎ et al. 42695, 54535 (MO); г. Ма аро Сат m, 3°20'S, 72*55'W, Vásquez & ps 11829 (MO, QCA) 11835 (МО); Quebrada Paparo, Río Manati, 110 m, 3°45'S, 72°55'W, Vásquez a Jaramillo 11654 (МО); n Llachapa Camp, rada Sucusari, below Mazán, 140 m ‚ Gentry et al. s (MO); Río Tamshiyacu, Caseria Alina. 130 m, my S, 72°58'W, Gentry et al. 29233 (BM, MO), Vás- z & Jaramillo 4376 (MO); Puerto Alianza, 160 m, 4208" S, 72°55'W, Vásquez & Criollo 1836 (MO); Prov. Requena, Río Tapiche, Yarina, 180 m, 5905'5, 73*50'W, Vásquez et al. 4882 (MO); Jenaro Herrera, 140 m, van Werff et al. 10061 (MO). Pasco: Prov. Oxapampa, ЊЕ 350 m, 10?11'S, 75?13'W, Smith 2881 (MO). SAN MARTÍN: Tar rapoto- Yurimaguas, km 55, NE of Pongo de Canarachi, 230 m, 6?15'S, 76°15’W, Gentry et al. 52272 (MO). Anthurium luteynii Croat, Selbyana 5(3-4): 324. 1981. TYPE: Panama. Veraguas: Rio Pri- mero Brazo, 2.5 km beyond Escuela Agrícola Alto Piedra, beyond Santa Fe, 700-750 m, Croat 25521 (holotype, MO 2827506-08; isotypes, B, C, CAS, CM, CR, DUKE, F, K, M, MEXU, NY, PMA, SEL, UCLA, US, VEN). Figures 18, 187-189, 195. Epiphytic or terrestrial; stem short; roots de- scending, whitish green, smooth to weakly pubes- cent, blunt at apex, 2-5 mm diam.; cataphylls subcoriaceous, elliptic, 10-28 cm long, minutely apiculate and inequilateral at apex, drying tan, persisting intact, soon dilacerating into reticulum of fibers. Leaves erect to spreading; petioles 9-50 cm long, 1.3-2.5 ст diam., + trapezoidal to thick- er than broad, flattened to weakly sulcate, some- times with weak rib diminishing toward the base adaxially, the margins prominently and sharply raised, sharply and prominently 1—3-ribbed abax- ially, the surface minutely pale-speckled; genicu- um much thicker and paler than petiole, 1-3 cm long; blades thickly coriaceous, oblong-elliptic to oblong-oblanceolate, gradually to abruptly acumi- nate at apex (the acumen apiculate), acute to ob- tuse at base, 50-125 cm long, 9-33 cm wide, broadest near or above the middle; upper surface — 678 Annals of the Missouri Botanical Garden weakly glossy to semiglossy, dark green, lower sur- face matte, paler; midrib flat at base, becoming convexly raised toward the apex, much paler than surface above, prominently higher than broad and sharply 1-3-ribbed at base below, gradually be- coming prominently and obtusely to acutely an- gular and eventually convex toward the apex, paler than surface; primary lateral veins 10—25 per side, departing midrib at 60—70? angle, slightly arcuate- ascending, weakly raised in grooves above, nar- rowly and convexly raised below; interprimary veins sunken above, flat or slightly raised and darker than surface below; tertiary veins obscure above, darker than surface below; collective vein arising from near the base, sometimes in the upper third of the blade, sunken or + obscure above, weakly raised to flat and darker than surface below, 3- 10 mm from margin. /nflorescences erect-spread- ing and slightly arching, equaling or shorter than leaves; peduncle 49-85 cm long, 1.3-2 cm diam., 0.5-3(6.6)X as long as petiole, green to violet- purple, with prominent rounded rib adaxially, more prominently ribbed abaxially; spathe erect to cu- cullate, hooding the spadix, coriaceous, dark pur- ple, narrowly ovate to oblong-lanceolate, 9-33.5 cm long, 2- cm wide, broadest in the lower third, inserted at 60? angle on peduncle, acuminate at apex (the acumen caudate, to 15 mm long), decurrent at base; spadix violet-purple to maroon, shortly tapered toward apex (occasionally toward both ends), 7-18 cm long, 8-28 mm base, 5-17 mm diam. near apex; flowers rhombic to slightly 4-lobed, 2-3.6 mm long, 1.8-3 mm wide, the sides straight to jaggedly sigmoid; ca. 15 flowers visible in either spiral; tepals matte to iam. near semiglossy, minutely punctate, covered with dust- .2-1.5 mm wide, the inner margins turned up against pistil; pistils exserted ca. 0.5 mm, green at base, purplish brown at apex; stigma 1 mm long, droplets copious, ap- pearing 3—4 weeks before stamens emerge; sta- mens emerging in a scattered pattern throughout spadix, lateral stamens emerging to apex before like waxy bloom; lateral tepals alternates start emerging; anthers creamy, 0.4– 0.5 mm long, slightly divaricate; pollen yellow-orange fading to white. Infructescence erect or spreading; spathe cucullate and withered; spadix 20-25 cm long, to 4 cm diam.; berries orange to yellow (B & K yellow 8/25), oblong-ellipsoid to obovoid, acute at apex, 8.9-19 mm long, 5-8 mm diam.; pericarp thick- ened, transparent in lower half with few raphide 0.6-1 mm wide; thecae ellipsoid, cells; mesocarp mealy, orange; seeds 2 per berry, tan, oblong, 4–6.5 mm long, 2-2.7 mm diam., ca. 2 mm thick, weakly beaked at both ends. Anthurium luteynii is known only from Pana- ma, on both slopes of the Continental Divide, in tropical wet and premontane rainforest life zones at elevations of 530 to 1,350 m This species is characterized by its coriaceous, more or less narrowly oblanceolate leaves and its somewhat trapezoidal petiole that is prominently three-ribbed abaxially and sharply sulcate adaxi- ally. Also characteristic is the stout, semi-erect inflorescence with a broad, often cucullate, purplish spathe and violet- p que and the long, sharp- ly pointed, orange berr Although initially udi with Anthurium sei- bertii because of its orange, pointed berries and coriaceous, persistent, intact cataphylls, А. lu- teynii is not believed to be closely related to that species or to any other known species. Anthurium seibertii differs from 4. luteynii principally in having thinner, strap-shaped leaves with the col- lective vein always arising from near the base, and in having a narrowly tapered, much paler violet- purple spadix. NAMA. BOCAS DEL TORO: Fortuna Road, Gualaca- Chiriquí Grande, 33 km from Gualaca, 530 m, 8°36'N, 2°19'W, Hoover 1334 (MO). coci£: N of El Valle de Antón, trail to Las Minas, 800-900 m, Luteyn 3162 (DUKE, МО); El Соре Region, sawmill 7 km NE of El 0 m, Antonio 1140 (MO); Alto Calvario above *36'W, Croat 68773 (МО); El Valle N of El Valle, 800-900 m, 8?40'N, 80 Dressler 4881 (MO); El Valle de Anton, La Mesa, Folsom 2785 (MO); El Valle de Antón, 2 km W of Cerro Pilón, La Mesa, 860-900 m, Croat 37416 (MO); beyond La Mesa, toward Los Llanos, Luteyn 3163 (DUKE). DARIEN: m, Croat 27120, 27156, 27169 (MO). PANAM Jefe, ridge above Altos de Azul to Rio Chagres, 9°15'N, 79°30'W, McPherson 11900 (MO). VERAGUAS: Santa Fe Region, vic. Escuela Agricola Alto Piedra, 735-770 m, Croat & Folsom 33993 (MO), Hammel 4669 (MO), Mori & Kallunki 4751 (МО); Rio Santa Мапа, road from Santa Fe, Witherspoon & Dressler 8923 (MO); Cerro Tute, 830-1,150 m, Са: 48909 (МО); D'Arcy 14982 (MO); Rio Primero Brazo, 2.5 km beyond school 700- 750 m, Croat 25521 (B, С, "uci CM, CR, DUKE, F, K, А, SAR, SEL, UCLA, US, VEN), 48990 (MO), 49071 (FTG, МО); NW of Santa Fe, W fork of road beyond school, 1,300 m, Croat 49044 (CM, ENCB, IBE, K, MO, VDB). Anthurium machetioides Matuda, Anales Inst. Biol. Univ. Nac. Mexico 22: 379, fig. 7. 1952. TYPE: Mexico. Oaxaca: Chinantia, 1,200 m, Ramírez 144 (holotype, MEXU). Figure 190. Description based on dried material only. Stem short; cataphylls thin, to 6.5 cm long, drying brown, Volume 78, Number 3 1991 Croat 679 Anthurium sect. Pachyneurium persisting as linear fibers, probably deciduous. Leaves with petioles 6–7 cm long, 4-5 mm diam., broadly sulcate adaxially, rounded abaxially; ge- niculum - m long; blades, oblanceolate, narrowly acuminate at apex, narrowly cuneate at base, 45—48 cm long, 6-8 cm wide, broadest above the middle, both surfaces matte; midrib promi- nently and narrowly raised above and below; pri- mary lateral veins 10—15 per side, departing midrib at 20-30” angle, arcuate to the collective vein; reticulate veins prominulous; collective vein arising from one of the lowermost primary lateral veins, 3—4 mm from margin. Inflorescences equaling ог longer than leaves; peduncle 40-48 cm long, to 5 mm diam., terete; spathe narrowly lanceolate, in serted at 45° angle on peduncle; spadix greenish, glaucous, long-tapered, 20-24 cm long, 13 mm diam. near base, 6 mm diam. near apex; flowers rhombic, 2-2.4 mm long, 2.4-2.6 mm wide; 13- 14 flowers visible in principal spiral, 9-10 in al- ternate spiral; pistils not emergent; stigma ellipsoid, slitlike, 0.5 mm long; stamens emergent, held at edge of tepals; thecae ovoid. Anthurium machetioides is endemic to central Mexico on the Atlantic slope. The type was from the region of Chinantla in northeastern Oaxaca (a region comprising parts of four districts south of Valle Nacional), at 1,200 m in “зе]уа baja cadu- cifolia," and a second collection was made rela- tively nearby at 200 m in the District of Tuxtepec (between Valle Nacional and Tuxtepec). This species is characterized by its oblong-linear leaves with a collective vein arising from near the base, primary lateral veins that are scarcely more prominent than the tertiary veins, and a very elon- gate, gradually tapered spadix. n describing this species Matuda posited its анам as sect. Leptanthurium; however, de- spite the fact that it has a collective vein arising from the base, its affinities are more likely with Anthurium schlechtendalii, which often has a col- lective vein from near the base in juvenile leaves. It differs from A. schlechtendalii in its propor- tionally narrow leaves, the presence of a collective vein, and in its proportionally much longer spadix. MEXICO. OAXACA: Chinantla, 1,200 m, Ramírez 144 (MEXU); Dist. Tuxtepec, Chiltepec, 200 m, Martínez- Calderón 844 (LL, MICH). Anthurium maguirei А. Hawkes, Bull. Torrey Bot. Club 75: 633. 1948. ТУРЕ: Surinam. Table Mountain, 564 m, Maguire 24218 (ho- lotype, NY; isotypes, F, US). Figure 191A. Description based on dried material only. Ter- restrial or epiphytic; stem moderately elongate, horizontally creeping; internodes short, 0.8-1 c diam.; roots moderately few, brownish, smooth, moderately slender and elongate, ca. 2-3 mm diam.; cataphylls broadly triangular, 1.5-3.5 cm long, acute at apex, reddish brown to dark brown, per- sisting intact. Leaves with petioles 4.5-21 cm long, ca. 2- apparently subterete or C-shaped, sharply and narrowly sulcate adaxially, rounded abaxially; geniculum slightly thicker than petiole, 0.3-0.7 cm long; blades coriaceous, ob- long-ovate to broadly elliptic, obtuse at apex (the acumen apiculate), rounded (occasionally obtuse) at base, 9.5-16 cm long, 2.8-5.7 cm wide, broad- est at or below the middle; both surfaces matte, yellowish brown; midrib apparently acutely raised above, convexly raised below; primary lateral veins 2-8 per side, departing midrib at 30—45° angle, arcuate, probably flat or weakly sunken above, mm diam., raised below; interprimary veins almost as con- spicuous as primary lateral veins; tertiary veins weakly raised below; collective vein arising from near the apex, sunken above, raised below, equally as prominent as primary lateral veins, 2-6 mm from margin. /nflorescences longer than leaves; peduncle 22-26.5 cm long, ca. 2-3 mm diam., 3.1-5.5X as long as petiole, terete; spathe prob- ably spreading or reflexed, subcoriaceous, green, lanceolate, 3-4 cm long, 0.6-0.8 cm wide, broad- est near the base, acuminate at apex (the acumen inrolled, 1-2 mm long), acute at base; stipe 2-2.2 cm long in front, 1.7-1.8 cm long in back; spadix green, cylindroid, 4.3-4.5 cm long, 4 mm diam.; flowers square, 2-2.4 mm in both directions, the sides straight to smoothly sigmoid; 5-6 flowers visible in principal spiral, 3-4 in alternate spiral; tepals apparently smooth; lateral tepals 1.2-1.5 mm wide, the inner р dena straight, the outer margins 2-sided; p posed; stigma linear, 0.2— 0.3 mm long; antha 0.4 mm long, 0.5 mm wide, inclined over and obscuring the pistil; thecae ob- long-ovoid, slightly divaricate. Infructescence not seen. Anthurium maguirei appears to be a local relict of the Guiana Shield flora, known only from Table Mountain in Surinam at 564 to ca. 1,000 m. It has been recorded as growing terrestrially in a "marshy" area as well as *'terrestrial/epiphytic, climbing," although the last attribution is dubious. This species is characterized by its very small ize, moderately elongate stem and more or less ellipti leaf blades. It is one of the most distinctive species within sect. Pachyneurium and is unlikely 680 Annals of the Missouri Botanical Garden to be confused with any other. Its presumed closest ally is А. bonplandii subsp. bonplandii, which is much larger in overall size and occurs further to the west at lower elevations. Anthurium yutajense is another relict species of the Guiana Shield which superficially resembles A. maguirei. The former is placed provisionally in sect. Urospadix because its leaf blades have a collective vein originating at the base, while leaves of А. maguirei have free-ending (except near the apex) primary lateral veins. SURINAM. Table Mountain, 564 m, Maguire 24218 , NY, 05); Tafelberg, Herbarium Suriname Forest Аре 16571 (МО, 0). Anthurium manabianum Croat, sp. nov. TYPE: Ecuador. Manabi: Portoviejo-Pichincha, 12 m of San Plácido, 450 m, Harling & Andersson 24750 aeu GB 1304-9, 1304-10). Figures 191, Aim terrestris aut epiphytica; petiolus 22 cm longus, 6-8 m iam., complanatus; lamina late oblanceolata, 95.5 cm 2i 27 cm lata; pedunculus 24 cm lo vod stipes 2.5-3 cm longus; spadix um 6.5-10 c longus, 4 mm diam., purpureo-brunneus Description based on dried material only. Ter- restrial; stem, roots and cataphylls unknown. Leaves spreading; petioles 22 cm long, 6-8 mm diam., apparently flattened adaxially, with sharply raised margins; geniculum slightly darker and scarcely thicker than petiole, 1 cm long; sheath 3 cm long; blades chartaceous (probably young), broadly ob- lanceolate, shortly acuminate at apex, long-atten- uate at base, 95.5 cm long, 27 cm wide, broadest in the upper third, the margins slightly sinuate; both surfaces nearly matte, gray-green; midri weakly and convexly raised toward the base, more prominently raised toward the apex above, prom- inently convex below; primary lateral veins ca. 27 per side, departing midrib at 40—50° angle, straight to the collective vein, convexly raised above, less conspicuously so below; interprimary veins absent except near the apex, almost as conspicuous as primary lateral veins; tertiary veins obscure above, weakly visible below; reticulate veins not visible; collective vein arising from near the base, as con- spicuous as larger tertiary veins, 2-8 mm from margin. Inflorescences spreading; peduncle 24 cm long, 2 mm diam., 1.1 X as long as petiole, green, terete; spathe erect-ascending, chartaceous, green with numerous, conspicuous raphide cells, oblong- elliptic, 9-10 cm long, 1.3-1.5 ст wide, broadest at or near the middle, acuminate-caudate at apex (the acumen 10 mm long), obtuse-rounded at base; stipe 2.5-3 cm long; spadix purplish brown, cyl- indroid, moderately curved, semi-erect, held at ca. 30-40? angle from peduncle, 6.5-10 cm long, 4 mm diam. throughout or 3 mm diam. near apex; flowers square, 1.2-2 mm long, 1.5-2 mm wide, the sides straight to smoothly or jaggedly ren —5 flowers visible in principal spiral, 5-8 in al- ternate spiral; tepals very weakly papillate; aa tepals 0.9-1.1 mm wide, the inner margins round- ed, weakly erose, the outer margins 3-sided; pistils not visible; stamens + erect; filaments exserted 0 mm; anthers brownish yellow, 0.5-0.6 mm long, 0.2-0.4 mm wide; thecae oblong-ellip- soid, not divaricate. /nfructescence not seen. Anthurium manabianum is known only from the type specimen, collected in Ecuador in Manabi Province (hence the name) in a premontane wet forest life zone at 450 m. his species is characterized by its broadly ob- lanceolate, thin, greenish drying blades which are shortly acuminate at the apex and attenuate at the base, and by its small, long-stipitate, weakly ta- pered, purplish brown spadix. Anthurium manabianum appears to be super- ficially similar to 4. lennartii; see discussion of that species for details. CUADOR. MANABÍ: Portoviejo- Pichincha, 12 km E of San Plácido, 450 m, Harling & Andersson 24750 (GB). Anthurium manuanum Croat, sp. nov. TYPE: Peru. Madre de Dios: Мапа, Salvación a Glo- ria, 710 m, Vargas 17747 (holotype, CUZ). Figure 192A. Planta epiphytica; internodia ad 1.5 cm diam.; cata- phyllum persistens intactum apud nodos superiores; pe: tiolus 4.5-9 cm lo ongus, a mm diam., D e sulcatus, margine et costa medi a; lamina dapi не Kr пне e 26-39 cm longa, m lata. Inflorescentia tam longa quam petioli; pedunculus 8. cm longus; spatha linearis- Таки 4-5 cm longa, 8 mm lata; spadix uri нона 4.2 ст longus, 3-4 mm diam. Baccae igno Description based on dried material only. Ер:- phytic; stem 1.5 cm diam.; roots brownish gray, densely woolly-pubescent, elongate, at least 3 cm long, 1-1.5 mm diam.; cataphylls probably sub- coriaceous, 3-7.5 cm long, acuminate at apex, brown to brownish gray, persisting intact at least at the upper nodes, probably persisting as a retic- ulum of fibers on the lower ones. Leaves erect to spreading; petioles 4.5-9 cm long, 3-4 mm diam., + D-shaped, sulcate with the margins and medial rib prominently raised adaxially, + rounded abax- ially; geniculum slightly darker than petiole, ca. Volume 78, Number 3 1991 Croat 681 Anthurium sect. Pachyneurium 0.5-0.9 cm long; sheath 1.5-2.3 cm long; blades subcoriaceous, oblanceolate, acuminate at apex (the acumen ca. 2 cm long, apiculate), long-attenuate at base, 26-39 cm long, 7-12 cm wide, broadest above the middle, the margins probably undulate; upper surface matte, lower surface matte to semi- glossy, both surfaces tan to dark brown; midrib prominently convex at base, becoming sharply acute toward the apex above, convexly raised and slightly darker than surface below; primary lateral veins 6–7 per side, departing midrib at 50—55° angle, straight, becoming arcuate-ascending to the margin or to the collective vein in the upper third of the blade, prominently raised above, flat to weakly raised below, slightly paler than surface; interpri- mary veins few, weakly raised above and below; tertiary veins prominulous, very weakly raised above and below; collective vein arising from near the apex, very weakly raised above and below, 4-7 mm from margin. /nflorescences erect, much shorter than leaves; peduncle 8 cm long, 1-2 mm diam., about equalling petiole, dark brown; spathe reflexed, subcoriaceous, brownish, late, 4-5 cm long, 0.8 cm wide, broadest near the base, acuminate at apex, acute at base; stipe ca. (2)7 mm long in front, 1-2 mm long in back; spadix grayish brown, glaucous, cylindroid, held at 150? angle from peduncle, 4.2 cm long, 3-4 mm diam.; flowers rhombic, 1.9-2.3 mm long, 1.5- 1.8 mm wide, the sides jaggedly sigmoid, sometimes straight; 6-7 flowers visible in principal spiral, 4— |; tepals densely papillate; lateral tepals 1.2-1. 4r mm wide, the inner margins straight, occasionally broadly convex, pale, erose, the outer margins 2—3-sided, a few 4- ipse pistils with ex- posed area 0.6-1 mm diam., + square; stigma + ellipsoid; anthers yellowish, ca. 0.6 mm wide. /n- fructescence not seen. linear-lanceo- 6 in alternat Anthurium manuanum is known only from Peru in the Department of Madre de Dios, in the Manü region (hence the name), at 260 to 840 m. It probably occurs in a subtropical moist forest life zone This species is distinguished by its short peduncle (about as long as the petioles), green, oblong-lan- ceolate spathe and short, oblong, glaucous spadix, with the tepals having more or less erose margins. Anthurium manuanum is most likely to be con- fused with, and is perhaps most closely related to, A. uleanum, which occurs further to the north. Anthurium uleanum differs in having much longer petioles and peduncles and a longer, more slender spadix. Anthurium manuanum may also be con- fused with А. ernestii, but differs mainly in having a conspicuously glaucous spadix. PERU. MADRE DE DIOS: Мапи, Salvación a Gloria, 710 m, Vergas 17747 (CUZ); Prov. Мапи, Cerro de Pantia- colla, Rio Palotoa, 10-15 km NNW of Shintuya, 840 m, 12%35'S, 71*18'W, Wachter 54 (Е); Мапа National Park, Cocha Cashu uplands, 400 m, 11%45'S, 7190", Nunez 5769 (MO); Tambopata, Comunidad Nativo de Infierno, Hermosa Chica, 260 m, 12?49'S, 69*20'W, Alexiades & Pesha 972 (NY). Anthurium martianum K. Koch & Kolb in K. Koch, Wochenschr. 11: 276. 1868. TYPE: Surinam? Cultivated. Engler's Araceae Exsic- catae #165 (lectotype, B; isolectotypes, GH, P). Figures 194, 194A, 196 Habit unknown; stem short; internodes to 2.5 cm diam.; roots spreading, rather short, tapered; cataphylls coriaceous, oblong-lanceolate, to 12 cm long, acuminate at apex, drying tan, persisting semi-intact. Leaves erect-spreading; petioles 11)20–37 cm long, 5-12 mm diam., obtusely and narrowly sulcate adaxially, rounded abaxially; ge- niculum thicker than petiole, 3-4 cm long; blades (dried) subcoriaceous, oblong-elliptic to oblong-lan- ceolate, acuminate at apex (the acumen thick and inrolled), rounded to truncate (34)43-103 cm long, (6)20-38 cm wide, broadest slightly below or at the middle, the margins weakly undulate; upper surface semiglossy, dark green, paler below, both surfaces drying greenish gray to brownish; midrib ~ raised above, slightly paler than surface, convexly raised below; primary lateral veins 18—24 per side, departing midrib at 55-85" angle, weakly curved to the collective vein, raised above and below; interprimary veins less conspicuously raised than the primary lateral veins; tertiary veins weakly raised above and below; collective vein arising from near the base, weakly raised above and below, 5- 35 mm from margin. /nflorescences (dried) with peduncle 25-60 cm long, 4-7 mm diam., са. 0.7- 1 X as long as petiole, terete; spathe spreading to reflexed, subcoriaceous to coriaceous, green, lan- ceolate, 9.5-17 cm long, 1.5-4 cm wide, broadest near the base, inserted at 60? angle on peduncle, acuminate at apex (the acumen tightly inrolled, ca. 6 mm long), rounded at base; spadix maroon, sub- sessile, tapered, weakly curved, held at 150- 165(185)* angle from peduncle, 9-15 cm long, 7-11 mm diam. near base, 3-4 mm diam. near apex; flowers rhombic, 1.9-2.6 mm long, 1.6-2.1 mm wide, the sides mostly straight to smoothly or jaggedly sigmoid; 11-12 flowers visible in principal spiral, 12-18 in alternate spiral; tepals minutely 682 Annals of the Missouri Botanical Garden and densely papillate when dried; lateral tepals 1.1-1.5 mm wide, the inner margins convex, weak- ly erose, the outer margins 2-3-sided; pistils ca. 0.8 mm long, 0.5 mm wide; stigma oblong, 0.4— 0.6 mm long; thecae oblong, slightly or not divar- icate; pollen fading to cream-white. /nfructescence not seen. Anthurium martianum is known from a single live collection, purportedly from Surinam, that was propagated at the botanical gardens in Berlin, Kiel, and Munich. The herbarium material is highly uni- form and appears to be the result of a single in- troduction. Because it is known only from culti- vation and no field-collected herbarium vouchers exist, there is the possibility that it is a cultivated plant of hybrid origin. However, because the taxon is clearly distinct, it will be retained here. This species is distinguished by its persistent spathe, short, yet very tapered spadix, moderately long petioles and its leaf blades, rounded at the base, with several pairs of congested lateral veins near the base, and with a collective vein arising from near the base. Anthurium martianum is closest in appearance to А. dombeyanum, which has generally shorter petioles and more undulate leaf blades that are broadest above the middle. A living collection at the Munich Botanical Gar- den (without number) labeled 4. martianum and vouchered as Croat 61178 perhaps represents this species; however, although it closely matches the plate in Gartenflora 20: t. 681 (1871), it is smaller and the primary lateral veins are fewer and not nearly as distinct as the type material. SuRINAM?. Cultivated. Engler's Araceae Exsiccatae #165 (B, СН, P). Cul at Munich, Croat 61178 ). Anthurium napaeum Engl., Bot. Jahrb. Syst. 25: 407. 1898. TYPE: Ecuador. Pichincha: Río Napac (“Мара”), Sodiro s.n. (holotype, B). Figures 197, 199, 200. Anthurium peripense Engl., PE: Ecuador. s.n. (holotype, B) Anthurium rircayanum Sodiro, Anal. Univ. Centr. (Qui- to) 22(156): 21. 1906. TYPE: Ecuador. Guayas: Río Rircay, Rimbach s.n. (holotype, B). Bot. Jahrb. Syst. 25: 442. Manabi: Río Репра, Sodiro Terrestrial, or caespitose on rocks, sometimes long, 1.5-2.5 cm diam.; roots numerous, dense, spreading to descending, white to green, whitish or brown on drying, + smooth, thick and short to slender and elongate, weakly tapered at apex, 2-4 mm diam.; cataphylls epiphytic; stems 9-30 cm subcoriaceous, 8-9(16) cm long, acuminate and apiculate at apex, pale green, drying tan, persisting intact, eventually as a reticulum of fibers. Leaves erect to erect-spreading; petioles (8)16—42 cm long, 5-15 mm diam., broadly triangular, convex adax- ially with weak medial ridge, the margins promi- nently and sharply raised, almost winged, bluntly to sharply angular abaxially, the surface conspic- uously pale short-lineate; geniculum slightly thicker and paler than petiole, 1-2 cm long; blades mod- erately coriaceous, broadly lanceolate to oblong- elliptic to oblanceolate, long-acuminate at apex (the acumen short-apiculate), acute to obtuse at base, (31)40-70(82) cm long, (5)9-28 cm wide, broad- est at or above the middle, the margins flat to broadly undulate; upper surface glossy, dark green, lower surface glossy to semiglossy, conspicuously paler; both surfaces matte when dried, green to brown; midrib prominently convex at base, becom- ing higher than broad toward the apex above, pale short-lineate and paler than surface, prominently and sharply acute-raised below, raised above and below when dried, paler than surface below, same as surface above; primary lateral veins numerous, ca. 20-30 per side, departing midrib at 50-75? angle, straight-ascending, weakly visible above and below when fresh, slightly raised on both surfaces when dried; interprimary veins almost as conspic- uous as primary lateral veins; tertiary veins to visible when dried; collective vein arising from nea the base or in the upper third of the blade, slightly more prominent than primary lateral veins when fresh below, equally as prominent as primary lateral veins when dried, 2-13 mm from margin. /nflo- rescences erect to spreading, shorter than leaves; peduncle 23-50(62) cm long, 3-7 mm diam., sharply 2-3-ridged, 2-3 x as long as petiole, green heavily tinged with red-violet at base, faintly red- violet toward the apex; spathe reflexed-spreading, subcoriaceous, green tinged with red-violet (В & K yellow-green 6/7.5), linear-lanceolate to lan- ceolate, 6-10 cm long, 0.4—1.5 cm wide, broadest near the base or in lower third, inserted at 60– 70° angle on peduncle, long-acuminate at apex (the acumen inrolled), acute at base, the margins meet- ing at 90° angle; spadix brownish purple (B & K red-purple 2/10) to dark purple, rarely dark olive- green, cylindroid to weakly tapered, sessile or stip- itate 1-2 mm, erect, 5.5-16 cm long, 5-6 mm diam. midway; flowers + square to 4-lobed, 1.9— 2.5 mm long, 2-2.4 mm wide, the sides jaggedly sigmoid; 4—6 flowers visible in principal spiral, 8- 10 in alternate spiral; tepals matte to semiglossy, densely papillate; lateral tepals 0.7-1.2 mm wide, the inner margins straight, the outer margins Volume 78, Number 3 1991 Croat 683 Anthurium sect. Pachyneurium 2-sided, sometimes weakly 4-sided; pistils raised but not emergent, or held slightly above the tepals, green; stigma ellipsoid, 0.4 mm long, brushlike; stamens emerging shortly above the tepals, laterals emerging almost to the apex before alternates emerge; filaments fleshy, translucent or purplish, .5-0.8 mm long, 0.4-0.8 mm wide; anthers yel- low to purplish, 0.3-0.6 mm long, 0.4-0.7 mm wide, inclined over the pistil; thecae oblong, drying ovoid, scarcely divaricate; pollen white to yellow- orange fading to tan. /nfructescence with spathe persisting; spadix 8-22 cm long, cm diam.; berries reddish to dark purplish violet, ovoid, ex- serted, acute at apex, drying 4.2-6 mm .9 mm diam ong, 2- А member of series Multinervia, Anthurium napaeum occurs in Ecuador on the western slopes of the Andes, at 200 to 1,400 (1 montane wet and premontane moist forest, and in arino, Colombia. One specimen from Cotopaxi, between Latacunga and Quevedo at 800 to 950 m, was reported to occur in montane moist forest. his species is characterized by its generally terrestrial habit, short, densely rooted stem, cata- phylls that persist as a reticulum of fibers, broadly triangular, conspicuously pale-speckled petioles, leaf blades which are usually glossy on both surfaces ,800) m in pre- when fresh, drying green and often conspicuously paler below, and by its cylindroid, brownish purple spadix with 4-6 flowers visible in the principal spiral. Anthurium napaeum is most closely allied to A. palenquense, which differs in having leaf blades with the margins conspicuously concave toward the ase. The names Anthurium peripense and А. rir- cayanum are synonymized here for the first time. The type localities of these species, together with that of A. napaeum are in the coastal lowlands of western Ecuador, draining into the Gulf of Gua- yaquil. The differences between the type specimens are slight, the main one being leaf shape: the types of А. napaeum and of А. rircayanum closely resemble each other in this respect, while that of A. peripense has straighter blade ao which are somewhat concave toward the An unusual collection (Croat 55782) exhibits stamens starting emergence at the apex of the spadix and progressing regularly to the base, with the laterals preceding the alternates by up to 18 spirals. Basipetal maturation of the stamens has not been observed in any other Anthurium. COLOMBIA. NARIÑO: La Planada, Salazar Finca 7 km above Ricaurte, 1,750 m, 1?08'N, 77*58'W, Gentry et al. 35197 (COL, MO). EcUADOR. CAÑAR: Azoques-El Triunfo, 1 km 5 of La Delicia, 2°27'S, 79°10'W, Croat 50876 (AAU, B, CAS, K, CM, M, MO, RSA, US); Azo- gues-El Triunfo, 5 km W of La Delicia, 8 km SE of El Truncal, 600 m, 2?27'S, 79°15'W, Croat 50904 (CM, MO). caRCHr: Peñas Blancas, 900 m, Christenson 1535 MO). CHIMBORAZO: Sibambe, 1,000-1,600 m, Solís 5342 (F); Rio Chanchán, 5 km N of Huigra, 1,667-2,167 m, Camp 3428 (NY). COTOPAXI: Quevedo- Latacunga, 800- 950 m, 0%55'S, 79107, Holm-Nielsen et al. 3021 (AAU); Rio Pilaló, Quevado- Latacunga, Tenefuerste, 750- 1,300 m, Dodson & Gentry 11987, 12297, 12797 (MO, SEL); 3 km E of El Palmar, 800 m, Dodson & Gentry 10267 (MO, SEL); 55.5 km from Quevedo, 23.5 km E of La Maná, 930-950 m, 0°53’S, 79°04'W, We 57035 (CAS, BG, K, MBM, MEXU, MO, P, PMA basa El Corazón, 1.9 km NW of El Corazón, 67. 5 km Quevedo, 1,225 m, 1%07'S, 79*06'W, ies 55829 (MO, QCA); Rio Angamarca, Quevedo- El Corazón, Las Juntas, 200 m, Harling & еј а 19025 (MO, GB); Que- vedo- El от T 4 km of Quevedo, 6 Km NW near Uie: above river, 1,030 ~ GUAYAS: Naranjal, Entable, below 500 m, mann s.n. ~ B). PICHINCHA: Quito- Santo Doming Chiriboga Road, Estación Los Fais m s, 12 km from Rio Pilatón, 1,400 m, Harling & roce 23128 (GB); е Pilatón, Alluruquin, 22 т anto Domingo, 950 ‚ 0915'5, 78*58'W, Hammel & Wilder 16073 (B, K, МО); Km 23 on Santa Domingo-Puerto Limón road, 100 m, 0?21'S, 79°22'W, Kvist 40676 (AAU); San Juan, La Palma, о m, Croat 38739 (F, MO); Alluruquin-Chiri- boga, 1,100 m, Madison 4069 (K, SEL); Río Blanco, below eris ce with Río Toachi, 300 m, Harling 4516 (5); Rio Napa, Sodiro s.n. (B) Anthurium narinoense Croat, sp. nov. TYPE: 78?10'W, Gentry et al. 55343 (holotype, MO 3486084; isotype, PSO). Figure 201. Planta epiphytica; petiolus 12 cm longus, 7 mm dia D-formatus; lamina oblongo-oblanceolata, basi и. ad 93 ст longa, 16 cm lata; pedunculus ad 40 cm longus; spatha oblongo-lanceolata, ad 11 cm longa, 1.5 cm lata; spadix atrorubens, 17 cm longus, 5-6 mm diam.; baccae otae. ja’ Description based on dried material only. Epi- phytic; stem and cataphylls unknown. Leaves erect- spreading; petioles ca. 12 cm long, 7 mm diam., D-shaped, slightly convex adaxially, apparently rounded abaxially; geniculum 1.5 cm long; blades subcoriaceous, oblong-oblanceolate, acuminate at apex, long-attenuate at base, 93 cm long, 16 cm wide, broadest above middle, yellowish, the margins flat; midrib convexly raised above, prominently convexly raised below; primary lateral veins ca. 30 Annals of the Missouri Botanical Garden per side, departing midrib at 60—70° angle, slightly arcuate to the collective vein, slightly raised above and below; interprimary veins numerous, almost as conspicuous as primary lateral veins; tertiary veins scarcely raised above and below, more prominent below; collective vein arising from near the base, nearly as conspicuous as primary lateral veins, 5— 9 mm from margin. Inflorescences erect-spread- ing; peduncle 40 cm long, ca. iam., ca. 4X as long as petioles, terete; spathe reflexed, subcoriaceous, green, oblong-lanceo late, atleast 11 mm cm long, 1.5 cm wide, broadest near base, inserted at 45? angle on peduncle, apex not seen, the base acute; spadix dark reddish, slightly tapered, + straight, 17 cm long, 4 mm diam. near base, 3 mm diam. near apex; flowers rhombic, 2.4-2.8 mm long, 1.6-2 mm wide, the sides smoothly sig- moid; 7-8 flowers visible in principal spiral, 5-6 in alternate spiral; tepals + smooth; lateral tepals ca. 1 mm wide, the inner margins straight, the outer margins 2—4-sided; anthers 0.2 mm long, 0.3 mm wide; thecae slightly divaricate. Јпјтис- tescence not seen. A member of series Multinervia, А. narinoense is known only from the type collection made in Narino (hence the name), Colombia, at 900 to 1,000 m in pluvial forest. Anthurium narinoense is distinguished by its large leaves (for the series), which dry yellow-green, and its slender, tapered spadix. It is apparently not closely related to any other species, but somewhat resembles А. fasciale and А. santiagoense from the eastern slopes of the Andes in Ecuador, which both have a more stubby spadix with more flowers in the principal spiral. Anthurium narinoense is unusual in being the only species of Pachyneurium known from pluvial forest. COLOMBIA. NARIRO: Junin-Barbacoas, 2-10 km N of Junin, 900-1,000 m, 1°30’N, 78?10'W, Gentry et al. 55343 (MO, PSO). Anthurium nervatum Croat, Monogr. Syst. Bot. Missouri Bot. Gard. 14: 141. 1986. TYPE: Panama. Veraguas: valley of Rio f'ercero Bra- zo, beyond Escuela о Alto Piedra, above Santa Fe, 500-700 m, Croat 27329 (holo- type, MO 2253314; isotypes, F, K, PMA, US). Figures 198, 202. Epiphytic, occasionally terrestrial; stem 10-75 cm long, 2.5-6 cm diam.; leaf scars 1-1.5 cm high, 2-3.5 cm wide; roots few, descending, green- ish, pubescent, moderately thick and short, slightly tapered, 3-7 mm diam.; cataphylls subcoriaceous, lanceolate, 16-35 cm long, rounded to emarginate at apex with subapical apiculum ca. 5 mm long, green, drying tan (B & K yellow-red 9/10), per- sisting semi-intact, eventually as a reticulum of fibers. Leaves erect to spreading; petioles 18-10 cm long, 7-14 mm diam., bluntly to sharply qua- drangular to D-shaped or subterete, prominently to weakly sulcate or flattened adaxially, the margins flat or raised, rounded abaxially, the surface pale- speckled; geniculum thicker and paler than petiole, becoming fissured transversely with age, 1.5-5 cm long; blades moderately to thickly coriaceous, tri- angular to ovate-triangular, ovate-elliptic, acute to acuminate at apex (the acumen weakly apiculate), subcordate to cordate at base, sometimes decurrent on petiole, 27-98 cm long, 14.5-46 cm wide, broadest in the lower fourth of the blade, the mar- gins broadly undulate; anterior lobe 25.5-90 cm long, the posterior lobes 4-19 cm long, directed downward and sometimes inward or outward; sinus arcuate with iui decurrent on dann d i spath- ulate; uppe k green, lower pice semiglossy, slightly Ка midrib bluntly angular-raised, becoming acutely to ob- tusely raised, and then weakly sunken at apex above, prominently and convexly raised below, pal- tat igl er than surface above and below; basal veins 3-6 pairs, often free to base, 2nd and 5th sometimes coalesced 1.5-2 cm, convexly raised above and below (less so below); primary lateral veins 19-25 per side, departing midrib at 50—60? angle, straight or weakly arcuate to the collective vein, promi- nently to weakly convexly raised above (more so below), paler than surface; interprimary veins sometimes present; tertiary veins obscure to weakly visible above, flat and darker than surface below; collective vein arising from one of the lowermost primary lateral veins ог lst to 2nd basal vein, weakly sunken above, weakly raised and darker than surface below, 1-5 mm from margin. Inflo- rescences erect to spreading; peduncle 10-35 cm long, 8-10 mm diam., 0.2-0.7 x as long as petiole, green, terete; spathe reflexed, sometimes recurled, coriaceous to subcoriaceous, green to green heavily tinged with purple-violet to purplish or maroon, lanceolate to lanceolate-elliptic, 10-33 cm long, 2.5-5.5 cm wide, broadest just above the base, inserted at 30? angle on peduncle, long-acuminate at apex (the acumen inrolled), rounded to cordate at base; spadix green to purple to maroon, slightly tapered, (6)18-28 cm long, 1-2 cm base, 8-10 mm diam. near apex; flowers rhombic to weakly 4-lobed, 2.2-3.2 mm long, 2.0-3.1 mm wide, the sides weakly to jaggedly sigmoid; 10-15 diam. near Моште 78, Митбег 3 1991 Croat 685 Anthurium sect. Pachyneurium flowers visible in principal spiral, 8-20 in alternate spiral; tepals matte, conspicuously punctate, mi- nutely papillate; lateral tepals 0.8-1.6 mm wide, the inner margins broadly convex to concave when dried, the outer margins straight; pistils not emer- gent except when dried, green tinged with violet- purple; stigma linear, slitlike, 0.5-1 mm long; stamens emerging from the base, lateral stamens followed by alternates in rapid succession, the lat- erals preceding the alternates by 5—8 spirals, the 3rd stamen preceding the 4th by 6—8 spirals, held over and obscuring pistil; filaments red-violet, 1.4— 1.6 mm long; anthers yellow, 0.4-0.7 mm long, 0.5-0.9 mm wide; thecae narrowly oblong to ovoid- oblong, weakly or not at all divaricate; pollen or- ange fading to tan or white. /nfructescence spread- ing to pendent; spadix to 50 cm long, to 5 cm diam.; berries orange, oblong-ellipsoid to oblong- ovoid, acute at apex, the tip relatively blunt, to 14. mm long, 3.5 mm diam.; mesocarp with linear to punctiform raphide cells; seeds 2 per berry, oblong- ellipsoid to ellipsoid, somewhat flattened, 4.5- mm long, 2.7-3.2 mm diam., 1-1.5 mm thick, with a thin appendage at apex only. Anthurium nervatum is endemic to Panama in Bocas de Toro, Cocle, Chiriqui, and Veraguas prov- inces at (130)500 to 2,100 m in tropical wet, premontane rain, and lower montane wet forest life zones. This species is керше by из long, triangular, shallowly lobed to oderately thick leaf blades with numerous primary lateral veins, its usually long-tapered, green to maroon spadix, and by its lanceolate, coriaceous spathe that is deep maroon on the inner surface, green heavily tinged with maroon on the outer surface and often forms a loose spiral at anthesis. Anthurium nervatum is most easily confused with А. ranchoanum, but that species has a shorter spadix with an erect, green spathe and larger flow- ers which are fewer per spiral. It may also be confused with А. colonicum, but that species has early exserted, acute pistils and fewer primary lat- eral veins. One collection from El Valle in Coclé Province (Croat 37443) is probably also this species, but has a smaller, more ovate-triangular, more coria- ceous blade with the posterior lobes sharply turned upward, and a deeper sinus; it also has a green rather than maroon spadix. PANAMA. BOCAS DEL TORO & CHIRIQUÍ: Cerro Colorado, Folsom et al. 4706 (МО); trail into Bocas and in woods on Pacific slope, from Chami station to ca. “9 mi. ions road, 1,100-1,750 m, 8?35'N, 81%54'W, Hammel & Trainer 14919 (MO, PMA). CHIRIQUi: above Santa Fe along mining road 18-27 mi., off Pan-Am Hwy. above Chami or turnoff to Escopeta, 1,200-1,500 m, Croat 33145 (DUKE, F, MO, РМА); ps mining road 31.6 km beyond bridge over Rio San Féli .6 km beyond turnoff to Escopeta, 1,690 m, Croat 371 77 (MO), 37169 (MO, PMA); upper mining road 20-28 mi. from San Félix, Continental Divide, 1,200-1,500 m, Dad: poda (MO), 33394 (MO, RSA, US); along mining road 20 above bridge over Río San Felix, near village of San Félix, 750 m, Croat 48459 (MO); Cerro Pate Macho, 1,500- 1,900 m, Stein et al. 1257 (MO); SE slopes, trail from Río Palo Alto, 4 km NE of Boquete, 1,700-2,100 m, Sytsma et al. 4830 (MO). состЕ: Cerro Caracoral, Duke & Dwyer 15118 (NY), Kirkbride 1096 (MO, NY); Cerro Moreno, Molejón-Coclecito, 13 km NW of Cascajal, 130- 250 m, 8?46'44"N, 8*31'54"W, Davidse & Hamilton 23713 (B, MO); El Copé Region, Alto Calvario, 800- 900 m, Folsom & Jaslon 2691 (МО); El Cope, W of sawmill, Hammel 2408 (MO); Continental Divide, lumber road N of El Copé, 10 km above El Copé, 2.2 km N of sawmill, 750-930 m, Croat 44717, 49169 (MO, PMA); Е Valle Region, El Valle de Antón, La Mesa, Croat 37443 (CAS, CM, F, LE, KYO, M, MO, PMA, US). vERAGUAS: Santa Fe Region, vic. Escuela Agricola Alto Piedra, 500- 5 m, Croat 33987 (CAS, MO, US), 34244 (MO, PMA), Croat & Folsom 34134 (MO); 1,150-1,450 m, McPherson 12070 (МО); Cerro Tute, 750-1,450 m, Croat 48912, 48925 (MO, PMA), Knapp & Dressler 5415 (MO, US), 5446 (MO), Knapp & Kress 4369 (MO), Knapp & Sytsma 2475 (MO), Sytsma et al. 4590 (МО); Rio Tercero Brazo, beyond school 500-700 m, Croat 27329 (F, К, MO, PMA, US). Anthurium nizandense Matuda, Bol. Soc. Bot. Mexico 24: 35, fig. 1. 1959. TYPE: Mexico. Oaxaca: Nizanda near the Isthmus of Te- huantepec, 16?40'N, 95°02'W, MacDougall n. (holotype, MEXU). Figures 207, 208. Anthurium kruseanum Matuda, Anales Inst. Biol. Univ. Nac. Mexico 36: 107, fig. 1. 1966. TYPE: Mexico. Guerrero: La Junta (at junction of Rio Omitlan and Rio Papagayo) S of Tierra Colorada, Kruse 881 (holotype, MEXU). Usually terrestrial or epilithic; leaf scars 0.7- 1.2 cm high, 1.3 cm merous, descending, greenish to tan, smooth to wide; roots moderately nu- somewhat pubescent, moderately elongate, ta- cataphylls subcoriaceous, 3-7 cm long, rounded and minutely apiculate at apex, drying pered; dark brown, persisting semi-intact, eventually as a reticulum of fibers. Leaves erect-spreading; peti- oles 6-24 cm long, (3)5-6 mm diam., quadran- gular to D-shaped to subterete, flattened to broadly and sharply sulcate adaxially, the margins sharply raised, rounded abaxially, the surface pale-speck- led, sometimes tinged with red-violet; geniculum slightly paler and thicker than petiole, 1-1.5 cm long; blades subcoriaceous, obovate-elliptic to el- liptic, acute to abruptly acuminate at apex (the 686 Annals of the Missouri Botanical Garden acumen apiculate), rounded to acute or attenuate at base, 25-57 cm long, 10.7-30 cm wide, broad- est at or above the middle, the margins weakly undulate; upper surface matte to semiglossy, me- dium green, lower surface matte, paler below; both surfaces with short, linear raphide cells and con- vexly raised epidermal cells, drying brown to green- ish brown; midrib acutely to convexly raised above, more prominently so below; primary lateral veins 6-11 per side, departing midrib at 50—80° angle, straight to weakly arcuate-ascending to the collec- tive vein, raised above and below, darker than surface below; tertiary veins obscure above, weakly visible below; collective vein arising from about the middle of the blade or near the apex, 3-9 mm from margin. /nflorescences erect, equalling or longer than leaves; peduncle 31-60 cm long, 5- mm diam., 1.8-2.7 х as long as petiole, terete; spathe reflexed, subcoriaceous, green, ovate-lan- ceolate, 8-9.5 cm long, 2.5-3 cm wide, broadest near the base, inserted at 45? angle on peduncle, acute at apex, apparently obtuse at base; spadix pale green becoming reddish brown, somewhat ta- pered, 12-14 cm long, 9-11 mm diam. near base, 4 mm diam. near apex; flowers rhombic, 1.8-2.5 mm long, 2.8-3 mm wide, the sides sigmoid; 7- 9 flowers visible in principal spiral, 5—6 in alternate spiral; lateral tepals 1.5 mm wide, the inner mar- gins broadly convex or weakly concave; pistils not emergent; stigma broadly ellipsoid; stamens emerg- ing rapidly in a complete sequence, emerging short- ly above the tepals, arranged in a circle around the pistil; anthers ca. 0.4 mm long, 0.6 mm wide; thecae ellipsoid, weakly divaricate. Infructescence pendent; spathe withered; spadix 6.5 cm long, 1.5 cm diam.; berries greenish white, ovoid, 8 mm long; mesocarp transparent, juicy, pulpy; seeds 1(2) per berry, pale yellow (B & K yellow 9/2.5), broadly obovoid, 5.2-5.4 mm long, 4.5-4.7 mm diam., 3.4-4 mm thick, with a minute, translucent appendage, + gnarled at apex on side, tinged with brown, enveloped by gelatinous substance. Anthurium nizandense is endemic to Mexico in southern Guerrero and southern Oaxaca, where it occurs at less than 850 m in seasonally very dry habitats on steep, rocky cliffs in “selva baja cadu- cifolia." It is known only from the vicinity of Tierra Colorada in Guerrero and Nizanda in Oaxaca. This species is recognized by its greenish white berries and thick, narrowly to broadly elliptic or oblanceolate-elliptic, relatively long-petiolate leaves that are matte on the lower surface and sometimes exhibit a dense array of minute, linear raphide cells on drying. The epidermal cells on both surfaces of the blade are convexly raised. Matuda reported the petiole to be subquadrangular, but this was possibly due to a misinterpretation of the dried specimen since this feature is not apparent on live material. In the area of Mexico where it occurs, А. ni- zandense could be confused only with A. schlecht- endalii subsp. jimenezii, which occurs in similar habitats and is more generally abundant. Anthur- ium nizandense is distinguished by its proportion- ally longer petioles and its more typically elliptic lade, which is distinctly matte on the lower sur- nizandense the blades are 1.3-2.5 times longer than the petioles. In А. schlechten- dalii subsp. Jimenezii, the leaf blades are usually semiglossy beneath and usually 6-10 times longer (rarely as little as 3 times longer) than the petiole. face. In MEXICO. GUERRERO: Río Omitlán, bridge over Rio Omit- lan at confluence of Rio Papagayo with Rio Omitlan, 850 m, Croat 45756 (B, BM, CM, K, M, MBM, MO, P, TEX, US); La Junta, 5 of Tierra Colorada, Kruse 881 (MEXU); Tierra Colorada, Kruse 1595 (MEXU). OAXACA: Nizanda, Isthmus of Tehuantepec, MacDougall s.n. (MEXU Anthurium obscurinervium Croat, sp. nov. TYPE: Ecuador. Cañar: along road between Azogues and El Triunfo, ca. 2 km S of Hwy. at Cedro Pampa, 800-1,000 m, 2?25'S, 79°10'W, Croat 50914 (holotype, MO 2828745; isotypes, AAU, BM, G, K, LL, M, MICH QCA, RSA, SEL, US). Figures 203, 204 ша ا‎ a aut epiphytica; caulis ad 30 cm longis, 1.3-3. am.; cataphyllum a in fibris ten- ге гы = 5)6-22 cm longus, 4-10 mm diam., e D-formatus; lamina anguste К шш elliptica aut e шышы. (18)30- 79 cm longa, (3.2)4- 7(11) cm lata; nervis мунан lateralibus obscuribus; pedunculus 20-45 longus; spatha (2.5)3-7 cm longa, 0.7-1.7 cm diam.; spadix viridis, (1.7)4–11(15) em longus, 3-6 mm diam.; baccae atropurpurae. Epiphytic or terrestrial; stem to more than 30 cm long, l- iam.; roots dense, short, ca. 3 mm um cataphylls subcoriaceous, 4-9 cm long, acute to acuminate at apex, drying whitish (B € K yellow 9/2.5 to yellow-red 9/7.5), per- sisting as linear fibers. Leaves spreading; petioles (3.5)6-22 cm long, 4-10 mm diam., sharply D-shaped, slightly thicker than broad, sharply flat- tened to broadly and shallowly sulcate adaxially, the margins sharply raised to curved inward, round- ed to obtusely or acutely angular, sometimes l-ribbed abaxially; geniculum somewhat thicker and much paler to darker than petiole, 0.7-1.7 cm long; blades subcoriaceous, narrowly oblong- Volume 78, Number 3 1991 Croat 687 Anthurium sect. Pachyneurium elliptic to oblong-linear, long-acuminate to short- acuminate (rarely acute) at apex, attenuate to nar- гому acute at base, (18)30- 79 cm long, (3.2)4— 7(11) cm wide, broadest at or near the middle, the margins flat, weakly revolute; upper surface semi- glossy, medium to dark green, lower surface glossy, much paler; midrib convexly raised at base, be- coming acutely raised toward the apex above, prominently and acutely raised at base, becoming less pronounced toward the apex below; primary lateral veins 25-55 per side, departing midrib at 40—50° angle, straight to the collective vein, ob- scure and flat above and below; interprimary veins numerous, almost as conspicuous as primary lateral veins; tertiary veins drying visible; collective vein arising from the base, slightly more prominent than primary lateral veins, 4-6 mm from margin. /n- florescences erect-spreading; peduncle 20-45 cm long, 2-4 mm diam., 1.6-4(5.5)X as long as pet- iole, pale green, sometimes tinged purplish, sub- terete to 3—4-ribbed; spathe spreading, rarely re- flexed, thin, green to pale green, sometimes tinged with purple at margins, lanceolate-elliptic to linear- lanceolate, (2.5)3-7 cm long, broadest near the base, short-acuminate at apex (the acumen somewhat inrolled), acute at base; stipe 2-13 mm long in front and in back; spadix medium green (B & K yellow-green 7/5) to pale green, cylindroid, scarcely tapered, curved, (1.7)4- 11(15) ст long, 3-6 mm diam. near base, 2-3 mm diam. near apex; flowers 4-lobed; 2-3 mm in both directions, the sides jaggedly sigmoid, 3-5 flowers visible in principal spiral, 5-7 in alternate spiral; tepals matte, weakly and minutely papillate, droplets present at anthesis; lateral tepals 1—1.8 mm wide, the inner margins straight to broadly 7 cm wide, rounded, the outer margins irregularly 3—5-sided; pistils weakly raised, green; stigma depressed, 0.2— 0.4 mm long; stamens emerging in prompt se- quence from the base, the laterals preceding the alternates by 3-11 spirals, held in a tight cluster above the pistil; filaments translucent, 0.4 mm long, 0.5-0.7 mm wide; anthers creamy white, 0.6-0.7 mm long, 0.7-0.9 mm wide; thecae el- lipsoid to ovoid, scarcely divaricate; pollen pale yellow fading to white. /nfructescence erect to pen- dent; spathe persisting, green, or absent; spadix 7-16 cm long; berries green becoming purplish to purple-black, obovoid-ellipsoid, quadrangular at apex, 6 mm long, 5 mm diam. A member of series Multinervia, Anthurium obscurinervium is endemic to the Pacific slope of Ecuador, at 50 to 1,330 (2,000) m, in premontane moist, premontane dry and premontane wet forest. This species is characterized by its more or less oblong leaf blades with the primary lateral veins obscure when fresh and obscure to weakly raised on drying (hence the name), green spadix with exserted stamens, and orange berries. Anthurium obscurinervium is most similar to A. carchiense, which differs in having a brownish spadix and purple to purple-black fruits. ECUADOR. BOLÍVAR: Charquiyacu, 600 m, Solís 6114 (Е). CAÑAR: Azogues-El Triunfo, 2 km $ of Hwy. at Cedro Pampa, 800-1,000 m, 2°25'5, 79*10'W, Croat 50914 (AAU, BM, G, K, LL, M, MICH, MO, NY, QCA, RSA, SEL, US). CARCHÍ: beris Chico, near Awá encampment, 1,300 m, 0°58'N, 78°16'W, Hoover et al. 2658 (MO). COTOPAXI: Rio Pilaló, Quevedo-Latacunga, 950-1,100 m, 0%53'S, 79°10'W, Holm-Nielsen et al. 3076 (AAU). EL ORO: 11 km W of Pinas, new rd. to Santa Rosa, 850 m, Dodson et al. 9033, 9037 (MO, SEL); Machala- Loja, 25 km SE of jet. to Piñas, 890 m, 4?15'S, 79°45'W, Croat 50715 (CM, MO, NY, SEL, US, TEX). GUAYAS: Naranjal-Machala, 13 km S of Naranjal, 50-150 m, Harling & Andersson 19301 (GB, MO). Los ríos: Alta be 29.5 km W of Patricia Pilar, 450-475 m 0°33’S, 79°22'W, Croat 50664 (AAU, MO, NY, RSA); Patricia Pilar-24 de Mayo, Centinela, Dodson et al. 8703 (MO, SEL); Rio Blanco, Santo Domingo de Los Colorados- Esmeraldas, Villa Hermosa, 3 km S of Km 24, 250 m, 0*5'S, 79°15'W, Croat 50694A (MO). PICHINCHA: 27 km 5 of San Juan, 12 km NE of Chiriboga, 2,000 m, 0°17'S, 78°42'W, Croat 50606A (IBE, К, MO, NY, RSA, US, VDB). Anthurium oerstedianum Schott, Oesterr. Bot. Z. 8: 180. 1858. ТУРЕ: Costa Rica. Naranjo, Oersted s.n. (lectotype, illustrated by Schott, Aroideae 328). Figures 205, 206 Anthurium TM Schott, Oesterr. Bot. Z. 8: 180. Я : Costa Rica, Oersted s.n. M Ший ү Schott, Aroideae 314). Terrestrial; stem to 30 cm long, 2-4 cm diam.; leaf scars, 0.8-1 cm high, 2-2.5 cm wide; roots moderately numerous, descending, tan, smooth, short, moderately thick, slightly tapered, 3-5 mm diam.; cataphylls subcoriaceous, 7-13 cm long, acute at apex and with subapical apiculum, medium green, drying brown, persisting intact, eventually deciduous. Leaves erect to spreading; petioles 2)30–73 cm long, 5-7 mm diam., quadrangular to subterete, flattened to weakly and broadly sul- cate adaxially, sharply to bluntly 1—3 ribbed abax- ially, the surface minutely pale-speckled; genicu- lum remote from the base of the blade 10-20 cm, 1.5-2 ст long; blades subcoriaceous, ovate to nar- rowly ovate or lanceolate, long-acuminate at apex, abruptly attenuate then acute to obtuse to truncate or subcordate at base, 20-66 cm long, 8.5-30 cm wide, broadest at or near the middle, the mar- ~ = 688 Annals of the Missouri Botanical Garden gins undulate; upper surface matte to weakly glossy, medium green, lower surface semiglossy to glossy, paler; midrib bluntly angular-raised above, below prominently higher than broad at base, becoming angular-raised toward the apex and paler than sur- face; primary lateral veins 11-20 per side, de- parting midrib at 40—60° angle, broadly arcuate, sunken to weakly raised in grooves above, prom- inently convexly raised below; interprimary veins almost as conspicuous as primary lateral veins; tertiary veins sunken above, raised below; collec- tive vein arising from near the base, sunken above, raised below, equally as prominent as primary lat- eral veins, 2-8 mm from margin. /nflorescences erect, longer than leaves; peduncle (30)70-130 cm long, equal to 1.2-1.5X as long as petiole; spathe pale green, ovate-lanceolate, 8-13 cm long, 2.5-4.8 cm wide, inserted at 45? angle on pedun- cle, acuminate at apex, rounded at base; spadix pale green, moderately tapered, 5-12 cm long, ca. 9 mm diam. near base, ca. 6 mm diam. near apex; flowers sub-rhombic to 4-lobed, 1.7—2.5 mm long, 2-2.7 mm wide, the sides straight to sigmoid; 8- 9 flowers visible in principal spiral, 12-18 in al- ternate spiral; tepals matte, minutely and densely papillate; lateral tepals 1.4—1.8 mm wide, the inner margins pede pistils weakly emergent, green; stigma m long, weakly exserted and brush- like, пива appearing 3—4 days before stamens emerge; stamens emerging in a scattered pattern, held against the pistil; filaments fleshy, greenish, translucent; anthers creamy white, 0.5-1 mm long, —1.1 mm wide; thecae ellipsoid, not divaricate; pollen yellow fading to creamy white, abundant. Anthurium oerstedianum is endemic to Costa Rica at 800 to 1,300 m, principally on the south- western, Pacific slope in premontane wet, tropical wet, and premontane rain forest. This species is distinguished by its strictly ter- restrial habit as an understory forest herb and by its geniculum, which is located 10-20 cm below the base of the blade. Other distinguishing features are its four-sided petiole and greenish spadix. This species is a somewhat atypical member of sect. Pachyneurium because of its elongate petioles, remote geniculum, and lack of a dense, rosulate habit. It is apparently not closely related to any other species. Anthurium cuspidifolium, which is synony- mous with A. oerstedianum, was reported by Мас- bride (1936) from Peru. However, the specimens cited by him (Macbride 5622 (F); м 572 (F)) are referable to А. flavescens Poep The Cartago collection locality Ше E. Schott in the type description is dubious because no mod- ern collection of this species has been made from anywhere on the Atlantic slope. Since Schott men- tions only “Naranjo,” perhaps the collection refers to the Rio Naranjo on the Pacific slope where it easily could have occurred. Costa RiCA. CARTAGO: vic. Naranjo, Oersted s.n. (il- lustrated by Schott, Aroideae 328). PUNTARENAS: 6 km S of San Vito de Java, Wilson's finca, 1,330 m, Raven 20922 (MO); Las Cruces Botanical Garden, along road to river below botanical garden, 1,300 m, Croat 44415 (MO); Osa Peninsula, Corcovado, Gilbert s.n. (MO). SAN JOSÉ: San Isidro de General-coastal town of Dominical, SW of San Isidro, 900-1,000 m, к 35267 (МО, I те (МО); 0.5 mi. above turnoff to Сапаап at Rivas, 900 m, Croat 43419 (МО); El ad 1,010 m, Skutch 2291 (MO); Cerro Pelon, N base, just E of main road, 0.5 km S of San Martin de Puriscal, 800 m, T4N, 22 4 1 и del Pacifico, Canaán-Chimirol, General Valley, 000 m, Burger & Liesner 7124 (MO). Anthurium ottonis K. Krause, Notizbl. Bot. Gart. Berlin-Dahlem XI: 609. 1932. TyPE: Bolivia. La Paz: Prov. Sud Yungas: Jirupaysi, near Yanacachi, Buchtien 387 (holotype, B). Fig- ures 209, 210. Terrestrial; stem moderately slender, to at least 11 cm long, 0.8- numerous, spreading, green, drying grayish to pale brown, = villous, especially when young, elongate, somewhat tapered, sometimes branching, ca. 3- 25 cm long, 2-8 mm diam.; cataphylls subcoria- ceous, lanceolate to broadly lanceolate, 7-13 cm long, acute to rounded at apex, yellow-green, tinged with red, drying brown to reddish brown (B & yellow-red 4/10), persisting intact to semi-intact, rarely as fine linear fibers, eventually deciduous. Leaves erect to spreading; petioles 22-42 cm long, 3—6 mm diam., erect to erect-spreading, subterete to broader than thick, bluntly and shallowly sulcate adaxially, rounded abaxially, the surface promi- nently pale-speckled; geniculum much paler and somewhat thicker than petiole, becoming fissured transversely with age, 0.5-2 cm long; blades co- riaceous, linear-oblong, narrowly acute and mu- cronate at apex, acute to obtuse to narrowly round- ed at base, 38-60(100) cm long, 2.2-7 cm wide, broadest at or near the middle (12.6-17.6 x longer than the petiole), the margins flat to somewhat undulate; upper surface matte to weakly glossy, dark green, lower surface matte, conspicuously paler; both surfaces drying matte, green to yellow- ish green; midrib convexly raised, paler than sur- face above, pale-speckled like petiole, higher than broad at base, becoming prominently convex to- cm diam.; roots moderately Volume 78, Number 3 1991 Croat 689 Anthurium sect. Pachyneurium ward the apex below, moderately paler than surface and pale-speckled like petiole; primary lateral veins 18-26 per side, departing midrib at 45—60° angle, straight to arcuate, weakly sunken to obscure above, darker than surface below; interprimary veins ob- scure above, almost as conspicuous as primary lateral veins below; tertiary veins obscure above, weakly visible below; reticulate veins not visible; collective vein arising from the base, equally as prominent as primary lateral veins, 1-4(6) from margin. /nflorescences erect, shorter than leaves; peduncle 27-50 cm long, 2-4 mm diam., 0.8-1.3x as long as petioles, pale reddish, drying greenish to brownish, terete; spathe spreading-re- curled, subcoriaceous, yellow-green tinged with red on the midline, oblong-lanceolate to linear, ca. 3.5- 6.5 ст long, 0.8-1.3 cm wide, broadest near the base, acute to abruptly acuminate at apex (the acumen apiculate, 4 mm long), acute at base; spa- dix maroon (B & K red-purple 2/5), dark brown when dried, slightly tapered, subsessile, somewhat curved, held at 150? angle from peduncle, 4- 10(12) cm long, 4–5(6) mm diam. near base, 2- 3 mm diam. near apex; flowers + square to almost 4-lobed, 2-2.4 mm in both directions, the sides + straight to jaggedly sigmoid; 3-6 flowers visible in principal spiral, 6—9 in alternate spiral; tepals matte, very minutely and densely papillate, punctate, with few droplets appearing as stamens emerge; lateral tepals 1— mm wide, the inner margins straight to convex, the outer margins 2- rarely 3-sided; pistils semiglossy, minutely papillate, emergent be- fore stamens emerge, dark maroon; stigma ellip- soid, 0.3-0.7(0.9) mm long, depressed medially; stamens emerging in a regular sequence from the base, arranged shortly above the tepals, the laterals preceding the alternates by 14—15 spirals, held in a circle around the pistil; anthers white, drying white to brown, 0.5-0.7 mm long, 0.6– m wide; thecae ovoid, not divaricate; pollen white. /nfruc- tescence not seen. A member of series Multinervia, Anthurium ottonis is known from Bolivia in the Department of La Paz and in Puno, Peru, at 890 to 1,700 (2,500) m in lower montane subtropical moist, lower montane subtropical wet, and subtropical moist forest life zones. This species is distinguished by its erect-spread- ing leaves, long, pale-speckled, subterete petioles and narrowly straplike blades, and long-peduncu- late inflorescence with a moderately stubby, ma- roon spadix. It is not confused with any other Pachyneurium species and apparently has no close relatives. Although on the majority of dried spec- imens the anthers appear orange to brown, the type specimen displays white-drying anthers. A collection (Bogner 903) cultivated at the Mu- nich Botanical Garden and represented by two sheets at the Kew Herbarium is similar to 4. ottonis in most respects, except that it has a prolonged, free- ending sheath 4-11.7 cm long. It may represent a new species. BOLIVIA. LA PAZ: Prov. Inquivisi, Inquivisi-Circuata, 2,500 m, Besse 2 " 656 (SEL); Prov. Loayza, Circuata- Miquilla, Km 28, 1,560 m, Besse et al. 1839 (SEL); Prov. Nor Yungas, Coroico, Polo-Polo, 1,100 m, Buch- tien 3659, 3661 (US), 3660 (GH, US), 3662 (HBG, US); Coroico (Yolosa jct.)-Caranavi, road parallel to Rio Coroico, 670-1,280 m, Davidson 4790 (MO); Rio Hua- rinilla, 4.5 km below Yolosa, then 14 km W on road up Rio Huarinilla, 1,200-1,300 m, 16?12' 5, 67°50'W, Sol- отоп 9390 (MO); below Yolosa, 1,450 m, Solomon 8652 (= TU "9 4 о = un" = = ne E d да £5 5 н“ e. lan 8 = о РА 3 я a — ч = "о @ d d E cachi, ipi ies i, 1,650 m, Buchtien 387 (В); Chula sal 50 towards Asunta, 890 m, Beck 12606 (MO); 26 km کک‎ Asunta, past Tajama, 1,300 m, Beck 12074 MO). PERU. cuzco: Urubamba, Machupicchu, 2,000 m, 903 (K). PUNO: San Gavón, Baker 4358 (culti- sed; at MO) (MO, NY). Anthurium oxycarpum Poeppig in Poeppig & ., Nov. Gen. Sp. 3: 83. 1845. TYPE: Peru. Loreto: Yurimaguas, Poeppig s.n. ава dis selected here, С). Figures 211-213 Anthurium strictum N. E. Br ex Engl., Monogr. Phan. 2: 638. 1879. ТУРЕ: Brazil. Acre: Rio Branco E. Brown s.n. (К, identified with Kew Negative #2805). Epiphytic or terrestrial, sometimes epipetric; stem short or elongate, to 60 cm long, 1-3 cm wide; eaf scars inconspicuous, obscured by root mass, 1 cm high, cm wide; roots numerous, dense, spreading, green to whitish, smooth to densely pu- escent when dried, short, bluntly tapered, 1-7. cm long, (1)2-4 mm diam.; cataphylls membra- nous to subcoriaceous, lanceolate, prominently l-ribbed throughout, 5.5-12 cm long, acuminate at apex with subapical apiculum 1-2 mm long, light green, drying thin, yellowish to pale tan (B & К yellow 7/5), persisting + intact, eventually as fine linear fibers. Leaves erect to erect-spread- ing; petioles (2.5)5-15 cm long, 4-9 mm diam., subterete to C-shaped to D-shaped, flattened to narrowly or obtusely sulcate adaxially, the margins blunt, rounded abaxially, the surface dark green, pale-speckled; geniculum slightly thicker than pet- iole, 1-2.5 cm long; sheath for 3-4 cm long; blades subcoriaceous, narrowly to broadly elliptic to broadly oblanceolate to narrowly obovate, acute to short- 690 Annals of the Missouri Botanical Garden acuminate at apex (the acumen downturned, in- rolled), obtuse to acute or attenuate at base, (12)22- 55 cm long, 10-19 cm wide, broadest at or above the middle, the margins flat to broadly undulate; upper surface glossy to semiglossy, occasionally matte, medium to dark green, lower surface matte to weakly glossy, moderately to conspicuously pal- er; midrib above raised and paler than to concol- orous with the surface, below prominently and ob- tusely to acutely raised at base, becoming convexly raised toward the apex and slightly paler than sur- face; primary lateral veins 10-25 per side, de- parting midrib at 40—70° angle, straight to weakly arcuate to the collective vein, weakly raised near the midrib, becoming sunken in grooves toward the margin; interprimary veins weakly sunken above, weakly raised below; tertiary veins weakly sunken above, raised below; reticulate veins obscure; col- lective vein arising from near the base, sunken above, raised below, paler than surface, 5-11 mm from margin. /nflorescences erect, equaling or lon- ger than leaves; peduncle 18.5-54.7 cm long, 3- 5 mm diam., -5X as long as petiole, green, terete; spathe spreading to reflexed at an acute to almost right angle from spadix, coriaceous to sub- coriaceous, light to dark green, sometimes tinged with red or maroon at margins, lanceolate to ovate- lanceolate, sometimes elliptic, 4.5-12.5 cm long, 1-3 ст wide, broadest usually in the lower third (sometimes at or near the middle), inserted at 35- 45° angle on peduncle, abruptly acuminate at apex (the acumen tightly inrolled, 6 mm long), acute to acuminate at base, the margins meeting at 60–80• angle; stipe to 17-20 mm long in front, 0-1 mm long in back; spadix bluish green to green to yellow- green, also reported as creamy and tan, tapered, erect, held at 140-170? angle from peduncle, 4- 8.5 cm long, 6-10 mm diam. near base, 3-7 mm diam. near apex; flowers square, 2.4-3.2 mm in both directions, the sides moderately straight par- allel to spiral, straight to jaggedly sigmoid perpen- dicular to spiral; 4—7 flowers visible in principal spiral, 9-11 in alternate spiral; tepals covered with thin, matte, bluish green, waxy bloom, few droplets present at anthesis; lateral tepals 1.3-2 mm wide, the inner margins broadly convex, the outer mar- gins 2—3-sided; pistils weakly emergent at anthesis, matte covered with waxy bloom, green, becoming brown; stigma oblong-ellipsoid, 0.6–0.7 mm long, brushlike and depressed medially before droplets emerge; stamens emerging in a complete sequence from the base, inclined slightly inward over the stigma; anthers creamy white to 0.5-0.6 mm long, 0.6-0.8 mm wide; the- cae oblong-ovoid, scarcely divaricate; pollen white. rompt, regular, orange, Infructescence with spathe persisting; spadix 6.5— 8.5 cm long, 1.8-2.5 cm diam., with berries scat- tered throughout; berries green, becoming red and eventually purple in the apical half, obovoid-oblong to ellipsoid-obovoid, long-acuminate toward the apex, rounded at apex with radial ridges, 7.5-12.3 mm long, 4.2-5 mm diam near base, toward the apex narrowing to 1.7-2.5 mm diam.; pericarp thin, transparent; mesocarp gelatinous with mod- erate number of raphide cells; seeds 2 per berry, yellowish, oblong-ovoid, flattened, 5.3-5.5 mm long, 3 mm diam., 1.5-1.8 mm thick, enveloped by transparent, sometimes amber, gelatinous, sticky substance Anthurium oxycarpum ranges from southeast- ern Colombia to Amazonian Ecuador, Peru, Bolivia, and Brazil at 100 to 870 (1,300) m. It occurs in tropical moist, premontane wet and tropical wet forest life zones. This species is distinguished by its rosulate habit, generally short, densely short-rooted stem, con- spicuously veiny leaves (with at least some of the tertiary veins sunken above), and especially by its erect, narrowly ovate-lanceolate spathe and fre- quently bluish green, slightly tapered spadix. The leaf blades characteristically dry thin and somewhat glossy on the lower surface. Some cultivated plants of this species have markedly bullate and weakly quilted leaf blades, a characteristic not yet observed in the fie Anthurium oxycarpum is closest to A. knap- piae, which differs in having petioles two to three times longer, an attenuate leaf base, a much longer peduncle, a nonglaucous spadix, and tepals with conspicuous raphide cells. See discussion of that species for details. The original type specimen designated by Poep- pig was collected in Brazil at Ega (now Tefe), lo- cated on the Rio Solimoes at the mouth of the Rio Јарига. Originally deposited at Vienna, this spec- imen is now lost, and a thorough search of all major herbaria has turned up no duplicates. Consequent- ly, a second Poeppig collection from Yurimaguas in Peru, cited by Engler in his 1905 revision, is here designated as the lectotype. The Yurimaguas collection is also the only one illustrated by Schott (Scott drawing 356; NYBG Negative #3873; mi- crofiche #15: С-7). Considerable confusion exists regarding the type of Anthurium strictum, now a synonym of A oxycarpum. In preparing a description of what he presumed to ђе А. dombeyanum for the Refugium Botanicum, Baker (1871) described and illustrated instead 4. oxycarpum. Upon realizing this error Volume 78, Number 3 1991 Croat 691 Anthurium sect. Pachyneurium and assuming the latter to be a new species, Engler (1879) redescribed it as А. strictum, attributing the name to . Brown at Kew. He cited as the type the same material used to illustrate the Re- fugium Botanicum article, namely a specimen from the Rio Branco in Brazil (at least an inflorescence of the type ipd had been sent to Engler in Berlin by N. E. Bro Engler also misidentified as Anthurium strictum an Ule collection (5598) from western Acre along the Rio Jurua Mirim. While photographing type specimens in Berlin, J. F. Macbride of the Field Museum incorporated into a single photograph an inflorescence of А. oxycarpum (labeled А. stric- tum) with the Ule collection misidentified by En- gler. The latter collection, which is complete, is actually 4. uleanum. The mixed collection rep- resented in this Field Museum photograph needs to be corrected. The specimen to the left represents A. uleanum, while the one to the right (inflores- cence only) represents А. oxycarpum. BOLIVIA. EL BENI: Rurrenabaque, Cárdenas 1167 (NY); Prov. Ballivian, n2 Fátima Rio Maniqui, ec hi & Hinojosa 1045 (LPB, MO). ТА Paz: Prov. Larecaja, 27.8 m N of Caranavi, Cual 51651 (МО); У Sul Yugas, Alto Beni Concesión de la Cooperació n José de Popoy, 600 m, Seidel & Schulte 2246 un MO). BRAZIL. ACRE: Abuna-Rio Branco, Forero et al. 6315 (COL, NY, US, INPA); Rio Branco, Zoobotanical Garden of Federal University of Acre, Lowrie et al. 170 (INPA, MG, MO, NY); Mpo. Caramari Amazonas, Rio Juruá, N of Cruzeiro do Sul, Lago da Cigana (Sao Luis), 5 Alvaro Nestrinho, 150 m, 7%37'S, 72*36'W, 624984 (CM, INPA, MO); Rio Juruá Mirim, Ule 5598 (B). AMAZONAS: Rio Javari, behind Estiráo de Equador, Lleras et al. P17284 (INPA, NY); Rio Purus, Monte Verde, Huber 4623 (MG). RONDONIA: NW of Rio Ma- deira, across from Matuparana, Calderón et al. 2828 (MO, NY, US); Mpo. Ariquemes, Mineracáo Mibrasa, sector Alto Candeias, Km 128, 10%35'S, 63?35'W, Tei- xeira et al. 467 (NY, INPA). COLOMBIA. Without 2 Fantz 4032 (FTG). amazonas: Leticia, Krukoff 514 (NY), Oldenburg 2812 (US); Rio Loretoyacu, 100 m, Schultes & Black 8407 (US, GH); Puerto Narino, 100 m, Plowman 3221 (GH, US, SEL), Madison 3743 (cult. plant of Plowman 3221 (SEL). ЕСЏАРОК. MORONA- SANTIAGO: 20 km from Limón, 700-900 m, Harling & Andersson 12875 (СВ). МАРО: Reserva de Producción Faunistica Cuyabeno, М of Laguna Grande, 265 m, Niel- sen 76226, 76516 (AAU), Poulsen 79692 (AAU); Puer- apo-Puerto Misahualli, 3.5 km E of Puerto Napo, 125. 77°47'W, Croat 58882 (C, CAS, CM, GH, MO, NY, ОСА, RSA, SEL); 8 km below Puerto Misahualli on Ко Маро, 1.5 km S, 1*4'S, 77*36'W, 450 m, Palacios et al. а m ‚ ОСМЕ), Palacios 2859, 2877 (MO), 2961 bs M, МО, ОСМЕ, W); Lago Agrio-Coca, 8.8 km 10 Жай uarico оп road to Coca, San Francisco de Orellana, Croat 585124 (MO); confluence of Rio Borja and Rio Quijos, Е bank, ca. 1,300 m, 0%25'S, 77?49'W, Holm-Nielsen et al. 26234 (AAU); Río Aguarico, Holm- Nielsen et al. 21115 (AAU, UEC); 30 min. by canoe below San Pablo, 250 m, 0*18'S, 76*25"W, Laegaard 51563 (QCA); Río Arajuno, Sola Cocha, 500 m, 1*7'S, 77°36'W, Palacios et al. 896 (MO); Rio Lagartococha, 190 m, 0*33'S, 75°13'W, Lawesson et al. 44409 (AAU); Rio Napo, Limoncocha, 240 m, Foster 3839 (F), Mad- ison et al. 5432 (MO, SEL, QCA, US, K); Jatün Sacha Biological Reserve, 8 km below Misahualli, 450 m, 1904'5, 77°36'W, Cerón 723, 1312 (MO), Cerón et v 1992 (MO, QCNE); 400 m, Cerón 6289 (MO, QCA), 5963 even MO, OOM); Rio Wai si aya, 1 km upstream from Rio Aguarico, Brandbyge et al. 33190 (AAU); NAPO-PASTAZA: Tena, Asplund 9199 (S); Tena-Napo, Asplund 10253 (S). PASTAZA: Rio Curaray, Curaray, 200 m, Harling & Andersson 17517 (GB). PERU. AMAZONAS: Prov. Leoncio Prado, 5 bin So ¿Ti Kennedy 5699 (SEL, МО); Rio Monzón, Schunke 5111 (GH, F, US, NY); Jacintillo, W of Tingo María, 800 m, Plowman & Kennedy 5772 (F, GH); Quebrada Las Pa- 02 (МО); Dtto. Ruppa Ruppa, Carpar Bella, Cave of Los Huarinos, 700-900 m, Schunke 9465 (CAS, CM, F, G, MO); Rio Huallaga, 750-800 m, Plowman & Ramirez 7569 (F); Tingo Maria- Monzón, Río Patay Rondos, 650 m, 9?21'S, 76?12'W, Croat 57946 (CM, MO). J JUNÍN: Puerto Bermúdez, 375 m, Killip & Smith 26604 (US, NY); Oserato Tambo, Weiss 152 (Е); Río Leco, Soukup 3516 (Е); Rio Perene, mouth of river, n ndoas, 180 m, 02%55'S, 16*25' W, Vásquez & Jaramillo 4566 (МО); Rio Huallaga, Santa Rosa, below Yurimaguas, 135 m, Killip & Smith 28714 (NY, US); Yurimaguas, 135 m, Killip & Smith 28014 (NY, US, F); Yurimaguas- рева: 135-150 т, Killip & Smith 28367 (US) Rio Huallaga (lower basin), Balsapuerto, 150-350 m, Killip & Smith 28628 (US, NY); Rio Napo, Entrada de Isla Inayuga, Croat 20549 (МО); Prov. Co- ronel Portillo, ere . C. Mariategui and Mejico, 300 m, 8?15'S, 73°45'W, Diaz et al. 756 (MO); Prov. Maynas, Iquitos Re egion, Soledad, Rio Itaya, 110 m, Kil- lip & Smith 29762 (NY, US); Río Amazonas, Explorama Inn, 1 km $ of Indiana, 130 m, 3°30’S, 73°01'W, Croat 61679 (AMAZ, MO), Gentry et al. 54623 (MO), 55986 (MO), 61772 (MO), 65765 (M P 3°28'S, 72?48'W, Gentry et al. 31429 (MO); Iquitos Region, Las Pebas, io Am- piyacu, Pijuayal Army Base, 3?10'S, 71%49"W, Plowman et al. 7131 (F, СН); сес of Rio Маро, Yana mono, 130 m, 3?28'S, 72°. MO), Gentry et al. 27509, PUN 42256 (MO), Gentry & Jaramillo 28062 (MO), Gentr 19428 (F, MO, USM), 20779, 20789 M Prov. Re- quena, Sinchicuy, 106 m, 3?35'S, 73*15'W, e et al. 7830 (MO). MADRE DE DIOS: Prov. Мапа, Мапа tional Park, Cocha Cashu Biological Station, 350-4 400 m, Foster 9902 (B, K, MO), праве 2720 (МО), 43638 692 Annals of the Missouri Botanical Garden (MO, SEL), Миле: 5537 (MO), 5738 (МО, W), 5809 (MO, US); Rio La Torre, confluence of Río Tambopata and Río La Torre, 39 km of Puerto Viger 12%50'S, 69*20'W, Barbour 4767, 5424 (MO), Sm et al. 132 (MO, US), 346 (US), 1387, 1391 (US). Pasco: A ERA 1 (MO); Rio San o Chuchurras draina win 00 m, 10909'5, 75°20’ W. Smith 4037 (M 0) UCAYALI: LSU bus camp, Quebrada Shesha (trib. of Río Abajao), 65 km Pucallpa, 250 m, 8°02’S, 73°55'W, Gentry & Diaz 58559 (MO) o ~ Anthurium oxyphyllum Sodiro, Anales Univ. Centr. Ecuador 15(108): 5. 1901. TYPE: Ec- uador. Pichincha: near Santo Domingo de Los Colorados, 400 m, Sodiro s.n. (holotype, Q). Figures 214-216 Epiphytic; stem short, (1)1.5-2 cm diam.; roots dense, numerous, ascending to spreading, green, smooth when young, soon becoming tomentose, slender and elongate, to 13 cm long, 2-6 mm diam.; cataphylls membranous, 7-10 cm long, nar- rowly rounded at apex with subapical apiculum ca. 8 mm long, drying tan (B & K yellow 9/2.5), persisting as fine linear fibers. Leaves spreading; petioles 6–25(30) cm long, 4-7 D-shaped, flattened to convex to weakly sulcate or occasionally with a medial rib adaxially, the mar- gins sharply raised, rounded abaxially; geniculum thicker and paler than petiole, (0.7)1-2 cm long; blades coriaceous, narrowly elliptic to sometimes narrowly oblanceolate, long-acuminate at apex (the acumen flat), long-attenuate at base, 45-70 cm long, (4)6-11 cm wide, broadest at or near the middle, the margins broadly undulate; upper sur- face matte, dark to bright green, lower surface glossy to semiglossy, paler; both surfaces drying matte, greenish to yellowish brown; midrib flat to convexly raised above, prominently and acutely raised and paler than surface below; primary lateral veins 25-35 per side, departing midrib at 30—60° angle, + straight to the collective vein, scarcely raised to flat or weakly sunken above, weakly raised to + obscure below, drying slightly raised above and below; interprimary veins numerous, almost as mm diam., conspicuous as primary lateral veins, obscure when fresh, weakly raised when dried; tertiary veins vis- ible when dried; collective vein arising from near the base, equally as prominent as primary lateral veins, raised when dried, 3-6 mm from margin. Inflorescences pendent to erect-spreading, shorter than leaves; peduncle (21)24-46 cm long, (3)4- 5 mm diam., 1.2-6X as long as petiole, light green tinged with purple or maroon at least at pane, terete to subterete and flattened adaxially; spath to reflexed, held at 130-160° angle to peduncle, often weakly twisted, subcoriaceous, green to green tinged with purple or red (B & К yellow-green 7/10), lanceolate to oblanceolate, (5)7.2—10.5 cm long, (1.8)2.3-3 cm wide, broadest near the base, inserted at 30-90? angle on peduncle, abruptly acuminate at apex (the acumen inrolled), obtuse to rounded at base, the margins meeting at 100- 140? angle; stipe 3-7 mm long or absent; spadix green to yellow-green (B & K yellow-green 6/10), cylindroid, semi-erect to nodding, usually slightly curved, held at (150)130-60? angle from pedun- cle, 5.8-11 cm long, 3-7 mm diam. near base, 3-5 mm diam. near apex; flowers + rhombic to 4-lobed, 3-4.5 mm long, 3.7-4.3 mm wide, the sides straight to smoothly or sometimes jaggedly sigmoid; 5-7 flowers visible in principal spiral, 6– 8 in alternate spiral; tepals brown, matte to se- miglossy, minutely papillate, with abundant drop- lets present; lateral tepals 1.3-2.5 mm wide, the inner margins broadly convex to straight, scarcely turned up against the pistil, the outer margins 2-sided; pistils prominently emergent, exserted and papillate, the exposed portion squarish, green be- coming dark purple; stigma linear to ellipsoid, ca. 0.4 mm long, weakly raised, copious droplets ap- pearing before the stamens emerge; stamens emerging in a regular sequence throughout, held well above the tepals, lateral stamens emerging almost to apex before alternates begin to emerge, arranged in a circle around the pistil; filaments transparent, prominently exserted, thin and flat- tened, 1.5-1.7 mm long ве, becoming pinkish or reddish brown, ca. 0.8 mm long, 0.9 mm wide; thecae oblong-ellipsoid, scarcely divar- icate; pollen orange, fading to white or tan. /n- fructescence with spathe persisting; spadix 9.5- 11.5 ст long, 1-1.5 cm diam., with berries scat- tered throughout; berries orange, ellipsoid, acute and with radial ridges at apex, 6.8- 7 mm long, 3.2-3.7 mm diam.; mesocarp mealy; seeds 1 per berry, yellow-brown when dried, + ovoid, truncate at both ends, 3.5-5 mm long, 2.2-3 mm 1.5-1.6 mm thick, enveloped by gelatinous, trans- lucent, amber substance. iam., A member of series Multinervia, Anthurium oxyphyllum is endemic to the Pacific slope in Ecuador from Carchi to Cotopaxi at 300 to 1,300 m. This species is ecologically variable, with spec- imens collected both in a lower montane dry forest life zone (Esmeraldas) and a montane rainforest life zone (Carchi). Despite occurring in different life zones, the specimens themselves do not differ significantly in any aspect. However, the type col- Volume 78, Number 3 1991 Croat 693 Anthurium sect. Pachyneurium lection is somewhat unusual in having leaf blades with apices cuspidate-acuminate rather than long- acuminate. is species is characterized by its epiphytic habit, generally oblong-elliptic, coriaceous, green- ish-drying blades with numerous primary lateral and interprimary veins, long peduncles (1.2—6 x longer than the petioles) and flowers with promi- nently exserted stamens. However, the best distin- guishing character is the nodding inflorescence with the spathe and spadix abruptly bent downward at the base. Anthurium oxyphyllum bears some resem- blance to A. carchiense and А. obscurinervium, both of which have less coriaceous leaf blades and erect inflorescences. Sodiro (1902) asserts that the leaf blades are not punctate, but rather “‘pellucid-punctate under a lens." This agrees with living material, in which the leaf blades cast a semi-“‘velvety”’ held at arm's length. This effect results from mod- ified epidermal cells (**pellucid punctations””) which scatter light in different directions. Most of Sodiro's sheen when descriptions were based on living material, and this observation was probably made on a living plant. ECUADOR. CARCHÍ: Chical, 1,130 m, 0%56'N, 78?14'W, Madison et al. 4507 (К, SEL, Е); below Maldonado, 1,100 m, Madison 6967 (SE Chico, pco Reserve, vic. of encampmen Ethnic, 1,330 m, Hoe et al. 3256 (MO). COTOPAXI: Quevedo-Latacunga, 55.5 km Penas Blancas, 2 km У» Dare a ison et al. 4992 (SEL, K, QCA Croat 69736 (cultivated at MO); Lita, 6 , Maas & Cobb 4743 (U, QCA). IMBABURA: vic. rM 600 m Cobb 35 (MO); Rio Mira, Ibarra-Lita, 2.5 km E of Lité, 750-775 m, Croat 38977 (M, MO, RSA, SEL). Anthurium pachylaminum Croat, sp. nov. TYPE: Peru. San Martin, Moyobamba-Cha- кабы Km 430-431, E of Naranjos at Rio Naranjos, 770 m, 5?21'S, 77°20'W, Croat 28161 (holotype, MO 3148985; isotypes AAU, B, CAS, F, GH, K, M, NY, P, SEL, U, US, USM). Figures 217-220. Planta epiphytica; petiolus sulcatus, 8-20(32) cm lon- longior quam petiolus; spatha subcoriacea, linearis- lan- ceolata, 10-23 cm longa, 1-2.5 cm lata; spadix atro- rubens vel atroviolaco-atropurpurea, cylindricus, 11-39 cm longus; cum fructu pendulus; baccae ferentes ad basim. Epiphytic, sometimes terrestrial; stem to 8 cm long, 1.5-6 cm spreading, the uppermost ascending, gray when dried, drying short-pubescent, 2-10 cm long, 2- m diam.; cataphylls coriaceous to subcoria- ceous, broadly lanceolate, 5-7 cm long, acute at apex, drying reddish brown (B & K yellow-red 04/ 05), persisting semi-intact at the upper nodes. Leaves erect-spreading; petioles 8–20(32) cm long, ca. 8 mm diam., D- to C-shaped, broadly to nar- rowly sulcate adaxially, with the margins blunt, rounded abaxially; geniculum 1-2 cm long; sheath to 6 cm long; blades coriaceous, elliptic to broadly eliptic or oblanceolate, acuminate at apex (the acumen to 2 cm long), occasionally obtuse at base, but mostly acute to decurrent, 40-112 cm long, (11)1 7-43 cm wide, broadest at or above the mid- dle, the margins flat to broadly undulate; upper surface glossy, dark green, lower surface glossy to matte, paler, often sparsely to moderately pustulate or glandular; both surfaces drying yellowish green; midrib flat at base, becoming convexly raised to- ward the apex above, obtuse to acute-raised below; primary lateral veins 6—9 per side, departing midrib —50° angle, mostly ascending, + straight to the margin, sometimes loop-connecting in the up- permost portion of the blade, raised above and below; tertiary veins etched above, weakly raised diam.; roots numerous, dense, below; collective vein arising from near the apex or absent, sunken above, raised below. Inflores- cences erect; peduncle 36-95 cm long, 3-10 mm diam., 2-9 X as long as petiole, olive-green, terete; spathe spreading-reflexed, recurled, subcoriaceous, green heavily tinged with maroon especially on the inner surface, with raphide cells, linear-lanceolate, 10-23 cm long, 1-2.5 cm wide, broadest near the base, acute at apex, decurrent at base; spadix dark red to dark violet-purple (B & K purple 2/7.5), sessile or stipitate to 10 mm, cylindroid to tapered, erect, 11-39 cm long, 6-7 mm diam. near base, 4 mm diam. near apex, broadest at the base or near the middle; flowers rhombic, ca. 2.4 mm long, ca. 1.4 mm wide, the sides straight; 7-9 flowers visible in principal spiral, 8-10 in alternate spiral; tepals matte, papillate; lateral tepals 1-1. wide, the inner margins + straight to broadly con- vex, the outer margins 2—3-sided; pistils not emer- gent; stigma linear, 0.4-0.6 mm long, appearing granular; stamens emerging in a regular sequence from the base, the laterals preceding the alternates by 28 spirals, the 3rd stamen preceding the 4th by 6 spirals, inclined over and obscuring pistil; 694 Annals of the Missouri Botanical Garden anthers 0.4—0.6 mm long, 0.6–0.8 mm wide; the- cae ovoid, slightly divaricate; pollen pale orange. Infructescence pendulous; spathe withered or ab- sent; spadix 22-53 cm long, 1.4-1.7 cm diam., mostly bearing berries in the basal portion only; stipe to 10 cm long; immature berries reported as brown. Anthurium pachylaminum is known from Lo- reto, San Martin, and Pasco in Peru at (240)320 to 770 m. It may be epiphytic or terrestrial in sandy, inundated soils and on steep slopes in trop- ical moist and premontane wet forest life zones. This species is distinguished by its unusually large, coriaceous leaf blades (hence the name), with the tertiary veins etched into the upper surface (when fresh) and the primary lateral veins depart- ing the midrib at an acute angle and running more or less straight to the margin. The leaf blades of А. pachylaminum are often pustular or glandular on the lower surface, and the spa ot wither at anthesis, but often persists and remains in a weathered condition on the fruiting spadix. Anthurium pachylaminum bears a similarity in leaf shape and inflorescence shape and color to 4. atropurpureum var. arenicola. The latter, how- ever, has mostly eglandular lower leaf surfaces, and the spathe withers at anthesis. It also generally has smaller leaves and shorter peduncles (1-4 х as long as petiole) than 4. pachylaminum (2-9 x as long as petiole). PERU. LORETO: Prov. Alto Amazonas, Rio Capihuari, 5 km NE of Andoas, near Ecuador border, 240 m, Gentry & Diaz 28226 (MO, NY); Prov. Maynas, Rio Tacsha- curaray, Croat 20389 (Е, MO, USM). Pasco: Prov. Ox- apampa, Iscozazin- Villa America, Palcazú Valley, 300 m, 10?12'S, 75?13'W, Smith 3855 (MO); Rio Iscozazin, Cabeza de Mono, 10 km SW of Iscozazin, Palcazü Valley, 320 m, 10°20'5, 75?18'W, Gentry et al. 41949 (MO). SAN MARTÍN: Moyobamba- Chachapoyas, Km 430-431, Croat 58161 (AAU, B, BM, CAS, F, NY, P, ОСА, RSA, SEL, U, US, USM); Ко Huallaga, m nd village of San 550-600 m, 8°10' S, 76°33'W, Croat 50990 (K, МО); Quebrada de Canuto, 500 m, се 10945 (СМ, МО); Quebrada de Ishichimi, Тосасће, 400 m, Schunke 10047 (MO); Rio Tocache, Puente Palo Blanco, 10 km W i Tocache Nuevo on rd. to Puerto Pizana, 550-650 8914'5, 76*36'W, Plowman et al. 11347 (Е); Puede Palo Blanco, 500-550 m, 8?14'S, 76°36'Ұ, Plowman & Schunke 7432 (F) Anthurium palenquense сон а поу. ТҮРЕ: uador. Los Rios: Rio q Cen- ter, along trails W of о and in vicinity of laboratory clearing, 210- m, Croat 38670 (holotype, MO 2395515; isotypes, AAU, B, К, NY, ОСА, PMA, RSA, US). Fig- ures 221, 223, 224. Planta terrestris; internodia brevia, 1-1.2 cm diam.; petiolus (6)10-30 c laris; lamina elliptica ad lanceolata ad oblanceolata, basi attenuata; nervis primariis lateralibus (9)1 4-30 utroque; pedunculus 10-60 cm longus, 2-10 mm latus, quadran- gularis vel 2-4-costatus; spatha жш lanceolata ad oblonga-lanceolata, 3-12 cm longa, 0.5-2.5 cm lata, viridis; stipes liber ad 1-5 mm; spadix contractus, (2)4- m longus, 6-9 mm latus prope basim, 3-4 mm latus prope MER flavo-virens; baccae ovoideae, purpureae, 5-11 mm longae, 5-6 mm latae. Terrestrial; loused, 15-30 cm long, 1-2.5 cm diam.; leaf scars obscured by root mass, 1 cm high, 1-1.5 ст wide; stem green turning brown, cal- roots numerous, descending, grayish or green, smooth or scurfy, blunt, 1-20 cm long, 2-5 mm diam.; cataphylls membranous to subcoriaceous, lanceolate, 4.5-17 cm long, narrowly acute and apiculate at apex, green tinged reddish at margins, drying pale tan (B & K yellow 5/10), persisting semi-intact or as coarse linear fibers or as bundles of linear fibers, often with the apex remaining in- tact. Leaves erect to spreading; petioles (6)10-30 cm long, 5-20 mm diam., sharply triangular, flat- tened to broadly convex adaxially, the margins winged, sharply angular abaxially; geniculum scarcely thicker and slightly paler than petiole, the angular ribs often conspicuously and minutely un- dulate, 1-3 cm long; sheath 1-2 cm long; blades subcoriaceous to coriaceous, elliptic to lanceolate or oblanceolate, long-acuminate at apex (the acu- men flat to slightly inrolled), attenuate with con- spicuously concave margins at base, (16)30–64(70) cm long, (7.5)10-23 cm wide, broadest usually near or below the middle, the margins broadly undulate; upper surface matte, velvety, the cells convex, sunken, forming an alveolate pattern when dried, dark green (B & K green 4/2.5), lower surface matte, paler (B & K green 7/2.5); midrib flat to obtusely angular at base, becoming sharply acutely raised toward the apex above, acutely raised (knife-edged), and higher than broad below; pri- mary lateral veins (9)14-30 per side, departing midrib at 40—65(80) angle, ascending straight to the collective vein, weakly sunken to weakly raised above, darker than surface and weakly raised be- low; interprimary veins numerous, scarcely sunk- en, + parallel to and less conspicuous than primary Volume 78, Number 3 1991 Croat 695 Anthurium sect. Pachyneurium lateral veins; tertiary veins obscure above and be- low; collective vein arising from near the base, weakly sunken or weakly raised above, weakly raised or flat below, prominulous when dried, equal- ly as prominent as primary lateral veins, 5-18 mm from margin. /nflorescences erect to slightly spreading or + pendent, shorter than leaves; pe- duncle 10-60 cm long, 2-10 mm diam., 0.3-3 x as long as petiole, green to dark purple or heavily tinged-mottled with red-purple (B & K red-purple 2/2.5), + quadrangular or with 2-4 irregular ribs and convex sides; spathe spreading and weakly twisted or reflexed, membranous to subcoriaceous, green tinged maroon at margins and along main nerves (B & K гед-ригрје 2/10), minutely pale- speckled (at least abaxially), narrowly lanceolate to oblong-lanceolate, 3-12 cm long, 0.5-2.5 cm wide, broadest near base, inserted at 30-90° angle on peduncle, acuminate at apex (the acumen in- rolled, (2)5-15 mm long), acute to decurrent at base; stipe 3-15 mm long in front, 1-5 mm long in back; spadix yellow-green tinged with purple, becoming purple-violet or dark green tinged pur- plish at anthesis, cylindroid, sometimes weakly ta- pered, erect to curved, (2)4-12 cm long, 3-5 mm diam. near base, 3-4 mm diam. near apex; flowers with a faint, sweet, fruity scent, square to rhombic, 2-3.8 mm long, 1-3.8 mm wide, the sides straight to gradually sigmoid; 2-5 flowers visible in prin- cipal spiral, 2-6 in alternate spiral; tepals matte to semiglossy, minutely papillate and weakly punc- tate; lateral tepals 1.4-2.6 mm wide, the inner margins broadly convex, turned up against pistil, the outer margins 2-sided; pistils exserted 0.4-1.5 mm, glossy, green, sometimes tinged with purple around the stigma; stigma ellipsoid, greenish, 0.4— 0.8 mm long, 0.25-0.4 mm wide, brushlike with blunt papillae, droplets appearing 1 week before stamens emerge; stamens emerging rapidly in a the length of the spadix), held erect above tepals and pistil; filaments translucent, exserted, 0.6-1 mm long, 0.4 mm wide; anthers orange (B & K yellow-red 7/10), 0.4-0.6 mm long, 0.5-0.7 mm wide, not obscuring pistil; thecae ellipsoid, 0.25 mm wide, weakly divaricate; pollen orange to yel- low fading to cream-white (B & K yellow-red 7/ 2.5). Infructescence with spathe persisting; spadix 5-15 cm long, 0.5-1.5 cm diam.; berries dark purple, ovoid, acute at apex, 5-11 mm long, 5- 6 mm diam.; pericarp thickened, with numerous raphide cells; seeds 1-2 per berry, with numerous raphide cells, rounded to ovoid, usually flattened on one side, 4–5.6 mm long, 3.2-3.8 mm diam., 1-2.8 mm thick, with a translucent gelatinous ap- pendage at both ends. А member of series Multinervia, Anthurium palenquense occurs in Ecuador in Esmeraldas, Co- topaxi, Los Ríos, and Pichincha provinces, at 80 to 400 m, typically in mature forest or in disturbed primary forest in tropical moist and premontane wet forest life zones. Most collections were made at the Rio Palenque Biological Station in Los Ríos, hence the name. Anthurium palenquense is distinguished by its terrestrial habit, coarse cataphyll fibers, triangular- winged petioles and closely veined, bicolorous blades which are attenuate at the base and matte to semi- velvety on the upper surface. In addition, the long- exserted stamens and the brushlike stigma are characteristic. An unusual feature is the pale raph- ide cells speckling the petioles, spathe, tepals, and even the filaments, as well as the pericarp and seed coat. This species (erroneously reported as А. buen- aventurae Engl.) is pollinated by Eulaema bom- boides (Dodson & Gentry, 1978). Two collections of Anthurium palenquense from Esmeraldas and Cotopaxi (Asplund 16306 and Sparre 17214), do not display the brushlike stigma (typical of the species), but agree in other char- acters with the present overall concept of 4. pa- lenquense. Anthurium palenquense is most closely related to 4. napaeum, which differs in having the leaf blades mostly acute at the base with convex to more or less straight (rather than concave) margins. ECUADOR. COTOPAXI: Quevedo-Latacunga, 44.7 km E O m, 0%51'S, 79°1 250 m, Sparre 17214, 17465, 17466 (S). ESMERALDAS: Quinindé, Rosa Cg Asplund 16306 (5); Rio Esme- i . Luis Vargas Torres- Est. Exp. Мише, 80 m, 0*52'N, 79933", Croat 55610 (MO, QCA, U). Los E Quevedo-Santo Domingo de Los Colorados, Río Palenque Biological Sta- tion, 150-250 m, 0935'5, 79°25'%, "rant 38670 (AAU, B, К, MO, NY, QCA, PMA, RSA, US), 38688 (AAU, MBM, MO, NY, ОСА, RSA), 49739 (MO), 50659 (CM, MICH, МО, ЗАК), Dodson 5173 (MO, ОСА, SEL, US), Dodson & Gentry 12830 (SEL), Dodson & McMahon 5059 (MO, SEL, US), Gentry & Dodson 41298 (MO), Lojtnant & Molau 15766 (AAU), Madison 3742 (cult. at SEL) (MO), Schupp 83 (SEL), Watson 267 (NY). PICHINCHA: Santo Domingo-Quinindé, 300 m, Solís 13952 Anthurium pallatangense Engl., Bot. Jahrb. Syst. 25: 385. 1898. TYPE: Ecuador. Coto- paxi: vic. Angamarca, Sodiro s.n. (holotype, B). Figures 222, 225 696 Annals of the Missouri Botanical Garden Anthurium р pecie Sp Univ. Centr. Ec- b 901. TYPE: Ecuador. Cotopaxi: An- ЈЕ ioe m, Sodiro s.n. (holotype, B). chincha: Nanegal, Sodiro s.n. (holotype, B). Terrestrial or epiphytic; stem ca. 30-50 cm long, 1.5-2 cm diam.; leaf scars obscured by root mass, 0.5 cm high, 1 cm wide; roots dense, de- scending, greenish, drying tan to grayish, smooth to somewhat pubescent, slender and elongate, ta- pered at apex, ca. 4-26 cm long, 1-4 mm diam.; cataphylls lanceolate, 4-11 cm long, acuminate at apex, drying reddish brown, persisting as pale, fine, linear fibers. Leaves spreading-erect; petioles 10- 33 cm long, sometimes alitusely flattened adaxially, rounded abaxially; geniculum obscurely flattened, scarcely darker than petiole, 0.6-2 cm long; blades sub- coriaceous, narrowly to broadly elliptic, occasion- ally oblanceolate, gradually acuminate at apex, slightly attenuate to narrowly acute to obtuse to subtruncate at base, 20-56 cm long, 4-15 ст wide, broadest at or near the middle; upper surface matte to semiglossy, dark to medium green, lower surface glossy to semiglossy, much paler; both sur- faces drying matte to semiglossy, bright green to 3-8 mm diam., terete to subterete, yellowish; midrib round-raised at base, becoming very narrowly and sharply raised (knife-like) to- ward the apex above, prominently and convexly raised with a sharp rib below; primary lateral veins 12-35 per side, departing midrib at 45-55? angle, + straight to the collective vein, prominently and sharply raised above, flat to raised below, very sharply raised above when dried; interprimary veins drying almost as conspicuous as primary lateral veins, sharply raised above; tertiary veins obscure, drying prominulous; collective vein arising from the base, scarcely visible above, flat below, less prominent than primary lateral veins when dried, —11 mm from margin. /nflorescences spreading- erect, shorter than leaves; peduncle 6.2-29(38) cm long, 2-6 mm diam., )1.9-2.7 x as long as petiole, pale green, tinged with violet-purple, subterete to angulate; spathe spreading to reflexed and recurled, subcoriaceous, green to green tinged with purple at margins and on veins, oblong-lan- ceolate, 2.5-7 cm long, 7-10 mm wide, broadest near the base, inserted at 65? angle on peduncle, acute to shortly acuminate at apex (the acumen apiculate), acute at base; stipe 7-10 mm long in front, 3-6 mm long in back; spadix maroon to purple, becoming brown (B & K red-purple 2/7.5), cylindroid, sometimes slightly tapered at apex, 3- 8 cm long, 3-6 mm diam. midway; flowers 4-lobed, 2.1-2.4 mm in both directions, the sides jaggedly sigmoid; 3—5 flowers visible in principal spiral, 6– 8 in alternate spiral; tepals matte, weakly papillate, with droplets; lateral tepals 1-1.3 mm wide, the inner margins straight to convex, slightly turned up against the pistil, the outer margins irregularly 3-4-sided; pistils raised, green; stigma ellipsoid; stamens emerging in a scattered pattern from the base of the spadix; filaments transparent, thick, exserted, soon retracting, holding anthers at level of the tepals, ca. 0.3 mm long, 0.4 mm wide; anthers yellowish heavily tinged with maroon, 0.6 mm long, 0.9 mm wide, inclined over and obscuring the pistil; thecae oblong, not divaricate; pollen or- ange, fading to tan. /nfructescence with spathe persisting; spadix 6 cm long, 1.3 cm diam.; berries purple, obovoid, truncate at base, rounded and somewhat mammilliform at apex, 6–6.8 mm long, 3.7-4.5 mm diam.; seeds 2 per berry, brownish, oblong, flattened, rounded at both ends, 3.2-3.5 mm long, 2.2-2.5 mm diam., 1.5-1.7 mm thick, enveloped by gelatinous, translucent, amber sub- stance. A member of series Multinervia, Anthurium pallatangense is endemic to Ecuador in Carchi, Pichincha, Cotopaxi, and Chimborazo provinces at (1,710) 2,000 to 3,000 montane dry, premontane dry, and montane moist forest life zones. m. It occurs in lower This species is distinguished by its broadly elliptic leaf blades with the primary lateral veins drying sharply raised (knife-edge-like) and a collective vein arising more or less from the base. In these char- acteristics it is very similar to A. penningtonil, probably its closest ally. Were it not for their geographical separation (either side of the Andes), А. penningtonii might be relegated to subspecific or varietal status. Anthurium penningtonii differs, however, in its green spadix (sometimes tinged pur- ple or brown), and occurs in wetter life zones as well. ECUADOR. CARCHÍ: 7.8 mi. SE of Maldonado, rd. to Tulcán, 2,400 m, 0%53'N, 78*5'W, Thompson & Raw- lins 888 (СМ); Maldonado-Tulcán, Km 205, 2,400- 2,600 m, Werling & Leth-Nissen 362 (QCA). CHIM- BORAZO: Pallatanga, Sodiro 24 (B). COTOPAXI: Angamar- ca, ca. 3,000 m, Sodiro s.n. (B). PICHINCHA: Aloág-Santo Domingo, San Ignacio, 2,000 m, Sparre 14687, 17742 (5); Machachi-Santo Domingo, 40.8 km E of Alluriquin, 2,050 m, 0°26' S, 78*40' W, Thompson & Rawlins 1094 (CM); Tandayapa-Mindo, 15.1 km beyond Tandayapa, 400 m, Croat 49374 (MO, QCA); Nono-Nanegal, N at Quito, 11-12 km ae of Nono, 2,200-2,270 m, Croat 38837 (МО); Nanegal, Sodiro s.n. (В); Mindo-Tanda- yapa, 10 km SW of Тапдауара, 2,200 m, 0°00'N, 18*32'40"W, Croat & Rodríguez 61545 (МО); Chiri- Volume 78, Number 3 1991 Croat 697 Anthurium sect. Pachyneurium boga Road, Chiriboga-Santo Domingo de Los Colorados, El Volante, 1,900 m, Asplund 17418 (Sy; Quito-Santo Domingo de Los Colorados, 19 km S of San Juan, 15 km NE of Chiriboga, 1,710 m, 0%17'S, 78?43'W, Croat 50621 (MO); Rio Alambi, Nono- Tandayapa, km 43-45, 2,000-2,500 m, Sparre 17016 (5); Quito- Puerto Quito, 13 km NW of Nono, 2,225 m, Luteyn & Lebrón-Luteyn 6523 (NY); Volcán Pichincha, W side, Tandayapa-Min- do, 2,300 m, 0?2'S, 78%42'W, Croat 50256 (MO, RSA, Anthurium paraguayense Engl. Bot. Jahrb. 1898. Syst. 25: 361. a. Anthurium paraguayense var. paraguay- ense. TYPE: Paraguay. Central: Asunción, Co- lonia Elisa, Lindman А1823 (holotype, B; isotype, S). Figures 226, 227, 231. Anthurium rodrigoi A. Hawkes, و‎ 3: 27. 1948. TYPE: Argentina. Chaco: Tap , Enrique Urien, Rodrigo 2690 (holotype, NY). Anthurium rusticum . Br. ex Engl., Pflanzenr. J 23B(Heft 21): 82. 1905. TYPE: Bolivia. La Paz Yungas, Coroico, Bang 2479 (holotype, NY; iso- types, GH, MO). Terrestrial or epiphytic; stem to 40 cm long, (1)5-6 cm diam.; roots numerous, dense, descend- ing or ascending, green or tan, smooth to pubes- cent, blunt, longitudinally fissured when dried, 3- 10 cm long, m diam.; cataphylls membra- nous to subcoriaceous, 3-9.5 cm long, acute and sharply 1-ribbed at apex (apiculate to 1 mm), pale green, drying dark brown to tan (B & K yellow 6/5), persisting as a reticulum of fibers at base. Leaves erect to spreading; petioles 3-28.5 cm long, 6-10 mm diam., D-shaped to sometimes nearly quadrangular, shallowly sulcate with a me- dial rib adaxially, the margins acutely raised, blunt- ly to sharply 1—3-ribbed abaxially, the surface green sometimes pale-speckled; geniculum thicker and paler than petiole, 0.7-2.5 cm long; sheath 0.9- 2.5 ст long; blades subcoriaceous, oblanceolate to oblong-oblanceolate, long-acuminate at apex (the acumen inrolled to flat), acute to obtuse to rounded or truncate at base, 22-100 cm long, 5-19.5 cm wide, broadest above the middle, the margins prom- inently undulate; both surfaces semiglossy to matte, dark to medium green above, paler or concolorous below; both surfaces drying usually greenish; mid- rib broadly raised at base with an acute medial rib, diminishing and becoming sunken toward the apex above, slightly raised when dried, convexly raised at base, becoming acute and then narrowly raised toward the apex above, slightly raised when dried; primary lateral veins 4 - 1U per side, departing mid- rib at (25)30- 70? angle, arcuate-ascending to the margin or loop-connecting to the primary lateral vein above it, raised above near the midrib, weakly sunken at the margin, raised below; interprimary veins obscure above, visible below; collective vein arising from near the apex, weakly sunken above, bluntly raised below, prominulous when dried, 5- 9 mm from margin. Inflorescences erect or spread- ing, shorter than leaves; peduncle 14—65 cm long, 8-9 mm diam., ca. 2-6X as long as petiole, dark green, terete to subterete; spathe spreading to re- flexed with the margins rolled under, coriaceous, dark green tinged with purple (B & K yellow-green 4/ 1.5), lanceolate to ovate-lanceolate, (2)6–15 cm long, (0.7)1.7-3 cm wide, broadest in the lower third, inserted at 40—50? angle on peduncle, acute to acuminate at apex (the acumen inrolled), acute to decurrent at base, minutely verruculose on inner surface; spadix green tinged with purple or gray- brown (B & K yellow 5/7.5), tapered, (2.4)5-15 cm long, 4-15 mm diam. near base, ca. 3-8 mm diam. near apex; flowers 4-lobed to rhombic, (1.7)2.5-4 mm long, (2.2)2-4 mm wide, the sides jaggedly sigmoid; 5-8 flowers visible in principal spiral, 7—10 in alternate spiral; tepals matte, some- times appearing punctate and papillate; lateral te- pals 0.9-1.5 mm wide, the inner margins convex or straight, the outer margins smoothly 3-4-sided; pistils not emergent or scarcely emergent, green tinged with purple; stigma ellipsoid to slit-like, 0.2— 0.5 mm long; stamens emerging in a regular se- quence, emerging shortly above the tepals, the laterals preceding the alternates by 10-12(21) spi- rals, the 3rd stamen preceding the 4th by 3-4 spirals; filaments translucent, with raphides, 1–1.5 mm long; anthers yellowish to red-purple (B & K red-purple 7/5), 0.5-0.7 mm long, 0.8-0.9 mm wide, inclined over and obscuring the pistil; thecae oblong-obovoid to oblong-ellipsoid, slightly or not at all divaricate; pollen pale yellow to cream, scent- ed like grass. Infructescence spreading-pendent; = persisting; spadix 17.5-20 cm long, to 3.5 m diam.; berries reddish to maroon (B & K red- le 2/7.5), obovoid, rounded at apex, 5-10 mm long, 3-6 mm diam.; pericarp moderately thickened, with raphide cells at base; mesocarp translucent to milky, thick-gelatinous; seeds 2 per berry, white, streaked with reddish or maroon, ovoid, flattened, 5-6.2 mm long, 2.5-3.4 mm diam., 1-2 mm thick, with a gelatinous appendage at both ends. Anthurium paraguayense is a common terres- trial, occasionally epiphytic species of southern Brazil (Mato Grosso Sul), Bolivia, Paraguay, and northern Argentina. It is mostly reported from dry 698 Annals of the Missouri Botanical Garden areas, on steep slopes and sandy or rocky soils at 150 to 1,700 m. In Bolivia it occurs in subtropical moist and lower montane subtropical wet forest life zones. This species is distinguished by its long, narrow blade with undulate margins and by its petiole that is bluntly 1-3-ribbed abaxially and shallowly sul- cate adaxially, with a prominent medial rib. Also characteristic is the peduncle, which is usually 2- 6 times longer than the petiole, and its persistent, coriaceous spathe, which is verruculose on the in- ner surface, appearing to have minute grains of sand spread across it. Other diagnostic features include the presence of raphides in both vegetative and floral structures (including the blade, petiole, peduncle, spathe, filaments, tepals, berry, and seed coats). Anthurium paraguayense var. paraguayense is similar to var. coroicoanum, with which it may occur, but has a proportionally longer peduncle compared to the petiole, and a persistent spathe that appears rather verruculose adaxially. Both taxa are represented by an unusually high pro- portion of fruiting specimens in herbaria. The var. coroicoanum is distinguished by its broadly elliptic leaf blades with reflexed basal primary lateral veins, e var. paraguayense has lanceolate to oblan- ceolate blades lacking any reflexed venation. The two taxa occur in similar habitats and elevations in close geographic proximity. A noteworthy collection is Beck 6996, from the Department of Beni, Bolivia, which has the low- ermost primary lateral veins spreading at a ca. 90° angle from the midrib, but otherwise demonstrates all the characteristics of the typical variety. ARGENTINA. CHACO: Tapenanga, Enrique Urien, Ro- drigo 2690 ауе Уефа, Меуег 2406 (СН). n Ignacio, San Ignacio, Schinini е Bang 1660 (Е), 2581 (МУ); Кизђу 2479 (Е, NY, US). BENI: Prov. Ва ујап, Rio Yacuma, Espiritu, 200 m, Beck 3296 (B, K, MO, TEX); Serrania de Pilón Lajas, 75 km from La Paz, San Borja, 950 m, Beck 6996 (MO). L4 PAZ: Canamina, 1,500 m, Rusby 375 (NY); Prov. Larecaja, Caranavi- "Guanay, 27.8 km N of a 865 m, 15?33'S, 67?45'W, Croat 516464 (МО); Ргоу. Loayza, Chulumani-Circuata, ‚ Mi- quilla, 1,530 m, Besse et al. 1852 (SEL), Christenson 1533 (MO); Prov. Morillo, Valle de Zongo, along trail from end of rd. at Kahua Power Plant, 1,660–1,690 m, Croat 51406 (MO); Prov. Nor Yungas, Coroico, Bang 2479(GH, MO, NY); Coroico, Polo-Polo, 1,100 m, Buch- tien 3654, 4516 (US); Millaguaya, 1,300 m, Buchtien 4284 (NY, US); Chuspipata- Yolosa, 20 km NE of Chus- pipata, 10.2 km SW of Yolosa, 1,700 m, 16?12'S, 67°47'W, Solomon & Uehling 12252 (MO); Unduavi- Caranavi, 82.6 km NE of Unduavi, 1,300 m, Croat 51559 (MO); Rio Huarinilla, 4.5 km below Yolosa, then О km W on rd. up Río Huarinilla, 1,450 m, 16°12’S, 67°50'W, Solomon 8560 (MO); Prov. Sud Yungas, 22 km toward new bridge on Rio La Paz, Irupana, 1,100 m, Beck 2962 (MO). SANTA CRUZ: Concepción-San Ја- vier, Evrard 8483 (BR); Chavez, El Carmen, 40 km 5 of Concepción on road to Lomerio, 750 m, Killeen 1447 (MO); Rio Ichilo, Valle Grande, Vallecito, 1,300 m, Саг- denas 4006 (US). BRAZIL. MATO GROSSO SUL: near Bela Vista on road to Caracol, Hatschbach & Silva dd (MBM, MO). PARAGUAY. WITHOUT LOCALITY: Palmer : (05); Cerro San Bernardino, Chodat 345 (G); Cordillera de Los Altos, Fiebrig 348 (G, MO), Hassler 1170, 1503, 3264 (G); Cerro de Tabati, Schinini 21254 (CTES); Co. de Caacupe, Schinini 23882 (CTES). AMAMBAY: Parque Nacional Cerro Cora, Schinini & Bordas 20315 (CTES, K); Cerro Muralla, 1.5 km N of administración, Brunner et al. 977 (МО); 7 km SW of administración, paso Nande- dependencia, Schinini 5644 (CTES), Schinini & Bordas 21238, 21234, 21235 (CTES); Rio и te tas А: Rio Tebicuari, Kuntz 2 (NY). CENTRAL: Asuncion, Colonia Elisa, pda yo (B); Rio Paraguay, Villa Elisa, Pedersen 58 (BR), 3150 (GH, US); Cerro Yaguaron, Yaguaron, Arenas 953 (CTES). CORDILLERA: Caacupe, 25?20'S, 57°10'W, Hahn 2028 (МО); Cerro Tabati, Zardini & Degen 3688 (MO, PY), 3692 (С, MO, РҮ); Serrania de Tobati, Meseta Ybytu Silla, 297 m, 25?12'S, 57%07'W, Zardini 8283 (CTES, ЕСО, С, MO); Colonia Rosado, Tobati, 200 m, 25?16'S, 57907 МУ, Schinini & Bordas 24882 (CTES). REEMBUCÓ: Pilar, Schulz 7865 (CTES). PARAGUARÍ: Cerro Асаћау, 25954'5, 57°9'W, 370 m, Brunner 1249 (МО); ee Chodat 346 (G); Parque Nacional Ybycui, Duré 4 (MO). PRESIDENTE HAYES: Estancia La Golondrina, a Hayes, 24°55'$, 57*40'W, Hahn 698 (MO); Trans-Cha- co Hwy., Km 127, 150 m, Gentry et al. 51990 (MO). SAN PEDRO: Río Apa- -Aquidabán, Alto Paraguay, Fiebrig 4149 (С, К); Rio Tapiracuai, Alto Paraguay, 20-28°М, 59-63°W, Woolston 1074 (К, NY, U, US); Ditto. Lima, Jejui, Estancia “Carumbe,” Pederson 9417 (К). b. Anthurium paraguayense var. coroicoan- um Croat, var. nov. TYPE: Bolivia. La Paz: between Caranavi and Guanay, 28.1 km N of Caranavi, near bridge over Rio Coroico, 15?27'S, 67°50'W, 820 m, Croat 51704 (ho- lotype, MO 2825694; isotypes, B, К). Figures 228. 229, 242. a var. typicum per laminum ellipticum, nervis primarii lateralis reflexis. Terrestrial to epiphytic; stem to 20 cm long, 1- 4 cm diam.; roots dense, greenish, pubescent, short, thick and blunt at apex, 2-5 cm long, 4-5 mm Volume 78, Number 3 1991 Croat 699 Anthurium sect. Pachyneurium diam.; cataphylls subcoriaceous, lanceolate, un- ribbed, 2.5-10 cm long, acute at apex, green (B & K yellow-green 9/2.5), drying dark to medium brown (B & K yellow 3/2.5), persisting semi-in- tact, eventually as coarse linear fibers or as a reticulum of fibers. Leaves erect to spreading; pet- ioles 5.5-26 cm long, 4-5 mm diam., D-shaped to C-shaped, flattened with blunt to sharp margins adaxially, sometimes with a medial rib, rounded abaxially; geniculum paler than petiole, up to 11 cm remote from the base of the blade, 0.6-1 cm long; blades subcoriaceous, elliptic to ovate-elliptic, acute at apex, acute to prominently attenuate at base, 24—43 cm long, 8-25 cm wide, broadest at or near the middle, the margins broadly undulate; upper surface semiglossy to glossy, lower surface weakly glossy to semiglossy, both drying matte, brownish to brownish green; midrib sharply and acutely raised above, conspicuously paler than sur- face, convexly to obtusely raised at base below, becoming flat toward the apex, paler than surface; primary lateral veins 4—9 per side, departing midrib at 55-85” angle, weakly arcuate in the middle of the blade, + reflexed at base, prominently raised above, raised below; interprimary veins weakly vis- ible; tertiary veins visible; reticulate veins obscure; collective vein arising from about the middle of the blade, flat above and below, slightly darker than surface below, 1-7 mm from margin. Inflores- cences erect-spreading, shorter than leaves; pe- duncle 12.5-26 cm long, 4 mm diam. when fresh, 1-2 mm diam. when dried, 0.8-3.8 as long as petiole, green (B & K yellow-green 8/7.5), terete; spathe reflexed to recurled, subcoriaceous,medium green, sometimes tinged with purple at margins (B & K yellow-green 6/7.5), lanceolate, 3-5 cm long, 0.8-2 cm wide, broadest near the base, inserted at 30-60” angle on peduncle, acute to short-acu- minate at apex (the acumen inrolled, ca. 2 mm long), acute at base, the margins meeting at 45? angle; spadix greenish tinged with reddish to pale pink, also reported as becoming purplish post an- thesis, long-tapered to nearly cylindroid, erect, 4.5- 6.5 cm long, 4-8 mm diam. near base, 3-5 mm diam. near apex, broadest at the base; flowers rhombic to 4-lobed, 1.6-2.5 mm long, 1.8-2.2 mm wide, the sides jaggedly sigmoid; 7-9 flowers visible in principal spiral, (6)14—15 in alternate spiral; tepals semiglossy to matte, with droplets; lateral tepals mm wide, the inner margins straight to convex, the outer margins usually 2- sided (meeting at ca. 90° angle), occasionally 3- 4-sided; pistils weakly exserted, dark brownish pur- ple; stigma slit-like, 0.5-0.6 mm long; stamens emerging rapidly in a regular sequence from the base, held shortly above the tepals, the laterals preceding the alternates by 13-14 spirals, the 3rd stamen preceding the 4th by ca. 2 spirals; anthers orange to reddish purple, 0.4–0.5 mm long, 0.6– .8 mm wide, inclined over and obscuring the pistil; thecae oblong, 0.3-0.4 mm wide; pollen yellow fading to white, weakly fruity-scented. /nfructesc- ence with purple berries. Anthurium paraguayense var. coroicoanum is known only from the vicinity of the type locality in La Paz Department, Bolivia, at 700 to 900 m in subtropical moist and subtropical wet forest life zones, and from Cusco, Peru, at 800 m in a sub- tropical rainforest life zone. Anthurium paraguayense var. coroicoanum differs from the typical variety in its more or less elliptic rather than oblanceolate leaf blades, and by having the lowermost primary lateral veins di- rected at an almost 90? angle from the midrib to prominently reflexed, vs. prominently ascending in the typical variety. Another character distinguish- ing the two taxa is berry color: purple in var. coroicoanum, dark red in var. paraguayense. The new variety is named for the type locality, near the Rio Coroico. BOLIVIA. LA PAZ: Caranavi-Guanay, 27.8-28.1 km М of Caranavi, near bridge over Rio Coroico, 820-865 m, 15?27-33'S, 67%45-50'W, Croat 51657 (MO), 51704 (В, К, МО, NY). SANTA CRUZ: Prov. Ichilo, Parque Na- cional Amboro, 15 km SE up the Río Pitasama from Rio Surutú, 700 m, 17%44'S 63%40'W, Solomon & Urcullo 14159 (MO). PERU. cUzco: Prov. Paucartambo, Río Tono, N of Patria, Cosnipata Valley, 800 m, 13?07'S, 71%12'W, Wachter et al. 141 (F, MO). Anthurium pendulifolium N. E. Br., Gard. Chron. 36: 362. 1904. TYPE: unknown lo- cality, possibly Colombia, Kew Н 457 / 83 (ho- lotype, K). Figures 9, 230, 233, 234, 255. Epiphytic, rarely terrestrial; stem pendent, 20— 30 cm long, 1.5-3.5 cm diam.; leaf scars reported as conspicuous, ca. 2.8 cm wide; roots moderately dense, descending, brown, 6 mm diam.; cataphylls moderately coriaceous, broadly lanceolate, scarce- ly 1-ribbed, 4-19 cm long, acuminate at apex with subapical apiculum, green, drying brown (B & K yellow-red 9/2.5), persisting as coarse linear fibers, the apex remaining intact. Leaves spreading-pen- dent to pendent; petioles 27-65 cm long, 5-10 mm diam., terete to subterete or thicker than broad, slightly flattened adaxially, rounded abaxially, the surface minutely pale-speckled; geniculum paler and thicker than petiole, 1.5-3.5 cm long; sheath 2.5-12 cm; blades coriaceous to subcoriaceous, oblong to oblanceolate, acuminate at apex, acute 700 Annals of the Missouri Botanical Garden to rounded (rarely subcordate) at base, 64- 190(207) cm long, 16-32(42.5) cm wide, broadest above the middle, the margins weakly revolute, broadly and weakly undulate; upper surface glossy to semiglossy, light to medium green (B & K yellow- green 5/10), lower surface matte, paler; midrib convexly to obtusely raised, paler than surface above, obtusely angled below; primary lateral veins (8)1 1-15 per side, departing midrib at 45? angle, arcuate-ascending to the margin, sunken in shallow grooves above, paler than surface, raised below; interprimary veins few, obscure; tertiary veins ob- scure above, scarcely raised and darker than sur- face below; collective vein arising in the lower third to about the middle of the blade, less prominent than primary lateral veins, (4)7—14 mm from mar- gin. Inflorescences erect-spreading; peduncle 35- 52 cm long, 0.5-1 cm diam., 0.5-1.7X as long as petiole, terete; spathe reflexed and twisted, sub- coriaceous, pale yellowish green to creamy white, violet at margins and at base (B & K yellow-red 9/10), oblong-lanceolate, (9)20-30(60) cm long, 1.5-4.8 cm wide, broadest just above the base, acuminate at apex (the acumen inrolled), acute at base; stipe 8-12 mm long in front, 2-5 mm long in back; spadix reddish violet to reddish purple (B & К purple 2/7.5 to red-purple 7/5-10), tapered, curved, 19.5-38 cm long, 4-9 mm diam. midway, 2-4 mm diam. near apex; flowers weakly 4-lobed to rhombic, 2.5-3.5 mm long, 1.7-3.3 mm wide, the sides straight to sigmoid; 12-16 flowers visible in principal spiral, 8-27 in alternate spiral; tepals matte and weakly punctate; lateral tepals 1.4-2.2 mm wide, the inner margins broadly rounded to somewhat concave, the outer margins 3-5-sided; pistils glossy, weakly exserted, green; stigma brush- like; stamens emerging irregularly; filaments fleshy, translucent, minutely papillate, exserted 0.5 mm; anthers creamy white to yellow, 0.5 mm long, 0.6– 0.7 mm wide, inclined over the pistil; thecae ovoid- ellipsoid to pyriform, weakly divaricate; pollen yel- low fading to white. /nfructescence pendent to spreading-pendent, sometimes coiled; spathe with- ered; spadix 27-55 cm long, 1-2 cm diam., with berries scattered throughout; berries lavender, be- coming white at maturity, obovoid-oblong, rounded at apex, 10-11 mm long, 5-6 mm diam.; pericarp thickened; mesocarp mealy, white; seeds 1-2 per berry, brown, oblong, (5.5)7.3-7.8 mm long, 2.8- 3 mm diam., 1.3 mm thick. Anthurium pendulifolium ranges from Colom- bia (Amazonas and Putumayo) to northern Peru as an epiphyte at 90-550 m in tropical moist and tropical wet forest. Most collections are from the upper Rio Napo and Rio Aguarico in northern Ecuador (Napo), and also from around Iquitos in Loreto, Peru. It has also been collected in northern Amazonas, in Peru along the Rio Cenepa and the Rio Santiago, as well as along the lower Rio Pastaza in northwestern Loreto. It is to be expected throughout lowland Ecuador and northwestern Lo- reto in Peru. This species is distinguished by its oblong-ob- lanceolate leaf blades which are pendulous and dry pale green or yellowish green (the type specimen, of cultivated origin, has pale brown leaves, but this may be due to its age of nearly 100 years). Also distinctive are the moderately long petioles (up to V$ as long as the blades) and pale purple berries which become whitish in age. Anthurium pendu- lifolium may be confused with А. atropurpureum var. arenicola, with leaves drying the same green color and occurring in the same area. The latter differs from А. pendulifolium in its erect-spread- ing, rosulate habit, broadly oblanceolate leaf blades, shorter petioles, and conspicuously sunken tertiary veins (when fres Anthurium pendulifolium was originally de- scribed from cultivated material thought to be from Colombia. While no Colombian material closely matches the type, much of the material collected in Ecuador matches it well. Common names for this species include **yakiya sugkip”” (Huambisa pu Tus Santiago, Amazonas, Peru); *kankur nuka" (R tiago); ““meko ико” (Siona Indians, bbs Colombia); “ате” (Rio San Miguel, Putumayo, Colombia); and **bagre panga" (Rio Lagartacocha, Napo, Ecuador). The mashed, cooked leaves are used in a water bath by the Siona Indians (Putumayo, Colombia) to re- lieve bone aches and rheumatism. WITHOUT LOCALITY: ponis at Kew, Kew H 457/ 83 (К). COLOMBIA. AMAZONAS: Puerto Narino, Parque acional d 100 m, 3?45'S, 70°15'W, Vásquez 6 (MO). PUTUMAYO: ipn Vista, Piaguaje Santa Rosa de Sucumbios, А А 0: Laguna Grande, Pica Saladero de Dantas, Cuyabeno, 0%00'00”, 16*12'45"W, Jaramillo 6852 (MO, QCA); 265 m, Niel- sen 76272 (AAU, МО); Lago Agrio-Coca, 26 km N of Соса, 450 m, 0%29'S, 76?55'W, Croat 50420 (МО); Coca-Hollin, Huaticocha, 500 m, 0%45'S, 77929'W, Pa- lacios et al 35 543 је Рауапипо, Кезегуа ия “El Chun А ‚ 0°0'S, 77%01'W, Cerón lacios M 0, OCNE): ridge W of Cotapino, да to Rio Cotapino, 360 m ‚ 0940'5, 77°20'W, Whitmore 718 (Ky; trail to Shushufindi, WSW of San Pablo de Las Secoyas, 300 m, 0?15'S, 76?21'W, Brandbyge & Asan- za 32892 (AAU, МО); Lago Agrio-Puerto El Carmen de Putumayo, 15 km E of Tarapoa, 91.7 km E of Lago Agrio, 260 m, 0%08'N, 76?23'W, Croat 58579 (MO, QCA); 28 km W of Tarapoa, 48 km E of Lago Agrio, Volume 78, Number 3 1991 Croat 701 Anthurium sect. Pachyneurium 240 m, 0%06'N, 76?33'W, Croat к МО, ОСА, US, W); rd. to Agua Pungo, Tena, 1.5 km past bridge over Rio Mishahualli, 550 m, 0°59 S, 77°47'W, Croat 58825 (MO); каза Biológica Jatün Sacha, 8 km E of Mishahualli, 4 & Hur tado 4012 (MO); Cantón Aguarico, lagunas de Garza Cocha, 200 m, Cerón & Gallo 4953 (MO, QCNE); ai ton Tena, headwaters of eer Huambuno, 6 km Ahuano, 440 m, 1900'S, 77°40'W, Kohn 1155 (MO): Rio Aguarico, E of Lago ОА road їо Сере Ferry Cross- ing, 450 m, 0%02'N, 76%50'W, Croat 50427 (МО), 58513 (MO. P ‚ ОСА); Santa Cecilia, 200 m, 0°04'N, 76*58'W, E 13043 (5); Rio Lagartococha, Redondo Cocha, m, 0?35'S, 75?15'W, Lawesson et al. 44424 (AAU). PASTAZA: Rio Curaray, Lagunas Patoamo, 230 m, 1%30'S, 76°30'W, Palacios & Neill 7 62 (MO, NY, QAME). PERU. AMAZONAS: Río Cenepa, Croat 57196 (MO); Rio Santiago, La Poza, 400 m (180 m), Huashikat 150 (МО), behind com- , 3°50'5, 77°40'W, Tunqui ). LORETO: vic. Saas. Explorama Camp, 130 m, 3?28'S, 72%50'W, Croat 61760 (AMAZ, МО), 66057 (МО); Prov. Alto Amazonas, Rio Pastaza, pios payaco-Rimachi, 4?20'S, 76?40'W, Díaz et al. 12 (MO); Prov. Loreto, Pucayacu, native community of Shi m, Rimachi 6490 (IBE); Rio Maniti, eni (s кайн. 115 m, 3°42’S, x ие: & Jaramillo 1123 (MO); Rio Momón, tributary of Rio anay, Iquitos, Davidson & Jones 9802 (LAM), a 116 m, 72°55'W, 3?20'S, Vásquez & Jaramillo 8247 (MO); 11 km from Rio Nanay, near Iquitos, Quebrada Momoncillo, 200 m, 3°43’S, 73°20'W, Goat 51224 (COL, GH, M, M "d (IBE); Rio Tacshacuraray, tributary of Rio Маро, 2?40'S, 73°30'-74°20'W, Croat 20380 (MO). SAN MARTÍN: Tar- apoto- Yurimaguas, Km 54, Lamas, 350 m, 6%23'S, 76°18'W, Knapp 8262 (MO). Anthurium penningtonii Croat, sp. nov. TYPE: Ecuador. Napo: between Quito & Baeza, 5 of road above jct. of road to Baeza & to Lago Agrio, 1,800 m, Croat 58491 (holotype, MO 3183150; isotypes, QCA, VEN, RSA, K, US). Figures 235, 236, 256, 257. Planta terrestris aut raro epiphytica; internodia brevia, 1-3 cm diam.; cataphyllum lanceolatum, persistens semi- intactum mox fibris stramineis Raden, petiolus Jed 31 cm longus, 4-8 mm diam., subte lamina late elliptica, · raro oblanceolata, 35- 61 cm ‘longa, 10-19 cm lata; о җе pedunculus 21-52 spadix viridis, 2.5-5.5 cm longus, 3-5 mm diam. Terrestrial, rarely epiphytic; stem to 50 cm long, 1-3 cm diam.; roots dense, drying whitish to tan, moderately elongate; cataphylls thin, lanceolate, 6-13 cm long, acuminate at apex, drying reddish brown, persisting semi-intact, quickly weathering to straw-colored fibers. Leaves erect-spreading; petioles 11-31 cm long, 4-8 mm diam., subterete to triangular, flattened to broadly convex adaxially, sometimes with the margins sharp, rounded or sharply 1-ribbed bebes tmi slightly к and scarcely thicker than petiole, (0.6)1- long; sheath 1.5-3 cm long; blades coriaceous to bestia, broadly elliptic, rarely somewhat oblanceolate, shortly acuminate at apex, acute to narrowly acute at base, 35-61 cm long, 10-19 cm wide, broadest near the middle, the margins flat to slightly revolute; upper surface weakly glossy to semiglossy to glossy, dark to medium green, lower surface semiglossy, conspicuously paler; mid- rib convexly raised at base, becoming sharply acute toward the apex above, slightly paler than surface, acutely raised below; primary lateral veins 24-43 per side, departing midrib at 40—60° angle, straight to the collective vein, usually prominently and sharply raised above, flat to sunken below; inter- rimary veins almost as conspicuous as primar lateral veins; tertiary veins obscure, prominulous above and below on drying; reticulate veins not visible; collective vein arising from the base, equat as prominent as primary lateral veins, 10-1 from margin. /nflorescences erect; peduncle 21- 52 cm long, 3-4 mm diam., 1-2.2X as long as petiole, green to purple, terete, sometimes with many weakly to prominently raised striations; spathe reflexed or rarely spreading, subcoriaceous, green, usually tinged ы purple, oblong-lanceolate, 3.5— 7 ст long, 0.6– roadest near the base, acute to foh os acuminate at apex, acute at base; spadix green, usually tinged with purple to purplish brown, sessile or stipitate to 2 mm, scarcely tapered, erect, 2.5-5.5 cm long, 3-5 mm diam. midway, 3-4 mm diam. near apex; flowers square to rhombic, 1.4-2 mm long, 1.4-1.8 mm wide, the sides straight; 5-9 flowers visible in prin- ара! spiral, 4—6 in alternate spiral; tepals matte, minutely papillate; lateral tepals 1.1—1.2 mm wide, the inner margins rounded, the outer margins 2-sided; pistils emergent, glossy; stamens exserted; anthers pink-orange, 0.4— mm long, 0.5 mm wide; thecae shina EK slightly divaricate; pollen pale orange fading to white. /nfructescence erect; spathe persisting; spadix 8-10 cm long, 1.3- 1.7 cm diam., with berries scattered throughout; berries (immature) green, emergent ca. halfway, obovoid, shortly beaked at apex on drying, ca. 6.5 m wi e, mm long, ca. 4 mm diam.; pericarp with raphide cells; seeds 2 per berry, oblong-obovoid, flattened, beaked at apex, 4-4.5 mm long, 2.5-2.7 mm wide, 1.5-1.8 mm thick, with a gelatinous appendage at one end. 702 Annals of the Missouri Botanical Garden A member of series Multinervia, Anthurium penningtonii ranges from the Amazon slopes of Ecuador to northern Peru on both slopes of the Cordillera Oriental and the Cordillera Central at elevations of (800)1,000 to 2,500 m. The species is ecologically variable. In Ecuador it is known from premontane moist, premontane wet, lower montane moist, and lower montane wet forest life zones, while in Peru it is known from premontane rain and montane rain forest life zones. This species is distinguished by its broadly ellip- tic, green-drying leaf blades with sharply raised (knife-edge) major veins, and by its rather remote collective vein (1-1.5 cm from the margin). Anthurium penningtonii is not confused with any other species on the eastern side of the Andes, but А. pallatangense, from the Pacific slope of Ecuador, is very similar, differing mainly in its dark purple to maroon spadix. It occurs in drier life zones as well. Some material from Morona-Santiago (Stein 2826, Madison et al. 3457, Madison 2605) has sharply triangular petioles. Considering the varia- tion in the vicinity of Baeza alone (petioles terete to flattened adaxially and sometimes sharply l -ribbed abaxially, sometimes with the margin sharp on one side), these collections are best placed in A. penningtonii and agree well in other characters. Another collection from Morona-Santiago (Stey- ermark 53610) differs in having the leaf blades attenuate at the base, rather than acute to narrowly acute. More collections from southern Ecuador and northern Peru are needed to determine the range of variability in this taxon. The new species is named in honor of T. D. Pennington, a member of the 1960 Oxford Uni- versity expedition to Ecuador and one of the first collectors of this species. E :UADOR. MORONA-SANTIAGO: Limón (General Plaza)- Macas, 8 km N of Limon, 1,100 m, 2%57'S, 78°25'W Pts 2826 (B. K, MO, QCNE, US); 7-8 km N of Cua: laquiza on road to Indanza, 1,450 m, Harling & An- dersson 24199 (GB, МО); Tumbez- Tucumbatz, Km 20 on road Gualaquiza-Indanza, 1,600 m, Harling & An- de son 24. 360 (GB); Cordillera de Cutucu, Rio Chihuasi, 5 km ,000 m, Madison 2605 . 2%46'S, 78?06'W, 3502 (SEL) Rio Tintas, Cam- panas- Arenillas, " leagues SE of El Pan, 2,195 m Steyermark 53610 (US, NY). маро: Lago Agrio- Baeza, 1,940-2,000 m, 0?15'S, 77945'W, Oellgaard et al. 35767 (AAU); Quito- Baeza 5 of rd. above jet. of rds. to Baeza & wo Agrio, 1,800 m, Croat 58491 (К, MO, QCA, RSA, US, VEN); Lago A grio- е 32.8 km E of Baeza turn- ds 1 ,940 m, 0?17'S, 46'W, Croat 58730 (B, CM, МО, ОСА); Baeza jct. ES бы to Lago Agrio, 2,500 m, d 49439 (MO, NY, QCA); Baeza- e 37.7 km 3aeza, 59 km N of Archidona: 1,700 m, 0?36'S, 77°51'W, Croat 58781 (CAS, MO, ОСА); Cantón Ar- chidona, Carretera Hollin- Loreto, Km 50, Guagua Su- maco, 1,000 m, Cerón & Hurtado 6565 (MO); Km 25, Challua Yacu, 1,200 m, Cerón & Hurtado 6455 (MO); Baeza-Lago Agrio, 19.7 km NE of El Chaco, 39 km NE of jct. of road to Tena, 1,750 m, Croat 58529 (MO, QCA); Rio Oyacachi, 1,620 m, 0%20'S, 77*55'W, Croat 50297 (M, MO); Km 142, 1,800 m, Croat 49454 (IBE, O, NY); Cerro Antisana, 1 mi. NE of Borja, NE of Baeza, 1,850 m, 0%25'S, 77*50'W, Grubb et al. 1267c (K) Cordillera de Guacamayos, S slope, Baeza- Tena, above Jondachi, 1,500 m, Harling & Andersson 16342 (AAU, МО); Volcán MARIO Cotundo-Coca, 15 km E of Baeza- Tena road, 1,300 m, 0940'5, 77%40'W, Pa- lacios & Neill 1559 (MO). TUNGURAHUA: Rio Negro, 1,200 m, 1°24’S, 78*13'W, Harling & Andersson 17240 Уш ZAMORA-CHINCHIPE: La Чан Yacuambi, 1 km N of Chapintza road, 1,100 m, Harling & Andersson 23868 (GB). PERU. AMAZONAS: Serrania de Bagua, km E of La Peca, 1,850-1,900 m, Gentry et al. 22969 (MO); Prov. Bagua, 20 km on trail E of La Peca, 2,190 m, Barbour 2735, 2746 (MO). CAJAMARCA: Cutervo, San n yobamba-Chachapoyas, Km 380, E of Río Nieva, 1,750 m, 5°44'5, 77*32'W, Croat 58215 (MO, USM); Rioja- Pomacocha, Km 291, Venceremos, 1,850 m, 5?45'S, 77°40'W, Gentry et al. 45504 (MO). Anthurium plowmanii Croat, Candollea 42(2): 811-813. 1987. TYPE: Brazil. Amazonas: Mpo. Manaus, Chacaras de Taruma, 16 km NW of Manaus, 60 m, 2%51'S, 59%50'W, Croat & Nelson 53563 (holotype, INPA; isotypes, B, С, К, IBE, MO, NY, RSA, US). Figures 237- 239, 240, 258. Epiphytic or epilithic; stem 10-30 cm long, 2.5-6 cm diam.; leaf scars ca. 4 cm wide; roots dense, ascending or descending, whitish green, shortly pubescent, short or elongate, tapered, 4— 6 mm diam.; cataphylls coriaceous, 5-20(27) cm long, obtuse to acute at apex, drying reddish brown (B & K yellow-red 4/5), persisting as a reticulum of fibers, the epidermis drying thin and flaking off, revealing the thin, reddish brown, hairlike fibers beneath. Leaves erect to spreading, reported to 2 m long; petioles (7.5)10-40(50) cm long, 4-20 mm diam., U- to C-shaped, den and obtusely sulcate adaxially with the margins blunt, rounded to rarely 1-2-ribbed abaxially, the surface pale- speckled; geniculum thicker than petiole, becoming fissured transversely with age, 0.5-2 cm long; sheath ca. 3 cm long; blades coriaceous, obovate to oblanceolate or elliptic to broadly elliptic, acu- minate to obtuse at apex (the acumen inrolle downturned), acute to attenuate (rarely long.at- tenuate), or obtuse at base, (24)40-125(200) cm Volume 78, Number 3 1991 roat 703 Anthurium sect. Pachyneurium long, (6)20-55(66) cm wide, broadest at or near the middle, the margins prominently and broadly undulate; upper surface matte to semiglossy or weakly glossy, dark green, lower surface matte to weakly glossy, concolorous with the upper surface or slightly paler; midrib flat to angular at base, becoming prominently and convexly raised (almost round-raised, with narrow sulcus along each side) and narrowly angular toward the apex above, prominently thicker than broad to weakly 2-3- ribbed at base, becoming prominently and convexly raised toward the apex below; primary lateral veins (4)7-15 per side, departing midrib at (30)45-65° angle, ascending + straight to the margin or slight- ly arcuate-ascending, prominently and convexly raised above and below, more so below; interpri- mary veins obscure; collective vein arising from near the apex or absent, if present flat to weakly sunken above, flat to weakly raised below, 4-6 mm from margin. /nflorescences erect to erect-spread- ing; peduncle (2)6-32 cm long, 3-9 mm diam., 0.2-0.3 X as long as petiole, plain green or heavily to slightly tinged with red, purple, or maroon, terete to ribbed abaxially; spathe semi-erect to spreading to reflexed or rolled up, subcoriaceous to coria- ceous, violet purple or green tinged with purple, lanceolate to broadly lanceolate, (4)5-26(29) cm long, (0.5)1-3.5 cm wide, broadest near the base, inserted at 40—55° angle on peduncle, acute to long-acuminate at apex (the acumen inrolled), ob- tuse to acute or sometimes slightly decurrent at base, the margins meeting unequally, thick, rolling down at edges; stipe 1-20 mm long in front, 1–3 mm long in back; spadix green to dull maroon to green tinged with violet to violet-purple, tapered to cylindroid, longer than peduncle, (7)10-46 cm long, 10-20 mm diam. near base, 3-10 mm diam. near apex, broadest at the base; flowers square to rhombic, 1.5-2.5 mm long, 1.1-2.5 mm wide, the sides + straight to jaggedly sigmoid; (8)10—16(18) flowers visible in principal spiral, 6-10 in alternate spiral; tepals matte, minutely and densely papillate; lateral tepals 0.6–1.2 mm wide, the inner margins straight to convex, the outer margins 2-sided; pistils emergent, weakly raised before the stamens emerge, reddish to dark purple; stigma linear, 0.2-0.6 long; stamens emerging in a regular sequence from the base, the laterals preceding the alternates by ca. 5 spirals; filaments holding anthers above the tepals; anthers orange, са. 0.6-0.8 mm long, 0.8- 0.9 mm wide, inclined over and obscuring the pistil; thecae ellipsoid to obovoid, not divaricate; pollen yellow to orange fading to cream, yeasty-scented. Infructescence erect to pendent; spathe deciduous, leaving a coarse scar; spadix 5-25 cm long, 1-3 cm diam., the apical Уз withered, fruits developing in the lower 25 only; berries red, oblong-obovoid, — rounded at apex (the stigma button-shaped), m long, 3-6 mm diam.; mesocarp juicy, sweet, pum aue seeds 1-2 per berry, brown, densely speckled with translucent dots in rehy- drated specimens, oblong to ovoid, 5-6 mm long, 2-3 mm diam., 1.5-2 mm thick, with sticky ap- pendage at the apex (forked in rehydrated speci- mens). Anthurium plowmanii ranges from Brazil (Acre, Amazonas, Mato Grosso, and Rondónia) to Bolivia (La Paz, Pando, and Santa Cruz) Paraguay (Amambay) and Peru (San Martin) at 50-900 m, typically in the dry forest life zones of Peru and the varzea and terra firme habitats of Brazil. A large epiphytic or epilithic species, А. plow- manii is unusual in sect. Pachyneurium by virtue of its spadix, which is longer than (or rarely as long as) the peduncle. This feature is shared only with А. solomonii (see discussion under the latter for distinctions from А. plowmanii). Also distinc- tive are the C- to U-shaped petioles with blunt adaxial margins, otherwise of rare occurrence in the section. Typical of A. plowmanii is the gradual disintegration of the distal (presumably unpollinat- ed) portion of the spadix in the fruiting stage with, at most, only the weathered axis remaining. Fruit- ing spadices are thus regularly much shorter than flowering spadices, quite opposite the normal sit- uation in the section Some atypical material of cultivated origin is included here under А. plowmanii, such as Casari 165, which differs in its unusually coriaceous leaf blades that are oblong-attenuate (vs. usually acute to attenuate) at the bases. The infructescence and other features are, however, typical. Croat 53701, received from and cultivated at the sitio of Roberto Burle-Marx in Rio de Janeiro, Brazil, has lanceolate (vs. obovate to elliptic) leaf blades with unusually prominent venation and notably olive-green col- oration on drying. Photos of this collection taken in Brazil closely resemble typical material, how- ever. BOLIVIA. WITHOUT ye gd bipes by F. Fuchs of d Florida, Croat 2 (СН, MO, UCLA). COCHABAMBA: Rio Grande, Town 7192 (NY). LA PAZ: Prov. i Solomon "169 40 (MO); Luisita, 180 m, 13%05'S, 67°15'W, Haase 266 (MO). PANDO: Lago Bay, black water lake formed by Río Arroyo, upstream from junction with Río Manuripi, 11%57'S, 68940 W, Sperling & King, 6564 (МО); Río Madeira, Cachoeiras Miseri- cordia- Madeira, Prance et al. 6598 (NY); 4 km above Ађипа, Prance et al. 6271 weary Rio Madre de Dios, Federico Roman, Loma Alta, 110 m. Solomon 17068 (MO); Manuripi, upstream from ве 220 m, 10955'5, 704 Annals of the Missouri Botanical Garden 66*8'W, Daly et al. 2042 (NY). SANTA CRUZ: Weddell 3518 (P); Andres Ibanez, 12 km E of Santa Cruz--Cotoca, 375 m, 17%46-47'S, 63°04'\/, Nee 34009 (NY); ca. 40 km S of Asunción dos Guarayos, Nufles de Chaves, 15955'5, 63°05'W, Hopkins et al. 218 (МО); Serranía de Santiago, N slope, Chiquito, 10 km ENE of Santiago de a 900 m, 18°20'5, 59?28'W, Daly et al. 2243 (MO). BRAZIL. ACRE: Mpo. Sena Мадигета. Rio Тасо, Cid & Nelson 2757 (INPA). AMAZONAS: cultivated, INPA cam- pus, Manaus, Nelson 1335 (MO); N of Coreiro, 10 km و‎ from Manaus, Leppard 1645 (K); Solimöes, mia, Kuhlman 1191 (RB); Lago do Janauari, Manaus, Coelho 638 (INPA); Manaus, 50 m, Croat 53563 (B, G, IBE, INPA, K, MO, NY, RSA, TEX, US); Rio Acre, behind Santa Maria, W bank of Rio Acre, Boca do Acre, Prance et al. 2361 (INPA), 2371 (INPA, NY, US); Rio Madeira basin; Humayta, near Livramento, on Rio Liv- ramento, Kruko 6767 (NY). МАТО GROSSO: Апрџапа, Cidade Humboldt, 10°12'S, 59°21'W, Rodrigues 9816 (INPA); Chapada dos Guimaraes, Hutchison Е Porto Espiridiao-Porto Velho, 286 km NW of Esperidiào, on BR-364-MT, — aep 8602 (MO, UEC), Urucum, Corumba, 19%00'S, 57, нај“ keler & Graziela 447 (RB); Serra то Fran 1598, 60°W, Windisch 1514 (К). вю DE ais a tivated, Rio de Janeiro, base of Pedra Bonita, Niemeyer 14, Casari 165 (GUA); cultivated by Bur Marx, San Carlos, Croat 53701, 537 10 (MO). DONE Rio Madeira, Riberáo, Cachoeira Misericordia, Prance et al. 6721(INPA, NY, US); Abunà-Penha Colorado, Prance et al. 8725 (INPA, NY, US); Mpo. Costa Marques, ca. 5 km NW of Costa Marques, 150 m, 12°25'5, 64°14' м. Nee 34548 (NY). PARAGUAY. АМАМВАҮ: Cerro Chanchi 22°26'5, 56°3'W, Schinini & Bordas 20398 (К); Сы. Memby, cultivated in Asunción, е - Schinini 28533 (K); Sierra de Amambay, Mon ncas, Pie- dra Esperanza, Hassler 10640 (G); Dus Nacional ји ro Cora, 300 m, 22*39'S, 56°03’W, Solomon et al. 698 (MO); summit of Cerro Muralla, Casas 3955 (NY). Peau. WITHOUT LOCALITY: cultivated at SEL, #81-76-10, and at MO, Plowman 6011 B (МО, NY, TEX). SAN MARTÍN: Tarapoto-Juanjui, Km 24-25, 300- 500 m, 6?40'S 76°20'W, Croat 50987 (F, MO), Plowman 6044 (GH); Río Huallaga, 29-31 km S at Km 15, Cunumbigue village, 350 m, 6°23’S, 76°39’ W, тош 51094 (CM, F, К, МО, RSA). Anthurium pranceanum Croat, sp. nov. TYPE: razil. Acre: Rio Moa between Cachoeira Grande and Serra de Moa village, Prance et al. 12640 (holotype, INPA; isotypes, F, K, NY, U, US). Figure 241. Planta terrestris; internodia brevia, 1-1.5 cm veu cataphyllum lanceolatum, persistens semi-intactum; pe- tiolus 35-65 ст longus, 4-8 mm diam „ Subteres, ы sulcatus; lamina ova to-elliptica ad ovata, 40-70 cm longa, 7-26 cm lata; geniculum 4-20 cm remotum; pedun- culus 24- чо; m longus; spatha ni e violaceus, c ca. , 1.2 em lata; age 1-10 cm longus; spadix 4 cm lon a ca. 5 cm longus, 5-6 mm Description based on dried material only. Ter- restrial; stem 1-1.5 cm diam.; roots densely and nely pubescent, ca. 2-4 mm diam.; cataphylls subcoriaceous, lanceolate, ca. 10 cm long, acute at apex, reddish brown, persisting semi-intact. Leaves with petioles 35-65 cm long, 4-8 mm diam., subterete, sulcate adaxially, rounded abax- ially; geniculum slightly darker than petiole, ca. 1 cm long, appearing remote from base of blade by 4-20 cm; blades moderately coriaceous, ovate- elliptic to ovate, shortly acuminate at apex, broadly rounded then abruptly attenuate at base, 40-70 cm long, 17-26 cm wide, broadest at or below the middle, lower surface dark glandular-punctate; both surfaces green to yellowish green; midrib convexly raised above, acutely raised below; primary lateral veins 6-9 per side, departing midrib at 50-60? angle, + straight, becoming arcuate near the mar- gin, apparently convexly raised above and below; tertiary veins weakly raised; collective vein arising from about the middle of the blade, apparently sunken above, raised below, less prominent than primary lateral veins, 7-24 mm from margin. /n- florescences erect, shorter than leaves; peduncle 24—40 cm long; ca. 2-5 mm diam., about half as long as petiole, terete; spathe recurled, membra- nous, violet, lanceolate, ca. 4 cm long, ca. 1.2 cm wide, broadest near the base, narrowly acute at base; stipe 15-28 mm long in front, 1-10 mm long in back; spadix violet, weakly tapered, ca. 7.7 cm long, 5-6 mm diam. near base, ca. 3 mm diam. near apex; flowers rhombic, ca. 2 mm in both directions, the sides + straight to smoothly sigmoid; 4—6 flowers visible in either spiral; tepals smooth when dried; lateral tepals 1 mm wide, the inner margins straight, the outer margins 2-sided. /n- fructescence not seen. Anthurium pranceanum is known only from the type collection made near the Rio Moa in Acre, Brazil, below 300 m. This species is distinguished by its ovate to ovate- elliptic leaf blades which are dark glandular-punc- tate below and have the geniculum situated 4-20 cm below the base of the blade Anthurium pranceanum is apparently most closely related to A. krukovii, which also has a remote geniculum but lacks dark glandular punc- tations on the lower surface of the leaf blade. It also bears some resemblance to А. bonplandii subsp. bonplandii, which occurs further north in the Amazon basin and often has similar dark glan- dular punctations. The latter differs in its propor- tionately narrower leaf blades which are acute to Volume 78, Number 3 1991 Croat 705 Anthurium sect. Pachyneurium attenuate at the base and do not have a remote geniculum. The species is named in honor of G. T. Prance, one of the foremost explorers of Brazilian Ama- zonia, who has collected numerous Brazilian Ara- ceae, including the type of Anthurium prancean- um. BRAZIL. ACRE: Rio Moa, Cachoeira Grande-Serra da Moa village. Prance et al. 12640 (F, INPA, NY, U, US). Anthurium prolatum Croat & R. A. Baker, Brenesia 16(Suppl. 1): 72. 1979. TYPE: Costa Rica. Heredia: Istarú Farm, Tirimbina, Sara- piqui, 220 m, Lent 2140 (holotype, F-1724125). Figures 242-245, 259. Epiphytic; stem short, 1-2.5 cm diam.; roots few, descending, раје greenish, + smooth, slender, 3-5 mm diam.; cataphylls moderately coriaceous, narrowly lanceolate, 7.5-36 cm long, acute at apex with subapical apiculum, medium green, drying tan, persisting as fine linear fibers. Leaves spread- ing to pendent; petioles 1-35 cm long, 3-6 mm diam., terete to subterete, weakly flattened to flat- tened or rounded adaxially, rounded abaxially, the surface weakly pale-speckled; geniculum thicker and paler than petiole, 0.5-1 cm long; blades sub- coriaceous to moderately coriaceous, strap-shaped to narrowly oblong-elliptic, long-acuminate at apex (the acumen weakly apiculate), narrowly acute to obtuse to narrowly rounded at base, (35)83-142 cm long, 3.5-10 cm wide, broadest at or near the middle, the margins weakly undulate; upper sur- face semiglossy, dark green, lower surface semi- glossy to glossy, paler; midrib above acutely raised at base, becoming weakly sunken toward the apex, below prominently higher than broad at base, be- coming convexly raised toward the apex and paler than surface; primary lateral veins more than 15 per side, departing midrib at ca. 55? angle, + straight, obscure to weakly sunken above, some- what raised and darker than surface below; inter- primary veins numerous, almost as conspicuous as primary lateral veins; tertiary veins obscure above, somewhat darker than surface below; reticulate veins prominulous above and below; collective vein arising from near the base, sunken above, weakly raised and darker than surface below, more prom- inent than primary lateral veins, 3-4 mm from margin. Inflorescences with peduncle to 39 cm long, ca. 9 mm diam., slightly longer than petiole, several-ribbed on drying; spathe spreading, sub- coriaceous, pale green, linear-lanceolate, to 29 cm long, ca. 1.5 cm wide, broadest near the base, inserted at 70° angle on peduncle, acuminate at apex (the acumen minutely inrolled), acute at base; spadix brown (fide labels) probably greenish purple, sessile, very narrowly уйи, id tapered at apex, to 60 cm long, m. near base, ca. 4 mm diam. near apex; ^s (dry) rhombic, (2.7)3.4–4.5 mm long, 1.8-2.2 mm wide, the sides jaggedly sigmoid; 7-8 flowers visible in principal spiral, 5-6 in alternate spiral; tepals epunctate, smooth; lateral tepals 2 margins convex; pistils rectangular, green; stigma oblong, ca. 0.7 mm long; filaments retracting on drying. Infructescence to more than 30 cm long, probably with orange berries. Anthurium prolatum is known from western Panama and Costa Rica (in Alajuela, Limón, He- redia, and Cartago provinces) in tropical wet forest at 350-1,000 m This species can be distinguished by the unusual, very long and narrow leaves, cataphylls, spadix, and spathe. It is most easily confused with А. protensum, differing in its spadix, which is more than twice as long as any observed for А. proten- sum; its spathe, which is considerably narrower and lacking the purple markings characteristic of A. protensum; and its flowers, which are twice as long as those of А. protensum. In addition, the much more common 4. protensum is most frequent at about 1,800 m CosTA RICA. ALAJUELA: ba i MEE d Zarcas, 8.5 km NE of Villa Quesada, "L У ан: 46979 (МО). CARTAGO: 1.5 m 10. 2 mi. NE of junction at Paraiso, Gl 47087 (MO). HEREDIA: Tirim- bina, Sarapiqui, Istarú Farm, Lent 2140 (F). LIMÓN: Lago Dabagri, W coast, toward Rio Llei, 1,000 m, 9°38'N, 83°18'W, Gómez et al. 23155 (В, K, MO); Rio Tor Amarillo, 6 km SW of Guapiles, 400 m, Jimenez 2113 (Е). SAN JOSÉ: Parque Nacional Braulio Carrillo, entrance- Rio Sucio, Estación Carrillo, 400-1,500 m, 10%5-10'N, 84°57-60 са. 8°48'N, 82°10'W, McPherson 9154 (MO) Anthurium protensum Schott, Oesterr. Bot. 2. : 181. 1858. a. Anthurium protensum subsp. protensum. TYPE: Costa Rica. Heredia: Volcán Barba, 2,500-2,700 m, Oersted 1992 (destroyed?; illustrated by Schott Aroideae 427). Figures 246-248, 260, 261. Anthurium Basic dn Engl., Bot. Jahrb. Syst. 25: 398. 1898. TYPE: Costa Rica. Cartago: above Agua Cali- 706 Annals of the Missouri Botanical Garden ente, 1,800 m, Lehmann s.n. (holotype, B; isotype, Anthurium joseanum Engl., Pflanzenr. ТУ 23B(Heft 21): 1905. TYPE: Costa Rica. Cartago: Las Vueltas, Tucurrique, 635 m, Tonduz 12818 (holotype, В; isotypes, P, US). Epiphytic; stem often pendent, less than 20 cm long, 1-2 cm diam.; roots descending, brownish to green, smooth, thick and moderately elongate, 5- 7 mm diam.; lanceolate, (8)1 1 -30 cm long, obtuse to acuminate and long-apiculate at apex, pale green, drying pale tan, persisting intact, soon dilacerating into a re- ticulum of fine fibers. Leaves spreading to pendent; petioles (6)10-35 cm long, 3-7 mm diam., sub- terete to bluntly angular, sometimes flattened or weakly sulcate adaxially, the surface sparsely pale- speckled; geniculum paler and slightly thicker than petiole, becoming fissured transversely with age, 1-2 cm long; blades subcoriaceous, strap-shaped to oblong-elliptic, long-acuminate at apex (the acu- men weakly apiculate), acute to obtuse to narrowly rounded at base, (18)32-80 cm long, (2.5)4-11.5 cm wide, the margins undulate; upper surface matte to semiglossy, dark to medium green, lower surface semiglossy, paler to almost concolorous with the upper surface, drying greenish and matte; midrib convexly raised at base, becoming narrowly raised cataphylls subcoriaceous, narrowly and then weakly sunken toward the apex above, higher than broad at base, becoming convexly raised toward the apex below, paler than surface above and below; primary lateral veins more than 15 per side, departing midrib at 40-70? angle, sunken to obscure above, flat to raised and darker than sur- face below; interprimary veins almost as conspic- uous as primary lateral veins; tertiary veins weakly visible above and below; collective vein arising from near the base, sunken above, raised and somewhat darker than surface below, equally as prominent as primary lateral veins, 2-5 mm from margin. Inflorescences spreading-pendent, shorter than leaves; peduncle (13)15-49 cm long, 6-7 mm diam., 0.6–1.8 as long as petiole, pale-speckled, terete; spathe erect to reflexed, subcoriaceous, yel- low-green to purple to maroon, oblong-lanceolate to lanceolate, (4.5)10–19 cm long, 1.7-4 cm wide, acuminate to long-acuminate at apex (the acumen inrolled), rounded to obtuse at base; spadix greenish to pale violet-purple to purplish gray, usually glau- cous, long-tapered, prominently curved, (4)7-19 cm long, 5-9 mm diam. near base, 2-4 mm diam. near apex; flowers rhombic to 4-lobed, 2.2-2.7 mm long, 1.8-2.2 mm wide, drying 1.5-1.8 mm diam.; 4—9 flowers visible in principal spiral, 6— 12 in alternate spiral; tepals densely and minutely papillate, weakly pale-punctate; pistils emergent, densely and minutely papillate, green; stigma lin- ear, droplets appearing 2-4 days before stamens emerge; stamens emerging from the base, the lat- erals preceding the alternates by 30-34 spirals, the 3rd stamen preceding the 4th by 5-6 spirals, held in a tight cluster above the pistil; anthers yellow to tan, 0.5-0.6 mm long, 0.6-0.7 mm wide; thecae ellipsoid, slightly divaricate; pollen pale or- ange fading to white. Infructescence with spadix 14-32 cm long; berries orange, ovoid to oblong- ovoid, acute to beaked at apex, (6.4)8-13 mm long, (3.9)4.5-8 mm diam.; seeds 2 per berry, yellow to pale brown, oblong to semi-ovoid, flat- tened, (3.4)3.7-6 mm long, (1.8)2.5-4.5 mm diam., (1.2)1.6-3 mm thick, with a gelatinous, amber appendage. This species is found in Costa Rica and Panama from 635 to 2,700 m, principally in premontane rain and lower montane rain forest life zones. Anthurium protensum is distinguished by its cataphylls weathering to a reticulum of fibers, its elongate, narrow leaf blades with many primary lateral veins, its more or less terete petiole, the long, ovate to lanceolate spathe, pale violet-purple to lavender or greenish spadix and orange berries. This species is closely related to 4. prolatum, A. brenesii, and А. seibertii, all with similar, elon- gate leaves. See the commentary following those species for distinguishing characteristics. Anthurium protensum comprises two subspe- cies. The typical subspecies occurs mostly in Costa Каса (also in western Panama), while subsp. ar- cuatum occurs only in Panama, ranging from Chi- паш to Veraguas at 1,300 to 1,750 m in pre- montane rainforest. The latter differs in being an erect plant with smaller leaves with fewer, more prominent, sunken veins and an acute lower midrib, an arching inflorescence with shorter, usually ovate spathe and a smaller spadix. A third subspecies may be represented as well. Some collections have slender green spadices some- times two or more times longer than the spathe, a generally green spathe and glabrous tepals. Ex- amples of this entity include Croat 60444, Mc- Pherson 8026 and 8685 from Bocas del Toro Province in Panama, and Croat 367 14 and Cha- con 7, from Cartago Province in Costa Rica. More typical collections have the spadix more short-ta- pered, frequently tinged purple, about as long as the spathe or even shorter than the spathe, and tepals covered with a thin, waxy layer. The spathe in this latter group is usually heavily tinged purple. Examples of these plants include Croat 44501, Volume 78, Number 3 1991 Croat 707 Anthurium sect. Pachyneurium 47113, Grayum 3867, and Lehmann 1756 (the type of А. costaricensis). Because the types of both А. protensum and А. costaricensis (consid- ered a synonym of the former) correspond to the more commonly collected variant, the plants with the longer spadices and glabrous tepals, if they prove to be distinct, would perhaps represent a new subspecies. Costa RICA. WITHOUT LOCALITY: Lehmann 1756 (В). ALAJUELA: Zapote, Smith 1268 (NY) 11 mi. NW of Zarcero, Finca Los Ensayos, 900 m, Croat 43533 (CM, MO); Zarcero, Smith 4692 (MO); Cordillera de Tilarán, Finca Penas Blancas, E slope Cerros Centinelas, Monte Verde Reserve, 1,300-1,450 m, 10?18'N, 84%47'W, Grayum 5364 (MO). ALAJUELA & PUNTARENAS: Monte- verde, Burger et al. 10792 (F); N side of Quebrada Cuecha near Sendero El Río, 1,560 m, 10918'30"М, 84%47'50"W, Hayworth 240 (WIS), 1,580 m, Hay- worth 222 (WIS). cARTAGO: Las Vueltas, се и 5 m, Tonduz 12818 (B, P, US); Moravia de Chirrip zn 900-1,000 m, Chacon ? (MO); SE of Platanillo) no Raiz de Hule, 1,200-1,400 m, Croat 367 14 (F, DUKE, ENCB, GH, K, MEXU, MBM, MO, NY, PMA, RSA, gh TEX, US); Carpintera, Brade 2513 (BR); La Cangreja, Williams et al. 24182 (NY); NE of Pacayas, | а 3250 (D UKE); 6 km past town of Río Macho on road to Humo, Continental Divide, 1,600 m, 9?17'N, 83°45'W, Hoover 1347 (MO); Rio Grande де Orosi, suspension bridge- -point opposite mouth of Quebrada Casa m, 9*46.5'N, 83*48'W, Grayum 3776 (MO). GUANACASTE: La Palma, Greenman Greenman 5465 (MO). HEREDIA: N of Vara Blanca, Maxon & Harvey 8304 bg. Skutch 3706 (MICH, US); mi. N of V m, Croat 35566 (MO); 513 (МО); М of Concepción, Primack et al. 276 (DUKE); Braulio Carrillo Park, Zurqui, 1,800 m, Gómez 20092 Cordillera Central, N slopes, Poas- Volcán Barba, Vara Blanca de Log og 1,650 m, Skutch 3652 (MICH, MO, МУ); Rio Para Blanco, 1,600 m, 10%3'N, 84°1'W, Lent 2874 (МО); Rio Socorro, Sarapiqui- Cariblanco, 600- HEREDIA & SAN 10 ‚80 W, Davidse et al. 25543 (MO); Monteverde, Kennedy 605 (US), veo 3406 i ER 1,500-1,550 m, Haber 1164, 3897 (MO); Rio San Luis valley, 900- 1,000 m, Haber et al. 8518 (CR); Cordillera de Тана, 1,550-1,720 m, Almeda et al. 4972 (CAS, МО); S dero El Rio, Montever vid 1,600 m, 10?18'N, 84%48' W, (MO); Mike Fogden property, Croat 46770 (B, jus Palmer 158 (NY); Santa Elena-Monteverde, 3.5 mi. from Santa Elena- Monte verde junction, near E edge of Monteverde Reserve, 1,350 m, Croat 47113 (CM, MO); Río Вага, upper part, 201 m, Gómez et al. 21523 (MO). SAN JOSÉ: Aserri- Tarbaca, Standley 41402 (US) Bajo La Hondura, Lent 1446 (WIS), Standley 36334 (US); 3 km N of Cascajal, 1,680 m, Almeda & Anderson 5284 (CAS, MO); La Palma, p 3301 (DUKE); Piedra, Luteyn 3291 (DUKE); Parque Nacional Braulio Carrillo, Bajo La Hondura, 1,100- 1,200 m, 10%4'N, 83*58'W, Davidse et al. 23190 (MO); Ко Claro-Paracito, Bajo La Hondura, 1,100-1,400 m, Croat 44501 (MO); Santa Maria de Dota, 1,850 m, Stork 1737 (MICH). PANAMA. BOCAS DEL TORO: Fortuna- = riqui Grande, 1.2 mi. N of Continental Divide, 5.3 m N of bridge over Fortuna Dam, 910 m, 8%44'N, 8291 TW. Croat & Стауит 60444 (MO, PMA). CHIRIQUÍ: above Boquete, rd. to Bajo Mono and Alto Quiel, 1,650 m 8%51'N, 82°29'W, McPherson & Merello 8360 (K, MO). b. Anthurium protensum subsp. arcuatum Croat, Ann. Missouri Bot. Gard. 14: 164, figs. 139 & 140. 1986. TvPE: Panama. Chiriqui: above San Félix along mining road, 25 mi. off Pan-American Highway, 1,500 m, Croat 33138 (holotype, MO 231113; isotypes, F, K, PMA, US). Figures 249-251 Usually epiphytic; stem less than 10 cm long, 7-1.8 cm diam., leaf scars 0.6 cm high, 1 cm wide; roots descending, greenish, pubescent, elon- gate, 3—4 mm diam.; cataphylls coriaceous, nar- rowly lanceolate, 3.4-7 cm long, green weakly tinged with red at margins, acute at apex, drying tan (В & К yellow 5/5), persisting intact, soon dilacerating into reticulum of fine fibers. Leaves spreading to pendent; petioles 10-32 cm long, 2- 4 mm diam., subterete to bluntly D-shaped to qua- drangular, sharply 4—5-ribbed and winged, flat- tened to weakly or prominently and narrowly sul- cate adaxially, the margins sharply raised to winged, rounded to bluntly angular to sharply ribbed abax- ially, the surface sparsely pale-speckled; geniculum slightly paler and thicker than petiole, 1-1.5 cm long; blades coriaceous to subcoriaceous, oblong to narrowly oblanceolate to narrowly elliptic or lan- ceolate, gradually long-acuminate at apex, obtuse ded at base, 19-51 cm long, 2.5-7.5 ст wide, broadest at or below middle; upper surface to roun weakly glossy to semiglossy, dark green, lower sur- mi y to glossy below, paler; midrib above convexly таа at base, becoming narrowly raised and eventually weakly sunken toward the apex, below acutely higher than broad at base, becoming sharply acute toward the apex, paler than surface above and below; primary lateral veins 7-12 per side, departing midrib at 45—50° angle, weakly arcuate to the collective vein, prominently to slight- — y sunken above, sharply to weakly raised and darker than surface below; interprimary veins 708 Annals of the Missouri Botanical Garden scarcely visible, sunken above, raised below; ter- tiary veins obscure above, sometimes weakly visible below; collective vein arising from the base, sunken above, moderately sharply raised below, equally as prominent as primary lateral veins, 2-5 mm from margin. /nflorescences erect to spreading, shorter than leaves; peduncle 13-38 cm long, diam., usually longer than petioles, 0.5-2 х as long as petioles, pale green sometimes tinged with red- dish violet, terete to sharply l-ribbed abaxially; spathe erect, curved over and hooding spadix, plain green or heavily tinged with reddish violet to red- dish purple, ovate to broadly lanceolate, 4.5-14 cm long, 1.7-3.5 cm wide, broadest near the base, inserted at 70? angle on peduncle, long-acuminate to caudate at apex, obtuse to rounded or subcordate at base; stipe 5-15 mm long; spadix white to lav- ender (B & K purple 7/7.5) (also reported as light green becoming tan), cylindroid to weakly tapered, weakly curved, 2-9 cm long, 4-6 mm diam. near base, 3—4 mm diam. near apex; flowers rhombic to 4-lobed, (1.3)1.5-3 mm long, (1.3)1.6-2.8 mm wide, drying 1.7-1.9(2.2) mm wide; the sides jag- gedly to smoothly sigmoid; 4—5 flowers visible in principal spiral, 5—8 in alternate spiral; tepals matte, minutely and densely papillate; lateral tepals 0.7— 1.7 mm wide, the inner margins convex, the outer margins 2-3-sided; pistils scarcely emergent, pa- pillate, pale green to pale violet-purple; stigma el- lipsoid, 0.3-0.5 mm long, brushlike, droplets ap- pearing briefly before stamens emerge; stamens emerging from the base, lateral stamens followed by alternates in a rapid succession, the laterals preceding the alternates by 1-2 spirals, held over and obscuring the pistil; anthers pale yellow, 0.4— 0.5 mm long, 0.9- 1 brown, slightly divaricate; pollen pale yellow fading to creamy white. /nfructescence pendent; spadix to 7 cm long, 2.5 cm diam.; berries bright orange, ovoid, beaked at apex, 6-9 mm long, 5-6 mm diam.; mesocarp pulpy, orange; seeds 2 per berry, creamy white, ovoid-ellipsoid, 3-3.2 mm long, 2.2- 2.3 mm diam., 1.5 mm thick, enclosed in trans- parent, sticky envelope ca. 5 mm long, extending further beyond the apex than the base. mm wide; thecae ellipsoid, Anthurium protensum subsp. arcuatum is known principally from western Panama in premontane rainforest at 1,100-2,000 m. A single collection from Monteverde is apparently also included in this taxon even though it has not been collected else- where in Costa Rica. This taxon is distinguished by its slender, short stem, more or less oblong leaf blades that are usually spreading and weakly arched, and by its usually arched inflorescence. Also characteristic is the usually lavender, greenish white or purplish violet spadix, and its ovate to broadly lanceolate, frequently purplish spathe that usually curves over and hoods the spadix. The few cultivated collections of Anthurium protensum subsp. protensum and subsp. arcuatum are markedly different, but herbarium material is more difficult to separate. Subspecies protensum differs from subsp. arcuatum in having typically much larger leaf blades, rarely less than 50 cm long, with the primary lateral veins more numerous and less conspicuously sunken on the upper surface (scarcely more prominently sunken than the in- terprimary veins). In addition, the lower midrib is convex and the plant typically has a lanceolate spathe, a much longer, narrowly tapered spadix (usually more than 9 cm long), and an infructes- cence usually 20-30 cm long. Subspecies proten- sum is also usually a pendent or nearly pendent plant, whereas subsp. arcuatum is usually erect. Subspecies arcuatum rarely has leaf blades more than 50 cm long, relatively fewer, sharply sunken primary lateral veins, an acutely raised lower mid- rib, a typically ovate spathe, with a spadix usually less than 5 cm long, and the infructescence less than 7 cm long. ie PUNTARENAS: Monteverde Cloud Forest of ја те Hammel 6254 (MO); Fortuna Road, road main Fortuna-Chiriqui Grande Hwy., 1.1 m, 8*44'N, 82°17'W, Croat E ?28'W, Hoover 1341 er Cerro Colorado, along mining. road 18-27 mi. off Pan-Am Hwy., above Chami or turnoff to Escopeta, rae San Félix, 1,200-1,5 00 m, Frou 33138 (F, K, MO, PMA, iqui and Bocas del Tor ro aban oad 20 of San Fe lix, 2,000 m, Croat 48461 (MO); on per mining road 20-28 mi. from San Félix, 1,200- 1, m, Croat 33369 (MO); 50 km N of San Felix on осы Divide, 1,200-1,500 m, Mori & Dressler 7770 (МО); along road above San Félix, 30 km above bridge over Rio San Félix, 1,610-1,670 m, Croat 37111, 37132 ; Cerro Horqueta, Dwyer 8749 (DUKE), Pittier 3160 (NY), 3161 (US); Cerro Punta, Wilbur et al. 13081, 15235, 15236, 15252, 15377 (DUKE); Fortuna Road, Fortuna Lake-Chiriqui Grande, 4.5-5 km М of dam over Fortuna Lake, 1,100-1,135 m, Croat & Grayum 60077 (B, CM, K, MO, РМА); trail along Continental Divide to W at Disodudta road, Fortuna Dam, 1,200-1,500 m, 8°47'N, 82°13'W, Churchill 5306 (MO); Río Chiriquí “~ Volume 78, Number 3 1991 Croat 709 Anthurium sect. Pachyneurium Viejo, upper part, Monte Lirio, Seibert 172 (K, NY). VERAGUAS: Santa Fe Region, Cerro Arizona, 1,330 m, ammel & Kress 8570 (MO); Cerro Tute, above Escuela Agricola Alto Piedra, 800-1,450 m, Croat 48964 (F, MO, PMA, US), Folsom & Mauseth 8344 (MO), Knapp & Dressler 5407 (MO), Knapp & Kress 4353 (MO), Sytsma & Andersson 4587 (MO). Anthurium pseudospectabile Croat, Monogr. yst. Bot. Missouri Bot. Gard. 14: 165. 1986. TYPE: Panama. Chiriqui: along road between Gaulaca and Fortuna Dam site, 8.3 mi. NW of Los Planes de Hornito, 8*44'N, 82?16'W, 1,260 m, Croat 49942 (holotype, MO 2769003-005; isotypes, B, K, NY, PMA, SEL, W). Figures 252-254. Usually = stem + slender, less than 30 cm long, 1.5-5 cm diam.; leaf scars 2—4 cm wide; roots few, O to еее, pale grayish to pale tan, occasionally greenish, somewhat pubes- cent, thick and elongate, tapered, 3-6 mm diam.; cataphylls coriaceous to subcoriaceous, lanceolate, stiffly erect, 19-37 cm long, acute to obtuse or + rounded at apex, with subapical apiculum to ca. 4 mm long, light green, drying orange brown to tan, persisting intact, eventually as a reticulum of fibers. flattened to weakly and broadly sulcate adaxially, rounded abaxially, the surface pale-speckled; ge- niculum thicker and slightly paler than petiole, 1— 3 cm long; blades subcoriaceous to coriaceous, oblong to oblong-elliptic, gradually acuminate at apex (the acumen weakly apiculate), decurrent to obtuse to truncate or subcordate at base, 45-150 cm long, 12-45 cm wide, broadest at or just below the middle, the margins weakly undulate; upper surface matte to semiglossy, dark green, lower surface semiglossy to glossy, slightly paler; midrib obtusely angular, becoming flat toward the apex above, prominently convex to round-raised, con- spicuously paler than surface below; primary lateral veins up to 40 per side, departing midrib at 35- 75? angle, straight to weakly arcuate to the col- lective vein, raised above in grooves, prominently raised below; interprimary veins almost as con- To as primary lateral veins, flat to weakly unken above, raised below; tertiary veins obscure E. slightly darker than surface and weakly raised below; collective vein arising from near the base, sunken above, prominently raised below, equally as prominent as primary lateral veins, 2- 5 mm from margin. Inflorescences spreading-pen- dent; peduncle 13-36 cm long, 5-9 mm diam., equaling up to 1.5X as long as petiole, terete to weakly quadrangular with blunt margins; spathe usually twisted or recurled, moderately coriaceous, green below, Tue tinged with purple above, lan- ceolate, 10-32 cm long, 2.5-7 cm wide, broadest near base, a at 45-60° angle on peduncle, gradually acuminate at apex, rounded to cordate at base; spadix yellowish green to green, sessile, narrowly tapered, 1 5-38 cm long, 8-12 mm diam. near base, 3-5 mm diam. near apex; flowers rhom- bic, ca. 3.5 mm long, 2.7-3 mm wide, the sides straight to weakly sigmoid; 4—7 flowers visible in principal spiral, 6-9 in alternate spiral; tepals matte, minutely papillate; lateral tepals 1.8-2 mm wide, the inner margins convex, turned up against the pistil; pistils early emergent, pale green; stigma linear, 0.6 mm long; stamens emerging + rapidly from the base, arranged in a loose cluster around pistil; filaments ca. 0.3 mm long, ca. 0.5 mm wide; anthers white; thecae oblong, scarcely divaricate; pollen white. /nfructescence pendent; berries or- ange, oblong-ellipsoid, 10-12 mm long, 5-6 mm diam. Anthurium pseudospectabile is endemic to Panama at (590)1,000 to 1,600 m in a region that is probably premontane wet or premontane rain forest. This species is distinguished by its pendent habit and its large, almost strap-shaped, many-veined blades with undulate margins. Also characteristic are the short-pedunculate inflorescence with a spreading, usually twisted and contorted green to purple spathe, the weakly protruding tepals, ex- serted stamens, early-emergent pistils, and orange berries. Anthurium pseudospectabile is closely related to А. spectabile, from Costa Rica, but the latter has proportionately broader leaf blades which gen- erally lack the undulate margins, and a petiole that is markedly quadrangular in cross section. PANAMA. BOCAS DEL TORO: Cerro Colorado, дф along quebrada 7.5 mi. from Сћапи, 1,220-1,250 °35'N, 81°45'W, McPherson 8853 (MO); Fortuna Dam area, Pls et al. 2664 (PMA); Oleoducto Road, Continental Divide, 1,000 m, Churchill et al. 4647 (MO), 4648 ‚ РМА); vic. Fortuna Dam, Continental Divide, trail west df highway, 1,250 m, McPherson 11626 (МО); Fortuna- oo Grande, 8.5 mi. N of bridge over For- tuna Lake, 4. N of а Divide, 590 m, Croat & Grayum ioo (CM, МО); road branching N off main Fortuna-Chiriquí Grande Highway, Continental Divide, 1,200 m, Croat & Grayum 60339 (CM, МО); 1.1-1.2 m N of Continental Divide, Croat 60363 (MO, PMA, RSA), 60446 (MO). CHIRIQUi: Fortuna Dam area, Ham- mel 2080 (MO, PMA), Mendoza 336 (MO, РМА). 1, 100- 1,200 m, 8°45'М, 8218", Croat 66592 (MO), Croat 48738 (MO); 1,200-1,600 m, Croat 48737 (MO), Fol- som et al. 5516 (MO), 8220 (MO, PMA), 5607 (B, MO, is 710 Annals of the Missouri Botanical Garden VEN); 5.9-8.6 mi. beyond Los Planes de Hornito, 1,260- 1,400 m, Antonio 5022 (MO), Croat 48732 (MO), 48777 10, NY, e over Rio Esti, 1,400 PMA), Croat 48674 (B, CAS, А ХО, МО, МУ, РМА, ЕЗА, UC, US, VEN); Fortuna Lake-Chiriquí Grande, Fortuna Dam, 4.5-5 km N of dam over Fortuna Lake, 1,100- 1,135, Croat & Grayum 60072 (K, MO). Anthurium purpureospathum Croat, Selbyana 5(3-4): 328. 1981. TYPE: Panama. Bocas del Toro: Changuinola—Almirante Railroad, Sta- tion Milla 7.5, 0-100 m, Croat 38094 (ho- lotype, MO 2388335; isotypes, B, BM, С, CAS, СМ, CR, DUKE, F, GH, IBE, K, LE, M, MISSA, NY, P, PMA, QCA, RSA, S, SEL, US, VEN, W). Figures 263, 264. Epiphytic, or terrestrial at higher elevations; stem ca. 2 cm diam.; roots dense, descending, brown to greenish, smooth to scurfy, moderately thick, somewhat elongate, 3-5 mm diam.; cata- phylls coriaceous to subcoriaceous, narrowly lan- ceolate, 7-14(25) cm long, obtuse to acuminate at apex, yellowish green, weakly tinged with red, drying dark tan (В 4 K yellow 5/2.5), persisting semi-intact, eventually as a reticulum of fibers. Leaves erect to spreading; petioles 12-37(66) cm long, 4-10 mm diam., quadrangular to D-shaped, flattened to broadly and sharply sulcate adaxially, the margins raised, rounded to 1-3-ribbed abaxi- ally, the surface pale-speckled; geniculum thicker and paler than petiole, 1-2.5 cm long; blades sub- coriaceous to moderately coriaceous, oblong-ob- lanceolate to oblong-elliptic, shortly acuminate at apex (the acumen apiculate), obtuse to acute or attenuate at base, 27-74(94) cm long, 7.5-40 cm wide, broadest at or above the middle, the margins broadly undulate; upper surface matte to semi- glossy, dark green, lower surface glossy to weakly. glossy, paler; midrib above obtusely angular-raised to convex at base, becoming narrowly raised toward the apex, below prominently higher than broad and l-ribbed at base, gradually becoming convexly raised toward the apex, paler than surface above and below; primary lateral veins (6)1 1—16 per side, departing midrib at 45—70° angle, arcuate-ascend- ing, loop-connecting in the upper half of the blade, convexly raised in grooves, paler than surface above, prominently and convexly raised and somewhat darker than surface below; interprimary veins sometimes present, almost as conspicuous as pri- mary lateral veins; tertiary veins weakly visible above, flat to weakly raised and darker than surface below; collective vein arising from about the middle or the base of the blade, weakly sunken above, raised and darker than surface below, less prom- inent than primary lateral veins, 4-8(13) mm from margin. /nflorescences erect to spreading; pedun- cle 3-38(50) cm long, 6-20 mm diam., rarely shorter than, mostly equaling or slightly longer than petiole, purplish or green, subterete; spathe sub- coriaceous, purple to purplish with green mottling especially near the base at the center, oblong to narrowly ovate, 4.5-17.5 cm long, 1.5-4 ст wide, broadest just above the base, acute at apex (the acumen cuspidate), obtuse to rounded at base; spa- dix dark violet-purple (B & K purple 2/7.5), cyl- indroid to slightly tapered, (4)5-22 cm long, 3- 0 mm diam. near base, 2-5 mm diam. near apex; flowers rhombic to weakly 4-lobed, 2-2.3 mm long, 2.5-3 mm wide, the sides smoothly to jaggedly sigmoid; ca. 14 flowers visible in principal spiral, ca. 7 in alternate spiral; tepals matte with small, scattered droplets; lateral tepals 1.4—1.6 mm wide, the inner margins flat against pistil; pistils slightly raised, purple becoming green and finally orange; stigma linear, ca. 0.5 mm long, droplets persisting for 4-10 days, drying with exserted papillae 4—6 days before stamens emerge; stamens emerging slowly from the base, held tightly in a circle around pistil; filaments transparent, flattened, soon retract- ing, holding anthers at level of tepals, 0.7-1 mm long, ca. 0.7 mm wide; anthers pale orange; thecae ellipsoid, usually not divaricate; pollen golden yel- low (B & K yellow 8/2.5) fading to creamy white. Infructescence erect to spreading; spathe usually persisting; spadix 9-19 cm long, to 4 cm diam.; berries orange, red-orange to red (B & K red 6/7.5), narrowly ovoid, acute with minute, mam- milliform style at apex, to 10 mm long; mesocarp pasty, transparent; seeds 1 or 2 per berry, pale greenish yellow, ca. 4 mm long, 2 mm diam. This species is known only from Panama and occurs principally from near sea level to 1,500 m in Bocas del Toro and Chiriqui provinces in wetter parts of tropical moist and premontane wet forest. А collection from San Blas (de Nevers 4207) is out of range, but does not differ significantly from other collections from western Panama. Anthurium purpureospathum is distinguished by its rosulate habit, large, thin, semi-intact cat- aphylls, more or less oblanceolate leaf blades, al- most quadrangular petioles which are often three- ribbed abaxially, its generally short peduncles, usually violet-purple, reflexed spathe, usually short violet purple spadix, early-emergent pistils, and by Моште 78, Митбег 3 1991 Croat 711 Anthurium sect. Pachyneurium its long, sharply pointed, orange to red-orange ber- ri This species is most closely related to А. upa- laense, which differs in having leaves with the collective vein arising from above the middle of the blade, a generally longer peduncle, and a more long-tapered spadix. n odd collection is McPherson 8572, because of its very large size. This collection counts for the parenthetical measurements for cataphyll, petiole, blade, and peduncle lengths in the description. PANAMA. BOCAS DEL TORO: 15 km S of Changuinola, PMA, ОСА, ESA, S, SEL, US, VEN, ub. b. mi. NW Ojo del Agua, 3-6 km W n 1 , McPherson 8572 (MO); 175 m, sit 66820 (MO); Isla Bastimentos, l-2 km N of Trouble Hole, 30-100 m Annable 6855 (US); Isla Colón, 0-120 m, von Wedel 125 (МО), 569 (GH, MO); Rio Changuinola, 1 mi. S o Changuinola, Lazor & Tyson 2661 (MO); Río Serpe, Finca St. Louis- Konkintoe, 10-50 m, Woodson et al. 1889 (MO, NY); Rio Guarumo, 5 mi. S of Chiriquicito, Lewis et al. 1985 (MO, NY); Water Valley, von Wedel N of village San Félix, 1 ,430-1,500 m, Croat 48483 (MO); along mining road 18-27 mi. off of Pan-Am Hwy., above Chamí or turnoff to Escopeta, above San Félix, 1,200-1,500 m, Croat 33139 (CM, MO, МУ); Bocas Road, 1,500 m, Folsom & Collins 1835 (MO); 28 mi. above San Félix, 9-10 mi. above turnoff to Escopeta, 1,200-1,500 m, Cai 33257 (MO, PMA); Fortuna Road, Chiriquí Grande-Fortuna, 7.7 mi. W of Chiriqui Grande, 1.5 mi. W of Punta Peña, 80 m, Croat & Grayum 60090 BLAS: El Llano-Cartí Road, Nusgandi, 19.1 km from Interamerican Hwy., 350 m, de Nevers 4207 (MO). Anthurium ranchoanum Engl., Bot. Jahrb. Syst. 21. 1898. TYPE: Costa Rica. Heredia: Rancho Flores in territorio Bacha, Volcán Bar- ba, 2,100 m, Pittier 2295 (holotype, B; is- otypes, BR, G). Figures 262, 265, 266. Anthurium valerii Standley, Publ. Field Mus. Nat. Hist., Bot. Ser. 18: 137. 1937. TYPE: Costa Rica. Heredia: Vara Blanca, Valerio 1600 (holotype, F). Anthurium vinicolor Standley & L. O. Williams, Ceiba 3: 106. 2. ТУРЕ: P po сви Jose: Cordillera de Talamanca near Copey, О m, L. О. Williams 16538 (holotype, EAP; Men F, US). Epiphytic or terrestrial; stem to 100 cm long, ; leaf scars occasionally con- spicuous, 071 е ст inh. 1.5-] moderately numerous, нано, greenish, pu- escent, sometimes smooth, moderately thick and elongate, blunt, 3-7 mm diam.; cataphylls sub- coriaceous, 6-30 cm long, subrounded at apex with subapical apiculum, green tinged with red, drying tan (B yellow-red 9/2.5), persisting semi-intact, eventually deciduous. Leaves spread- ing; petioles 11-150 cm long, 5-15 mm diam., erect-spreading, subterete to nearly D-shaped, shallowly and narrowly sulcate to flattened adaxi- cm wide; roots ally, rounded abaxially, the surface faintly pale- speckled; geniculum slightly thicker and paler than petiole, 1.5-3 cm long; blades moderately coria- ceous, somewhat pliable, narrowly ovate to ovate- deltoid to ovate-triangular, acuminate at apex (the acumen apiculate), shallowly or deeply lobed at base, 17-64 cm long, 9-34 cm wide, broadest between the base and middle, the margins undulate; anterior lobe 12-29 cm long, the posterior lobes 7-13 cm long; sinus arcuate to parabolic; both surfaces semiglossy to glossy, medium to dark green above, much paler below; midrib bluntly angular to convexly raised above (rarely acutely angular), much paler than surface, round-raised at base, becoming prominently and convexly raised toward the apex, paler than surface below; basal veins 3- 6 pairs, 151 usually free to base, 3rd to 6th co- alesced 1-2 cm, 4th to 6th coalesced 3.5-6 cm, raised above and below; posterior rib naked, turned up on outer margin; primary lateral veins 4—9 per side, departing midrib at 40—50° angle, broadly arcuate, weakly raised above and below, more so below; interprimary veins flat above, weakly raised below, less conspicuous than primary lateral veins; tertiary veins weakly visible above and below; col- lective vein arising from one of the primary lateral veins, sometimes from the first basal vein, raised to weakly sunken above, raised below, 2-10 mm from margin. /nflorescences erect to erect-spread- ing, much shorter than leaves; peduncle 13-100 cm long, 5 mm diam., 0.6–0.7 X as long as petiole, green to purplish, terete; spathe spreading, hoode over the spadix, recurled at apex, subcoriaceous, green to green tinged with purple to purple or maroon, lanceolate to narrowly ovate, 4-35 cm long, 2-6.5 cm wide, broadest near the base, acute and apiculate at apex (the acumen tightly inrolled), truncate at base; stipe 1.6 cm long in front, 1.3 cm long in back; spadix pink to red to dark violet- purple to maroon, weakly tapered, 3.5-19 cm long, 9-15 mm diam. near base, 6-9 mm diam. near apex; flowers rhombic, 2-3.2 mm long, 2.6- 712 Annals of the Missouri Botanical Garden 3(5) mm wide, the sides jaggedly sigmoid; 5-10 flowers visible in principal spiral, 7—20 in alternate spiral; tepals matte, sparsely punctate, moderately papillate; lateral tepals 1-1.8 mm wide, difficult to discern, the inner margins broadly rounded, the outer margins 3-4-sided; pistils not emergent, pale green; stigma slitlike, ca. 0.5 mm long; stamens emerging promptly throughout in a scattered pat- tern, becoming prominently exserted, lateral sta- mens emerging almost to apex before alternates emerge, surmounting tepals; somewhat flattened, translucent, ca. 0.5 mm long, filaments fleshy, ca. | mm wide; anthers creamy white to pale , 0.7 mm wide; thecae ovoid-ellipsoid, slightly i ahi pollen yellow to yellow-orange, fading to white, yeasty-scented. In- fructescence arching-pendent; spathe persisting; spadix to 50 cm long, to 3 cm diam.; berries orange, narrowly obovoid, acute at apex, 12-14 mm long, 6 mm diam.; mesocarp bitter; seeds 1l- 2 per berry. orange, ca. 0.6 mm lon Anthurium ranchoanum is found in Costa Rica and Panama at 500 to 2,300 m (usually above 1,000 m), usually in tropical wet, premontane rain and lower montane rain forest life zones. This species is an atypical member of sect. Pachyneurium, recognized by its obtusely sulcate petioles and thick, ovate to ovate-triangular blades with heavy, rather numerous veins. Particularly characteristic is the ovate to lanceolate, hooded spathe which is green tinged with purple and re- mains stiffly erect. Other distinguishing features include the dark violet-purple, relatively stubby, scarcely tapered, somewhat pendent spadix with tight clusters of white stamens, and the orange berries. Anthurium ranchoanum is most closely related to А. cotobrusii, which differs in having an erect, non-hooding spathe and a markedly tapered spadix which is 10-20 x (vs. less than 6x) longer than thick. Two collections worthy of mention are Grayum & Dickie 6587 and 6590 which differ from typical collections in having blades drying thinner with less conspicuous tertiary veins, an acute midrib on up- per surface and proportionally longer spadix at anthesis. They are perhaps subspecifically distinct. Costa RICA. ALAJUELA: Rio La Paz Grande, 7.5 km М 0 m, Croat 36052 (МО); da Siberia, ca. 2 km E of Villa Mills, ai Vara Blanca- Pueri Viejo, 5.5 km N of Vara Blanca, 1,200 m, Croat 36053 (MOy; headwaters of Rio Santo Domingo, m NE of San Rafael de Vara Blanca, N ey Volcán fare 2,060 m, 10°1.5'N, 84*07'W, Gray- m 7226 (МО); 1 km МЕ of Los Cartagos, 2,100 m no 1898 en Cerro Chompipe-SW flank of Volcán Barva, N of Hwy. 113 gr with is br 2,000-2,100 m, 10?7'N, 84? Baker et al. (MO); Monte Cristo, 1,720 m, p 511 (МО); Adi of Volcán Barva, above San Rafael and Uvita, along Río Segundo, 1,900-2,000, 10%05'N, 84°05'W, Berringer & Schatz 2039 (MO); Parque Nacional Braulio Carrillo, San Rafael de Vara-Blanca, 1,830 m, 10?11'50"N, 84*6'35"W, Herrera 242 (MO), 244 (MO); 1,800 m, 266 (MO); 2,000 m, 228 (MO); NNW of Heredia, Lel- linger & White 981, 1012 (US); road to Sacramento, Utley & Utley 4564 (DUKE); San Isidro, Pittier 14058 (US), Бесна & Valerio 50243, 50798, 51455, 51620, 52345 (US); N of San Rafael, Lems s.n. (K, NY); Braulio Carrillo Park, Zurqui, 1,700-2,000 m, cu 20103, 20221 (CR, MO), 20225 (B, CR, F, МО); Cerro Caricias and Cerro Zurqui, S slopes, bordering Quebrada Caricias, МЕ of Heredia, 1,600-1,800 m, Luteyn 3275 (MO); Cerro Chompipe, Luteyn р Wilbur 4422 (DUKE): N slope, San 10%5'N, 84°4'W, Stevens 13980 (M per Río Patria, E slopes of Volcán PARA Continental Divide, 2,000 m, 10°6’М, 84*4'W, Burge iesner 6369 (MO); Volcan € Pittier 2295 (BR), Skutch 3657 (MICH, US). PUNTARENAS: Cantón e Coto Brus, cultivated at Las Cruces Tropic al B | Garden, 1,200 m, Croat 44443, 57242 (MO); Monteverde, Rio Gua cimal, below Lechería, 1,500 m, 10?17'N, 84?48'W, Hammel & Trainer 13816 (MO); near E edge of Mon teverde Reserve, 1,350 m, Croat 47130 (МО); 1, 450 7 Isidro del General, Utley 407 (DUKE); 5 km SW of Empalme on ips е Santa Maria де Dota, 2,000 m, Antonio 721 (M of Santa Maria de Dota, Luteyn 3324 (DUKE), Sail 41767 (US), Standley & Val- епо 44077 (US); Santa Rosa del Copey, Tonduz 12266, 12267, 12269 (05); Rio Cascajal, 3 km NE of Cascajal, ,680 m, Lent 2181 (MO). PANAMA. BOCAS DEL TORO: Cerro Colorado, к гоад n Río San Felix and mining exploration camp, i. W of Chami, along trail through Guaymi АЦЕ 1, 500 m, Croat 69211 ( Cerro Pate Macho, NW ridge, from summit to Finca Serrano, 1,200-2,100 m, Sytsma et al. 4949 (MO). BOCAS DEL TORO & CHIRIQUI: ridge of Continental Divide, NE of Cerro Pate Alto, 2,200 m, 8%47'N, 82?21'W, Knapp & Schmalzel 4755 (MO). CHIRIQUÍ: Cerro Hor- queta, Cordillera de Talamanca, above Boquete, 1,860- 1,940 m, 8°49'N, 82°29'W, Cochrane et al. 6264 (МО); NW of Boquete, 1,660-1,930 m, Herrera 545 (МО); Cerro Hornito, 1,400-1,750 m, Folsom et al. 7224 (MO); 1,750-1,900 m, 8°41'N, 82°10'W Croat 67985 (MO); Сен Hornito summit, above Los Planes de Ног- nito, 2,100-2,230 m, Hammel 6217 (MO), Knapp et al. 4202 (MO); Cerro Horqueta, Dwyer et al. 545 (GH, МО, US); Cerro Pando, E slopes, 2,000-2,300 m, 8*55'N, 82*44'"W, Knapp 1653 (К, MO); Cerro Pate Macho, Continental Divide, NE of Cerro Pate Macho, Bocas del Toro border, above Palo Alto, 2,200 m, 8°47'N, 82?21'W, Knapp et al. 4240 (MOy; along Continental Divide, NE ps 7 mi. NW of Rio Chiriqui Viejo, Las Nubes, 2,200 m Volume 78, Number 3 1991 Croat 713 Anthurium sect. Pachyneurium Liesner 326 (MO, PMA); Fortuna Road, 1,300 m, Mc- Pherson 8779 (B, Kin NY, PM ortuna Dam, ridge es de e Hornito, 1,000 m, Knapp et am, 5.9 mi. NW of , Croat 49846 (МО); P xd m, Hammel 2113 (МО); Río Palo Alto i пина 0. 5 km E of Cerro Pate Macho, 1,800-2,100 m, 8*47'N, 82°21'W, Knapp et al. 2120 (MO); Dtto. Bugaba, Cerro Pando, Santa Clara, 8*50'N, 82*44'W, van der Werff & Herrera 7275 (MO, PMA, RSA) Anthurium reflexinervium Croat, sp. nov. TYPE: Peru. Huánuco: Prov. Leoncio Prado, Dtto. Rupa Rupa, Tingo Maria, vic. of airport, 750— 800 m, Plowman & Ramírez 7585 (holotype, MO 2743714; isotypes, F, K, NY, SEL). Fig- ures 267, 268, 2771. Planta epilithica; internodia brevia, 1-3 cm diam.; 26(35) utroque, manifeste retrorse arcuati; p 18-40 cm longus, 3-6 mm diam.; spatha lanceolato- ovata aut oblongo-lanceolata aut oblongo-elliptica, 4.5-9 cm longa, 0.7-2.1 cm lata; stipes ad 5 mm liber; spadix contractus, 6-11 cm eae basi 5-9 mm diam., apice mm diam., rubellus-violaceus ad atrorubellum-pur- pureum. Baccae rubro-violaceae, 8 mm longae. Epilithic; stem usually less than 15 cm long, 1- 3 cm diam.; leaf scars 0.7 cm high, 1 cm wide; roots dense, ascending to descending, brown, smooth to tomentose, weakly tapered, 3-4 cm long, 3-5 mm diam.; cataphylls subcoriaceous, broadly lan- ceolate-triangular, (2)4—4.5 cm long, acute at apex, drying dull brown, persisting semi-intact, quickly dilacerating and persisting as fine linear fibers, split- ting at apex. Leaves erect to spreading; petioles 3-8 cm long, (4)6-12 mm diam., broadly an sharply triangular, sometimes sharply D-shaped, flattened, sometimes with a medial rib adaxially, the margins winged and undulate, prominently and sharply angular to rounded abaxially; geniculum scarcely to moderately thicker than petiole, 0.5— 1 cm long; sheath 2-3.5 cm long; blades coriaceous to subcoriaceous, broadly oblanceolate-elliptic, acute at apex, obtuse to shallowly cordate at base, 41- 72 cm long, (10.5)13-18 cm wide, broadest at or above the middle, the margins sometimes concave toward the base, moderately to strongly undulate; upper surface matte to semiglossy, dark green, lower surface semiglossy, paler; both surfaces strongly bullate and quilted; midrib flat and with a sharp rib at base, becoming acutely raised and higher than broad toward the apex above, acutely raised, higher than broad and 1-ribbed to winged below; primary lateral veins 22-26(35) per side, departing midrib at acute to obt ‚ Spreading retrorsely at 30—70° angle, a arcuate to the collective vein, narrowly raised in deep grooves above, less so below, knife-edge-like when dried below, less so above; interprimary veins few, almost as conspicuous as primary lateral veins; tertiary veins weakly raised to flat below, weakly visible above and below when dried; reticulate veins obscure to conspicuous when dried, flat; collective vein arising from near the apex, about the middle of the blade or near the base, equally as prominent as major tertiary veins, 2-5 mm from margin. Inflorescences erect; peduncle 15-40 cm long, 3- 7 mm diam., 2.8-8(11)X as long as petiole, olive- green to dark green, terete with a diminishing rib, firm; spathe ascending to reflexed, sometimes re- curled, subcoriaceous to coriaceous, olive-green to pale olive-green, tinged with red to uniformly dark green, lanceolate-ovate to oblong-lanceolate to ob- long-elliptic, 4.5-9 cm long, 0.7-2.4 cm wide, broadest near the base, inserted at 45—80? angle on peduncle, abruptly acuminate at apex (the acu- men tightly inrolled and cuspidate, 2-3 mm long), narrowly acute to obtuse to rounded at base, the margins meeting at 80? angle; stipe ca. 6-13(20) mm long in front, 1-6 mm long in back; spadix red-violet to dark red-purple (B & K purple 4/7.5), tapered, erect, somewhat curved, 6-11 cm long, 5-10 mm diam. near base, 2-5 mm diam. near apex, broadest at the base; flowers + square, 2.0- 2.5 mm long when fresh, drying 1.9-2.4 mm long, 2.5-3.7 mm wide when fresh, drying 1.7-1.9 mm wide, the sides straight to smoothly or jaggedly sigmoid; 5-9 flowers visible in principal spiral, 4— 14 in alternate spiral; tepals densely and minutely papillate; lateral tepals 1.1—2 mm wide, the inner margins + straight to broadly convex, weakly erose, the outer margins 2—3-sided; pistils weakly raised, greenish, somewhat papillate; stigma slitlike, be- coming broadly ellipsoid to nearly circular, 0.4— 0.5 mm long, depressed medially with the margins somewhat raised when fresh, drying slightly sunken and blackish; stamens emerging rapidly in a com- plete sequence, the laterals preceding the alternates by 6-14 spirals, 3rd preceding 4th by 1-2 spirals, held at the sides of the pistil; anthers yellow, 0.5— 0.9 mm long, 0.7-1.0 mm wide; thecae ovoid- ellipsoid, not divaricate. Infructescence with spathe green, persisting; berries reddish violet, 8 mm long. Anthurium reflexinervium is known only from the vicinity of Tingo Мапа in Huánuco, Peru, а! 650 to 800 m, on rocky slopes near the Rio Hua- llaga, in a premontane tropical moist-tropical moist forest transitional life zone. 714 Annals of the Missouri Botanical Garden This species is easily recognized by its strongly quilted-bullate leaf blades with numerous, promi- nently retrorse, arcuate primary lateral veins (hence the name). Prominently retrorse primary lateral veins are otherwise found only in 4. latissimum and А. paraguayense var. coroicoanum, but these lack strongly bullate leaf blades and have retrorse veins only toward the base. Other species may occasionally have a few retrorse primary lateral veins near the base. Anthurium reflexinervium is a stunning horti- cultural subject, at present cultivated at several botanical gardens and in a few private collections. PERU. HUÁNUCO: Tingo María Region, Tingo Maria- Pa tay Rondos, Cuevas de Guácharo, 650 °12'W, Croat 57949 (BM, CAS, CM, МО, NY, RSA, SEL, U, US, USM); Rio Huallaga, Croat 21080 (МО); Prov. Leoncio Prado, Rio H 700-800 m, Plowman 5823А (СН); Prov. Leoncio Pra = Dtto. Rupa Rupa, Tingo Maria, 750 ‚ Plow п & Ramirez 7585 (F, K, MO, NY, SEL). Anthurium remotigeniculatum Croat, sp. nov. TYPE: Peru. San Martin: along road between Tarapoto and Yurimaguas, 530 m, Croat 58114 (holotype, MO 3183169; isotypes, B, GH, K, M, NY, QCA, RSA, SEL, US, USM). Figures 272, 275. Planta terrestris, internodia brevia 1.5-6 cm diam.; cataphyllum anguste trigonum; petiolus 10-34 cm lo A 5-10 mm 1- motum; lamina ovata ad obovata vel 10.5-12 cm longa, 0.5-1.5 cm lata, viridis; stipes usque ad 4 mm liber; spadix grs et үнү" contractus, 10.5-25 cm longus, basi 6-7 mm diam., -4 mm diam., atroviolaceo- Lebe "pin rubello ERE ca. mm longae, 2.8-3 m dia Terrestrial, occasionally epilithic; stem short, 1.5-6 cm diam.; roots few, descending, greenish to brown, smooth to densely pubescent, elongate, 4-6 mm diam.; riaceous, narrowly triangular to broadly lanceolate, unribbed, 3-7.5 cm long, acute at apex, reddis purple, drying brown (B & K yellow 3/2.5), per- sisting intact, eventually deciduous. Leaves erect to erect-spreading; petioles 10-34 cm long, 5-1 cataphylls coriaceous to subco- mm diam., D-shaped, narrowly and obtusely sul- cate adaxially with the margins blunt, not raised, rounded abaxially, the surface pale-speckled; gen- iculum thicker, slightly paler than, and more shal- lowly sulcate than petiole, becoming fissured trans- versely with age, 1.5-2 cm long, remote from the base of the blade ca. 2-16 cm, rarely contiguous; blades coriaceous, ovate to obovate to broadly el- liptic, abruptly acuminate, sometimes acute at apex (the acumen weakly apiculate and downturned), long-attenuate at base, (20)27- 100 cm long, (9)1 4- 28.5 cm wide, broadest at or above the middle, the margins weakly to prominently undulate; upper surface matte to weakly glossy, medium to dar green, lower surface matte, conspicuously to mod- erately paler, appearing minutely pale-speckled and with translucent, alveolate pattern on high mag. nification when fresh; both surfaces drying yellow- ish to yellowish green; midrib above flat to convexly raised at base, becoming more convex toward the apex and paler than surface, below prominently higher than broad at base, becoming convexly raised toward the apex and slightly paler than surface; primary lateral veins 6–8 per side, departing midrib at 30-60? angle, straight near the midrib, becom- ing arcuate-ascending near the margin, weakly convexly raised and scarcely paler than surface above, prominently raised below, drying promi- nently raised above and below; interprimary veins weakly visible and sunken above, slightly darker than surface, scarcely raised below; tertiary veins sunken above, weakly raised and darker than sur- face below; collective vein arising from about the middle of the blade or in the upper third, slightly sunken above, raised and darker than surface be- low, 5-15 mm from margin. /nflorescences erect to spreading, equaling or longer than leaves; pe- duncle 63.5-96 cm long, 7-17 mm diam., 2-4x as long as petiole, green to green heavily tinged with purple; spathe spreading-recurled, coriaceous to subcoriaceous, plain green or heavily to slightly tinged with purple, linear-lanceolate, 10.5-13.5 cm long, 0.5-3 cm wide, broadest near base, acute at apex (the acumen inrolled), acute to decurrent at base; stipe 1.3-2.5 cm long in front, to 4 mm long in back; spadix maroon to dark violet-purple (B & K purple 2/2.5), long-tapered, curved, 10.2- 25 cm long, 6-8 mm diam. near base, 3-4 mm diam. near apex; flowers + square, 1.8-2.5 mm in both directions, the sides straight to smoothly sigmoid; 4-8 flowers visible in principal spiral, 7- 13 in alternate spiral; tepals matte, densely and minutely papillate, scarcely punctate; lateral tepals 0.9 m wide, the inner margins convex, the outer margins irregularly 2—3-sided; pistils emer- gent, densely and minutely papillate, reddish pur- ple, becoming green; stigma linear, slitlike, 0.5- 0.7 mm long; the lateral stamens preceding the alternates by 3—4 spirals, the 3rd stamen preceding the 4th by 3-5 spirals, held in a tight cluster above the pistil; anthers yellow, 0.4-0.6 mm long, 0.8- mm wide, scarcely exserted; thecae ovoid, slight- Volume 78, Number 3 1991 Croat 715 Anthurium sect. Pachyneurium ly or not divaricate; pollen yellow fading to white. Infructescence with spathe absent; spadix at least 13 ст long, 1.5 cm diam.; berries obovoid, reddish purple, white at base, rounded at apex (truncate when dried), ca. 6.5-7.5 mm long, 5-6 mm diam., mesocarp juicy, translucent; seeds 2, 4.5-5.0 mm long, 2.5 mm diam., scarcely flattened, reddish purple, minutely pale-speckled. Anthurium remotigeniculatum is recorded only from Peru, where it occurs in the provinces of San Martin and Amazonas at 530 to 1,160 m, in trop- ical moist and tropical premontane wet forest life zones. This species is characterized by its terrestrial habit, short cataphylls that remain semi-intact, D-shaped petioles which are obtusely and bluntly sulcate adaxially, and its elongate, long-tapered, dark purple spadix. Especially significant is the leaf blade base, which is long-attenuate (especially in older plants) and eventually naked, making the geniculum seem remote sometimes for as much as 16 cm from the base of the blade, hence the specific epithet. Also characteristic are the coriaceous leaf blades which are ovate, obovate, or broadly elliptic, dark green above and conspicuously paler below, appearing speckled and with a translucent, alve- olate pattern under high magnification on fresh plants. Anthurium remotigeniculatum is most closely allied to 4. pachylaminum, which differs in having more coriaceous leaf blades lacking the remote geniculum. Although most of the specimens of А. remoti- geniculatum display the remote geniculum char- acteristic very clearly, there are some exceptions. Specimens lacking this feature otherwise show re- markable similarity to more typical material. Ob- servations on living material show that the same plant may produce the characteristic leaves with a remote geniculum and “normal” leaves at the same time, and such aberrant specimens have thus been included in the species concept. PERU. AMAZONAS: Prov. Bagua, 15.6 km E of main llaga, Tarapoto- Yurimaguas, Croat 58114 (B, K, GH, M, MO, NY, QCA, SA SEL. TEX, US, USM). Anthurium salvadorense Croat, Selbyana 5(3- 4): 333. 1981. TYPE: El Salvador. Ahuacha- pan: 1.5 mi. SW of Tacuba on road to San Francisco Menéndez, riverbank of Rio Cho- kama, 700 m, Croat 42169 (holotype, MO 2583750; isotypes, CAS, F, ITIC, K, LE, M, MEXU, RSA, SEL, TEX, US). Figures 269, 273. Terrestrial or epilithic; stem short, 2-5 cm diam.; roots numerous, dense, descending to spreading or ascending, pale green to tinged reddish, smooth to pubescent, short and moderately thick, bluntly pointed at apex, 4-10 mm diam.; cataphylls mod- erately coriaceous, shortly lanceolate, 5-12.5 cm long, pale yellowish green, acute and apiculate at apex, drying light to dark brown, persisting par- tially intact, eventually deciduous. Leaves erect to spreading; petioles 9.5-46 cm long, (4)6–10 mm iam., + quadrangular to D-shaped, broadly and shallowly to deeply sulcate, sometimes with medial rib adaxially, the margins blunt to sharp and raised, rounded to multi-ribbed abaxially, the surface pale- speckled; geniculum thicker than petiole and paler than petiole, 1-2 cm long; blades subcoriaceous, elliptic to broadly elliptic to oblanceolate, gradually to abruptly acuminate at apex (the acumen down- turned, ca. 15 mm long), acute to obtuse to nar- rowly rounded at base, 25-65 cm long, 8-26 cm wide, broadest at or above middle, the margins undulate; upper surface matte to semiglossy, lower surface matte; medium green above, slightly paler below; midrib above flat with weak medial ridge at base, becoming obtusely angular and eventually sunken toward apex, below higher than broad, l- 4-ribbed at base, becoming prominently and con- vexly raised or convexly raised with a rib toward the apex and slightly paler than surface; primary lateral veins 7-14 per side, departing midrib at 45-60” angle, straight, loop-connected in the up- per half, convexly raised above and below, slightly paler than surface; interprimary veins few, raised above and below; tertiary veins sunken to flat above, raised below; collective vein arising from about the middle of the blade or higher, sunken above, raised below, 5-12 mm from margin. Inflorescences erect to spreading; peduncle 13-57 cm long, 3-7 mm diam., 1.4-2.8X as long as petiole, light green, terete, weakly multi-ribbed; spathe erect to spread- ing, becoming recurled, moderately coriaceous, pale yellow-green (B & K yellow-green 8/10), ovate- lanceolate to oblong-lanceolate, 4.5-10 cm long, 1-4.3 cm wide, broadest just above the base, in- serted at 155-170? angle on peduncle, acuminate at apex (the acumen apiculate and inrolled), sub- cordate and decurrent on peduncle at base; spadix pale green (B & K green 8/5), sessile, cylindroid 716 Annals of the Missouri Botanical Garden to weakly tapered, somewhat curved, 2.5-11 cm long, 7-12 mm diam. near base, 4-7 mm diam. near apex; flowers + irregularly 4-lobed, 2-3.3 mm long, 2.5-3.6 mm wide, the sides prominently jaggedly sigmoid; 6– 500 flowers visible in prin- cipal spiral, 9—12 in alternate spiral; tepals smooth, matte; lateral tepals 1.4-1. Tu mm wide, the inner sometimes margins rounded, somewhat erose, turned up against pistil, the outer margins irreg- ularly 2—4-sided; pistils weakly emergent, promi- nently raised before stamens emerge, dark green; stigma circular to ellipsoid, 0.3-0.4 mm long, erect and brushlike, droplets appearing 2—4 days before stamens emerge, stamens emerging rapidly in a complete sequence from the base, lateral stamens almost to apex before stamens emerge; anthers pale green, 0.8-1.2 mm long, 0.7-1.3 mm wide, contiguous, inclined over pistil; thecae weakly di- varicate; pollen pale yellow, soon fading to white, sweetly fruit-scented. /nfructescence spreading; spathe persisting and erect; spadix 6.5-10.5 cm long, ca. 2 cm diam.; berries orange-red (B & K yellow-red 5/2.5 to red 4/10), broadly obovoid, rounded to flat or weakly indented at apex, 5.5- 9 mm long, 4-9 mm diam.; pericarp moderately thickened, with numerous punctiform raphide cells especially in the apical half; mesocarp pulpy or mealy, with numerous raphide cells; seeds 1 or 2, yellowish white to tan, obovoid to broadly obovoid, 6-6.5 mm long, 5-5.8 mm diam., 3 mm thick, densely covered with punctiform raphide cells. Anthurium salvadorense is known only from El Salvador and Guatemala at 500 to 800 m. It was first collected by Sixto Alberto Padilla in 1922 in the Department of Ahuacapán in El Salvador. The species was later collected by Paul C. Standley in the adjacent Department of Jutiapa in Guatemala. This species is a typical member of the section, and earlier collections were confused with 4 schlechtendalii, to which it is related. Anthurium salvadorense differs in being a usually well-rooted plant with fewer, very large fleshy roots and a usually almost elliptic blade, but especially by its inflorescence, which has a pale green, ovate-lan- ceolate spathe that is weakly convolute at the base and erect, enshrouding the usually short, pale green spadix. Anthurium schlechtendalii differs in hav- ing usually oblanceolate blades, a coriaceous, more lanceolate, prominently reflexed, usually purplish spathe, and a more elongate, purplish spadix. EL SALVADOR. AHUACHAPÁN: San Francisco Menendez- n. (MO), pios okama, 1.5 m S of Tacuba, 700 m, Gál 42169 (CAS, F, ITIC, K. LE, M, MEXU, МО, RSA, SEL, TEX, US). GUATEMALA. JUTIAPA: Jutiapa, Standley 76313 (F, US); SW of San Cristóbal, Dunn et al. 23222 (ОМО); Monjoy Canyon, mi. W of San Cristóbal, 530 m, Dunn et al. 19 (MO). Anthurium salviniae Hemsley, Diagn. Pl. Nov. Mexic. 36. 1878. TYPE: Guatemala. de Fuego, Salvin s.n. (holotype, K). Figures 15, 20, 270, 274, 276. Мојсап Anthurium со Matuda, Madroño i 169. 1950. YPE: Mexico. Chiapas: Escuintla, Salto de Agua, Matuda 18043 fades MEXU; isotype, Anthurium enormispadix Matuda, € Soc. Mex. . Nat. 11: 94. Dec. 1950. ТҮРЕ: Same as Ап- ern Prasad Matuda. Epiphytic, rarely epilithic; stem to 6 cm diam.; roots dense, usually ascending, some descending, moderately numerous, greenish to white, pubes- cent, moderately short and tapered, 3-6 mm diam.; cataphylls thick, fistlike, cucullate, rounded at apex, drying brown, persisting intact, ultimately becom- ing fibrous at base. Leaves erect to spreading; petioles (3)5-23 cm long, 10-20 mm diam., + p shaped to terete, flattened to sulcate adaxially, n prominent, rounded abaxially (once réported as prominently 5—7-ribbed), the њи pale-speckled; geniculum paler and thicker than petiole, 0.7-2 cm long; blades moderately coria- ceous, obovate to oblanceolate to broadly elliptic, acuminate to obtuse at apex, attenuate to rounded or truncate (rare subcordate) at base, (32)39-180 cm long, (9)1 3-50 cm wide, broadest usually above (sometim slightly undulate; both surfaces semiglossy, medi- um green ( yellow-green 4/7.5) above, slightly paler below; both surfaces drying greenish to brownish green; midrib convexly raised at base, becoming acute toward the apex above, convexly raised at base, becoming obtusely raised toward the apex below; primary lateral veins 9-24 side, convexly raised above and below; interpri- mes at or near the middle), the margins mary veins rarely present; tertiary veins weakly raised to obscure above, weakly raised to flat and darker than surface below; collective vein arising from near the apex or absent (often from below the middle in small plants), weakly raised, 3-5 mm from margin. Inflorescences spreading to usually pendent, shorter than leaves; peduncle 16-80 cm long, 2-5 mm diam., 3.1-10.3 х as long as petiole, terete; spathe spreading to recurled, coriaceous to subcoriaceous, purple or sometimes green tinged with red or purple (B & K yellow-green 6/5), lanceolate-linear, (4)10.5-40(50) cm long, 1-2.5 cm wide, inserted at 45-90? angle on peduncle, Volume 78, Number 3 1991 Croat 717 Anthurium sect. Pachyneurium gradually acuminate at apex (the acumen cuspi- date, inrolled), rounded to truncate or acute at base; spadix pale green when immature, soon lav- ender to dull violet-purple to purple-maroon, glau- cous, long-tapered, slightly curved, (6)10-34(47) cm long, 4-15 mm diam. near base, 2—7 mm diam. near apex; flowers rhombic, 2.4-2.5 mm in both directions, the sides straight to sigmoid; 6-8 flowers visible in principal spiral, 4—12 in alternate spiral; tepals matte; lateral tepals 1.2-1.3 mm wide, the inner margins broadly rounded; pistils emergent but not raised, green; stigma linear to oblong-el- lipsoid, 0.6-0.9 mm long; stamens emerging from the base in a slow, regular progression; anthers orange to pale yellow to creamy white, ca. 1 mm long, 0.9 mm wide, held at edge of tepals and inclined over pistil; thecae ellipsoid; pollen pale yellow-orange fading to white. Infructescence with spadix to 60 cm long; berries red, oblong-ellipsoid, rounded at apex, 10—15 mm long; mesocarp trans- parent, gelatinous; seeds 2 per berry, oblong, flat- tened, ca. 3-5 mm long. Anthurium salviniae ranges from western Mex- ico (Chiapas), along the Pacific slope of Guatemala, and along the Atlantic and Pacific slopes from Nicaragua to Panama. There is also a single col- lection from Honduras (Atlantic slope) that appears to belong to this species. The species also occurs in Colombia, where it has been collected from Chi- gorodo and Dabeiba in Antioquia and Chocó de- partments. Anthurium salviniae occurs from near sea level to 1,600 m in wetter parts of tropical moist forest, in premontane wet, tropical wet, and also, rarely, in premontane rain forest. is species is characterized by its cucullate cataphylls, large, mostly oblanceolate leaves with circinate vernation, usually pendant inflorescence with a slender peduncle and slender, long-tapered, pale lavender spadix and red berries. Anthurium salviniae has long been called А. tetragonum, but that пате is a synonym of А. schlechtendalii, a species ranging from Mexico to Honduras or Nicaragua on the Atlantic slope. Al- though leaves of the two species are very similar, А. salviniae can be distinguished by its long-ta- pered, pale lavender, glaucous spadix, long-ta- pered, spreading spathe, and by its fist-shaped, cucullate cataphylls. In addition, the leaves of А. salviniae usually dry darker and thinner. Anthur- ium schlechtendalii usually has a shorter, blunter spadix and a short, thick spathe that is usually reflexed and often twisted. The spadix is darker violet-purple, and is not at all glaucous at anthesis. A collection in cultivation at the Missouri Bo- tanical Garden from the Rio Маје region of Рапа- ma, in Panamá Province, is unusual in having dark violet-purple rather than red berries. COLOMBIA. ANTIOQUIA: Quebrada La Puerca- Malagon, 10 m, Rentería et al. 4776 e з Claro, Pto. Triunfo- Medellin, S of Hwy., 5%54'N, 74*51'W, Croat 56579 (МО); Ко Sucio, Dabeiba, нь & Barkley 1822 (COL, US), Johnson & Barkely 18C411 (US); Dabeiba- Chigorodó, 4 km NW of Dabeiba, 180-400 m, Callejas et al. 4783 (MO); Mpo. Chigorodo, Plowman 3175 (GH). 6?16'N, negro, Rio La Vieja, La María, 1,000 m, Arbeláez et al. e oun Mpo. Turbo; Km 11 on Carretera Tapón del n, Leon- Lomas Aisladas area, 10-20 m, Brand & Күш 727 (MO). Costa RICA. ALAJUELA: М of vd Croat 43640 (MO); Naranjo-Quesada, Hwy. 15, 3.2 m of Zapote, 1,560 m, Croat 46916 (MO); Cañas- TN Road, N of Bijagua, Croat 36480 (COL, MO, NY); Río Cacao, 800 m, Gómez 1 Zarcero, Finca L 84°27'W, tanillo, Croat 36808 (MO); NE of Tapanti, Lent 1034 (GH). GUANACASTE: Tilaran, Standley & Valerio 44445 (US); Cerro Nubes, W slope, 2 km E of Silencio de Tilaran, 900 m, 10?28'N, 84*53'W, Grayum et al. 4999 (МО); Lake Arenal, Río Chiquito- Rio Cano Negro, 550-850 m, 10?27'N, 84°50'W, Hammel et al. 15146 (МО); Parque Nacional Rincón de la Vieja, SE slopes of Volcán Santa Мапа, above Estación Hacienda Santa María, 900- 1,200 m, 10?47'N, 85°18'W, Davidse et al. 23335 (CR, о i e Croat 32898 (MO); Cordillera de Tala- ‚ Rio Guineal, Helechales, 1,100-1,200 m, 9204: "30'N, 85°05'W, Davidse & Баага 26240 (МО); Monteverde, Rio Guacimal, Lechería, 1,500 m, 10%17'N, 84°48'W, Hammel & Trainer 13830 (MO). GUATEMALA. ESCUINTLA: Escuintla-Sta. Lucía Cotz, Standley 63469 (F). RETALHULEU: Ајаха, Standley 88227 (Е); W of Re- talhuleu, Standley 87499, 88391 (F). SACATEPÉQUEZ: Las Lajas, e ee 58294 (F). SAN MARCOS: La Trinidad, Croat 40902 (CM, К, MEXU, MO, USCG), 40903 (MO). SANTA ROSA: Cuilapilla, —— 78060 (F); Naranjo, 1,200 yde & G, K, NY, US) SUC TEPÊQUEZ: Е of Мага Pereda Croat 32786 (MO). Номо ), Маи .п. (МЕ ‚ 16377 (Е, МЕХ И, MO), 16776 (Е, MEXU), 18043 (МЕХО, ОСТА), 18381 (05, МЕХО, РМА), 18519 (DS, MEXU, UCLA); Finca California-summit of Monte Ovando, Croat 47572, 47573 (МО); Guatimoc, Miranda 1763 (MEXU). Nic- ARAGUA. JINOTEGA: Las Camelias-La Salvadora, rd. from Hwy. 3 through La Fundadora, 1,100-1,150 m, 13°05- 06'N, 85°53-54'W, Stevens & Grijalva 15326 (MO); La Palestrina-La Fundadora, rd. from H ~ 13 de a al, 21 219 (МО). MATAGALPA-JINOTEGA: Que- brad El Quebradón, E s de Penas B Hacienda San Marti 9501,40 00 m, 13?14-15'N, 85°39'W, Stevens & ruber 20892.b, 21035 (MO). nío 718 Annals of the Missouri Botanical Garden SAN JUAN: San Juan del Norte, Smith 87 (MO). zELAYA: Cerro El Inocente, S slope, Stevens 6793 (MO). PANAMA. WITHOUT LOCALITY: Duke 11427 (OS). CANAL AREA: Bal- boa, Standley 28564 (МО); Gatün, hills W of Canal, Standley 27210 (USy Culebra, Gaillard 2698 (US); Frijoles, Pittier 2679 (05); Gorgona-Gatün, Pittier 2278 (05); Las Cascadas, Dodge & Hunter 8651, 8659a, 8662 (MO); Obispo, Standley 31719 (US); Salamanca Hydrographic Station, Dodge et al. s.n. (MO); Summit Gardens, Croat 10281 (MO), 32983 (M, MICH, MO, RSA, TEX); Ancon Hill, Orchid Garden, Bartlett & Las- ser 16609 (MICH, MO); Barro Colorado Island, Bailey & Bailey 72 (BH), Croat 4231, 4554, 6511, 7292, 8495, 8512, 9534, 11325 (MO), 7920, 10195 (MO, PMA), Ebinger 181 (MO), Shattuck 638 (MO); Gatun Lake, Standley 31343 (US); Madden Dam, Azote Caballo Road, Dodge 16579 (МО); Quebrada Ancha, Steyermark & Allen s.n. (MO); Rio Casaya headwaters, E of Gamboa, Nee 9024 (MO); Rio Indio de Gatün, Pittier 2801 (US). CHIRIQUÍ: SW of Portobelo, uid 1075 (MO, US); mining rd. above San Félix, 18-27 mi. off Panamerican Hwy. (above a 1,200-1,500 m, Coat 33085 (B, C, K, MBM, MO, US); Cerro Colorado, Antonio 1527 (MO), Croat 33498 (MO, UCLA); Puerto Armuelles re- gion, W of Puerto Armuelles, San Bartolo Limite, Busey 593 (MO). COLON: Portobelo Region, Rio Guanche, 1-4 km S of Portobelo hwy., 0-50 m, 9?30'N, 79°40'W, Knapp & Schmalzel 3598 (MO), Knapp et al. 4609 (MO); Rio Boquerón, E of Salamanca, 50 m, 9?35'N, 79*32'W, Knapp et al. 5826 (MO, B). DARIÉN: line СС, Duke 5235 (МО); NE of Jaque, Sytsma & D'Arcy 3433 (K, MO); Cerro Sapo, Río San Antonio at base о сегго, 5 km S of Сагасћте, 130 m, 7%59"'М, 78°25'W, Hammel et al. 14802 (MO), 14811 (M, МО); Cerro Tacarcuna, slope, Gentry & Mori 13892 (MO); Serrania de Pirre, Croat 3775 1 (МО). PANAMA: S of Alcalde Diaz, Nee 8566 (MO); Alajuela, Pittier 2344 (US); Torti жне Folsom et al. 6584 ( ); Cerro Brewster, SW of mountain, Lewis et al. 3475 (МО); Rio Маје, Croat 34441 (MO), 34600 (F, K, MO, 05); Serrania de Маје, Folsom & Collins 1701 (МО); confluence of Río Ambroino and Rio Ipeti, 100 m, 8%57'N, 79*32'W, Churchill & de Nevers 4479 (MO) Parque Nacional del Darién, ridge between Rio Topalisa and Rio Pucuro, ca. 13 km E of Pucuro, 8?03'N, 77°20'W, 450-600 m, Hammel et al. 16193 (MO). Anthurium santiagoense Croat, sp. nov. TYPE: Ecuador. Morona-Santiago: Proveduria, con- fluence of Rio Bomboiza and Rio Zamora, 600 m, 3?25'S, 78?27'W, Palacios 1492 (holo- type, MO 3420759; isotypes, NY, QAME). Figure 283. Planta terrestris; internodia brevia, 1.5-2 cm diam.; бе 24.5- em “longus, 1-1.2 cm diam., triangularis; lamina late или ad обопро-е риса, 75-97 ст ird 21.5-29 m lata; pedunculus 20-48 cm longus, 5-11 mm diam., с 1-3 mm longus; spadix кш a 6.5-7 cm longus, 5-6 mm dia Description based on dried material only. Ter- restrial; stem 1.5-2 cm diam.; roots few, pale yish, pubescent, elongate, to 4 mm diam.; cat- iiis probably coriaceous, 10-13.5 cm long, acute at apex, yellow-green, persisting as fine, pale, linear fibers; petioles 24.5-27 cm long, 1-1.2 mm diam., triangular, flattened to broadly sulcate adax- ially, the margins acute, acutely angled abaxially; blades subcoriaceous, + broadly elliptic to oblong- elliptic, abruptly acuminate to long-acuminate at pex (the acumen 15-20 mm long), narrowly acute at base, 75-97 cm long, 21.5-29 cm wide, broad- est just above the middle, matte to weakly glossy, greenish; midrib prominently raised above, higher than broad and paler than surface below; primary lateral veins 25-33 per side, departing midrib at 60-75? angle, straight-ascending to the collective vein, slightly raised above and below, slightly paler than surface below; interprimary veins almost as conspicuous as primary lateral veins, drying raised above and below; tertiary veins visible when dried, raised; collective vein arising from near the base, drying raised above and below, equally as promi- nent as primary lateral veins, 7-20 mm from mar- gin. Inflorescences with peduncle 20-48 cm long, 5-11 mm diam., drying greenish; spathe reflexed, subcoriaceous, green, broadly lanceolate, 7-8 cm long, 1.8-2.6 cm wide, broadest near the base, abruptly acuminate at apex, 4 mm long; stipe 7- 7 mm long in front, 1-3 mm long in back; spadix deep purple, vba erect, sometimes slightly curved, 6.5-7 c , 9-6 mm diam. midway; flowers + square, i ies 21 mm in both directions; 7-10 flowers visible in principal spiral, 6-7 in alternate spiral; lateral tepals 0.9-1.2 mm wide, the inner margins + straight, the outer margins 2-sided; pistils somewhat exserted; stigma ellipsoid, 0.3 mm long, droplets drying as abundant crystals and persisting; stamens emerging well above tepals; filaments translucent, flattened, 0.7 mm wide; an- thers yellow, 0.4–0.5 mm long, 0.6 mm wide; 0.2-0.3 mm wide, slightly divari- cate; pollen fading to tan. /nfructescence not seen. A member of series Multinervia, Anthurium santiagoense is endemic to Ecuador, where it is known from the Río Santiago watershed (hence the name) in the Serranía de Cutucú and in Morona- Santiago Province at 600 to 1,830 m, in premon- tane moist and tropical moist forest life zones. This species is characterized by its large, un- usually broadly elliptic leaf blades which dry green and have numerous primary lateral veins, its rel- atively long, more or less triangular petioles, and its deep purple, cylindroid spadix. Characteristic also are the exserted stamens and yellow anthers. Anthurium santiagoense is readily recognizable and not likely to be confused with any other species. The only other member of series Multinervia on Volume 78, Number 3 1991 Croat 719 Anthurium sect. Pachyneurium the eastern slope of the Andes which approaches А. santiagoense in size is А. fasciale, which has more or less oblong leaf blades with more irregular primary lateral veins, and petioles that are rounded abaxially. Anthurium narinoense, from the Pacific slope in Colombia, is remarkably similar in overall appearance and differs mainly by its more slender spadix EcUADOR. MORONA-SANTIAGO: Rio Bomboiza and Rio Zamora confluence, Proveduria, 600 m, 3?25'S, 78?27"W, Palacios 1492 (MO, NY, QAME); Cordillera de Cutucú, W slopes, Logrorio- Yaupi, 2°46'5, 78*6'W, Madison et al. 3382 (SEL) Anthurium sarukhanianum Croat & Haager, sp. nov. TYPE: Mexico. Guerrero: 2-3 km N of Zihuatanejo, dry slopes, less than 100 m, Haager s.n. (holotype, MO 3582582; iso types, MEXU, РВ). [Cultivated in Praha and Brno. ] Figures 277, 284 Planta аа, internodia brevia, ad 3 cm diam.; petiolus U-formatus, adaxialiter sulcatus, abaxiliter rotun- datus, 6-10 cm е ngus, 12-17 mm diam.; lamina oblan- ceolata ad anguste obevata, 40-75 cm longa, 17-23 cm lata; nervis primariis lateralibus 9-12 pii аш pedun culus 20 cm longus; spatha erecta, viridis multu ша purpurea introsum, -5 cm longa, 2.5-3. 5 cm lata; spadix clavatus, 4-5 cm longus, 1.3-1.7 mm diam Epiphytic, to less than 1 m tall; stem short, 3 cm diam.; ense, 3-5 mm diam., whitish (green when moistened); cataphylls triangular, 7 cm long, weathering into brown fibers. Leaves erect- spreading; petioles 8-12 cm long, 12-18 mm diam., obtusely U-shaped, deeply sulcate adaxially, the margins erect to incurved (in plants cultivated un- roots der epiphytic conditions), rounded abaxially; gen- iculum shaped like and paler than petiole, 1.2-1.7 cm long; blades moderately coriaceous, oblanceo- late to narrowly obovate, shortly acuminate at apex, acute to narrowly rounded at base, 40-75 cm long, (12)17-23 cm wide, the margins broadly and con- spicuously undulate; upper surface weakly quilted, glossy, medium green, lower surface matte, slightly paler; midrib above convexly raised (with obscure medial rib on drying), becoming convexly raised toward the middle, below prominently convexly raised and weakly speckled; primary lateral veins 9-12 per side, departing midrib at 40—45? angle, moderately straight to the margin, then ascending, convex on both surfaces; interprimary veins lack- ing or few; major tertiary veins weakly sunken above, weakly raised and slightly darker than sur- face below; reticulate veins visible only on lower surface; collective vein absent. Inflorescences erect, much shorter than leaves; peduncle 20 cm long, 5 mm diam., drying 3.5 mm diam., 2-3X as long as petiole, terete; spathe erect, broadly ovate, 3.5- 5 cm long, 2.5-3.5 ст wide, green, heavily tinged or mottled with purple on both surfaces, especially on the inner surface and along margins outside, inserted at 30° angle on peduncle, rounded and retuse at apex (the acumen short apiculate), round- ed to subtruncate at base, the margins tightly re- curled, meeting at 180? angle, then decurrent for a short distance; spadix medium green, clavate, sessile, held at 160? angle from peduncle, 5 cm long, 7 mm diam. near base, 1.3-1.7 mm diam. near apex, broadest just below the apex, broadly rounded at apex, evenly tapered to the base; flowers rhombic to sub-4-lobed, 2.3-2.7 mm long (fresh), 1.8-2.2 mm long (dry), 1.8-2.3 mm wide (fresh), 2-2.5 mm wide (dry), the sides almost straight to smoothly sigmoid parallel to spiral, jaggedly sig- moid perpendicular to spiral; 13-16 flowers visible in principal spiral, to 20 or more in alternate spiral; tepals semiglossy, minutely papillate, with ine sub- rounded inclusions visible through the epidermis, drying matte, light reddish brown, weakly warty, with a thin, loose layer of wax; lateral tepals 1.2- 1.5 wide, shield-shaped, the inner margins almost straight, the outer margins weakly acuminate; pis- tils green, umbonate, protruding weakly into the space between the tepals, the exposed area 0.5 i soon completely obscured by the sta- mens; stigma linear, drying 0.35 mm long; stamens emerging slowly beginning at the middle of the spadix and proceeding toward both ends, the lat- erals preceding the alternates, clustered tightly over the pistil; anthers whitish, 0.5 mm long, 0.6 mm wide, held at the level of the tepals; thecae narrowly ovoid, mm diam., somewhat divaricate. /nfructescence not known. Anthurium затикћатапит is narrowly endem- ic to western Mexico, known only from the type locality in the state of Guerrero, where it is locally common. It can be confused with А. schlechtendalii subsp. jimenezii or А. halmoorei, which occur in similar habitats in western Mexico, but can be distin- guished from either by its conspicuously clavate spadix, a feature unique among Central American species of Pachyneurium. This species was first collected in 1977 by Jiri Haager of Fraga; арынан, and is named in honor of José Sa ‚ Director of the Instituto de Biologia of the Gaiveridad Autonoma де Мех- ico, who assisted in the preparation of Haager's expedition in Mexico. 720 Annals of the Missouri Botanical Garden MEXICO. GUERRERO: 2-3 km N of Zihuatanejo, near sea level, Haager s.n. (MO, MEXU, PR). Anthurium schlechtendalii Kunth, Enum. Pl. 3: 75 а. Anthurium schlechtendalii subsp. schlechtendalii. TYPE: Mexico: Hacienda de La Laguna, Schiede & Deppe s.n. (not seen). Figures 278, 285, 286. Anthurium mexicanum Liebm., Vidensk. Meddel. Dansk Naturhist. Foren. Kjobenhavn 1: 21. 1849. TYPE: Mexico. Veracruz: between Colipa & Misantla (not s Anthurium kunthianum Liehm., Ann. Sci. Nat. Bot. ser. 2. Anthurium tetragonum Hook. ex Schott, Prodr. Aroid. Drawing #541 (Schott pe 541 serves as type). Anthurium fortinense Engl., : Aa : Mexico. а ен Fortín, Kerber 9b (holo e, B). аы Soie de Lundell, Wrightia 3:161, fig. 55. 1966. TYPE: Guatemala. Petén: Tikal National Park, on temple in Group **H," Mar. 1961, Lundell 18198 (holotype, LL). Epiphytic or epilithic; stem usually short, but reported to 40 cm long, 2.5-5.5 cm diam.; leaf scars obscured by root mass, to 4.2 cm wide; roots numerous, dense, ascending to descending, green- ish to tan, smooth, short to moderately elongate, tapered, 3-8 mm diam.; cataphylls moderately co- riaceous, curved, 7.5-16 cm long, caudate-apic- ulate at apex, drying brown, initially persisting intact, but eventually as a reticulum of fibers. Leaves erect; petioles (2.5)10-23 cm long, 4-20 mm iam., quadrangular to trapezoidal, rarely D-shaped or terete, mostly widest adaxially, flattened to shal- lowly sulcate and with sharp margins adaxially, sharply to bluntly ribbed abaxially, rarely rounded; geniculum somewhat thicker and paler than petiole, 1-2 cm long; blades coriaceous, obovate-elliptic to broadly oblanceolate, acute to short-acuminate at apex (the acumen apiculate), acute to obtuse at base, (16.5)30–140(1 75) cm long, (4.5)10-60 cm wide, broadest near or above middle, the margins prominently undulate; both surfaces matte to semi- lossy, medium green above, paler below, drying greenish; midrib flat to weakly raised at base, be- coming more acute (sometimes weakly ribbed), di- minishing and sunken at apex above, raised and square at base below, convexly raised at apex; primary lateral veins 15-16 per side, departing midrib at 50° angle, straight almost to the margin, then arcuate toward the apex, sharply to convexly raised and paler than surface above and below; tertiary veins flat, scarcely visible above, darker than surface below; collective vein arising from near the apex, sunken above, raised below, 2-5 mm from margin. /nflorescences erect to spread- ing, sometimes pendent, shorter than leaves; pe- duncle 10-43 cm long, (4)6-13 mm diam., 1.6- 7.4% as long as petiole, plain green or tinged with violet-purple, terete to weakly flattened and ribbed; spathe spreading to strongly reflexed, coriaceous, plain green or heavily tinged with violet-purple to red or entirely purple, narrowly triangular to lan- ceolate, (8)10-28 cm long, (0.7)1.5-5 cm wide, inserted at 70? angle on peduncle, acuminate at apex (the acumen inrolled), obtuse at base; spadix green to gray to brown to red to purplish violet, tapered, curved, (5.5)8-29 cm long, (7)1 7-20 mm diam. near base, 4-6 mm diam. near apex; flowers square, (2)2.2-2.6 mm in both directions, the sides straight to weakly sigmoid; 7-17 flowers visible in principal spiral, 10-14 in alternate spiral; pas matte, purplish punctate; lateral tepals 1-1.3 m wide, the inner margins convex and turned up against pistil; pistils emergent to 1 mm, dark purple to brown; stigma linear, 0.8 mm long, droplets appearing ca. 4 days before stamens emerge; sta- mens emerging rapidly from the base, lateral sta- mens emerging to midway before alternates emerge; filaments translucent, soon retracting, 0.2-0.5 mm long, 1 mm wide; anthers yellow to pale orange, 0.9 mm long in both directions, inclined over the pistil; thecae + oblong, scarcely divaricate; pollen orange fading to cream, yeasty-scented at anthesis. Infructescence arching-pendent; spathe persisting; peduncle to 4 cm diam.; spadix 15.5-65(80) с long, 4—6(9) cm diam.; berries bright red, obovoid to oblong to ellipsoid, rounded to acute at apex, 10-29 mm long, (5)6- 7 mm diam.; pericarp thick- ened; mesocarp pulpy, white, with numerous raph- ide cells; seeds 2 per berry, greenish white, ovoid to oblong, flattened, 3-6 mm long, 1.5-3 mm diam., 1-1.5 mm thick, astringent to taste. Anthurium schlechtendalii consists of two sub- species. Subspecies schlechtendalii ranges from Mexico (central Veracruz) to Nicaragua on the Atlantic slope, from near sea level to 1,600 m (most common below 1,000 m) in various life zones. Anthurium schlechtendalii is characterized by its trapezoidal to quadrangular petioles, coriaceous spathe, tapered spadix, and bright red berries. is taxon is most easily confused with 4. schlechtendalii subsp. jimenezii, from the Pacific slope of Mexico, which differs by occurring in sea- Volume 78, Number 3 1991 Croat 721 Anthurium sect. Pachyneurium sonally drier habitats, and in its generally smaller leaves and frequently narrowly ovate, usually erect spathe. Anthurium schlechtendalii subsp. schle- chtendalii may be confused with А. upalense in Nicaragua, where they both occur. See the dis- cussion under that species for differences. Anthurium schlechtendalii merges into the range of А. salviniae only in a few areas in the mountains of southeastern Chiapas, such as on Monte Ovando. Specimens collected in the muni- cipios of Siltepec, Escuintla, and Acacoyagua tend to have leaf blades approaching the texture of those of A. salviniae, but the inflorescences and cata- phylls are those of typical А. schlechtendalii. See the discussion following A. salviniae for separation characters for these two taxa. In discussing Mexican aroids, Bunting (1965) included an illustration of a plant cultivated at Cornell University and reputedly collected by H. Moore in Mexico. This is almost certainly A. crena- tum (L.) Kunth and perhaps represents a mixing of labels in the greenhouse. Two vouchers presum- ably prepared from this plant (Bunting 1579 and Nicolson 603) tend to confirm this notion. Ап- thurium crenatum, as far as is known, is restricted to Puerto Rico and the Virgin Islands. Studies of populations of А. schlechtendalii in Mexico and elsewhere show most morphological characters to be variable. The petioles, althoug usually trapezoidal in cross section with the angles moderately acute, may be broadly rounded abax- ially, such as those illustrated by Schott for A. brachygonatum (see Schott drawing No. 473 (W)). Although Bunting (1965) considered the latter dis- tinct from 4. schlechtendalii, it is here believed to be one form of this unusually variable species. eg ы FURTHER LOCALITY: Schipp 5402 (MICH, А UC); Camp Cuello, Lincoln 36 (MO). BELIZE: E 11325 (MICH); Gracie Rock, 1.1-4 mi. S o 100 m, Croat 23851, 23896 mi. 22 on Western Hwy., (MO), Liesner & Dwyer 1472 era Northern River, Gentle 1291 (F, К); Western hway, Gracie Rock, t Whitefoord 317 1 (BM). cayo: Ялан 53 (MICH, МО); W of Hummingbird Hwy. at point 7 mi. S of jct. with Western Hwy. e Transect 1, 90 m, Spellman & Newey 1984 (MO): Waterhole camp, Vaca, Gentle 2567 (MO); Mt. Pine Ridge, Blancaneaux Lodge, Dwyer 12658 (MO); La F i. S of Grano de La Flor, 570-670 m, Croat 23811 (MO). COROZAL: Arnason Lambert 17186 (MO), Gentle 597 (MICH); Tiger Sa vana, 12 km W of Little Belize, 1 , 18°11'N, 88*17'W, Davidse & Brant 32604 ( M, MICH, MISSA, MO, PMA, ANGE WALK Walk-San Pablo, N of August Pine Ridge, road to Trinidad, 100 m, dm idse & Brant 32792 (MO, RSA). TOLEDO: Moffredye Lagoon, Gentle 5372 (LL, TEX) Rideau Camp, 40 m, Whitefoord 1801 (BM); Salamanca, 40 m, Whiteford 1968 (BM); 1.5 mi. S of Mayan village of San José, 5 mi. W of Colombian Forest Station, Croat 24297, 24333 (MO); 6.7 mi. N of Co- lumbia Forest Station, Croat 24431 (MO), Vanderveen 658 (MO); Maya Mts., Boutin & Schlosser 5144 (MO); vic. of junction of Richardson Creek and Bladen Branch, 80-420 m, 16°32'N, 88°45'W, Davidse & Brant 32025 (K, MO, NY, US); Rio Grande, Gentle 4723 (TEX). GUATEMALA. ALTA VERAPAZ: Tucurú—El Estor, above Pa- palha, 15 mi. W of Teleman, Finca Argentina, 250-650 m, Croat 41536, 41551 ( cra La Tinta, Smith 1530a (US); Lago Izabal, Tucurü-El Estor, Route 7E, 150-300 m, Croat 41525, 41529 2 ESCUINTL U m 7 mi. S of Puerto Barrios, 50 m, Croat 41802 (MO); 1 mi. E of Santo Tomás, 4 mi. of Puerto Barrios, 50 m, Croat 41839 (MO); E of El Estor, Jones & Facey 3456 (NY, TEX); Quiriguá Viejo, Standley 24053 (NY, US); Virginia-Lake Izabal, ie ваја 38769 (F); Lago Izabal, Jones & Facey 3437 (NY, TEX). PETÉN: 5 mi. S of entrance to Tikal National Park, Croat 24753 (MO); Dolores, Contreras 2254 (LL); El Paso, Herman 14664, 1467 1 (MICH); Lundell 1573 (MICH); La Libertad, Lun- dell 2873 (MICH); Santa Teresa, Lundell 2708 (MICH), 2889 (MICH, NY); Uaxactun, Bartlett 12295, 12296 (MICH); Lago Macanche, Fallabon road, Contreras 920 (LL) Tikal National Park, Tikal, Bernhardt T6 (MO), Contreras 1618 (LL), Lundell 15410 (LL, TEX), 18198 to Petén Morales, Finca Agua Епа, 150 m, ne 41879 (MO). HONDURAS. COLÓN: Trujillo, Clewell et al. 4335 (MO). COMAYAGUA: Lago de Yajoa, Zonigo 495 (TEFH), 590 (TEFH), 661 (TEFH), 867 (TEFH). coPÁN: E of Copán, Croat 42508, 42521 (MO); 35 km E of Copán, Blackmore & Chorley 3772 (BM). CORTES: 2- 3 mi. SW of Omoa on road from Puert border, 0 m, Croat 42564 (МО); 7 mi. S of 'Portrerillas on Hwy. , 100 m, Croat 42706 (МО); Guatemalan border, Co- Hobo: Nelson et al. 2904 (МО); La Pimienta, Molina 5623 (F). GRACIAS A DIOS: Rio Patuca, 175 m, Clewell 4564 (MO). OLANCHO: Rio Chiquito, 2 km above Río Grande, Blackmore & Heath 1843 (BM, МО); Rio Olan- cho, Gualaco-San Bonito Oriental, 7.4 mi f San Estéban, 540 m, 15?20'N, 85?42'W, Creat & Hannon 64362 (MO). SANTA BARBARA: road on N i p Pedro Sula, ju pui of = San Pedro Sula- - Tegucigalpa Hwy wy 50 15?18'N, 88°25'W, Croat z Hannon pe (MO). YORO: Coyoles, Yuncker et al. K, MO, NY, US); se F uncker et al. ыт ii F, ; Las Min Ruiz 116 (NY); slope above Lake Yojoa, Clewell 31 le 4 poton, 0-100 m, Ma Huixtla- Motozintla de Mendoza, 25 km bon Huixtla, 650 m, Croat 47225 (MOy; Chiapa de o Pichucalco, j . 190, 1,000 . SW of El Jocote on road to Motozintla, 700-900 m, Croat 40706 (МО); 20 722 Annals of the Missouri Botanical Garden mi. N of Ocozocoautla, road to peres near Km 31, 700 m, Croat 40653 (K, n 13 mi. N of Ocozocoautla, road to Арпрас near r Km 900 1 m, Croat 40626 (МО); mi. N of Ocozocoautla, е to Арпрас, 1,000 m, Croat W of Pa- 40587 (MO); Palenque-Ocosingo, 27 mi. of lenque, Hwy. 199, 210 m, Croat 40319 (МО); Palenque- Bonampak, 89-90 alenque, 350-370 m 0 a mi. Croat 40215 (MO); Hwy. 195, 8 mi. N of Pichu- mi of Chiapas border, 80 m, Croat 40074 (MO); San Cristóbal- -Palenque, 97 km NE of San Cris- tóbal, San Cristóbal de Las Casas, 1,130 m, 17%27'N, 92°4'W, Hammel et al. 15616 (MO); Palenque, 170 m, Davidse et al. 20334 (MO); 2 mi. S of Chiapas border along Hwy. i. i 40086 (МО); Motozintla de Mendoza-Siltepec, 2 mi. W of Motozintla, 1,400 m, Croat 47250 (МО); ruins of Bonampak, 670 m, Hoover 282 (MO); Motozintla- Huix- tla, 15 mi. S of Motozintla, 900 m, Croat 40766 (МО); din to Finca Tres de Mayo, 3 mi. NE of El Triunfo, 13 . NE of Escuintla, 350 m, Croat 43879 (F, GH, K, MO, PMA, VDB); Acacoyagua, Matuda 19646 (DS); Angel Albino Corzo, Ton 3880 (MAD, DS); Bochil, Roe et al. 1132 (WIS), Ton 2587 (DS); Chiapa de Corzo, Breedlove 26872 (DS); Escuintla, Matuda 17973, 19645 (MEXU); Huixtla-Motozintla, Breedlove 28609 (DS); Ixtapa, Laughlin 723 (DS, MICH, NY), Laughlin 1599 (DS, MEXU, TEX), Jitotol, Thorne & Lathrop 41341 (DS); Las Margaritas, Breedlove 33199, 33446 (DS); Siltepec, Matuda 18798 (MEXU); ruins of Yaxchilan, Ocosingo, Breedlove 33872 (DS); Ocozocoautla de Es- pinosa, Breedlove 29004, 36555 (DS), 30358 (DS, MO), MacDougall H261 (NY); Pichucalco, Breedlove 35048 (DS), Gilly & Hernández 182 (MICH); Nuevo Acosta Solistahuacan, Breedlove 26042 (DS), Roe 2 a 1239 (US); Raudales Mal Paso, Quintero 3441 (MEXU); Te- nejapa, apris 6242 (DS, F, MICH), uide a MICH, TEX), 350 (DS); Tuxtla Gutiérre a 13 ыз 20 km S of Palenque, road to "осо i 20 m, Mayo 32 (K); Mpo. Allende, El Bosque, 1,219 m, i Sn 51732 ape DS); Mpo. Mapastepec, Sierra E E nusco, road t ara Gutiérrez from Hwy. 200, 5.5 mi. i NW of mot o Mapastepec, 15°32’N, 92°48’ v Croat & pope. (B, MO, NY, US); Mpo. Ocosingo, 5 km SW pe Santo Domingo, 120 km SE of Palenque on road to Bo- nampak, m, Davidse et al. 20429 (МО); Mpo. Ocozocoautla de Espinosa, 3 km N of Ocozocoautla, road 900 m, Breedlove & Smith 21982 (DS, singo, 300 m, Breedlove & Davidse 55398 (CAS). OAXACA: no Sin locality, Calderón 145 (GH); Parque Nacional de Laguna Temazcal, Temazcal, nd of rd. across dam, 150 18?25'N, 96?25'W, Hammel & Merello 15455 (MO): Teotitlan del Camino- Chilchotla, 1.2-3.8 mi. past turnoff to Huatla de Jimenez, 1,265 m, Croat 48374, 48396 (MO); Teotitlan del Cami- no-Huatla de Jiménez, near junction of road to Chilchotla, 1,320 m, Croat 48296 (МО); hill 5 km S of Tuxtepec, 100 m, Conrad & Conrad 3251 (MO); 8.5 mi. NE of Valle Nacional, Pia et al. 231 (ОМО); "ep ca Oaxaca, 0.5-4.5 mi. S of Valle Nacional, 120-4 Croat 39707, 39708, 39738 (МО); 8 km S of e Romero, 100 m, Davidse et al. 30195 (CM, MO); Tux- 6 mi. W of Valle Nacional, (MEXU); Matatengo Gorge, Croat 32712, 32717 (МО); Santo Domingo Tehuántepec, Williams 70, 83 (MICH); Temascal, Sousa 936, 1037, 1801 РА Tuxtepec, Calderón 844 (MICH), Miranda 4274 (MEXU), Moore & Bunting 8907 (BH), Quintero 1820 (MEXU), Rze- dowski 25480, 25483 (US), Sousa 1809, 1825 (MEXU), ај es xri qure 15400 (DAV); queis Rio Ma- i. S а зан гајда, Охрапара, 90-110 05); Тгарісһе де Іа соно 1,000 m, Liebman s.n. (K); Cerro El Cacao (Sta. Cruz), Concepión del Progreso, Риа, 1,020 m, Mendoza 1516-А ( 250 m, Lorence & Jrigos 3093 (МО); Рио. Juchitán, Mpo. Matías Romero, La Laguna- Sarabia, 7 m W of Esmeraldas, 200 m, 17?7'N, 94°49'W, Wendt et al. 3634 (MO); Mpo. Santa Maria Chimalapa, Santa Maria, 300 m, 16%55'N, 94?40.5'W, Hernández 1167 (CHAPA, MO); Dtto. Tuxtepec, E arm of Presa Miguel Alemán, Temascal, 200 m, Neill 5400 (MO); Mpo. Acatlán, Cerro Buenos Aires, intersection 2 km S of Corrales on road to Capilla, 4 km NW of Lindero, 70 m, 18°28'N, 96°38’W, Gereau et al. 2209 (МО); 8 mi. from Acatlán on ranch Campo Chico, Butterwick 59 (LL); Mpo. Soyaltepec, 1 km S of S end of main dam of Presa Miguel Aleman, 140 m, 18?14'N, 96?23'W, Ge- reau et al. 2237 (MO). QUINTANA ROO: 6 km N of Xel- ha, Téllez & Cabrera 3233 (MEXU, MO). тавазсо: SE of Teapa, Km 4 on road to Tacotalpa, 325 m, Croat 47894 (MO); 3 km E of Teapa along road to Jalapa, 40 m, Croat 40109 (MO); Km 34 S of Villa Hermosa, 55 m, Croat 40069 (K, MO, RSA, US); Baláncán, opi 2028, 2052 (ХАТ); NW of Cárdenas, Barlow 26/10 (WIS); 9 km E of Teapa, road to Jalapa, 200 m, us & Madison 343 (K); Cerro Las Campanas, 3 km E of Teapa, 50 km S of Villahermosa, 50-100 m, Conrad et al. 2866 (MO); Río Teapa, 2 km SE of Teapa, 20 m, Davidse et al. 29522 (MO, ЕЗА); Mpo. Tacotalpa, 0.4 km E of Tacotalpa, Cowan et al. 3421 (MO). VERACRUZ: no further locality, Orentt 2950 (MO); Fortin, Kerber 9 (В); 3 km М of Catemaco, 400 m, 18?30'N, 95910", Solheim & Powers 851 (WIS); Cervecería Moctezuma hydraulic plant, 1,000-1,150 m, Croat 39414 (DUKE, ENCB, HBG, МО, ОСА, WIS); Catemaco, Sousa & Sousa 14 (MEXU), Menéndez 7599 (MEXU); von Catemaco, Boege 1257 (CAS); Mun. Cuitlahuac, 250 m r & Nee (MO); Hidalgotitlan, 17°47'N, 94° 1165, 1285, 1674 (MO), 275, 667, 907, 931, 1196 (MEXU, MO, XAL), 458, 1094 (MEXU, XAL), Vázquez 582 (BM, MO, XAL); Hacienda Álvaro Obregón, 149 m, 17°47'N, 94°38'W, Valdivia 291, 421 (MEXU, MO); Laguna, Hidalgotitlán, 160 m, 17°16'N, 94°33'W, Val- divia 689 (МО); 5 km from La Estación de Biología de Los Tuxtlas, Laguna Escondida, 300 m, Soto et al. 16 (MEXU, MO); 170-200 m, Gentry et al. 32391 (MO), Dillon et al. 1838 (MO), Menéndez 157 (MEXU, MO), Cochrane & Cochrane 8625 (MO), Calzada 455 (MEXU, MO); Valle de Córdova, и 1787 (MEXU, MO, P); Córdoba-Veracruz, 1 mi. off Hwy., above San Juan de Gracia, 750 m, Croat 39611 (MO), Miranda 4894 МЕХО), Bourgeau s.n. (К); El Mirador, 1,200 m, Croat 44000 (B, MO, NY, WIS); Mun. Coatepec, 670 m, Cas- tillo & Tapia (MO); Actopán, Ortega- Ortiz 563 (F, XAL), Ortega 556 (MO); Atoyac, Rosas 1126 (F, XAL); Barra Platanar, Dorantes et al. 1287 (MEXU); road to Jalapa, ~ Volume 78, Number 3 Croat Anthurium sect. Pachyneurium 723 Dorantes 505 (CAS, F, MEXU); SE of Jalapa, Barkley et al. 25 96 (MEXU, TEX); Cordillera, Galeotti 6055 17?13'N, 94?13'W, Wendt et al. 4828 (MO); 13. 7 km of La Laguna, 130 m, 17?19'N, 94°22'W, Wendt et al. 5809 (MO), Gilly & Hernandez 253 (MICH); Ori- zaba, Mueller 1335 (K, NY), Botteri & Lumichrast 1602 (P), Mueller 1314, 1335 (NY); Cerro de Nogales, 1,700 m, Matuda 38590 (CAS); Rio La Palma, Laguna де Sontecomapan, Catemaco, 0 m, 18°33’М, 95?00'W, Me- néndez 111 (МЕХ О, MO); Barra de Sontecomapan, Her- nandez 608 (MEXU); Mun. Tapalapa, 500 m, Gomez- Pompa 5105 (МО); vic. of Playa Escondida, 10 km N of Sontecomapan, 1-50 m, Nee 26699 (Е, МО, XAL); Rio Coatzacoalcos, Williams 8940 (MICH, US); Nautla, Fay & Calzada 925 (XAL); Rancho Viejo, Purpus 15711 (UC); Zacualpan, Purplus 1130 (MY, UC); San Andrés Tuxtla, Dressler & Jones 8(MICH, NY, UC), Hernandez & Vásquez 552 (MEXU), Moore & Cetto 6232 (BH, MEXU); Santiago Tuxtla, Sousa 2158, 2373 I Valle de Rio Pescados, Techacastla, 1 km SSE of Jal comuloco, 480 m, 19?21'N, 96?45'W, Iltis et al. 28970 (MO, WIS). YUCATÁN: Darwin & Sundell 2137 (MO); ca. 2 mi. S of Kalcetac at Actum Spukil, Butterwick 88 (LL, MO, TEX); 11 km S of Xcalacoop, Utley & Utley 6495 (МО); Mun. Dzemul, 15 km N of Dzemul, 5 m, Davidse & Davidse 29469 (МО); Chichen Itza- Mérida, 25 mi. W of Chichen Itza, Thompson 437 (MO), Сеет 1124 (MICH); Izamal, Gaumer 741 (Е, МО); S of Mérida, Schott 638 (Е); N of Muna, Luteyn 2528 (DUKE). NICARAGUA. BOACO: Cerro Mombachito, 500-900 m, 12?24'N, 85%32'W, Stevens 16318 (MO), Stevens & Grijalva 14768 (МО). CHONTALES: М of Сиара, Juigalpa- La Libertad, Stevens 4030 (MO). ESTELi: N of Esteli, Kukamonga, 800-840 m, 13?14'N, 86°21'W, Moreno T (MO); 15.8 km N of entrance to Esteli, km 167 n Hwy. 1, 825-840 m, 13°15'N, 86?22'W, mo 5786, 9095 (MO). RÍO SAN JUAN: Río Oyate-San guelito, 100 m, Grijalva & Almanza 3581 (МО). ZELAYA: Kurinwacito, 80-100 m, 13%8'N, 84%55'W, Moreno 23782 (MO); Siuna, Calera, 300-345 m, 13?46'N, 84°46'W, Ortiz 1773 (MO); El Recreo, S of Siuna, Neill 4226, 4521 (MO); Cerro Kana Coperna, 30 km E of Siuna, 250 m, Меш 4527 (МО); Cerro La Calera, 4 km N of Siuna, 350 m, Меш 4289 (MO); Cerro Waylawas, 100-268 m, 13*39'N, 84*48-49"W, Pipoly 4199, 4300, 4360, 4361, 4372 (MO), Stevens 737, 7383, 7386, 8736, 8753, 8755 (MO). b. Anthurium schlechtendalii subsp. jime- nezii (Matuda) Croat, Ann. Missouri Bot. Gard. 70(2): 377. 1983. Figures 279, 280 Anthurium jimenezii Matuda, Anales Inst. Biol. Univ. zu c. Mexico 32: 147. 1962. TYPE: Mexico. Mexico: nca de Zacualpan, 1,300 m, Matuda 37245 оа, МЕХ U). Usually terrestrial ог epilithic; stem less than 30 cm long, ca. 3-4 cm diam.; roots numerous, dense, ascending to descending, tan to greenish, usually smooth, short and thick, tapered, 5-10 mm diam.; cataphylls subcoriaceous, broadly lanceolate, 6-7 cm long, acute to obtuse and weakly apiculate at apex, light green tinged with red, drying brown, persisting + intact at the upper nodes, eventually deciduous. Leaves erect-spreading; petioles 2–17 cm long, 8-10 mm diam., subquadrangular to trap- ezoidal, flattened to broadly and sharply sulcate adaxially, the margins somewhat raised, 2—3-ri abaxially, the surface minutely pale-speckled; ge- niculum paler and thicker than petiole, becoming fissured transversely with age, 1-2 cm long; blades moderately coriaceous, oblanceolate to oblong-ob- lanceolate, acute to acuminate at apex (the acumen apiculate), acute to obtuse to narrowly rounded at base, 36-104 cm long, 6-32 cm wide, broadest well above the middle, the margins undulate; both surfaces glossy to semiglossy, dark green above, paler, sometimes bluish green below; midrib above flat at base, becoming obtusely to acutely angular and then weakly sunken toward the apex, below prominently higher than broad and sharply 2-ribbed at base, becoming prominently and convexly raised toward the apex and paler than surface: primary lateral veins 9-14 per side, departing midrib at 50-70? angle, weakly arcuate-ascending, raised at the midrib, becoming sunken toward the margin above, prominently raised and darker than surface below, much more prominent than interprimary veins; interprimary veins weakly sunken above, prominulous below; tertiary veins weakly sunken above, weakly raised and darker than surface be- low; collective vein arising in the lower half or in the upper third of blade or absent, flat to weakly sunken above, prominulous below, 2-4 mm from margin. Inflorescences erect, shorter than leaves; peduncle 10-54 cm long, equalling or 2-3.4X as long as petiole, medium green weakly tinged red- dish, terete; spathe spreading, subcoriaceous to moderately coriaceous, green weakly tinged with purple (B & К yellow-green 6/5), lanceolate, 4- 8.5 ст long, 1-2.5 cm wide, broadest just above the base, inserted at 30? angle on peduncle, oblique and narrowly acuminate at apex (the acumen in- rolled), subcordate at base; spadix greenish tinged with purple (B & K yellow-green 6/5), weakly and bluntly tapered, curved, 3.5-14 cm long, 5-17 mm diam. near base, 2-5 mm diam. near apex; flowers rhombic to 4-lobed, 2.1-2.6 mm long, 2.7- 2.9 mm wide, the sides straight to jaggedly sigmoid; 10-12 flowers visible in principal spiral, 6-8 in alternate spiral; tepals densely and minutely pa- pillate, sparsely punctate, with numerous droplets at anthesis; lateral tepals 0.8-1.5 mm wide, the inner margins straight to very broadly convex, the outer margins 2—4-sided; pistils weakly emergent, 724 Annals of the Missouri Botanical Garden medium green, darker than tepals; stigma oblong ellipsoid, 0.3-0.5 mm long; stamens emerging promptly in a regular sequence from the base, the laterals preceding the alternates by 6 spirals, the 3rd stamen preceding the 4th by 3 spirals, held just above tepals in a circle around the pistil; an- thers conspicuously white to pale orange (B & K yellow 9/5), 0.7-0.8 mm long, 0.8-1 mm wide, inclined over the pistil; thecae oblong-ellipsoid to ovoid, 0.5 mm wide, slightly or not divaricate; pollen bright yellow fading to white. Infructescence pendent; berries red, oblong-ellipsoid, rounded at apex, 11 mm long, 9 mm diam.; mesocarp with numerous dense raphide cells; seeds 2 per berry, tan, oblong-ellipsoid, flattened, 6 mm long, 4 mm diam., 2 mm thick, with pale punctiform raphide cells, with a sticky, gelatinous apical appendage. Anthurium schlechtendalii subsp. jimenezii is endemic to the Pacific slope of Mexico and is restricted to Guerrero, Oaxaca, and Chiapas in seasonally very dry forest (“‘selva baja cauduci- folia" and "bosque del pino-encino") at 250 to 1,500(2,325) m. Subspe sively terrestrial or epilithic habit, thick, oblanceo- late leaf blades with usually free-ending primary lateral veins, purplish spadix, green, lanceolate ' spathe, and red berries. This subspecies is geographically isolated from the typical subspecies of А. schlechtendalii, which is found in Veracruz and northern Oaxaca and also on the eastern side of the Isthmus of Tehuantepec, but not in western Oaxaca. Subspecies jimenezii differs from the typical subspecies in being gen- erally smaller, occurring exclusively on rocks and in habitats seasonally much drier, and in having usually thicker leaf blades. Matuda (1961) stressed that the taxon has peduncles relatively much longer than those of А. schlechtendalii. While it is gen- erally true that the inflorescences of subsp. jime- nezii are commonly as long as or longer than the leaves, they may also be much shorter. Subspecies Jimenezii is also similar to Anthu- rium halmoorei, which differs in having an ovate to ovate-elliptic spathe and mature berries that are pale greenish yellow. Also similar is А. nizandense, which is endemic to southern Guerrero and south- ern Oaxaca. Anthurium nizandense differs in hav- ing proportionately much longer petioles and ellip- tic to oblong-elliptic blades that are matte on the lower surface. In А. nizandense the blades аге 1.3-2.5 times longer than the petioles, whereas in A. schlechtendalii subsp. jimenezii they are usu- ally 6-10 times longer (rarely only as little as 3 times longer). cies Jimenezii is recognized by its exclu- MEXICO. CHIAPAS: Mpo. Ángel Albino Corzo, Finca Cux- tepec, 1,380 m, Breedlove 50763 (MO); Mpo. Cintalapa de Figueroa, Colonia Francisco I. Madero-Colonia A. López Mateos, 560 m, Breedlove 50560 (MO). eripe Acahuizotla-Agua de Obispo, Moore 5120 (BH, UC) above Acahuizotla, Moore 6957 (BH) Acapulco- -Chil- pancingo, is 5099 (BH) Atoyac-Filo de Caballo, Nueva Deli, 1,300-1,500 m, 17°24’N, 100?17'W, Mil- ~ ~ n of Tierra Colorada, Hwy. 95, 0 m, Croat 45694 (CAS, F, K, M, MO, MEX, US); ил S of Zacualpán, Matuda 37245 (MEXU); Milpillas- Atoyac .7 mi. W of turnoff on road to Chichihualco, 2,325 m, Croat 45621B (MO); Motozintla- етан са. 11 mi. 1 Hwy x 800 ft., Utley & Utley 6813 (MO); е Злий Коа- riguez 51 (US), Rzedowski 22809 ( Н, US); pne Colorada-Agua de Obispo, El Ocotito, Sa & Willis 199 (MICH); Dtto. Mina, Trincheras, Hinton 10125 (GH, K, US). oaxaca: La Galera, along Hwy. 175, 2.1 mi. N of turnoff to Pluma Hidalgo, 1,340 m, Croat 4614 2 (K, MO); Oaxaca-Pochutla, 55.1 mi. 12.9 mi. S of Suchixtepic, Hwy. 175, 46093 (MO); Oaxaca- Puerto Esc ondido, km Rosa, e 19592 (M Juan Guichicovi, 200- а 300 m, 16*58'N, 95%04'W, Wendt et al. 4816 (MO). Anthurium schottianum Croat & R. A. Baker, Brenesia 16(Suppl. 1): 83. 1979. TYPE: Costa Rica. Limón: ca. 1 mi. N of Bribri, ca. 4 mi. SW of Limón, 50-100 m, Croat 43247 (holotype, MO 2584489; isotypes CR, F, K, M, NY, SEL, US). Figures 281, 287. Terrestrial, to 1 or more m tall; stem usually less than 50 cm long, 2.5-5 cm diam.; leaf scars conspicuous, 2-4 cm wide; roots few, descending, green, with flaky or scaly epidermis when dried, often short and blunt, ca. 5 mm diam.; cataphylls coriaceous, hooked, faintly 1-ribbed at apex, 9- 12 cm long, apiculate at apex, dark purple to maroon (B & K purple 2/7.5), drying brown tinged with reddish purple, persisting intact or semi-intact, eventually deciduous. Leaves erect to spreading; petioles 40-150 cm long, 5-10 mm diam., stiff and firm, bluntly D-shaped to subterete, weakly and shallowly sulcate to flattened adaxially, round- ed (sometimes with few faint ridges) abaxially, the surface green, sometimes with dark purple tinge spreading from the base; geniculum thicker and paler than petiole, 2-3 cm long; blades subcoria- ceous, ovate, gradually acuminate at apex (the acumen downturned, long-apiculate), cordate at base, 44—100 cm long, 33-50 cm wide, broadest below or near the middle, the margins shallowly undulate, flat to revolute; anterior lobe 32-80 cm Volume 78, Number 3 roat 725 Anthurium sect. Pachyneurium long, the posterior lobes 12-21 cm long, directed inward, rounded at apex; sinus usually narrowly triangular to closed, sometimes with overlapping lobes; upper surface matte to semiglossy, medium green, lower surface semiglossy, slightly paler; mid- rib above acutely raised at base, becoming sharply and narrowly raised toward the apex and paler than surface, below convexly raised at base, be- coming sharply acute toward the apex; basal veins pairs, not coalesced, sharply and narrowly raised above, less so below; primary lateral veins 9-20 per side, departing midrib at 35-50” angle, slightly arcuate-ascending, faintly loop-connecting, sharply raised above, less prominently and paler than surfaces below; interprimary veins few, nar- rowly raised above and below, paler than surface above; tertiary veins flat to weakly sunken above, somewhat raised below, paler than surface above and below; reticulate veins scarcely visible above, scarcely raised below, prominulous when dried on both surfaces; collective vein arising from 2nd basal vein, slightly sunken above, slightly raised below, obscure above when dried, slightly raised and paler than surface below, 2-6 mm from margin. nflo- rescences erect, much shorter than leaves; pedun- cle sometimes absent or to 20 cm long, 4-5 mm diam., 0.2-0.4x as long as petiole, dark purple, terete; spathe spreading-reflexed to twisted, cori- aceous, dark purple tinged with green (B & K red- purple 2/7.5), ovate-lanceolate, 6.5-15 cm long, 1.6-3.5 cm wide, broadest near base, inserted at 45° angle on peduncle, abruptly acuminate at apex (the acumen inrolled), rounded to subcordate at base, the margins meeting at 90-180? angle and stiffly rolled under; spadix deep red-violet to purple (B & К purple 2/7.5), cylindroid-tapered, 4-14 cm long, 7-9 mm diam. near base, 4-5 mm diam. near apex; flowers rhombic to square, 2.6-3 mm long, 1.5-3.5 mm wide, the sides + straight; 4- 6 flowers visible in principal spiral, 5—8 in alternate spiral; tepals matte, densely greenish punctate, mi- nutely papillate; lateral tepals 1.9 mm wide, the inner margins straight to convex; pistils not emer- gent, rectangular, minutely and barely exposed and somewhat glossy, violet-purple (also reported as white); stigma slitlike; stamens emerging slowly from the base, in a regular progression, weakly exserted then retracting, grouped in a tight cluster above pistil; flaments flattened, translucent, 0.3-0.5 long, 0.6-0.7 mm wide; anthers yellow, 0.5-0.6 mm long, 0.7-0.9 mm wide, inclined over pistil; thecae ellipsoid, scarcely divaricate; pollen yellow (B & К yellow 9/5), sweet-scented; Infructescence with spathe usually persisting; spadix to 35 cm long, 2.5 cm diam., bearing berries in the basal portion only; berries pinkish with metallic sheen, darker in the apical third, obovoid, acute to round- ed at apex, 9-10 mm long; pericarp thin; seeds 2 per berry, light brown, flattened, 6 mm long, 5 mm diam., 2 mm thick Anthurium schottianum is known only from a region of premontane wet forest-basal belt transi- tion in extreme southwestern Costa Rica at less than 100 m. It will no doubt be found elsewhere in Costa Rica and Panama on the Atlantic slope. This species is distinguished by its large, thin, ovate leaf blades with scarcely any posterior rib (i.e., all basal veins are free to the petiole), and by its short peduncle, purple spadix and ovate-lan- ceolate, often purplish, frequently twisted spathe. An unusual feature is the seasonal appearance of several inflorescences at the same time. The species has no known relatives. Costa Rica. LIMON: 1 mi. NE of Bribri, 40 mi. SW of Limón, 50-100 m, Croat 43247 (CR, F, K, M, MO, NY, RSA, SEL, US); Reserva Biológica Hitoy Cerere, Rio Cerere to 1 km upstream from Quebrada Barrera, SW of Valle La Estrella, 90-200 m, 9?40.5'N, 83%2'W, Grayum & Hammel 5764 (B, MO); Río Catarata, N of Bribri, Utley 5500 (DUKE); Rio Sixaola, Bribri-Carib- bean coastal plain, 50-100 m, 9°37'N, 82°49'W, Baker & Burger 69 (MO), Burger et al 10391 (Е, MO), 10300 (MO). Anthurium seibertii Croat & R. A. Baker, Bre- nesia 16(Suppl. 1): 85. 1979. ТҮРЕ: Panama. Chiriqui: wooded slopes on Cerro Horqueta, 1,650 m, Croat 26976 (holotype, MO 2251853; isotypes, B, BM, CM, CR, F, GH, К, M, NY, PMA, RSA, US). Figures 282, 288, 291. Epiphytic or epilithic or terrestrial; stem thick, short, 3-6 cm diam.; leaf scars mostly inconspic- uous, ca. 2 cm wide; roots numerous, spreading to descending, pale green, pubescent, moderately thick and elongate, blunt, 3-6 mm diam.; cataphylls moderately coriaceous, lanceolate, prominently 1-ribbed near the apex, 12-21 cm long, acute at apex with subapical apiculum to 5 mm long, light green, drying tan (B & K yellow-red 8/5), per- sisting semi-intact, eventually deciduous. Leaves erect-spreading; petioles 10-55 cm long, 6-10 mm diam., stiff, firm, пи remera flattened adaxially, the margins sometimes raised, ribbed abaxially, the surface pale-speckled; pe thicker and paler than petiole, becoming fissured transversely with age, 1.5-3 cm long; blades co- riaceous to subcoriaceous, oblong-elliptic, gradu- ally acuminate at apex (the acumen apiculate), acute to obtuse to narrowly rounded at base, 45- 85 cm long, 4-16 cm wide, broadest at or near the middle, the margins flat to slightly undulate; 726 Annals of the Missouri Botanical Garden upper surface glossy to semiglossy, medium to dark green, lower surface matte to glossy, conspicuously paler; midrib convexly raised above, paler than surface, sharply and obtusely raised below, slightly paler than surface; primary lateral veins many per side, departing midrib at 50-60? angle, arcuate- ascending, prominently loop-connecting, promi- nently sunken above, raised and darker than sur- face below; interprimary veins almost as conspic- uous as primary lateral veins; tertiary veins obscure above, weakly visible below; reticulate veins ob- scure above, mostly obscure below, drying partly prominulous on both surfaces; collective vein aris- ing from near the base, equally as prominent as primary lateral veins, sunken above, raised and darker than surface below, 2-7 mm from margin. Inflorescences erect to spreading, shorter than leaves, 1-5 per plant; peduncle 30-60 cm long, 6-10 mm diam.; 1.5-3X as long as petioles, green, sometimes tinged with purple, terete with single dorsal rib; spathe reflexed, often becoming re- curled, subcoriaceous, plain green or tinged with purple (B & К yellow-green 7/10), oblong-lan- ceolate, 10-17 cm long, 2.5-4.3 cm wide, broad- est near the base, inserted at 50? angle on peduncle, gradually and narrowly long-acuminate at apex, rarely short-acuminate (the acumen inrolled), rounded (scarcely decurrent) at base, the margins meeting at ca. 140? angle; stipe 13 mm long in front, 1 mm long in back; spadix violet-purple (B & К blue-purple 2/10), cylindroid-tapered, 10- 20 cm long, 7-13 mm diam. near base, 5-6 mm diam. near apex; flowers 4-lobed at apex, 2-2.5 mm in both directions, the sides jaggedly sigmoid; 7-10 flowers visible in either spiral; tepals matte; lateral tepals ca. 1.5 mm wide, the inner margins straight; pistils rectangular, orange-yellow; stigma rectangular, ca. 0.7 mm long, droplets persisting for several days before stamens emerge; stamens emerging + rapidly from the base, held just above tepals at anthesis; filaments not exposed; anthers pale orange, ca. 0.5 mm long, 1 mm wide; thecae ellipsoid, scarcely divaricate; pollen white. /nfruc- tescence pendent; spathe often deciduous; spadix 2-30 cm long, ca. 2 cm diam.; berries orange, obovoid, flat to rounded at apex, 10-12 mm long; mesocarp mealy, orange, sweet-tasting but p gent; seeds 2 per berry, = oblong with rounded corners, ca. 5 mm long, 4 mm diam., 2 mm thick. un- Anthurium seibertii ranges from central Costa Rica to western Panama at 1,000 to 3,000 m in premontane wet, lower montane wet, and lower montane rain forest life zones. This species is characterized by its oblong-elliptic leaf blades, subquadrangular petioles, violet-purple spadix, and orange berries. Anthurium seibertii is very similar to А. pro- tensum. These species share similar strap-shaped leaves, purple spadices, and orange berries. Ап- thurium ргоіепѕит differs in having a nearly terete petiole, while 4. seibertii has petioles that are quadrangular in cross section and prominently ribbed abaxially. Leaves of Anthurium protensum are also usually pendent and generally smaller than those of 4. seibertii. a m, CosTA RICA. ALAJUELA: San Carlos, Smith de Botanical Garden, m of San m, Croat 57259 (МО); Cerro Burt, upper slopes, 2,00 00 m, 9°0’N, 82°49'W, Davidse et al. 23789, 23790, 23791, 23831 (MO); Cordillera de Pes dis Cerro Frantzius- Cerro Pittier, Rio Canasta, 9.4 air Wo Agua Caliente, 1,500-1,600 m, 9?2'N, 82°50'W, Da- vidse et al. 28395 (МО); Cerro Frantzius to Valle de Silencio, 2,000-2,500 m, 9%3-6'N, 82*58-59'W, Da- vidse et al. 28565 (CR, МО); headwaters of Rio Bella Vista-Sitio Coton on Rio Coton, 1,800-2,200 m, 9°49- 57'N, 82°46-49'W, Davidse et al. 25535 (MO); Tres Colinas, 1,800-1,850 m, 9*7'N, 83°4’W, Davidse et al. 25602 (K, MO, RSA), 25613 (CR, K, MO), 25636 (B, MO); Tres Colinas-Cerro Bekom, 2,300 m, 9?7-10'N, 83*4'W, Davidse & Herrera 261 79, 26183 (CR, MO); Sitio Coto Brus, 1,800-1,900 m, 8°59'М, 82*46'W, Da- vidse 24528, 24534 (МО); Monteverde Reserve, 11?20'N, 84%40'W, Hepper 110 (BM); Rio Виго, re part, 2,010 m, Gómez et al. 21800 (CR, MO); Rio Cotoncito, Las Tablas, Zona Protector a, Chacón et al 1768 (DUKE, General, 1,0 , Skutch 2626 (F, MO). PANAMA CHIRIQUÍ: 1 of Cerro Punta, 1,700 m, McDaniel 10161 (IBE); Bambito, 2,000 m, Cor 14 on road through Bajo Grande, 2,250-2,400 m, 8°50'N, 32'W, ME rid (MO, РМА); Volcán, Antonio 1059 (MO); 3 E of Boquete, E of road Río Palo Alto, нетте sad Maren ú d end of high ri of s 000. mel et al. 691 6 (MO) Boquete Region, en Road, 10 iai above Boquete, , Proctor 31816 (11); Bajo Mono, read Cliquer, 1,500 m, Woodson & Schery 599 (MO); Palo Alto, E of Boquete, 1,670 m, Stern et al. 1076 (МО); Boquete, Monte Rey, Croat 15743 (МО, NY), 15844 (MO), Woodson & Schery 265 (MO); Bo- quete, La Popa, 1,500-2,500 m, D'Arcy 10853 (MO); Boquete, Folsom 2194 (MO); Cerro Horqueta, 1,500- 1,830 m, Croat 26976 (B, CM, CR, F, GH, K, M, MO, NY, PMA, RSA, US), Dwyer 8749 (MO, DUKE); 1,870- 2,470 m, Averett et al. 1084 (MO); NW of Boquete, E slope above Quebrada Horqueta, 8?49'N, 82°29'W, Cochrane et al. 6243 (MO, WIS); Cerro La Trompeta, Palo Alto, 4.5 mi. NE of Boquete, 2,070 m, Hammel 7464 (MO); Cerro Pate Macho, 1,630- Volume 78, Number 3 1991 Croat 727 Anthurium sect. Pachyneurium 1,780 m, 8°46'N, 82?25'W Croat 66358 (МО); m waters of Rio Palo Alto, above Palo Alto, 1,700-2,10 m, 8°47'N, 82?22'W, Knapp «€ Schmalzel 4794 e USM); М fork of Rio Palo Mi Pate Macho, 6 km NE of Boquete, 1,500-1,700 m, Grayum et al. 6361 (CM, MO, NY), Stein et al. 1183 (MO); Cerro Punta, Raul Castro's house, Guadalupe, 2,000-2,330 m, Folsom 6034, 6035 (MO); 0.5 mi. SE of Entre Ríos, 1 mi. b road from Cerro Punta, 2,000 m, Croat 48568 (K, MO, PMA, SEL); 2.2 mi. below Cerro Punta, W side of Volcán Chiriqui, 200 m, Luteyn 902 (MO); Alto los Guerra, road W of Bambito, 1,800-2,000 m, 8°53'N, 82°37'W, Ham- ilton & Krager 3894 (B, K, МО); E of Bajo Choro, N slope of Barú, 2,000-2,170 m, Hammel 2994 (МО); Quebrada Bambito, S of Cerro Punta, Lazor 2727 (MO); Las Cumbres, W of Cerro Punta, 2,470 m, D'Arcy et al. 13196 (MO); Casita Alta, Volcán de Chiriquí, 1,500- 2,000 m, Woodson et 882, 982 (MO); Bambito- Cerro Punta, Volcán de Chiriquí, Croat 1055 1 (MO); Las Nubes, 2,000 m, Croat 26427 (MO, “a Croat 26494 (B, CAS, CM, DUKE, K, MO); N of Audubon к Croat 13618 (МО); along road to сг Punta, W of Boquete, 5.5 km of Alto Quiel, 1 o m, Nee 9945 (МО); E of Cerro Punta, road to Paseo Res- pingo, 2,330 m, Hammel et al. 6623 (MO); Rio o Bajo Mono, NW of Boquete, 2,130 e ras 12761 (MO); Rio Chiriqui 2 upper part, Monte Lirio, 1,300-2,000 m, Allen 1506 (MO), вн 172 (МО, NY), Seibert 289 (МО); jd Boquete, Bajo Chorro, Davidson 225, 276 Mon Dtto. Bugaba, Cerro vos Anthurium selloum K. Koch, Ind. Sem. Hort. Berol. App. 8. 1855. TYPE: St. Johns, Virgin Islands, Krebs s.n. (lectotype, C). Figures 193A, 289. p perde је ih а e Jp z Ind. Occ. : 269. 1806. TYPE: Jamaica? Not s а macrophyllum би) Schott (1829) non D. Don in Sweet (1839) nec Endl. ex Griseb. (1864), Prodr. ы. 516. 1860. Description based on dried material only. Epi- . m., mostly ading, apparently D- shaped, ‘slightly flattened la Eg. sulcate adaxially, with the margins acute- ly raised, rounded abaxially; geniculum thicker and paler than petiole, 1–1.5 cm long; sheath 4 cm long; blades subcoriaceous, ei some- what triangular, acute to weakly and shortly acu- minate at apex (the acumen ca. 10 mm long), broadly and shallowly cordate at base, 36-46 cm long, 14-19 cm wide, broadest in the basal third, the margins conspicuously and broadly undulate; apex of posterior lobes rounded; sinus parabolic- arcuate, 3 cm deep; both surfaces weakly glossy, yellowish brown and greenish in part; midrib ap- parently convexly raised above and below; major veins convexly raised, paler than surface and yel- owish above and below; basal veins 3—4 pairs, 2nd & 3rd to 4th coalesced for up to 1.3 cm; primary lateral veins 3—5 per side, departing midrib at 25— 45? angle, mostly straight, sometimes arcuate to the margin; interprimary veins sometimes present, ranging from not as conspicuous as to almost as conspicuous as primary lateral veins; tertiary veins conspicuous above and below; reticulate veins ob- scure above, weakly visible below; collective vein arising from uppermost У; or / of the blade, less prominent than primary lateral veins, to 11 mm from margin. /nflorescences apparently + spread- ing; peduncle 39-73 cm long, 2-4 mm diam., 1.1-2.1X as long as petiole, pale green, subterete; spathe semi-erect to reflexed, thinly coriaceous, pale green, oblong-lanceolate, 7.5-13.5 cm long, 0.8-1.4 cm wide, broadest near the base, inserted at 40—50? angle on peduncle, о acuminate to caudate at apex (the acumen 10-20 mm long), acute to obtuse at base; spadix dark maroon Supe to brownish purple, long-tapered, sessile, held at 150-170? angle from peduncle, 14-27 cm long, 6-8 mm diam. near base, 2-3 mm diam. near apex, broadest near the base; flowers rhombic, 1.6- 2.4 mm long, 1.7-2.8 mm wide, the sides jaggedly to smoothly sigmoid; 7—8 flowers visible in principal spiral, 4—6 in alternate he ind tepals roughened, lateral tepals 1.1-1. wide, the inner margins usually broadly convex, sometimes straight, the outer margins 2-sided; stigma broadly ellipsoid, 4–0.6 mm long, densely brush-like; filaments apparently flattened, 0.4–0.5 mm long, 0.5-0.7 mm wide; anthers yellow, 0.5-0.6 mm long, 0.4— 0.5 mm wide; thecae oblong, not divaricate. In- fructescence not seen. Anthurium selloum is endemic to the West In- British Virgin Islands (St. John and Tortola, re- — y). 5 species is distinguished by its somewhat triangular, shallowly cordate leaf blades broadest in the lower third and with 3-4 pairs of basal veins, and by the long-tapered, dark maroon-purple spa- dix. It is unlikely to be confused with any other species in the West Indies, or anywhere else. Engler (1905) questionably placed Pothos mac- rophylla Sw. in synonymy with 4. selloum, per- haps based on Schott's (1860) placement of 4. selloum in synonymy with A. macrophyllum Schott. BRITISH VIRGIN ISLANDS. TORTOLA: High Bush, 75 m, Britton & Shafer 714 (NY, US). U.S. VIRGIN ISLANDS. ST JOHN: Raunkiaer 2669 (C); Bordeaux, 300 m, Britton 728 Annals of the Missouri Botanical Garden & Shafer 560 (NY, US), Krebs s.n. (C). WEST INDIES. WITHOUT LOCALITY: Martens s.n. (BR), Brown s.n. 5 Anthurium simpsonii Croat, sp. nov. TYPE: Peru. Tumbes: Dtto. Matapalo, road to Campo Ver- de, Simpson & Schunke 392 (holotype, F 1771582). Figure 290. Planta epiphytica; internodia ca. 1.5 cm diam.; cata- phyllum persistens velis fibrarum reticularum; petiolus 6 cm longus, ca. 6 mm diam., D-formatus; lamina oblongo- oblanceolata, 66 cm longa, 8.5 cm lata; pedunculus 23- 5 cm longus; spatha oblanceolata, ca. 5 cm longa, ca. 8 mm lata; spadix purpureus, cylindricus, ca. 5 mm diam. Description based on dried material only. Epi- phytic; stem ca. 1.5 cm diam.; roots numerous, dense, gray-brown and shortly woolly-pubescent, moderately thick and elongate, somewhat tapered at apex, 3-5 mm diam.; cataphylls subcoriaceous, 11.5-12 cm long, acuminate at apex, tan, per- sisting intact, eventually as a reticulum of fibers. Leaves erect or somewhat spreading; petioles 6 cm ng, ca. 6 mm diam., D-shaped, probably flattened Mu. rounded or multi-ribbed abaxially; ge- niculum slightly darker than petiole, ca. 1 cm long; ades subcoriaceous, oblong-oblanceolate, acute at apex (the acumen apiculate), acute at base, 66 cm long, 8.5 cm wide, broadest at or above the middle; upper surface glossy, lower surface semiglossy; both surfaces matte, grayish green; midrib flat to con- vexly raised above, concolorous with the surface, prominently and convexly raised below, paler than surface; primary lateral veins ca. 22 per side, de- parting midrib at 50—60° angle, + straight to the collective vein, raised above and below, slightly paler than surface below; interprimary veins prom- inulous below; reticulate veins visible below; col- lective vein arising from near the base, raised above and below, less prominent than primary lateral veins, 1-3 mm from margin. Inflorescences with peduncle 23-25 cm long, ca. 3 mm diam., ca. 5 X as long as petioles, terete; spathe, yellowish brown, oblanceolate, ca. 5 cm at base, the margins meeting at an acute angle; stipe 7 mm long in front, 3 mm long in back; spadix purplish, cylindroid, erect, ca. 5 mm diam.; owers + square; 5-7 long, 0.8 cm wide, acute owers visible in principal spiral, 7-9 in alternate spiral; tepals deep purple with light tips; lateral tepals 0.9-1.2 mm wide, the inner margins straight, the outer margins 2-sided; pistils glossy, wine-red; stigma probably slitlike; anthers 0.4 mm long, 0.4-0.5 mm wide; thecae oblong, 0.2 mm wide, slightly or not at all divar- icate. Infructescence with spathe persistent; spadix 12.5 cm long, 1.8 cm diam.; berries purple, ovoid, acute at apex, 6.5 mm long, 5.5 mm diam.; me- socarp translucent, gelatinous, orange; seeds 2 per berry, yellow-brown, flattened, 3.5-3.7 mm long, mm thick, enveloped by gelatinous, translucent substance. mm diam., 0.9- Anthurium simpsonii is known only from the type collection from the Department of Tumbes in northern coastal Peru, at 600 to 800 m in the subtropical dry forest life zone. ts small size, very short petioles, and oblong-oblanceolate blades with collective veins arising from near the base, as well as by the long, rather slender spadix, deep purple flowers, and wine-red fruits. It is not likely to be confused with any other species. The onl other species of sect. Pachyneurium that occurs in this part of Peru is Anthurium barclayanum, a much larger plant with leaves drying dark brown (not grayish) and collective veins arising from near the apex. The new species is named in honor of Donald Simpson, formerly of the Field Museum of Natural History (F), who collected the type when he headed the Flora of Peru Project. is species is characterized by i PERU. TUMBES: Prov. Zarumilla, Dtto. Matapalo, Bos- que National de Tumbes, Campo Verde, 600-800 m, Simpson & Schunke 392 (F) Anthurium solitarium (Vell. Conc.) Schott, Prodr. Aroid. 478. 1860. Figures 292, 293, 295 Pothos solitarius Vell. Conc., Fl. Flum. 9: t. 123. 1825 1829]. TYPE: Brazil. Rio de Janeiro. (The plate cited Anthurium е Hook. f., Bot. Mag. t. 6833. 1885. PE: Brazil. Rio de Janeiro, Glaziou 17333 (ho- lot e B, photo seen; isotype, Ра nobile Engl., Bot. Jahrb. Syst. 25: 366. 1898. TYPE: Brazil. Rio de Janeiro, Glaziou 9039 (holo- type, B; MO, photo). Epiphytic or epilithic; stem 1.5-4 cm diam.; roots descending, velutinous, appearing inflated or with air spaces inside when dried, sharply tapered at apex, 2-9 cm long, 3-10 mm diam.; cataphylls lanceolate, 6-11 cm long, acuminate at apex with subapical apiculum to 5 mm long, drying brown, persisting semi-intact at apex and eventually weath- ering into coa linear fibers. Leaves erect- spreading; petioles (2)6-28(32) cm long, 5-9 mm diam., spreading, C-shaped to thicker than broad, narrowly and deeply sulcate with obtuse to acute margins adaxially, rounded abaxially, the surface dark green, pale-speckled; geniculum slightly thicker and paler than petiole, 1–1.5 cm long; sheath 2.5-4 cm long; blades subcoriaceous to rse, Моште 78, Митбег 3 1991 Croat 729 Anthurium sect. Pachyneurium coriaceous, obovate to oblanceolate, often some- what elliptic, short-acuminate to acute and apic- ulate at apex, obtuse to semi-rounded or weakly subcordate at base, 26-70(80) cm long, (6)10- 23(27) cm wide, broadest at or near the middle, the margins weakly and broadly undulate; upper surface semiglossy, medium green, lower surface weakly glossy, paler below; both surfaces drying matte to semiglossy, yellowish brown; midrib flat to obtusely raised or obtusely sulcate at base, be- coming convexly raised, sometimes sharply acute toward the apex above, obtusely to acutely raised, sometimes faintly 1-3-ribbed at base below; pri- y lateral veins 5-12 per side, departing midrib at at (25)40- 60(70)° angle, + straight, becoming arcuate near margin, raised in grooves or weakly to prominently raised on upper surface, promi- nently raised and stout below, paler than surface; interprimary veins weakly raised above and below when visible, darker below; tertiary veins promi- nulous, weakly raised to weakly sunken above, moderately visible or obscure above, very weakly raised below; collective vein arising from near the apex or absent, if present sunken above, raised below, 10-20 mm from margin. Inflorescences erect or erect-spreading, or somewhat pendent; peduncle (15)30-102 cm long, 3-5 mm diam., 1.3-16X as long as petiole, green sometimes tinged with purple, terete, pale- to dark-speckled; spathe spreading, coriaceous, green tinged with purple adaxially, oblong to lanceolate, 6-25 cm long, 1.5- cm wide, broadest near the base, acuminate at apex (the acumen inrolled, 5-10 mm long), obtuse to acute at base, often decurrent; spadix dark ma- roon or purple, mostly subsessile (rarely stipitate to 30 mm), tapered, curved, 12-33 cm long, 4- 14 mm diam. at base, 3-7 mm diam. at apex, broadest at the base; flowers rhombic, 1.8-3.6 mm long, 1.2-2.6 mm wide, the sides + straight to sigmoid; 6-10 flowers visible in principal spiral, (6)10-12 in alternate spiral; tepals matte; lateral tepals 1-1.6 mm wide, the inner margins broadly convex, the outer margins 2-sided; pistils not emer- gent; stigma linear becoming ellipsoid, possibly be- coming circular, sometimes appearing somewhat caviform when dried, 0.4-0.7 mm long; stamens emerging in a regular sequence; the laterals pre- ceding the alternates by 13-20 spirals, held at edge of the tepals, inclined over and obscuring the pistil; filaments not exserted; anthers 0.4–0.5 mm long, 0.8-0.9 mm wide; thecae ovoid, 0.3-0.4 mm wide, slightly divaricate; pollen golden. /n- fructescence with spathe persisting; spadix 19-40 cm long, 11-18 mm diam.; berries purple-red, more or less oblong, the basal tepalar fibers strongly adherent, 5-6 mm long, 2-3 mm diam.; pericarp with raphide cells; mesocarp + dry; seeds 1-2 per berry, covered with raphides, ellipsoid, 4-4.4 mm long, 1.8-2 mm diam., 0.8-1 mm thick, with a gelatinous appendage. Anthurium solitarium is known from the states of Espirito Santo, Goiás, Minas Gerais, and Rio de Janeiro in Brazil from 300 to 1,300 m, where it is epiphytic or epilithic in mainly primary forest habitats This species is recognized by its thickly coria- ceous, obovate or oblanceolate blades, with the primary lateral veins running straight from the midrib to the margin, and then steeply and ar- cuately rising along the margin. Also distinctive is the maroon to purple spadix and prominently de- current spathe. Anthurium solitarium is not easily confused with any other Brazilian species or with species from other areas. Several collections are noteworthy. Glaziou 9039, the type of Anthurium nobile (here syn- onymized), is unusual in having a prominently stip- itate (nearly 3.5 cm) spadix, instead of the more common sessile to weakly stipitate condition. Croat 23794, a sterile collection from cultivation, may possibly represent a distinct species. It differs in having a thicker blade with less conspicuous ter- tiary veins, as well as having the primary lateral veins and the midrib drying pale and wrinkled. Croat 53699, also of cultivated origin, is unusual in having a collective vein arising from one of the primary lateral veins almost at the middle of the blade. Generally, the collective vein of А. solitar- ium 18 lacking or strictly apical. BRAZIL. BAHIA: Santa Cruz de Cabrália (cultivated at Sitio Burle-Marx), 16?18' : с we Наг rley et al. 20316 (К). ESPÍRITO SANTO: Domingos ns, property of Rob- erto Kautsky, Croat 57165 (la (MO), 61808 (MO, R), 61817 (MO, R); Dom ns-Santa Мапа olatina- Me 45 km E и Colatina, Belem 3833 (NY), Croat 6 din В); Domingos Martins-Santa Maria, 845 m 194 C d at Kew, #275-79-02469) с GOIÁS: 12 km 5 of Caiaponia, Hutchison 8505 (MO). MINAS GERAIS: Paraibuna, ipi 7660 (RB, ОВ); 17 km > of Cam buquira, 1 to Caxambu, 950 m vidse & Ramamoorthy 300700407. north of са Clayton 9/22/1951 (NY). RIO DE JANEIRO: Glaziou 9039 (B), 17333 (K); cultivated by Burle-Marx, San Carlos, Croat 53699 (MO, NY, RB), 53708, 53711, 53712, 53713 MO, RB); Cultivated by Luis Gurken, San Carlos, Croat 53794, 53792 (MO); Friborgo, 1,000 m Croat 53793 (MO); Serra da Carioca, Rio de Janeiro, Macico da Tijuca, 300 m, Croat 53670 (MO); Serra dos Orgáos, Luetzel- burg 6061 (RB); E of Rio Maio, Teresopolis, Vidal 953 ~ 730 Annals of the Missouri Botanical Garden (RB); Neves Armond, Carmo, RB 43978 (RB); Mpo. Petropolis, Rodovia Washington Luis, 420 m, Martinelli 3080 (RB); Reserva do IBDF, Mata do Facao, estrada Fazenda Inglesa para Pati do Alferes, 1,100-1,200 m, Mpo. Teresopolis, Parque Nacional da Serra dos Orgaos, 1,300 m, Vidal 5557 (R). е solomonii Croat, sp. nov. ТҮРЕ: Bo- ivia. az: Prov. Yungas: 32.1 km N of Yo qa on road to Caranavi, Solomon & Es- cobar 12494 (holotype, MO 32475329-30; isotype, LPB). Figures 296, 345, 346. Planta epilithica aut terrestris; internodia brevia, ad 5 cm diam.; cataphyllum ad 19 cm longum, persistens semi- intactum demum fibris linearis findens; petiolus 7-21 cm longus, 8- 10 mm diam., D-formatus, 1-costatus adaxile, 1 -2-costatus abaxile; lamina oblanceolata ad oblongo-ob- lanceolata, (39)65-94 cm longa, (9)18-35 cm lata; pe- dunculus (4)10-26 cm longus, ca. 5 mm diam.; spatha lanceolata, utra 6 cm longa, ca. 1.7 cm lata; spadix longeus contractus vel cylindricus, 7.5-19.5 cm longus, 9 mm diam.; baccae laete roseo-purpureae. Description based on dried material only. Epi- lithic or terrestrial; stem to 5 cm diam.; roots numerous, gray, finely pubescent, short, 2-4 mm diam.; cataphylls probably subcoriaceous, to more than 19 cm long, brown-gray, persisting semi-in- tact, eventually as fine linear fibers. Leaves with petioles 7-21 cm long, 8-10 mm diam., D-shaped, with an obtuse medial rib and acute, somewhat raised margins adaxially, 1-3-ribbed abaxially; geniculum drying darker than petiole, 0.7-1 cm long; sheath 6 cm long; blades subcoriaceous, ob- lanceolate to oblong-oblanceolate, acute at apex (the acumen apiculate), narrowly rounded at base, (39)65-94 cm long, (9)18-35 cm wide, broadest above the middle, the margins undulate; both sur- faces glossy, matte to glossy, brown to greenish; midrib sharply and prominently acute above, below higher than broad and obscurely 3-ribbed at base, becoming convexly raised toward the apex; pri- mary lateral veins (10)15-19 per side, departing midrib at 40—90° angle, straight, sometimes broad- ly arcuate, convexly raised on both surfaces; in- terprimary veins few; tertiary veins raised above and below; reticulate veins raised below; collective vein arising from lower third of the blade or absent in young plants, less prominent than primary lateral veins on both surfaces, 2-7 mm from margin. Inflorescences shorter than leaves; peduncle (4)10— 26 cm long, са. 5 mm diam., 0.5-2* as long as petiole; spathe subcoriaceous, presumably lanceo- late, more than 6 cm long, ca. 1.7 cm wide, de- current at base; stipe ca. 2.7 cm long in front, 4 mm long in back; spadix color unknown, long- tapered to + cylindroid in small plants, weakly curved, held at 160? angle from peduncle, 7.5- 19.5 cm long, 9 mm diam. near base, 4 mm diam. near apex; flowers rhombic, 2.1-2.4 mm long, 1.4-1.7 mm wide, the sides straight to smoothly sigmoid; 9-11 flowers visible in principal spiral, 5-6 in alternate spiral; tepals roughened; lateral tepals 0.9-1.5 mm wide, the inner margins straight to broadly convex, the outer margins 2-sided; stig- ma linear, 0.5-0.7 mm long; anthers pale yellow, 0.5-0.6 mm long, 0.7-0.8 mm wide, obscuring the pistil; thecae oblong, slightly divaricate. /n- fructescence appearing to be + erect; spathe de- ciduous; spadix ca. 21 cm long, 2 cm diam., reddish brown (excluding tepals); berries bright pinkish pur- ple, oblong-ovoid, probably acute and with radial ridges at apex, becoming sunken in on drying, 7.5- 8 mm long, 4.6-5 mm diam.; pericarp thickened, with numerous raphide cells; seeds 2 per berry, reddish brown, + ovoid, flattened, 3.4-3.6 mm long, 2.2-2.4 mm diam., 0.8-1 mm thick, with а gelatinous, translucent, amber appendage at each en Anthurium solomonii is known only from La Paz Dept. in Bolivia at 1,000 to 1,400 m in wet subtropical or wet lower montane subtropical forest life zones (according to the Holdridge Life Zone Map for Bolivia). This species is distinctive, with its spadix longer than the peduncle, a character otherwise known in the section only in Anthurium plowmanii. The petiole of A. plowmanii is narrowly sulcate adax- ially with blunt margins, whereas in A. solomonii there is a prominent medial rib adaxially and prom- inently raised, acute margins. In Bolivia, А. plow- manii is known only from Pando and Santa Cruz departments and occurs at less than 900 m, mostly less than 300 m. The species is named in honor of James Solomon, who has collected extensively in Bolivia and was involved in making all known collections of 4. solomonii. BOLIVIA. LA PAZ: Prov. Murillo. Valle de Zongo, 45 km below the dam at Lago Zo pos Cahua hydroelectric plant, 1,200-1,400 m, 16?3'S, 68*1'W, Solomon 13000 (МО); Prov. Nor Yungas, Yolosa-Caranavi, 32.1 km N of Yo- а, 1,000 m, 16?2'S, 67°39'W, Solomon & Escobar 12494 (LPB, MO); 10 km by road N and above Caranavi, 1,400 m, 15?47'S, 67°32'\/, Nee & Solomon 30311 (NY). 5 D Anthurium soukupii Croat, sp. nov. TYPE: Peru. Cuzco: Urubamba, Machupicchu, 0.5 km N of the union of Rio Sayacmarca and Rio Ao- bamba, 2,370 m, Peyton & Peyton 1486 Volume 78, Number 3 1991 Croat Anthurium sect. Pachyneurium (holotype, MO 3024625-6). Figures 294, 297, 298 Planta terrestris; caulis elongatus, ad 50 cm longus; internodia 1.5-2.5 cm diam.; cataphyllum Кезге, persistens in fibris linearis; petiolus 6.5-11 ст ongus, ca. 5-7 mm diam., triangularis; lamina late ad anguste lat oblanceolato- -elliptica, (22.5)28-63 cm ee (3 .8)4 4.7- 10 ст lata; pedunculus 22-40 cm longus, 2-4 mm diam.; spatha lanceolata, 4.5-8 cm longa, 7-14 mm lata; spadix atropurpureus leviter contractus, 3.5-6 cm longus, ca. mm diam.; baccae globosae, 3.2-4 mm longae Terrestrial; stem elongate, to 50 cm long, 1.5- 2.5 cm diam.; roots moderately numerous, dense, white, smooth, thick, + elongate; cataphylls lan- ceolate, probably subcoriaceous, to 11 cm long, acute at apex, drying yellowish tan, persisting as fine linear fibers. Leaves spreading; petioles 6.5- 11 cm long, ca. 5-7 mm diam., triangular, flat- tened adaxially with the margins moderately to sharply raised, sides weakly convex, obtusely rounded to acutely angular abaxially; geniculum not at all or only slightly thicker than petiole when dried, 0.7-1 cm long; blades coriaceous, broadly to narrowly oblanceolate-elliptic, acute at apex, acute to obtuse at base, ( 8-63 cm long, (3.8)4.7–10.7 cm wide, broadest above the middle, the margins straight to weakly undulate; upper surface semiglossy, medium green, paler below; both surfaces drying matte, occasionally semiglossy below, green to yellowish; midrib convex- to round- raised above, sharply acute below; primary lateral 20 per side, de- straight, obscure veins numerous, to more than parting midrib at 50—65? angle, + to raised and darker than surface above, flat to weakly raised below; interprimary veins numerous, almost as conspicuous as primary lateral veins on both surfaces; tertiary veins not visible on either surface; collective vein arising from near the base, sometimes in the upper third of the blade, equally as prominent as primary lateral veins on both sur- faces when dried, 3-12 mm from margin. /nflo- rescences erect to spreading; peduncle 22-4 long, 2-4 mm diam., 2.6-4.2 x as long as petiole, apparently ribbed abaxially; spathe reflexed, sub- coriaceous, green, lanceolate, 4.5-8 cm long, 0.7— 1.4 cm wide, broadest near the base, acuminate at apex, acute at base; stipe to 1 cm long in front, to 3 mm long in back; spadix dark purple, weakly tapered, stipitate to sessile, ca. 3.5-6 cm long, ca. diam. near base; flowers + square 3.1- 4.3 mm in both directions, the sides straight to smoothly sigmoid; 5-6 flowers visible in principal spiral, 6-7 in alternate spiral; lateral tepals 2.3- 3 mm wide, the inner margins straight, slightly turned up against the pistil, the outer margins 2-sided; pistils drying blackened, emergent well above the tepals, green; stigma linear, 0.5-0.6 mm long; anthers pinkish, ca. 0.8 mm long, ] mm wide; thecae ovoid, slightly divaricate. Infructesc- ence with spathe absent; spadix 3-13.5 cm long, 0.8-1 cm diam.; berries globose, rounded at apex, 3.2-4 mm long, 3.7-4.5 diam.; pericarp thickened, with numerous pale raphide cells; me- socarp translucent, gelatinous; seeds 2 per berry, reddish brown, ovoid, flattened, dried 3-3.5 mm long, 2.3-2.4 1.1-1.3 mm thick, en- veloped by gelatinous, sticky, amber substance. mm mm diam., A member of series Multinervia, Anthurium soukupii ranges from La Paz in Bolivia to Zamora- Chinchipe in Ecuador, at (1,700) 2,000 to 2,400 m in subtropical lower montane moist forest, lower montane wet forest, subtropical montane wet for- est, and high montane wet forest life zones. This species is characterized by its elongate stem, subtriangular petiole, oblanceolate-elliptic blades with more than 20 primary lateral veins and a collective vein arising from near the base, and by its slightly tapered, purple spadix with 5-6 flowers per spiral. It is probably most easily confused with Ап- thurium fasciale, also from the Atlantic slope in southern Ecuador. That species, in addition to oc- curring at elevations below 1,600 m, differs in having 8-11 flowers visible per spiral and pistils which dry yellow-brown, lack raphide cells on the surface and are round (not at all protruding) at the apex. Anthurium soukupii may also be confused with A. penningtonii, another species from the Atlantic slope, but the latter differs in having more broadly elliptic blades with the primary lateral veins sharply elevated. There is an even closer resemblance be- tween А. soukupii and А. angustilaminatum, from the Pacific slope. See that species for a discussion of the differences. The new species is named in honor of Jaroslav Soukup, author of a book on vernacular names of Peruvian plants and the first to collect А. soukupii. BOLIVIA. COCHABAMBA: Prov. Chapare, San Onafre, 1,700 m, Steinbach 9419 (GH). La Paz: Prov. Sud Yun- gas, Huancane, 13 km toward San Isidro, N of Chulumani, 2,200 m, Beck 8756 (MO) Loja-Zamora, 13 k ve Lago Pomacocha, Moyobam b irs rov. Cutervo, Parque Nacional de Cuervo, 2. 300- 2,400 m, Díaz et al. 3964 732 Annals of the Missouri Botanical Garden (MO). cuzco: Río Lucumayo, La Convención, 6 km from Incatambo, 2,350 m, Peyton & Peyton 994 (MO); Prov. Urubamba, 2,100 m, Nuñez 8928 (MO); Machupicchu, on slope 0.5 km N of confluence of Río Sayacmarca and Río Aobamba, 2,370 m, Peyton & Peyton 1486 (МО); 2.5 km from Machupicchu, above lst waterfall of Río Mandor, 2290 m, Peyton € Peyton 455 (MO). JUNÍN: Prov. Oxapampa, Soukup 2324 (GH). Pasco: Cordillera Yanachaga, trail to summit of Cordillera via Rio San Daniel, 2,400 m, 10?23'S, 75°27'W, Smith et al. 7929 (MO). Anthurium sparreorum Croat, sp. nov. TYPE: Ecuador. Los Ríos: Centro Cientifico Rio Pa- lenque, W of laboratory & vic. of laboratory clearing, 210-250 m, Croat 38666 (holo- type, MO 2387864-65; isotypes, B, CAS, CM, К, M, NY, ОСА, ВЗА, 05). Figures 12, 19, 299, 300-302. Planta plerumque epiphytica, subinde terrestris; cata- da longs (9.5)15 10 mm Ший, .; spatha lanceolata vel ovata, (9)12-16 m longa, 2.8-4.5 cm lata; stipes 5 mm longus; spadix ры ad violaceus, снн 5.8-20 cm lon- gus, 8 mm diam.; baccae rubrae Usually epiphytic, occasionally terrestrial; stem short and thick, roots numerous, descending, bright green to green- ish white, densely pubescent, elongate, 3-5 mm diam.; late, conspicuously l-ribbed, 9-18 cm long, acute at apex with subapical apiculum, green, drying pale yellow-tan (B & K yellow-red 9/7.5), persisting semi-intact, eventually dilacerating into reticulum of fine fibers. Leaves erect to spreading; petioles (11.5)15-38 cm long, 6-15 mm diam., U-shaped, flattened and sulcate to weakly convex m long, 2.5-5 cm diam.; cataphylls subcoriaceous, broadly lanceo- with the margins bluntly to onn ipu: ачу, rounded to 1 -3-ribbed abaxially, th ly pale-speckled; geniculum thicker and ee than petiole, 1–2.5 cm; blades coriaceous, oblong-ellip- tic to narrowly oblanceolate to oblong-lanceolate, acuminate to obtuse-rounded at apex (the acumen weakly apiculate), acute to obtuse or cuneate at base, 46-96 cm long, (9.5)15-22 cm wide, broad- est at or near the middle, the margins slightly undulate; upper surface matte to semiglossy, me- dium to dark green, lower surface glossy to semi- glossy, paler; both surfaces drying matte, green to pale yellow; midrib above prominently convex at base, becoming l-ribbed toward the apex, paler than surface and somewhat speckled, below ob- tusely raised at base, becoming prominently convex toward the apex and paler than surface, dryin convex on both surfaces; primary lateral veins (15)20–30 per side, departing midrib at 40—70° angle, straight or arcuate, prominently convex and paler than surface above, less convex to flat and slightly darker than surface below; interprimary veins almost as conspicuous as primary lateral veins, raised above, flat to weakly raised below; tertiary veins obscure above and below, slightly raised when dried on both surfaces; collective vein arising from near the base, occasionally in the upper third of the blade, weakly sunken and less prominent than primary lateral veins above, raised and equally as prominent as primary lateral veins below, 4-15 mm from margin. /nflorescences erect to erect- spreading to spreading, shorter than leaves; pe- duncle 11-34 cm long, 5-10 mm diam., 0.5-2x as long as petiole, medium green, terete, sometimes 1-ribbed below the spathe; spathe reflexed-spread- ing, slightly twisted, subcoriaceous, green (B & K green 6/2.5), broadly to narrowly lanceolate or ovate, 9-16 cm long, 2.8-4.5 cm wide, broadest near the base, inserted at 60? angle on peduncle, acuminate at apex (the acumen inrolled), obtuse at base, the margins meeting at 110° angle; stipe 7-10 mm long in front, to 5 mm long in back; spadix dark purple to reddish violet (B & K purple 3/2.5), cylindroid to short-tapered at apex, + erect, 5.8-20 cm long, ca. 8 mm diam. near base, 4-5 diam. near apex; flowers irregularly 4-lobed to rhombic, 1.8-2.6 mm long, 1.8-2.8 mm wide, the sides jaggedly sigmoid to straight; 9-13 flowers visible in principal spiral, 8-10 in alternate spiral; tepals matte, densely and minutely punctate and papillate; lateral tepals 1-1.2 mm wide, the inner margins straight; pistils scarcely emergent, glossy, papillate, with exposed portion + фр жын, green; stigma linear, dark purple, 0.4–0. long; stamens emerging + regularly and и from the base, laterals emerging throughout, or at least to midway, before alternates emerge, borne at the edge of tepals; filaments translucent to green- ish, minutely purple-spotted, soon retracting, 0.9 mm wide; anthers pale orange with purple punc- tations, 0.5-0.7 mm long, 0.7-0.8 mm wide; the- cae oblong-ovoid, scarcely divaricate; pollen bright orange fading to lavender or pinkish white, faintly yeasty-scented. Infructescence with spathe per- sisting; spadix 20-22 cm long, ca. 2 cm diam.; berries red, exserted slightly beyond the tepals, subglobose, ca. 5 mm diam.; tepals becoming fleshy, white, enlarged, elongate and 3-sided in cross sec- tion with the outer angle of the apex raised and acute. Volume 78, Number 3 1991 Croat 733 Anthurium sect. Pachyneurium Anthurium sparreorum is nearly restricted to the Province of Los Rios in Ecuador, at 210 to 250 m in a tropical moist forest life zone. Ап additional collection is known from Cotopaxi. This species is characterized by its short stem, dense, bright green to greenish white roots, cata- phylls drying pale yellow-tan and persist semi-in- tact, coriaceous leaf blades drying matte and green and reddish violet, cylindroid spadix. Most "os icant is the infructescence, with the berrie beyond the tepals. The latter become fleshy, whit- ish, enlarged and elongated, with the outer edge raised and acute. Anthurium sparreorum bears some resem- blance to both А. сатри and А. bucayanum, which also have large, green-drying leaves with the collective vein arising from near the base. An- thurium campii differs in its green spadix, more coriaceous leaf blades, and shorter petioles. Ап- thurium bucayanum differs in its very thin leaves and prominently stipitate spadix. Anthurium spar- reorum also has a shorter peduncle than either of these species. The new species is named in honor of the late Benkt Sparre, formerly curator of the Regnellian Herbarium in Stockholm and co-editor of the Flora of Ecuador Project, and his wife, Bride. Together they made concentrated efforts to collect Araceae and discovered numerous novelties in the process. EcuADOR. Cultivated by Tom Fennel, Homestead, Flor- ida, Croat 57195 (Е, MO). COTOPAXI: Rio Guapara, 20 km NW of El Corazón, 250 m, Sparre 17248 (S). 105 Rios: Rio Palenque Biological Station, Quevedo-Santo ues то 0 =" Croat 38666 (B, CAS, CM, K, M, МО, NY, QCA, RSA, US), Dodson & Tan 5389 (US, SEL), hs E 5513 (K), 6959 (MO, SEL, US). Anthurium spathiphyllum N. E. Br., Gard. Chron. I. 652. 1877. TYPE: Origin unknown, Hort. Kew, 8 Nov. 1876 (holotype, K). Fig- ures 303, 304 Epiphytic; stem ca. 10 cm long, 2 cm diam.; roots moderately numerous, descending, greenish gray, prominently pubescent, gradually tapered and elongate, 3-5 mm diam.; cataphylls subcoriaceous, lanceolate, 6-8 cm long, cuspidate at apex with subapical apiculum ca. 15 mm long, green, drying pale tan (B & K yellow 9/7.5), persisting as linear fibers. Leaves spreading; petioles 6-26 cm long, 2-8 mm diam., triangular, sometimes 3—5-ribbed or flattened with the margins prominently raised adaxially, sharply 1-2-ribbed abaxially; geniculum conspicuously thicker and slightly paler than pet- iole, 0.7-1 cm long; blades subcoriaceous, linear to narrowly oblong-elliptic or narrowly oblong-ob- lanceolate, more than 5x longer than wide, acu- minate at apex (the acumen apiculate), narrowly acute at base, 46-80 cm long, 5.5-12.5 cm wide, broadest at or above the middle, the margins broad- ly undulate; upper surface matte to glossy, dark to medium green, lower surface matte to weakly glossy, paler below; midrib flat at base, becoming angular-raised toward the apex above, prominently and sharply raised and slightly paler than surface below; primary lateral veins 20-30 per side, de- parting midrib at 40—45° angle, straight to the collective vein, deeply sunken above, sharply raised below; interprimary veins almost as conspicuous as primary lateral veins, sunken above, raised below tertiary veins obscure above, raised and digits darker than surface below; collective vein arising from near the base, sunken above, sharply raised below, equally as prominent as primary lateral veins, 3-10 mm from margin. Inflorescences spreading, shorter than leaves; peduncle 33-48 cm long, 4- 15 mm diam., 2X as long as petiole, green, prom- inently 2-winged-angled on one side, rounded to sharply acute on the other (or triangular or few- angled near base); spathe erect, hooded over the spadix, subcoriaceous, green, lanceolate to broadly lanceolate, 6.5-8 cm long, 1.7-3 cm wide, broad- est near the base, about twice as long as spadix, inserted at 50° angle on peduncle, acuminate at apex (the acumen apiculate, 4 mm long), obtuse to rounded at base; spadix pale yellow to creamy white, cylindroid to clavate, short and stubby, erect, (1.7)2-4 cm long, 6-11 mm diam.; flowers 1.8- 2.2 mm long, 1—1.2 mm wide, the sides smoothly to jaggedly sigmoid; 7-12 flowers visible in prin- cipal spiral, 11—15 in alternate spiral; tepals matte, densely and minutely papillate; lateral tepals 0.7— 0.8 mm wide, the inner margins straight to convex, the outer margins irregularly 2-4-sided; pistils somewhat emergent, matte, pale yellow; stigma linear, 0.1-0.2 mm long; stamens emerging in a scattered pattern throughout, lateral stamens al- most to apex before alternates emerge, inclined over and obscuring pistil; filaments fleshy; thecae oblong, not divaricate; pollen pale yellow fading to creamy white. Berries translucent white. Anthurium spathiphyllum is known from Nic- aragua to Panama mostly from sea level to 380 m (rarely to as high as 1,350 m) on the Atlantic slope in wetter parts of tropical wet forest and in premontane wet and tropical wet forest life zones. Anthurium spathiphyllum can be recognized by its epiphytic habit, chartaceous leaves, usually triangular, sometimes three- to five-ribbed petiole, 734 Annals of the Missouri Botanical Garden lanceolate to broadly lanceolate, cucullate spathe, and short, scarcely tapered, pale yellow spadix. Anthurium spathiphyllum can be confused with A. bradeanum, which has a similarly clavate spa- dix, but that species has oblanceolate-elliptic blades less than five times longer than wide (vs. more than five in A. spathiphyllum), with (5)10-15 pairs of primary lateral veins that are not deeply impressed, a peduncle which is not prominently winged, a deflexed, ovate spathe about as long as the spadix, and petioles that are commonly quadrangular in cross section. OSTA RICA. ALAJUELA: 3 mi. N of San Miguel, 380 m, od 35651 (MO, RSA). HEREDIA: 4 mi. N of Vara Blanca, 1,350 m, Croat 35571 (MO); Finca La Selva, junction with Rio Sarapiqu 34194 (DUKE), Folsom 8950 (CAS); Puerto Viejo de Sarapiqui, along El cda жар 100-150 m, Croat 44309 (MO); Puerto Viejo Region, Rio Puerto Viejo, Burger & Mata 4258 (F); Zona с La Selva, 6 km by road from Río Peje crossing, 5 km of Magsasay, 10°21'N, 84°4'W, 340 m, Schatz & Grayum 644 (DUKE). LIMÓN: Pandora, 100 m, Ocampo 1987 (MO); Hacienda Ta- pezco-Hacienda La Suerte, 29 air km d 40 m, 10%30'N, 83?47'W, Davidson & Donahue 8 (MO); Quebrada Mata de Limón headwaters, Finca jm Sixaola, 25-40 m, 9%35' °39'W, Стаушт 4515 (МО); Ко Рабина W of Dos Bocas, Lent 2444 (F); Talamanca, Cocles, 150 m, Ocampo 1906 ( с. en Corcovado, campam MEXU, MO). SAN JOSÉ: Parque Nacional Braulio Carrillo, Fila Carrillo, 600-700 m, Cha- cón 338 (MO); Chasse, Talamanca, m, Ocampo 1625 (CR); 300 m, Ocampo 1624 (CR). NICARAGUA. ZELAYA: 0.3-1.9 km N of Limbaika, 8-10 m, 13?29'N, 84°13'W, Stevens 8255 (MO); Risco de Бар 40 m, о Stevens 8504 (M nola- Amira, Milla Porter 1 6 (MO), Kennedy 1273 (DUKE); 15 km S of ос a 300 m, Antonio 3150 (MO); Almirante— Changuinola Canal, Blum 1403 (FSU, SCZ). состЕ: at par an of Rio Toabre and Rio Coclé del Norte, Boca del Toabre, Lewis et al. 5569 (MO). и spectabile Schott, Oesterr. Вог. 2. 181. 1858. ТУРЕ: Costa Rica. Cartago: v Turrialba, Wendland 506 (holotype, СОЕТ). Figures 305, 319. Anthurium longispathum Schott, Prodr. Aroid. 520. 1858. PE: Costa Rica. Turrialba, Oersted s.n. (lectotype, shamed by Schott Мше 707). Epiphytic; stem erect; cataphylls coriaceous, 17— 51 cm long, lanceolate, obtuse to mucronate at apex, light green, drying light tan (B & K yellow- red 9/10), persisting with the apex remaining in- tact. Leaves spreading; petioles 14-63 cm long, 7-12 mm diam., quadrangular, flattened to broadly and sharply sulcate adaxially, the margins sharply raised, sharply 2-3-ribbed abaxially; geniculum thicker and paler than petiole, becoming fissure transversely with age, (1.3)2-7 cm long; blades coriaceous, oblong-lanceolate to oblong or oblong- triangular, rounded to obtuse to acuminate at apex, rounded to truncate at base, 37-140(160) cm long, 8-38(45) cm wide, the margins undulate; upper surface matte to semiglossy, medium green, lower surface semiglossy to glossy, paler; midrib obtusely angular at base, becoming narrowly raised toward the apex above, prominently higher than broad at base, becoming prominently convexly raised toward the apex below, paler than surface above and below; primary lateral veins more than 20 per side, departing midrib at 45-75? angle, prominently and narrowly raised above and below; tertiary veins obscure above, raised and darker than surface below; collective vein arising in the lower half of the blade, sunken above, raised below, less prominent than primary lateral veins, 2-3 mm from margin. Inflorescences spreading, shorter than leaves; peduncle 13-25 cm long, X as long as petiole, terete; spathe usually recurled and twisted, prominently reflexed, subcoriaceous, green or violet-purple, narrowly ovate to oblong-lanceo- late, 10.5-32 cm long, 3-7 cm wide, inserted at 45? angle on peduncle, acuminate at apex, rounded to subcordate at base; spadix yellow-green, long- tapered, 16-38 cm long, 9-12 mm diam. near base, 6-7 mm diam. near apex; flowers rhombic to 4-lobed, 2.5-4 mm in both directions, the sides sigmoid; 7-11 flowers visible in principal spiral, ca. 9 in alternate spiral; tepals matte to semiglossy; lateral tepals ca. 2.7 mm wide, the inner margins convex; pistils emergent, green, violet-purple at apex; stigma ellipsoid, ca. 1 mm long, stamens emerging promptly throughout length of spadix in a complete sequence, grouped in a circle around the pistil; anthers 1-1.1 mm long, 1-1.2 mm wide; thecae oblong to narrowly ovoid, slightly divaricate; pollen cream. /nfructescence with spadix to 50 cm long; berries probably orange at maturity, oblong- linear, acute to blunt at apex, 10-1 m long, a. 3.5 mm diam.; seeds 2 per berry, 3.4-3.6 mm long, ca. 1.6 mm diam., with a minute appendage at both ends. Anthurium spectabile is endemic to Costa Rica, where it occurs at 300 to 1,300 m in premontane wet forest and tropical wet forest life zones. This species is recognized by its elongate, pen- Volume 78, Number 3 1991 Croat 735 Anthurium sect. Pachyneurium dent leaf blades with numerous, prominent primary lateral veins and erect, four-sided petioles. Also distinctive are the short peduncle, large, narrowly ovate, often violet-purple spathe, and yellow spa- dix. Anthurium spectabile might be confused with A. ranchoanum, which has similar leaves (in dried herbarium material) and a similar inflorescence. Anthurium ranchoanum, however, has a subterete petiole and a more or less erect leaf blade in line with the petiole. Anthurium spectabile also has an elongate, more or less spreading spathe, while the spathe of А. ranchoanum is shorter and usually stiffly erect and hooding the spadix. Anthurium spectabile is extremely similar to А. pseudospectabile from Panama, but the latter has proportionally longer blades with more undulate margins, terete petioles, and occurs in premontane rainforest. Costa RICA. ALAJUELA: Cariblanco, En 3543 (Е); NE of Villa Quesada, Molina et al. 17287 (F); Naranjo- Zarcas, Hwy. 15, 8.5 km NE of Villa Quesada, 600 m, Croat 46968 (Е, К, M, MO, NY, RSA, SEL); Rio Sarapiqui at bridge on road to Colonia Virgen del Socorro, 830 m, Croat 68303 (MO). HEREDIA: Rio Gua- cimito, Magsasay, 500 m, Chacón 167 (MO); Rio Peje- Rio Sardinalito, Seil slope of Volcán Barva, 700-800 m, 10?17.5'N, 84° 5'W, Grayum iine заң x Turrialba- Limón, 11 mi of Siquirres, Hwy. 3 m, Croat 43325 (МО); Guápiles, Зтић 4982 E road from Rio Frio to Limón, vic. of Río Blanco, W of Guápiles, 360 m, 10?12'N, 83°49'W, Croat 68428 (MO); Zona Protectora, Barbilla (SE of Squirres), 600-660 m, 1090'N, *28'W, Grayum et al. 7939 (CR, MO). PUNTARENAS: ciitivated at Las Cruces Botanical Garden, 1,300 m, Croat 44431 (МО). SAN JOSÉ: Carrillo Station, Gómez et al. 21188 (MO). сл о: Anthurium standleyi Croat & R. A. Baker, Brenesia 16(Suppl. 1): 88. 1979. TYPE: Costa Rica. San José: ca. 5 mi. SW of Сапаап along gravel road from Rivas, ca. 900 m, Croat 43439 (holotype, MO 2582987; isotypes, COL, CR, DUKE, F, K, LE, MEXU PMA, SEL, US, VEN, W). Figures 306, 307, 320. Terrestrial; stem to 90 cm long, 2-5 cm diam.; roots descending, greenish to tan, smooth to fuzzy, slightly tapered, 4-12 mm diam.; cataphylls mod- erately coriaceous, lanceolate, 13.5-30 cm long, obtuse at apex with subapical apiculum, light green, drying tan (В & К red 9/10), persisting semi- intact, eventually as linear fibers. Leaves erect to spreading; petioles 43-120 cm long, 7-16 mm diam., stiff, firm, terete to subterete, sometimes weakly flattened adaxially, surface minutely pale- speckled; geniculum slightly thicker and paler than petiole, 3.5-6.5 cm long; blades subcoriaceous, ovate to ovate-triangular, acuminate at apex (the acumen apiculate, downturned), cordate at base, 29-75 cm long, 19-56 cm wide, broadest at the point of petiole attachment or slightly above it, the margins conspicuously undulate; anterior lobe 23- 55 cm long; the posterior lobes 7-21 cm long; sinus spathulate to obovate, often closed with the lobes overlapping, 3.5-12 cm deep, the inner edges turned up; upper surface weakly glossy to semi- glossy or matte, dark green, lower surface semi- glossy to matte, slightly paler; midrib triangular- raised at base, becoming sharply and narrowly raised toward the apex above, prominently and convexly raised and paler than surface below; basal veins 4—6 pairs, 4—5 coalesced 1—4 cm, two lowermost sunken, others raised above, convexly raised below; posterior rib slightly curved, naked, sharply turned up on the outer margin; primary lateral veins 4— 9 per side, departing midrib at 45-80? angle, straight, becoming arcuate-ascending, usually only the uppermost loop-connected, convexly raised above and below, prominently raised below on dry- ing; interprimary veins obscure or occasionally raised in grooves above, raised below; tertiary veins obscure above, slightly darker than surface below; reticulate veins slightly raised on both surfaces on drying; collective vein arising from about the mid- dle of the blade, slightly sunken above, raised and slightly darker than surface below, slightly raised above and below on drying, 4-11 mm from margin. Inflorescences erect to spreading, much shorter than leaves; peduncle 29-76 cm long, 4-9 mm diam., 0.3-0.8 X as long as petiole, green, terete; spathe erect-spreading to reflexed, subcoriaceous, green to yellow-green (B & K yellow-green 6/7.5), oblong-lanceolate, 10.5-18 cm long, 2-4 cm wide, broadest just above the base, inserted at 60—80? angle on peduncle, narrowly acuminate at apex (the acumen inrolled and hooked), obtuse to round- ed at base, the margins meeting at 90-180? angle; stipe 11-70 mm long in front, 6-60 mm long in back; spadix heavily tinged with violet-purple at anthesis, becoming yellowish brown, long-tapered, 10-30 cm long, 7-10 mm diam. near base, 4-6 mm diam. near apex; flowers rhombic to weakly 4-lobed, 3.5-3.6 mm long, 2.9-3.3 mm wide, the sides = straight to smoothly sigmoid; 5-10 flowers visible in principal spiral, 5-7 in alternate spiral; tepals semiglossy; lateral tepals 1.9-2 mm wide, the inner margins broadly convex; pistils weakly raised, medium green; stigma linear at anthesis, 0.4-0.7 mm long, droplets appearing before sta- mens emerge; stamens emerging from the base in 736 Annals of the Missouri Botanical Garden prompt sequence; anthers orange to greenish yel- low, 0.6-1 mm long, ca. 1 mm wide, inclined over and obscuring the pistil; thecae ellipsoid, not di- varicate; pollen orange fading to creamy white. Infructescence arching; spathe reflexed; spadix to 30 cm long, ca. 2 cm diam.; berries orange, broadly ovoid, rounded at apex; mesocarp pulpy, bitter- tasting; seeds 1 or 2 per berry, white, to 5 mm long, ca. 2.5 mm wide, 1.5 mm thick. Anthurium standleyi is known for certain only from central and southern Costa Rica. It was first collected by Paul Standley at two sites, La Verbena and San Sebastian south of San José in what was probably premontane wet forest. It has been col- lected more recently in the region northeast of San Isidro del General. The species ranges from 200 to 1,600 m and may be restricted to premontane wet forest. This species is distinguished by its cordate leaf blades, long-tapered, purplish violet spadix and or- ange berries. Among Central American species with cordate leaf blades and orange fruits, 4. standleyi is unique in having an ovate leaf blade and a collective vein originating above the middle of the lade. Because of its prominently cordate leaf blades, Anthurium standleyi is likely to be confused only with 4. watermaliense, which differs in having the collective vein arising from the first or second basal vein and a proportionally broader, often maroon spathe. A cultivated collection from the University of Hawaii was reportedly based on a Dressler collec- tion from Achiote, a region of tropical moist forest in the Isthmus of Panama. The species has never been collected in Panama, and the Hawaiian col- lection is possibly mislabeled. COSTA RICA. PUNTARENAS: Rio Coto Brus, 23 km М of La Unión, Panama border, Croat 26674 (MO); cultivated at Las Cruces Botanical Garden, 1,300 m, Croat 44436 (MO). SAN JOSÉ: Santa Eduviges de Escazu, residential neighborhood at W side of Rte. 105, 1,150 m, 9°55'N, 84°08.5'W, Grayum 9510 (CR, MO); 0.5 mi. above turnoff to Canaán at Rivas, 900 m, Croat 43431 (MO); 4.5 mi. SW of Canaàn, road from Rivas, 900 m, Croat 43439 (B, CAS, COL, CR, DUKE, F, GH, HUA, K, LE, M, MEXU, MO, NY, PMA, QCA, RSA, SEL, UC, US, VEN, Wy La Verbena, Standley 32243 (US); San Se- bastian, Standley 49296 (US); Talamanca, less than 200 m, Tonduz 8721 (NY) Cerro Tapezco, Jiménez 627 (05); Río Buena Vista, S of Buena Vista, NE of San Isidro del General, Luteyn 3287 (DUKE); Rio Herradura (trib- utary of Río Chirripó del Pacifico), NW of Canaan, Valle General, 1,600 m, 9°29’N, 83°37'W, Burger & Liesner 7093 (MO, US). Anthurium superbum Madison, Selbyana 5: 94. a. Anthurium superbum subsp. superbum. TYPE: Ecuador. Napo: La Primavera, Rio Napo, Madison 5516 (holotype, SEL 2727370; is- otypes, F, K, MO, QCA, U, US). Figures 308, 309 Epiphytic; stem 4-8 cm long, 3-4 cm diam.; roots numerous, descending, rose-colored, veluti- nous, 5-7 mm diam.; cataphylls subcoriaceous, (3)8-10 cm long, acute at apex, dark green, drying brown, persisting intact or semi-intact. Leaves erect; petioles 4-12 cm long, 5-8 mm diam., + D-shaped, with the medial rib and the margins raised adaxi- ally, 3-5-ribbed to rounded abaxially, the surface dark green to sometimes dark purple-green; gen- iculum slightly thicker than petiole, ca. 0. long; blades coriaceous, stiff, elliptic to oblong- elliptic, obtuse-rounded at apex, obtuse at base, 34-65 cm long, 12-18 cm wide, broadest just above the middle, the margins revolute; upper sur- face matte to semiglossy, bullate, dark green, lower surface matte, sometimes tinged reddish or purple; midrib prominently acute above and below; primary lateral veins 7-10 per side, departing midrib at 40—50? angle, very long-arcuate to the margin or slightly arcuate to the collective vein, sunken to weakly raised above, raised below, drying slightly raised above and below; interprimary veins few, less prominent than primary lateral veins; tertiary veins sunken above, raised below, drying slightly raised; collective vein arising from about the middle of the blade or absent, sunken above, raised below, slightly raised above and below when dried, nearly as prominent as primary lateral veins, 10-20 mm from margin. /nflorescences erect, shorter than leaves; peduncle 7-18 cm long, 5-12 mm diam., ca. 2X as long as petiole, dark green, terete; spathe spreading to reflexed, subcoriaceous, green, lan- ceolate-elliptic, 8-9 cm long, 2-3 cm wide, broad- est near the base, acute at apex (the acumen slight- ly inrolled), acute at base; stipe to 5 mm long; spadix creamy white, sometimes faintly tinged with purple, cylindroid, sessile or shortly stipitate, erect, straight to slightly curved, 8-15 cm long, 8-12 mm diam. midway; flowers = rhombic, 1-1.5 mm long, 1.5-2 mm wide; 16-17 flowers visible in dies spiral, 23-24 in alternate spiral; tepals utely puberlent to densely papillate; lateral tepals q .9 mm wide, inner margins + straight, outer margins 2—4-sided; pistils prominently exerted, pa- pillate; stigma circular, 0.3 mm long, brushlike. Моште 78, Митбег 3 1991 Croat 737 Anthurium sect. Pachyneurium Infructescence with the spathe persisting; berries lavender with the apical 1⁄4 white, obovoid, sharply apiculate at apex, 6-7 mm long, 2.5 mm diam Anthurium superbum is known to occur in trop- ical moist to tropical wet forest life zones, at 180 to 350 m. It comprises two allopatric subspecies, the typical one in Napo, Ecuador, near La Pri- mavera on the Rio Napo, and subsp. brentberlinii in Amazonas, Peru, in the Rio Cenepa area. Both subspecies are characterized by their stiffly erect, dark green leaves with short petioles and bullate, coriaceous blades often tinged red or purple below, and by the short inflorescence with a whitish spadix with many flowers per spiral. Neither subspecies is likely to be confused with any other member o the section. While subsp. brentberlinii is not known in cul- tivation, subsp. superbum has been cultivated for about 10 years in a few botanical gardens and by a few hobbyists. It is a stunning plant and will, hopefully, become more widespread in cultivation in the future. ECUADOR. NAPO: Rio Маро, La Primavera, Madis 5516 (F, K, MO, QCA, SEL, U, US); Laguna Dt 250 m, Christenson 1529 (MO), cultivated. = . Anthurium superbum subsp. brentber- linii Croat, subsp. nov. TYPE: Peru. Amazo- nas: Rio Cenepa, E of community of Huam- pami, 300-330 m, Berlin 644 (holotype, MO 2251387; isotype, NY). Figure 310. Differt ssp. typicum caulibus longioribus, laminas an- gustioribus, spadice angustiore cum seriebus spiralibus paucioribus florum Epiphytic; stem to 16 cm long, 7-14 mm diam.; roots dense, 3-5 cm long, 3-4 mm diam.; cata- phylls subcoriaceous, linear-lanceolate, 3-13 cm long, drying reddish brown, persisting semi-intact. Leaves erect; petioles 9-17 cm long, 5-10 mm diam., D-shaped, with the medial rib and margins raised adaxially, angular abaxially; geniculum somewhat darker than petiole when dried, 0.5-1 cm long; sheath 2.5—4 cm long; blades coriaceous, oblong-elliptic, acute to obtuse at apex, attenuate to acute at base, 45-70 cm long, 9-17 cm wide, broadest just above the middle; upper surface ap- parently quilted, lower surface often tinged red or purple; midrib acutely raised above and below; primary lateral veins 4—9 per side, departing midrib at 50—60° angle, very long-arcuate to the margin or slightly arcuate to the collective vein, raised to obscure above, raised below when dried; interpri- mary veins usually present, less conspicuous than primary lateral veins; tertiary veins weakly raised to obscure above and below when dried; collective vein arising from about the middle of the blade or absent, rarely from near the base, sunken above, raised below when dried, equally as prominent as primary lateral veins, 10-15 mm from margin. Inflorescences erect; peduncle 11-29 cm long, 2- 5 mm diam., 1-2.6X as long as petiole, terete; spathe reflexed, subcoriaceous, green, oblong-lan- ceolate, 4-10 cm long, 1-2 cm wide, broadest near the base, acute at apex, acute at base; spadix white, cylindroid, + straight, 5-7 cm long, 4—6 iam. midway; flowers = rhombic, ca. 0.8 mm in both directions; ca. 12-15 flowers visible in either spiral; tepals densely papillate; stamens held erect above the tepals; filaments exserted 0.2 mm; anthers 0.2 mm in both directions. /nfruc- tescence spreading-pendent; spathe persisting or absent; spadix 20 cm long, 2 cm diam., with berries scattered throughout; berries purple, tipped asi (B & K purple 5/5), obovoid-ellipsoid, obtuse apex, ca. 7 mm long, ca. 4 mm diam.; pericarp with numerous raphide cells; seeds 1-2 per berry, yellow-brown when dried, oblong, 2 mm long, 1 mm diam. Anthurium superbum subsp. brentberlinii is en- demic to northern Peru in Amazonas, where it is known only from the Rio Cenepa area at 180 to 350 m in a tropical wet forest life zone. This subspecies is characterized by its slender, densely rooted stem, short petioles, oblong-elliptic leaf blades with steeply ascending primary lateral veins and often tinged red or purple below, and cylindroid, white spadix with very small flowers. Anthurium superbum subsp. brentberlinii dif- fers from the typical subspecies because of its nar- rower leaf blades, more slender stem, and more slender spadix with fewer flowers per spiral. The new subspecies is named in honor of Brent Berlin (Department of Anthropology, University of California at Berkeley), who coordinated all known collections of this taxon. PERU. M. Ee Cenepa, Cenepa area, 250 m, Kayap 1252 (МО); Cenepa-Quebrada Kayamas, 250 m, Berlin 513 (МО); Квн Quebrada Shimpunts, 250-350 m, M 886 (MO); кеңе -400 m, 4?30'S, 78°30'W, Berlin 644 (MO, NY); Kayap 2007 (MO), Кыйы 162 (МО). ағ 7 tarapotense Engl., Pflanzenr. IV. 3B(Heft 21): 74. 1905. ТУРЕ: Peru. San Mn Tarapoto, Ule 5p (holotype, B). Fig- ures 311, 312. 738 Annals of the Missouri Botanical Garden Terrestrial, occasionally epiphytic; stem short, 1.5-4 cm diam.; roots dense, spreading to de- scending, pale green, velutinous, short, tapered, 2-6 mm diam.; cataphylls lanceolate, subcoria- ceous, 5-12 cm long, acuminate at apex, drying dark reddish brown, persisting semi-intact, even- tually deciduous. Leaves erect to spreading; peti- oles 7-19 cm long, 3-10 mm diam., erect, sharply D-shaped, broadly convex to 1-ribbed to flattened or somewhat sulcate with the margins sharply raised or sharp but not raised adaxially, rounded to weakly l-ribbed abaxially; geniculum moderately thicker and paler than petiole, becoming transversely fis- sured with age, 0.5-3 cm long; sheath 2-5 cm long; blades subcoriaceous to moderately coria- ceous, broadly elliptic to broadly oblanceolate to oblanceolate, long-acuminate at apex (the acumen apiculate), long-attenuate to attenuate, rarely acute at base, (15)30-60(90) cm long, 8-27 cm wide, broadest near or above the middle, the margins undulate; upper surface weakly glossy to semi- glossy, medium green (B & K green 2/2.5), lower surface semiglossy, conspicuously paler, both dry- ing brown to greenish brown; midrib flat and 1-ribbed at base, becoming acutely angular toward the apex above, higher than broad at base, becoming ob- tusely angular and then convexly raised at apex below; primary lateral veins 5-10(16) per side, departing midrib at 40—80? angle, straight to less often arcuate to the margin, paler than surface, convexly raised above and below, more so above; tertiary veins obscure above, prominulous and darker than surface below; collective vein arising from about the middle of the blade, near the apex or absent, prominulous on both surfaces, to 12 mm from margin. /nflorescences erect to spreading; peduncle 23-50 cm long, 3-4 mm diam., 2-4.6 x as long as petioles, yellowish green tinged with reddish in the lower half or purple, terete; spathe spreading to reflexed, subcoriaceous, concolorous with the petiole, sometimes tinged with purple or purple throughout, ovate-lanceolate to oblong-lan- ceolate, 4.5-15 cm long, 0.9-1.9 serted at 60? angle on peduncle, acuminate at apex (the acumen inrolled), rounded to obtuse at base; spadix purple to maroon (B & K red-purple 2/10) to yellowish green, sessile or stipitate to 5 mm, cylindroid to slightly tapered, erect to somewhat curved, (3)4-16 cm long, 4-6 mm diam. near base, 3-4 mm diam. near apex; flowers 4-lobed, 1.5-2.5 mm in both directions, the sides straight to jaggedly sigmoid; 4—6 flowers visible in principal spiral, 4—8 in alternate spiral; tepals matte, white- ша: lateral tepals 0 inner margins broadly rounded, the outer margins cm wide, in- 2-3-sided; pistil weakly raised; stigma slitlike to ellipsoid, 0.4-0.6 mm long; stamens emerging ir- regularly, the laterals preceding the alternates by up to 16 spirals, the 3rd stamen preceding the 4th by 3—4 spirals, borne at edge of the tepals; anthers pinkish to reddish or orangish, 0.4-0.7 mm long, 0.6–0.8 mm wide; thecae ellipsoid, divaricate; pol- len yellow to pale yellow, sometimes fading to white and then turning lavender. /nfructescence with spathe persisting, peduncle 38-44 cm long; berries deep purple, obovoid; seeds 2 per berry. Anthurium tarapotense is endemic to Peru, where it is apparently confined to premontane moist forest at (300) 360 to 700 m, in the vicinities of Tocache and Tarapoto in Dept. San Martin. This species is characterized by its obovate to obovate-oblanceolate blades, which are attenuate in the lower third and merge gradually with the petiole, and by the short, D-shaped petioles, fre- quently more or less intact cataphylls, and tapered or cylindroid, usually maroon or purple spadix. Two individual live plants, from near Tabalosos, Peru, collected together sterile and bearing the same collection number (Croat 58130a and 581306), differ substantially and perhaps герге- sent parts of a mixed collection. One of the two, 58130b, perhaps deserves subspecific recognition. It shares nearly identical leaves with typical 4. tarapotense (58 130a representing the species well), but differs in having cataphylls weathering prompt- ly to fibers and a spadix with pistils protruding above the anthers at anthesis and an ellipsoid, caviform stigma bearing stigmatic papillae well be- low the upper rim. The anthers, immediately after dehiscence (when they are most prominent), are relatively inconspicuous compared to the pistils. In contrast, the pistils of typical А. tarapotense are not at all exserted at anthesis and consist of a narrow slit more or less obscured by the anthers. Also, the spadix of Croat 58130b is grayish i not maroon or yellowish green as in typical 4. tarapotense. nthurium tarapotense is probably most closely related to А. llewelynii, which occurs in the same region, but in more exposed sites in tropical dry forest. Both species share similar long-pedunculate inflorescences with moderately short, bluntly ta- pered spadices and blades markedly arched along the midrib. Anthurium llewelynii differs in having larger, narrowly oblanceolate leaves, which end abruptly at the narrowly rounded base, and a red- dish olive-green spadix. PERU. MARISCAL CÁCERES: Dtto. Tocache, Nuevo, Que- brada de Mantención, road to Tocache Progresso, 500- Volume 78, Number 3 1991 Croat 739 Anthurium sect. Pachyneurium 600 m, Schunke 13648 (IBE, MO). a MARTÍN: hills above Chazute, 200-300 m, 6°34'S, 76°12'W, Knapp 8349 (MO); Tarapoto, Ule 5 (B); 360- pos m, Kennedy 3529 (F); Tarapoto- Yurimaguas, 10.6 km from junction with road to Вю Huallaga, 530 m, 06°30'5, 76°21'W, Croat 58115 (IBE, JBGP, K, MO, NY, USM), 58116 (AAU, MO, P, QCA, ЖЗ U, UCLA, USM, VDB), 58117 T CAS, CM, G, GH LE, M, MO, US, USM, VEN, WISy; Rio Mayo, ар Moyobamba, between Km n and 563, 15 km W of Tabalosos, 57 km of Tarapoto, 600 m, 6°15'$, 76*41'W, Croat 58130 (CM, IBE, KYO, MO, NY, SEL, USM); Рио. Tarapoto, Tara- poto- Yurimaguas, near Km 10, 700 m, Rimachi 5369 (IBE). Anthurium tenaense Croat, sp. nov. TYPE: Ecua- dor. Napo: along road between Tena and Puyo, 58.1 km М of Puyo, 600 m, Croat 49631 (holotype, MO 2738143). Figures 313, 314. Planta epiphytica; caulis erectus; pa 18-19 cm longus, ca. 5-6 mm diam., D-formatus; lamina oblongo- oblanceolata ad oblongo- pe 37- die m longa, 9- 9.5 cm lata; pedunculus 19.5 cm longus; e. 8 cm longa, 1 cm lata; spadix estad 6 cm longus, 3 mm diam. Epiphytic; stem erect; internodes short, roots numerous, drying whitish; cataphylls unribbed, 6 cm long, drying grayish brown, persisting intact, eventually shredding at base. Leaves with petioles (8-9)18-19 cm long, ca. (3-4)5-6 mm diam., + D-shaped, with obtuse medial rib and with the margins sharply raised adaxially, rounded abaxi- ally; geniculum drying slightly darker and narrower than petiole, (0.4)0.6-0.9 cm long; sheath 2.5- 5.5 cm long, free-ending; blades coriaceous, nar- rowly oblong- elliptic, acute to long attenuate at apex, acute t ) at base, (23.2)32.5- 69 cm long, (5.0-5. 6)9- —9.5 cm wide. broadest at or above the middle, the margins weakly undulate; both surfaces drying matte, reddish brown, slightly paler below; midrib obtusely angular to broadly convex at base, becoming narrowly acute toward the apex above, more prominently raised below, bluntly acute, becoming narrowly convex toward the apex, darker than surface when dried; primary lateral veins 5-8 per side, departing midrib at 65- 90? angle, prominently arcuate to the margin, dry- ing convexly raised and darker than surface above and below; interprimary veins not numerous, less prominent than primary lateral veins; tertiary veins darker than surface when fresh, prominulous; re- ticulate veins visible on either surface; collective vein arising from near the apex, probably sunken above, raised above and below when dried, 4-10 mm from margin. Inflorescences with peduncle (19.5)22.5—60 cm long, (2-3)6-9 mm diam. when dried, ca. (1.5)2.5-3X as long as petiole; spathe green, subcoriaceous, lanceolate with acumen in- rolled, (7.6)9-15 cm long, 1-2.5 cm wide, acute at apex, long attenuate at base; spadix green, cy- lindric to long-tapered, 6-18 cm long, 3-7 mm diam. near base, 1-2 mm diam. near apex; flowers rhombic to 4-lobed, (1-1.5)1.8-2.3 mm long, (0.8— 0.9)1.6-1.8 mm wide, the sides moderately straight to jaggedly to smoothly sigmoid; (6-7)8-17 flowers visible in principal spiral, 5 in alternate spiral; tepals minutely granular-papillate, lateral tepals (0.6)0.8-10 mm wide, + shield-shaped, the inner margins straight to broadly rounded, the outer mar- gins irregularly (2—3)3—4-sided; exposed portion of the pistil rectangular; stigma slitlike, 0.2-0.4 mm long. Anthurium tenaense is known from Napo Prov- ince, Ecuador, near Tena (hence the name), in a tropical wet forest life zone at 400-600 m. This species is characterized by its thick, more or less elliptic blades, which dry brown, and by its D-shaped petiole, which is rounded abaxially with a broad, obtuse rib, and with raised margins adax- айу. Anthurium tenaense is not easily confused with any other species, although it bears a resemblance to А. uleanum, which has thinner leaves mostly drying dark brown, and to A. harlingianum, which has broader leaves, more primary lateral veins (1 2— 23) per side, and a slightly to conspicuously curved, salmon-pink to pale violet spadix. ECUADOR. NAPO: Tena-Puyo, 58.1 km М of Puyo, 600 m, Croat 49631 (MO); Tena, Estación Biológica Jatun Sacha, 8 km E of Misahuallí, 1%04'S, 77°36'Ұ, 400 m, Cerón 6073, 6074 (MO) Anthurium uleanum Engl, Pflanzenr. IV. 23B(Heft 21): 74. 1905. a. Anthurium uleanum var. uleanum. TYPE: Brazil. Amazonas: Rio Juruá Mirim (“Мігу”), Ule 5731 (holotype, B; isotypes, G, MC). Fig- ures 315-317. ipe on аран 73. 1905. ТҮРЕ: Brazil. Amazonas: Rio Jurua Mirim (“Miry”), Ule 5597 йите В; serm Anthurium tenuispadix Engl., i К. vores дада Вог. Сап. 2 E 9: 260. 1925. : Peru. Loreto: Prov na, Cumaria, Rio Ucayali 225 m, Tess- mann 23351 (holotype, B). расна tessman Berlin Epiphytic; stem to 15 cm long, ca. 2.5 cm diam.; roots dense, green to whitish, smooth to velutinous, 2-5 mm diam.; cataphylls lanceolate, straight to curved, subcoriaceous, 2-11.5 cm long, acute to acuminate at apex, drying reddish brown (B & K 740 Annals of the Missouri Botanical Garden yellow-red 4/10), persisting semi-intact, weath- C ering to reticulate fibers, apex semi-intact. Leaves erect-spreading; petioles (3)10— 36 cm long, 2-10 mm diam., D-shaped, sulcate adaxially with medial rib and acute margins, round- ed or 2- 7-ribbed abaxially; geniculum shaped like petiole, paler and thicker, 0.3-2 cm long; sheath 2.5-6 cm long; blades subcoriaceous, oblong-ellip- tic to oblong-oblanceolate to obovate, acute to acu- minate at apex (the acumen flat), attenuate to acute at base, rarely truncate or subcordate, 20-65 cm long, 5.5-24 cm wide, broadest at or above the middle, the margins broadly undulate; upper sur- face semiglossy, medium green, drying gray to greenish brown, lower surface matte to semiglossy, slightly paler, drying greenish to gray-green to brown; midrib acutely raised above, acutely to ob- tusely or convexly raised below; primary lateral veins 5-11 per side, departing midrib at 30-85? angle, straight to arcuate to the margin, acutely raised; tertiary veins prominulous, raised on both f drying; collective vein arising from near the apex or absent, rarely from below the middle, sunken above, prominulous below, 4-8 mm from margin. Inflorescences erect spreading or spreading-pendent, usually shorter than leaves; pe- duncle 14-38 cm long, 3-6 mm diam., rarely less on drying, 0.5-3.7(6)X as long as petioles, terete, firm, flexible; spathe spreading to reflexed, rarely erect, subcoriaceous, matte abaxially, medium to pale green, sometimes tinged with red at margins, lanceolate to oblong-lanceolate, 3.5-14 cm long, 6-2.7 cm wide, broadest near the base, inserted at 30—50° angle on peduncle, acute to acuminate at apex (the acumen inrolled), acute to obtuse (sometimes decurrent 5-17 mm) at base; stipe 3- 7 mm long in front, 2-5 mm long in back; spadix usually pale reddish to purplish (B & K red 8/10), sometimes white to pale green, usually appearing glaucous, cylindroid to slightly tapered, curved, erect, (3)5-15 cm long, 3-5 mm diam. midway, -3 mm diam. near apex; flowers rhombic to 4- lobed, scarcely distinguishable, margins obscure, -2.5 mm long, 1.7-2.5 mm wide, the sides Pie to jaggedly sigmoid; 6-10 flowers visible in principal spiral, 6-8 in alternate spiral; tepals matte, weakly papillate, pale olive; lateral tepals (0.5)2-2.2 mm wide, the inner margins broadly rounded to straight, the outer margins 2-3-sided; pistils weakly raised, not emergent, white to pale pinkish; stigma ellipsoid, 0.3 mm long; stamens emerging promptly in a regular sequence from the base, held well above the tepals then retracting to the tepal level, the laterals preceding the alternates by ca. 4–5 spirals, the 3rd stamen preceding the Ath by ca. 2 spirals; anthers white to pinkish, 0.4 mm long, 0.4-0.5 mm wide, inclined over and obscuring the pistil; thecae oblong, divaricate or not; pollen pale yellow, drying white. /nfructesc- ence + pendent; spadix to 25 cm long, 1 cm diam.; berries reddish violet, 5 mm long, 2-4 mm diam.; seeds 1-2 per berry, 2.2-2.5 mm long, 0.8-0.9 mm thick, 1.2-1.3 mm wide. Anthurium uleanum is endemic to the upper Amazon basin, ranging from Ecuador and southern Colombia (Meta, Amazonas, and Vaupés), to Peru (Loreto, Huánuco, Amazonas, and Ucayali), and western Brazil (Acre and Amazonas). It occurs be- low 550 m in tropical moist, premontane wet, and subtropical wet forest life zones. This species is distinguished by its long-petiolate, brown-drying leaves, long peduncle, spreading spathe and slender, mostly pale purple, glaucous padix. Anthurium uleanum is closest to 4. ernestii, differing in the presence of a waxy, dustlike bloom on the tepals giving the spadix a matte, pale ap- pearance. А difference that can be used to deter- mine fruiting collections is the slender, often curved, tapered spadix of the present species, rather than a more or less cylindroid, stubby one. Anthurium uleanum also has proportionately longer petioles, frequently up to half as long as the blade, and the blades tend to dry brown to dark brown. Anthurium ernestii typically has petioles much less than one- third as long as the blade and leaf blades, most often drying greenish with orangish or brownish primary lateral veins. Anthurium uleanum has cataphylls which tend to persist intact or semi- intact for a long period before ultimately weath- ering to fibers. Often there are 2-3 more or less intact cataphylls on a plant. In contrast, the cat- aphylls of A. ernestii persist quite characteristically un as an intact network of fine, very pale fibers, and frequently there is little evidence of intact cata- phylls on dried specimens. Another species which may be confused with 4. uleanum due to having a similar waxy, grayish spadix is 4. manuanum. See the discussion under that species for distinction from А. шеапит. e new variety of А. uleanum, var. nanay- ense, is a morphologically homogeneous taxon re- stricted to Loreto, Peru, mainly in the Iquitos re- gion. See discussion under the variety for further details. Synonymized here for the first time is А. tenui- spadix. Its type (Ule 5597), collected at the same locality as that of А. uleanum, differs slightly from the latter in having leaf blades attenuate at the Volume 78, Number 3 1991 Croat 741 Anthurium sect. Pachyneurium base rather than ending abruptly. Engler separated the two species in his key on presumed differences in the shape of the lateral tepals, but these differ- ences in shape are probably due to their relative ages. Neither the difference in blade shape nor the tepal shape warrants separation of A. tenuispadix. The type of Anthurium tessmannii (Tessmann 3351), considered here as a synonym, is also to- ward the larger end of the spectrum in terms of leaf blade size (ca. 65 cm long, 19 cm wide). Some collections made in the northwestern part of the range of 4. шеапит in the Province of Napo, Ecuador, differ from Peruvian material and collections from further south in Ecuador in having broader, larger leaf blades with frequently obtuse bases. Brandbyge & Asanza 30425 and Pinkley 90 are examples. There is, however, an overall gradation to “typical” Peruvian material within Ecuador. Several other collections deserve mention. The leaf blades of a collection from the Río Ampiyacu in Loreto, Peru (Plowman et al. 6541), are un- usual in having only a few, very steeply ascending primary lateral veins (parallel to the margin for most of their length) and a geniculum noticeably shrunken to a narrower diameter than the attending midrib or petiole. The leaf blade, in color and venation, bears some resemblance to that of А. superbum subsp. brentberlinii. Harling & Ап- dersson 17571, Mexia 6142a, and Wurdack 2109 are unusually large plants. The following measure- ments, contrasted with those of typical material, circumscribe the main differences: petioles 45 cm long (vs. up to 36 cm); leaf blade 85 cm long (vs. up to 65 cm) and up to 31 cm wide (vs. up to 24 cm); peduncle 43-51 cm long (vs. 14-38 cm). The spadix does not differ significantly from that found on those of other plants which are vegeta- tively more typical and have shorter peduncles. BRAZIL. ACRE: Cruzeiro do Sul-Rio Branco, Е km E of Cruzeiro do Sul, Fazenda da Dona Cabúca, 50 m, 7°38'5, 7235 У, Croat 62334 (INPA, MO, к Ко Moa, Igarape Ipiranga- Aquidaba, Cruzeiro do Sul, Prance et al. 12049 (INPA, NY, U, US); Cruzeiro do Sul, 150 m, 7°37'S, 72°37'W, Croat & Rosas 62304 (СМ, С, IBE, INPA, K, MBM, МО, NY, QCA, RSA, US); Mpo. Caramari Amazonas, Rio Juruá, N of Cruzeiro do Sul, Lago da Cigana (бо Luis), S of Porto Alvaro Nestrinho, 150 m, 7°37'5, 72?36'W, Croat 62498 (CAS, CM, GB, INPA, К, MO, RSA, SAR, TEX); Mpo. Cruzeiro do Sul, n em do T Rio Branco, INCRA headquarters and forest reserve, 160 m, 7°38’S, 72*35'W, Croat 62629 (INPA, MO). AMAZONAS: Rio Cunhua, basin of Rio Purus, Deni Indian village, 6%43'5, aea Prance et al. 16534 (INPA, NY); Rio Javari, N of Palmeiras airstrip, 5%08'S, 72°49'W, Lleras et al. P17 101 (INPA); Rio Juruá Mirim ("Миу"), Ule 5597, 5731 (B, С, MG); Manaus-Porto Velho, 8 km S of de Jutai, Prance et al. 22916 ca US). CoLoMBIA. AMAZONAS: Rio Loretoyacu, vic. P Магто, Zarucchi & ada. 1 069 (CH). 100 m, Schultes 8468 (GH). META: Río Guayabero, 235 m, Garcia-Ba- rriga & Mejia 17071 (COL); 10 km from Саћо Lozada, 350 m, Pinto et al. 222 (COL, P); vauPEs: Río Apaporis, Rio Pacoa, 300 m, 0%20'N, 71°20'W, Schultes & Ca- brera 15221 (COL); Río Pacoa-Río Kananari, 250 m Schultes & Cabrera 13703 (COL, US); Soratama, du mouth of Rio Kananari, 300 m, 0%5'N, 70?40'W, Schultes & Cabrera 15164 (US). ECUADOR. NAPO: Parque Na- cional Yasuní, 200 m, 0%55'S, 76°11'Ұ, Cerón & Hur- tado 4015 (MO); Pozo petrolero Daimi 2, 200 m, 0°55'$, 76°11'W, Cerón & Hurtado 4177 (MO); Lago Agrio- Coca, 4.7 km N of Coca, Croat 50408 (MO); 15 km N of Coca, 450 m, 0°30'5, 76°56'W, Croat 50412 (AAU, F, СН, К, MO, WIS); San Pablo de Los Secoyas, W- SW of Shushufindi, 300 m, 0?15'S, 76°21'W, Brandbyge E Asanza 32894 (AAU, MO); Concepcion, Hacienda Co- tapino, 550 m, 0%48'S, 77°25'W, Harling et al. 7111 (GB); Coca-El Auca, SE of Francisco de Orellana, road to El Auca, 14.6 km past bridge over Río Napo, 450 m, 0°37'S, 76?40'W, ON 50382 (M, MO, NY, QCA, RSA); Lago Agrio, 250 m, 0%6'N, 76%55'W, Brandbyge & Asanza 30425 (AAU, MO); Lago Agrio-Francisco de Orellana, 26 km М of Coca, 450 m, 0%29'S, 76%55'W, Croat 50422 (MO, ОСА); Lago Agrio-Coca, along Сере Ferry Rd., 8.8 km S of Río Aguarico on road to Coca (San Francisco de Orellana), Croat 58519 (MO); Lago а Bs El Carmen de Putomayo, 2.2 km SE of Guar 240 m, Croat 58617 (MO, NY); Tena-Puyo, pen 'N of Puyo, 500 m, Croat 49661 (MO); Río ر‎ e from San Pablo de Las Las Secoyas, 230 m, 0 76°11'W, Holm-Nielsen et al. 21095, 21098, 21099 (AAU); Shushu&ndi. 244 m, Vickers 220, 254 (F); Dureno, Pinkley 134, 290 (ECON); | W of Lago Agrio, 340 m, 0*8'N, 76*58'W, Macbryde & Dwyer 1405 (MO); E of гт Agrio, road to Сере Ferry Crossing, 450 m, 0?2'N, "W, Croa і 50426 (МО); Río Cuyabeno, 10 km и dead from Aguas Negras, 220 m, 0°10’S, 75*58'W, Holm Мине et al. 2117 1 (AAU); Laguna Grande vic., 265 m, Nielsen 76207 (AAU); Rio Napo, Limoncocha, SEL 78-1155 (cultivated) (MO); Estación pe ат Sacha, 8 km , 450 m, 1?04'S, 77°36’ К Сегоп & p TJ Gentry 9826 (МО); Rio de ge Baeza- "amd El Car- men de Putumayo, 2.1 km W of р 43.7 km SE of Lago Agrio, 240 m, 0*07'N, 76°39'W, Croat 58593 (MO, QCA); Lago Agrio- Río San Miguel, 12.5 km N of Lago Agrio, 0%06'N, 76%50'W, Croat 50321 (MO); Rio San Miguel, 21.2 km М of Lago Agrio, 470 m, 0%8'N, 76°50'W, Croat 50374 (MO); Yasuní Nacional Parque 230 m, 0%52'S, 76?05'W, Cerón 3402 (MO); Rio Yasuni, Lago Garza Cocha, 200 m, 1°05'S, 75°47'W, Lawesson et al. 43345 (AAU) Puerto Napo-Misahuallí, 3.8 km W of Misahualli, 370 m, 1%02'S, 77°42'W, Croat 58888 (MO); at е of Río Tivacuno and Rio Tiputini, 220 m, 0%45'S, 76°20'W, Coello 216 (MO, NY). PASTAZA: Rio Cur ка Сигагау, 2 Кт W of the school, 250 m, 1°22'5, 76°58'W, Holm-Nielsen et al. 21892 (MO); Suc- umbios, Reserva Faunistica Guyabeno, S of Laguna Gar- zacocha, 265 m, 0901'5, 76°11'W, Balslev et al. 84679 (AAU); between Laguna Añangueno and Río Cuyabeno, 265 m, 0%01'S, 76°11'Ұ, Balslev et al. 84882 (AAU); 742 Annals of the Missouri Botanical Garden Rio Cuyabeno, Puerto Montüfar, 230 m, 0°06' $, 76°01'W, Holm-Nielsen et al. 21302 (A 15 km 1915'5, 8 (МО), Hurtado a Neill 1506, 1521, 1572, 1573 о к а & г 18В (МО); Valle де La Muerte, Cur 240 129 S, 76%52'W, Holm ‘Nielsen » al. 22467 (AAU) PERU. AMAZONAS: Prov. Bagua, Rio porate, гони Pinglo, > Pongo де Manseriche, 250-3 ‚ Wur- dack 2109 (BH, US); Rio Santiago, i “Pongo de ааа 200 m, Mexia 6142а (Е, СН, К, UC). HUÁNUCO: Rio Pachitea, Puerto Lira, 300 т, Killip & Smith 26833 (NY, US). LORETO: Iquitos, Amazonas Lodge, E 8463 (MO); Rio Corrientes, almost at Ecua- n border, upriver from Tnte. Lopez oil camp, 280- 300 m, Gentry et al. 1897 1 (МО); Rio Huallaga, Santa Rosa, below Yurimaguas, 135 m, Killip & Smith 28769, 28922 (NY, US); Yurimaguas- Balsapuerto, 135-150 m, Killip & Smith 28285 (NY, 5); Rio Samaria, Атап Cocha, Napo, im E 72%54'"W, Martin et al. 174 Prov. Lor San José de Parinari, Ач m, 4°32! 5. 74°30'W, Våsquez & Jaramillo 3354 (МО); Río Cor- rientes, Valencia, Platanoyacu, са. 3°10'5, 75?41'W, McDaniel & Marcos 11131 (MO); Rio Macusari, 220- 300 m, 2%55'S, 76°15'W, Lewis et al. 11044 (МО); Prov. Maynas, Ecuador border, 1-5 km from Puerto Peru (military post, 8 km from mouth of Rio Gueppi, trib. of Rio Putumayo) on trail to Rio Маро, 200 m, Gentry et al. 22038 (F); Iquitos Region, Las Pebas, Rio Ampi- yacu, 3°10'$, 71?49"W, Plowman et al. 6541 (GH); Rio Yaguasyacu, affluent of Вю Ampiyacu, Brillo Nuevo, 2*40'S, 72*00'W, Plowman et al. 6884 (GH); Rio Ya- vari, Caserio Paumari, Revilla 2203 (MO); Prov. Re- quena, Río Ucayali, Cumaria, 225 m, Tessman 3351 (B). UCAYALI: Pap Las Cachoeiras, Quebrada Sapallal, trib- utary of Q. Shesha, near Peru-Brazil border, 260 m, 8225. 13*55'W, Gentry & Díaz 58467 (МО); vic. LSV base camp, Quebrada Shesha (trib. of Rio Abajao), 65 km NE of Pucallpa, 250 m, 8%02'S, 73*55'W, Gentry & Díaz 58571 (B, K, MO, 05); Prov. Coronel Portillo, Bosque Nacional А. V. Humboldt, 8*40'S, 74°45'W, Vás- quez 3891 (MO). b. Anthurium uleanum var. nanayense Croat, var. nov. TYPE: Peru. Loreto: Mishana, Rio Nanay, 120 m, Solomon 3562 (holotype, MO 2627652). Figure 318. a var. typicum pro ratione habitu parvioro et seriebus spiralibus ferentibus floribus paucioribus. Epiphytic, sometimes terrestrial; stem moder- ately slender, elongate, creeping, up to 100 cm long, 0.7-1.7 cm diam.; leaf scars obscured by root mass, 0.4-1 cm high, 0.9-1 cm wide; roots forming a dense, contiguous mass, ascending or descending, pale green to white, when dried grayish to brownish gray, smooth, fleshy, elongate, tapered, to 10 cm long, (1)2-3 mm diam.; cataphylls sub- coriaceous, lanceolate, 3-6 cm long, acuminate at apex, drying brown to reddish brown, persisting semi-intact, eventually deciduous. Leaves spread- ing-pendent to spreading; petioles 5.5-19 cm long, —4 mm diam., subterete to C-shaped or sometimes D-shaped, with medial rib and margins raised adax- ially, rounded abaxially; geniculum slightly thicker and paler than petiole, sometimes up to remote from base of the blade, 0.3-1 cm long; sheath 2-2.5 cm long; blades thinly coriaceous, somewhat chartaceous, broadly elliptic to oblan- ceolate or broadly oblanceolate, shortly acuminate and apiculate at apex (the acumen downturned, 0-15 mm long), usually attenuate (sometimes acute to obtuse) at base, 22.5-43 m long, 8- 17.5 ст wide, broadest at or above the middle, the margins sometimes concave near the base; upper surface matte to semiglossy, medium green, lower surface matte, slightly paler; both surfaces drying matte to sometimes weakly glossy, grayish green to brown, with reddish brown major veins; midrib bluntly to acutely raised at base, becoming narrowly raised toward the apex above, promi- nently convex to higher than broad and slightly paler than surface below; primary lateral veins 5- 7 per side, departing midrib at 40—55° angle, straight then arcuate to the ma convex above, occasionally raised in shallow val- rgin, prominently leys, more prominently raised below; interprimary veins if present almost as conspicuous as primary lateral veins, weakly sunken above, raised below; tertiary veins obscure above, flat to slightly raised and somewhat darker than surface below; reticulate veins slightly raised or obscure below when dried; collective vein arising from near the apex or absent, if present weakly sunken to flat above, raised below, and raised on both surfaces, ca. 3-7 mm argin. Inflorescences much shorter than leaves; peduncle (4.8)16-28 cm long, drying 1-3 mm diam., 1.2—4.4 as long as petiole, + terete; spathe spreading, subcoriaceous, green, linear-lanceolate, 4—9.4 cm long, 0.6-1 cm wide, broadest near the base, acuminate at apex (the acumen apiculate, occasionally cuspidate), narrowly acute to almost decurrent at base, the margins meeting at 20° angle; stipe 6-17 mm long in front, 1-2 mm long in back; spadix purplish becoming grayish to white, sometimes reported as green, oblong, weakly ta- pered, curved, 5-7 cm long, 3-4(5) mm diam. near base, 2-3 mm diam. near apex; flowers + rhombic, (1.4)1.9-2.1 mm long, (1.1)2.4-2.6 mm wide, the sides smoothly to jaggedly sigmoid; 3-4 flowers visible in principal spiral, 4—6 in alternate spiral; tepals densely and conspicuously papillate to granulose, sometimes appearing glaucous; lateral tepals (1-1.7 mm wide, the inner margins broadly Volume 78, Number 3 1991 Croat 743 Anthurium sect. Pachyneurium convex, the outer margins 2—3-sided; pistils slightly raised; stigma broadly ellipsoid, 0.3-0.4 mm long; stamens grouped in a tight cluster above the pistil; anthers 0.4-0.5 mm long, 0.3-0.5 mm wide; the- cae oblong, slightly or not at all divaricate. /n- I with spathe persisting, withered; spa- 7.5-9.8 cm long, 0.6-0.9 cm diam.; berries P + ovoid, ca. 3.5 mm long, 3.4 mm diam.; seeds drying yellowish brown, ovoid, darker at the ends, ca. 1.7-2 mm long, 1.2-1.4 mm diam., 1 mm thick. Anthurium uleanum var. папауепзе is endemic to Peru in Loreto Department at 100 to 160 m, in the tropical moist forest transitional to tropical wet forest life zone. All collections have been made in the Iquitos region in Maynas Province, many from the Rio Nanay (hence the name). This variety is characterized by its relatively small size, rather elongate stem, broadly oblance- olate to almost obovate leaf blades drying brown and attenuate to narrowly acute at the base and slender, tapered spadix apparently starting as pur- plish and eventually becoming grayish white (ap- pearing glaucous). Variety nanayense is likely to be confused only with var. uleanum, which occurs in the same area and differs in having longer petioles, a longer spa- dix, more oblong leaf blades, and fewer flowers per spiral. Variety nanayense is smaller in overall size than var. uleanum, and the uniformity of herbar- ium specimens of the former is notable. PERU. LORETO: 17 km SW of Iquitos, Croat 184604 (MO); Rio Nanay, Iquitos, 100 m, Plowman 2482 (F, СН); Mishana, 120 m, Solomon 3559, 3562 (МО); Rio pores Santa Elena, McDaniel & Marcos 11261 (МО); Pro aynas, Iquitos Region, ca. 120 m, Revilla 3597 (MO), 3685 (AMAZ, F, MO); Rio Momón, 130 m, Croat 19929 (MO), pud & Jaramillo 45 (MO), нары 7711, 7856, 7858 (IBE, MO); Mishana, 130 m, 3%55'S, 73935", er et al. 5319 (MO); Puerto Almendras, 12 , 3°48'S, 73°25'W, Vásquez & Jaramillo 3159, 6100, 661 d 7590 (MO); Rio Nanay, Caseria Mishana, 30 km S uitos, Foster 4215, 4227, 4350, 4437 (F); pu om ria de Nanay, Mishana, 1 3?50'S, 7 "W, r Emmons 38015 (MO), Gen try et al. 28905, 31632, 36437 (MO); Santa Maria de Na 130 m, Schunke 3 (F, NY); Rio Putomayo, nay, 130 m, 245 Ri Ecuador border, Guarnición Militar de pus Díaz 356 °00'W, Plowman et al 6854 Prov. ap ren Dtto. Iquitos, Iquitos Region, Puerto Al- mendras, 120 m, 3%48'S, 73?25'W, Croat 51207 (MO); Iquitos-Puerto Almendras, 15 km from Iquitos, 120 m, 3%47'S, 73°25'W, Croat 51190 (MO); Rio Nanay, Ca- rretera de Picuruyacu, 160 m, Rimachi 5464 (IBE); Prov. Requena, Río Ucayali, 4*55'S 73%45' W, Gentry et al. 56361 (MO). e Anthurium upalaense Croat & R. A. Baker, Brenesia 16(Suppl. 1): 97. 1979. TYPE: Costa Rica. Alajuela: between Canas (Guanacaste) and Upala, near Rio Zapote, 1.8-2.7 km S of Rio Canalete, ca. 1,000 m, Croat 36342 (holotype, MO 2381219; isotypes, CR, F, K). Figures 321, 323, 324 Epiphytic; stem less than 20 cm long, 1.5-3 cm diam.; leaf scars obscured by root mass and persistent cataphylls; roots moderately dense and rown to green, scurfy to pubescent, moderately thick and elongate, mostly 2-6 mm diam.; cataphylls lanceolate, coriaceous, acutely l-ribbed throughout, 11-25 cm long, obtuse to acute to narrowly acuminate at apex, plain green to slightly tinged with red, drying brown, persisting intact, eventually deciduous. Leaves erect-spread- ing to spreading; petioles (6)10-47 cm long, 7- subquadrangular, broadly sulcate numerous, mm diam., with the margins sharply raised adaxially, 3-5- ribbed abaxially, the surface pale-speckled; geni- culum thicker and paler than petiole, sometimes reddish abaxially, 1-2 cm long; blades subcoria- ceous to coriaceous, elliptic to narrowly oblong- elliptic to oblong-lanceolate, obtuse to acute to gradually acuminate at apex (the acumen down- turned and weakly apiculate), usually acute to nar- rowly acute (rarely obtuse) at base, (30)40-106 cm long, (4.5)8-34 cm wide, broadest at or near the middle, the margins prominently undulate; up- per surface semiglossy, dark green, lower surface semiglossy to glossy, slightly paler; midrib flat to obscurely angular at base, becoming prominently and sharply raised toward the apex and paler than surface above, prominently higher than broad and 3-ribbed at base, becoming prominently and acute- ly raised toward the apex below; primary lateral veins 10-20 per side, departing midrib at 60-75? angle, almost straight to the collective vein, loop- connected in the upper half, prominently and con- vexly raised above and below, more so above, paler than surface; interprimary veins less prominent than primary lateral veins; tertiary veins weakly visible, paler than surface above, darker than sur- face below, flat above and below; reticulate veins partially visible above, darker than surface below, drying prominulous above and below; collective vein arising in the upper half or in the upper third of the blade, flat, when dried slightly raised on both surfaces, 3-5 mm from margin. /nflorescences erect to spreading or arching-pendent; peduncle (18)40-54 cm long, 4-10 mm diam., equaling to 1.4-2.4X as long as petiole, light green, weakly pale-speckled, sometimes tinged with purple at base, 744 Annals of the Missouri Botanical Garden terete or l-ribbed below opening of the spathe, sometimes faintly many-ribbed; spathe erect to re- flexed, often rolled up, subcoriaceous, yellow-green, sometimes heavily tinged with violet-purple (B & K yellow-green 6/7.5 to 7/2.5), oblong-lanceolate to lanceolate, (4.5)8-20 cm long, 1-4 cm wide, broadest near the base, inserted at 60-75? angle n peduncle, short-acuminate (the acumen mi- nutely inrolled), acute to decurrent on peduncle at base, the margins meeting at 20—60? angle; spadix green (B & K yellow-green 5/6) to maroon (B & К red-purple 2/7.5), long-tapered, 9-30 cm long, 6-15 mm diam. near base, 4-6 mm diam. near apex; flowers square to rhombic, 2.3-3.4 mm long, 1.7-2.8 mm wide, the sides shallowly sigmoid to straight; 10-20 flowers visible in principal spiral, 7-15 in alternate spiral; tepals matte, inconspic- uously pale punctate, with droplets appearing usu- ally after 2 or more anthers emerge; lateral tepals 0.8-1.8 mm wide, the inner margins straight to slightly convex; pistils raised, rectangular to bluntly pyramidal, green to purplish; stigma ellipsoid, green, 0.4-1.1 mm long, brush-like, droplets appearing 12-22 days before stamens emerge, persisting for 2-15 days; stamens emerging + rapidly from the base or near the base, grouped in a tight cluster above the pistil; filaments retracting, short, ca. 0.2 mm long, 0.5-1 mm wide; anthers pale orange, 0.5-0.6 mm long, 0.7-1 mm wide, contiguous; thecae ellipsoid, scarcely divaricate; pollen orange fading to pale orange to lavender to white or tan. Infructescence pendent; spathe reflexed; spadix to 55 cm long, 6 cm diam., with berries emerging + uniformly throughout; berries red, oblong, acute at apex, tapered to both ends, to 15 mm long; mesocarp mealy, white, sweet; seeds 2 per berry, white, oblong, to 4 mm long Anthurium upalaense is known from the At- lantic slope of Nicaragua and Costa Rica, from near sea level to usually less than 700 m (rarely to 1,300 m) in tropical wet forest, premontane wet forest, and wetter parts of tropical moist forest. This species is distinguished by its persisting, intact cataphylls, moderately long, subquadran- gular, sulcate petioles which are 3-5-ribbed abax- ally, long-tapered, usually violet-purple spadix and red fruits. Anthurium upalaense is most closely related to А. purpureospathum. See that species for distin- guishing characters. It is also apparently related to A. concolor from Panama, but differs in having the inflorescence pendent, the spadix long-tapered and more or less green at anthesis, the spathe longer and recurled, and the fresh anthers pale orange. In Nicaragua Anthurium upalaense can be con- fused with A. schlechtendalii, but that species differs in having most cataphylls weathering to fibers, as well as generally shorter, less conspicu- ously tapered spadix and thicker, more prominently reflexed spathe which is more heavily tinged with purple. Costa Rica. Cultivated at Kew: Brown s.n. (K), Bull 467/1873 (К). ALAJUELA: Florencia-Fortuna, 1 km S of Rio Penas Blancas bridge on road 142, 80 m, 10°22'N, Hammel et a 14054 ee ыш of е Arenal Volcano, una, Taylor & Taylor W of For (NY); San Juan de Lajas, Smith 14 498 (NY); А of Ciudad et al. 17264 (ЕАР, Е); Canas- Upala Road, 3-4 km of Bijagua, along new road to Upala, 400-450 m, 10?45'N, 5*3'W, Baker & Burger 9859 (CR, Е, MO), Croat s (MO); Rio Zapote, 1.8-5 km S of Rio Canalete, 100-200 m, Burger & Baker 9964 (F), 10001 (CR, F, MO), Croat 36380 (MO), Croat 36342 (CR, F, K, МО); Lago Arenal, Alto de la Carpintera, Tilarán, Brenes 18194, 18195 (CR); Rio Frio, Los Chiles, 30-40 m, 11?2'N, 4°44’W, Holm & lltis 771 (NY), 811 (B, К, P, U). CARTAGO: 24 km NE of Turrialba on road to Limón, 450- 525 m, 9*58'N, 83?34'W, Liesner et al. 15363 (B, MO, ы. 5 km МЕ of Rio Grande de Orosi at Tapanti, 1,300 ‚ Utley & Wd 5062 (MO); Tucurrique, Las Vueltas, aida 13319 (US); Forests of Tuis, Turrialba, Pittier 12720 (US); Lake Bonilla, Pacheco 92 (F). GUANACASTE: El Arenal, Standley & Valerio 45328 (US); Rio Agua- cate, Laguna de Arenal, 500 m, Khan et al. 1085 (BM); Río Chiquito, slope above Lake Arenal, 600-700 m, 10%25'N, 84%46' W, Haber et al. 4819 (MO). HEREDIA: Puerto Viejo Region, Finca La Selva, Folsom 9558(TEX), Croat 61218 (MO); on Rio Puerto Viejo just E of its junction with Rio Sarapiqui, 100 Chacón 462 (MO), Croat 44263, 44297 (MO), McDowell 389, 1058 (MO); Ко Sucio, 20 m, Croat 35708 (MO); Puerto Viejo-San Jose, SW of Puerto Viejo, Burger & Mata 4343 (F, CR). L со MO, MSC, У mp at Manzanillo de Talaman 8230W, Grayum & Burton 4324 (MO Puerto sel e Talamanca-Manzanillo, 10-20 m, 10°38-39'N 82%40-44"W, Стауит 3638 (MO); Hone Creek-NW on trail paralleling Rio Carbon, Utley & Utley 660 (F); Limón, Burger & Burger 8446 (CR, F); Limón-Puerto Viejo, Baker & Burger 174 (Е); hills 3.5 air km S of Islas Buena Vista in Rio Colorado, 16 air km SW of Barra del Colorado, 10-120 m, 10?39'N, 83?40'40"W, Dav- idse & Herrera 31271 (МО, W); Cerro Coronel, E of Río Zapote, 10-100 m, 10?40'N, 83%40'W, Stevens 23923 (MO); E of Laguna Danto, 20-170 m, 10?41'N, Volume 78, Number 3 1991 Croat 745 Anthurium sect. Pachyneurium 83938", Stevens 23726 (MO); Cordillera de Talamanca, Reserva Biol. rtuguero, 4 m, 10? 531" . бо ag 29 Anai, Sixaola, 25-40 m, 9? 82°39'W, a 4501 (MO); Rio Catarata, Bribri, 50 m, Croat 43221 (F, K, MBM, MO, o RSA Madre, 11 mi SW of Limón, Hwy. 43305 (AAU, CR, G, GH, K, LE, us NY, P, RSA, UCLA, W, WISy; Rio Reventazón, Cairo, Hambur poo d Valerio 48869 (US). f Coto Brus, Las Cruces bres Botanical Garden, 6 Es W of San Vito de Java, 1,200 m, Croat 57257 (CR, alva & Soza 5147 п Oluma, 600-700 m, 12°18’М, 85?23'W, Nee 28359 (MO). JINOTEGA: Rio Coco, Rapido Plis, 200 m, 14°0- 10'N, 85?20-30'W, Stevens et al. 16823 - o Musun, Rio Bilampi, 500-800 Neill 1782 (МО); "Wan nawas, 200-500 m E BM, LE, MBM, МО), 2744 (МО); 4 N, me oe Te ти Las , Stevens & Morenó 19715 (MO). RÍO SAN bo. e м, Cruz- Cano Santa Crucita, La Palma, 40-60 m, 1 km W of San Juan del Norte, 0-2 a е 83 Riviére 222 (МО); Rio San Juan ua Cano la Barca, 0-10 m еј вата а 3283, 3302, 3401 (MO), Martinez & Riviere 1952 ney 1957 (MEXU, МО). zELAYA: Waslaya-Siuna, of La Cuesta El Balsamo, Rosa Grande, 260-300 m, 13%38'М, 85°2'W, Mercedes, 160-180 m, Инес 328A (MO); Monkey Point, 3 km above Caño El Pato, 10 m, 11°35'N, 83?42'W, Moe 12466 (МО); Siuna- Matagalpa, 31.4 km beyond Rio Uli, 8.9 km beyond Rosa Grande La Balsamo, 200 m, Stevens 7524 (MO); Сапо Monte Cristo, La Grupera, 10 m, 11?33'N, 84°48’W, Moreno & Sandino 14688, 14655 (МО); village of Germán Pomares, Las Benitas, 100 m, 11?36'N, 83°51'W, Moreno & Sandino 14933 (MO); Cerro Waylawas (Penas Blancas), 5 k S of Wany, 13*30'20"N, 84?45'25"W, Grijalva € Burgos 1685 (МО); Rio Punta Gorda, Atlanta, 10 m, 11934'М, 84°1'W, Moreno & Sandino 12755 (MO); Rio Sucio, 0.5 km E of lst suspension bridge E of Bonanza, 140 m, 14°1'N, 84°34’W, Stevens 12317 ( Anthurium validifolium K. Krause, Notizbl. Bot. Сагі. Berlin-Dahlem 11: 607. 1932. TYPE: Panama. Chiriqui: forests along Rio Ladrillo, above El Boquete, 1,200-1,300 m, Pittier 3052 (holotype, US). Figures 322, 325-327, 231 Epiphytic; stem to 16 cm long; leaf scars 2.5 cm wide; roots numerous, descending, green, smooth, 5-15 mm diam.; cataphylls moderately coriaceous, 20-30 cm long, long-acuminate at apex, medium green, sparsely pale-speckled, matte, dry- ing tan ( yellow-red 9/10), persisting + intact. Leaves erect-spreading; petioles 20-27 cm long, 7-10 mm diam., quadrangular to bluntly D-shaped, flattened to broadly and shallowly sul- cate adaxially, rounded abaxially, surface раје- speckled; geniculum thicker and slightly paler than petiole, becoming fissured transversely with age, 3—4.5 ст long; blades coriaceous, oblong to ob- long-oblanceolate or lanceolate, acute at apex (the acumen apiculate), acute and decurrent to rounded at base, 47-69 cm long, 10-19 cm wide, broadest at or near the middle, the margins broadly undu- late; both surfaces matte, dark green above, paler below; midrib flat at base, becoming obtusely an- gular and narrowly raised at apex above, bluntly 2-ribbed and higher than broad at base, becoming prominently and convexly raised toward the apex elow; primary lateral veins numerous, departing midrib at 60—70° angle, += straight, arcuate near the margin, sunken in grooves above, raised below, paler than surfaces; interprimary veins obscure above, darker than surface below; tertiary veins obscure above, darker than surface below; collec- tive vein arising in the upper third of the blade, weakly sunken above, raised below, less prominent than primary lateral veins, 5-6 mm from margin. Inflorescences spreading, shorter than leaves; pe- duncle 33-39 cm long, 9-12 mm diam., 2x as long as petiole, subterete; spathe reflexed, coria- ceous, dark green yellow-green 5/10), ovate-lanceolate, 16-17 cm long, ca. 4 cm wide, inserted at 45? angle on peduncle, acuminate at apex, rounded at base; stipe 7-13 mm long in front, ca. 3 mm long in back; spadix pale lavender to violet-purple (B & K purple 6/5 to 5/2.5), tapered, 15-23 cm long, (4)6-12 mm diam. near base, 2-5 mm diam. near apex; flowers rhombic, 3-3.5 mm long, 3.1-3.6 mm wide, the sides и to weakly sigmoid; 7-8 Брега visible in prin spiral, 8-10 in 1; tepals matte; iam tepals 1.5-1.8 mm wide, disi inner margins broadly convex; pistils scarcely emergent, green; stigma ellipsoid, ca. 0.6 mm long, droplets appearing for са. 1 week before stamens emerge; stamens emerg- ing rapidly from near the base or sometimes from 746 Annals of the Missouri Botanical Garden near the middle, lateral stamens emerging to mid- n peus alternates appear; anthers yellow-or- e, ca. 1 mm long, 1.2 mm wide, inclined over к ы, thecae oblong-ellipsoid, scarcely divari- cate; pollen yellow-orange (B & K yellow-red 8/5). Infructescence pendent; spathe twisted; spadix 24— 26 cm iam.; berries са orange, obovoid, flattened at apex, 0.8-15 m long; pericarp thin, + transparent; mesocarp fleshy, orange, with minute raphide cells; seeds 1-2 per berry, pale yellow, ovoid, with a somewhat flattened appendage at both ends, enveloped by somewhat dry, sac-like structure. long, Anthurium validifolium is known from Chiriqui Province in Panama and adjacent Puntarenas Province in Costa Rica. Most collections have been made in lower montane rain forest, from 1,200 to 2,300 m. This species is characterized by its large, mod- erately thick, lanceolate leaf blades, lavender to purple spadix, and bright orange, obovoid berries. Anthurium validifolium is probably most close- ly related to А. seibertii and А. protensum, whic share a similar inflorescence and orange berries. It differs from both in having broader and thicker leaf blades. Its subquadrangular petiole is somewhat intermediate between the terete petiole of А. pro- tensum and the sharply quadrangular petiole of А. seibertii. Although this suggests А. validifolium may a hybrid between А. protensum and А. seibertii, its thicker, ia shorter blades are not a feature of those species. The specimen from Puntarenas was cultivated at Las Cruces Tropical Botanical Garden, but was reported to have been collected locally by Robert Wilson, founder of the garden. CosTA RICA. PUNTARENAS: Cantón of Coto Brus, Las Cruces Tropical Botanical Garden, 6 km of San Vito de Java, Croat 44445, 57258, 57279 (MO). PANAMA. CHIRIQUÍ: 2 mi. N of El Hato del Volcán, Croat 10667 , MO, NY, PMA, RSA, US); 2 300 т, 8*54'N, aged W, Knapp 1513(F, Pelota, pes slopes, 1,800-2,000 m, 8*53'N, 8243", Knapp 1497 (K ‚ МО); ridge S of Cerro Pelota, 1,700- 2,200 m, 8*53'N, 82°43’ W, Knapp a 1523, 1526, 1528, 1529 (MO); Cerro Punta, 4.7 mi. E of Volcán, 1,600 m, 8?49'N, 82°38'W, Croat 66204 (МО): Volcán- Rio Sereno, 1,520 m, Croat 66321 (MO); Bambito- Cerro Punta, Croat 10553 (MO); Croat 10554 (F, MO); Bo- W of Boquete, 1,870 m, Davidse & D'Arcy 10309 (MO, PMA); Cerro Pando, NW side, Croat 1597 1 (MO, NY); Rio Colorado, 1,200- 1,400 m, 8*50'N, 82%43"W, Hamilton & Stockwell 3532 (МО); Río Ladrillo, El Boquete, 1,200-1,300 m, Рипег 3052 (US). Anthurium vaupesianum Croat, sp. nov. TYPE: Colombia. Vaupés: vic. Mitú, 2-3 km below Urania, Rio Vaupés, Zarucchi & McElroy 1172 (holotype, GH; isotype, COL). Figures 329, 330. Planta epiphytica; internodia brevia, 1-2 cm diam.; д lanceolatum; petiolus 3 9 cm ш 3-5 daxile, abaxile r A ad trinervis: cin coriacea- а а, ођ- lanceolata-obovata, abrupte attenuate in dimidio inferiore, 20-62 cm yn 9-26 cm lata; nervis primariis late- ralibus 6-12 utroque; pedunculus (16)36- 70 cm longus; spatha а linearis, viridis-olivacea, 9-20 cm longa, 0.6-1.5 cm lata; spadix brunneo-rubellus aut P rirellus, sessilis vel stipitatus, ad 3.5 cm, longe contractus, 20- 35 cm longus, 3-4 mm diam.; baccae violaceae. Epiphytic; stem to 9 cm long, 1-2 cm diam.; roots dense, ascending, fuzzy, soon glabrous, 3-4 mm diam.; cataphylls subcoriaceous, lanceolate, 3— 5 ст long, acuminate at apex, drying dark reddish brown, semi-intact, persisting briefly. Leaves erect- spreading; petioles 4-9 cm long, 3-7 mm diam., D-shaped, flattened and conspicuously and obtusely ribbed adaxially, rounded to weakly angulate or weakly 3-ribbed abaxially; geniculum thicker than petiole, 0.5-1 cm long; sheath 3-7 cm long; blades coriaceous to subcoriaceous, oblanceolate-obovate, acute at apex, narrowly attenuate, ending narrowly acute to obtuse at base, 20-62 cm long, 9-26 cm wide, broadest above the middle, the margins usu- ally flat, sometimes weakly and broadly undulate; upper surface semiglossy, medium green, lower surface weakly glossy, paler; both surfaces drying grayish to dn brown; midrib convexly raised at base, becoming sharply raised (knifelike) toward the apex above, acutely raised at base, becoming convex toward the apex below; primary lateral veins 6-12 per side, departing midrib at 30-50? angle, straight to arcuate-ascending to the margin, raised in shallow grooves above; tertiary veins con- spicuously sunken above, conspicuously raised be- low; collective vein arising from near the apex or absent, sunken above, raised below, 6-12 mm from margin. /nflorescences erect to pendent, + equal- ing ue peduncle (16)36-70 cm long, (1)3-4 mm diam., (2.7)4- 7.7 х as long as petiole, terete; spathe spreading to reflexed, subcoriaceous, green to brownish green, sometimes mottled purple, ob- long-linear, 9-20 cm long, 0.6-1.5 cm wide, in- serted at 40? angle on peduncle, acute at apex, attenuate at base; spadix brown to reddish to green- ish or gray-brown (once reported as white), long- tapered, sessile or stipitate to 3.5 cm, somewhat curved, 20-38 cm long, drying 3-4 mm diam. Volume 78, Number 3 1991 Croat 747 Anthurium sect. Pachyneurium midway, 2 mm diam. near apex; flowers rhombic, 1.4-1.8 mm long, 1.2-1.4 mm wide, the sides straight to smoothly or jaggedly sigmoid; 3—5 flow- ers visible in principal spiral, 3 in alternate spiral; tepals semiglossy, d papillate; lateral tepals 0. , the inner margins straight to rounded, the рела, margins 2-sided; anthers 0.4 mm in both directions, inclined over and obscuring the pistil; thecae ellipsoid, slightly divaricate. /n- fructescence with spathe persisting + intact; spadix 0.9 cm diam.; berries violet. mm wi Anthurium vaupesianum is endemic to south- ern Colombia in Vaupés (hence the name) and Amazonas departments, where it is an epiphyte at 130-250 m in tropical premontane wet forest. This species is characterized by its small size, short petioles, obovate-oblanceolate leaf blades dry- ing grayish-brown and usually rather abruptly at- tenuate in the lower half with conspicuously con- cave margins, and by its very slender, long-tapered spadix. It is probably most closely allied to А. anorianum from Antioquia Dept. in northern Co- lombia. Although it has a similar long, slender spa- dix, the latter has larger leaves that have convex margins in the lower half and are obtuse to acute at the base, and peduncles that dry reddish. It occurs at 300 to 750 m, in a region separated from 4. vaupesianum by two major mountain ranges. COLOMBIA. AMAZONAS: Río Caqueta, La Padrera, 240 m, Garcia-Barriga 14605 (COL). AMAZONAS-VAUPÉS: Rio Apaporis, Rio Pacoa-Rio Cananari, Soratama, 250 m, Schultes & Cabrera 13535, 13796 (ECON), 12772, 13621 (US); Jinogoje, ome of Rio Piraparana, 230 m 0915'5, 70°30’ s & Cabrera 19820 (US). VAUPÉS: 2 km SE of Mitt, 200 m, 1?14'N, 70°14’W, Croat 96798 (AAU, COL, Е, JAUM, К, M, МО, NY, ; Rio Kubiyu, vicinity of Мий, Zarucchi & Divis 1185 (COL, GH); Río Piraparana, middle course, Schultes & Cabrera 17128 (US); Rio Vaupés, Mita, 2- 3 km below Urania, Zarucchi & McElroy 1172 (COL, Anthurium venosum Griseb., Cat. Pl. Cub. 219. 866. TYPE: Cuba. Westindische Province, Wright 3209 (holotype, MO 2058854; iso- types, NY, US). Figures 328, 349, 350. Terrestrial or epilithic. Stem short, ca. 1.5-2 cm diam.; roots dense, grayish, moderately to densely Канага. | modenstely elongate, ca. 3- mm diam.; I anceolate, 3— 6 cm Ки acute at apex, brown (B 4 K yellow 4/2.5), persisting semi-intact, eventually as a re- ticulum of fibers. Leaves with petioles (5)15-40 cm long, ca. (3)6-10 mm diam., terete to weakly sulcate, medium green; geniculum 1 cm long; blades coriaceous to subcoriaceous, oblong-ovate, acute to shortly acuminate at apex (the acumen shortly apiculate), + rounded to subcordate or weakly cordate at base, (17)24.5—46 cm long, (12.5)15.7- 21.5 cm wide, broadest at or near the middle, the margins broadly undulate; anterior lobe (15)22.5- 40 cm long; the posterior lobes (3.5)5—6 cm long, (4)5-6 cm wide, directed downward and rounded at apex; sinus mostly triangular, rarely arcuate with the blade decurrent on petiole to shallowly triangular, 1-3 cm deep; both surfaces glossy, moderately bicolorous, drying yellowish brown to yellowish green; midrib convexly raised above, pal- er than surface, more prominently raised below; basal veins 3—4 pairs, all free to base, raised above and below; primary lateral veins 5-9 per side, departing midrib at 40—65? angle, straight to broadly arcuate, prominently raised and paler than surface above and below; interprimary veins almost as con- spicuous as primary lateral veins; tertiary veins prominulous; reticulate veins raised; collective vein arising from near the apex or absent. /nflores- cences spreading; peduncle 12-84 cm long, ca. -8 mm diam., 1.4-2.9 x as long as petiole; spathe reflexed, subcoriaceous, green, lanceolate to linear- lanceolate, 3-10 cm long, 1-1.5 cm wide, broadest near the base, acuminate at apex (the acumen sometimes cuspidate, 6-7 mm long), acute to de- current at base; stipe 15-17 mm long in front, 5 mm long in back; spadix brownish green to purplish brown, tapered, + erect to curved, 2-15 cm long, –6 mm diam. near base, 2-3 mm diam. near apex; flowers rhombic, 2.5-2.9 mm long, 1.4-2 mm wide, the sides straight to smoothly sigmoid; 5-6 flowers visible in principal spiral, 8-9 in al- ternate spiral; tepals weakly roughened; lateral tepals 1.3-1.7 mm wide, the inner margins broadly convex, the outer margins 2-sided; pistils pale green; stigma = ellipsoid, 0.3-0.4 mm long; anthers 0.3- 0.4 mm long, 0.5-0.7 mm wide; thecae + ovoid, not at all or only slightly divaricate. /nfructescence with spathe persisting; spadix to 17 cm long, to 2 cm diam.; berries orange-red, + ovoid to obovoid, 6-6.5 mm long, 4.5-6.2 mm diam.; seeds 1-2 per berry, brown with numerous pale raphide cells on the surface, 3.7—4 mm long, 1.7-2.8 mm diam., 1.2-1.8 mm thick. Anthurium venosum is known only from Cuba and is apparently restricted to Pinar del Rio Prov- 748 Annals of the Missouri Botanical Garden ince, especially around Vinales, where it occurs on limestone cliffs at less than 300 m This species is distinguished at once by its rather small, oblong-ovate leaf blades with conspicuous basal veins (3 to 4 pairs) and more or less rounded to shallowly cordate leaf bases, and brownish green, tapered spadix. Anthurium venosum is a distinctive species with- in the section, but has been confused with other species in the West Indies. Most notable perhaps is a population of plants on Jamaica in St. James Parish at 700 m, which resembles both А. тал- cuniense C. D. Adams and the present species. This was reported by Adams (1972) for Jamaica based upon Proctor 7977 and Proctor & Stern 11748. Although these specimens look more like А. venosum, a re-collection (Croat 69757) proves that it belongs to neither species and probably represents a new species not belonging to sect. Pachyneurium. CUBA. NO FURTHER LOCALITY: Wright 3209 (MO, NY, 05); Cultivated at Havana Botanical Garden, Croat 69756 (MO). ied DEL RÍo: 250 m, Rutten-Pekelharing & 5 (B), Hermann 3183 (NY, US); Mogote de Gillavry 42 la Ba 14378 (NY); Mogote de La udera Vinales, Leon Jagua, La Jagua, Consolación del Norte, Roig & Asary 2540 (NY) San Vicente, een "Es 22746 (05); Sumidero, Shafer 13378 (A, BM, NY, US); Barios San Vicente, Britton et al. 7314 (NY, as Vinales, Killip 13583 (US); Vinales, Mogote de Santo Tomas, Arias s.n. (MO); Sierra de Viñales, Ekman 165396 (NY, US). Anthurium vinillense Bunting, Phytologia 64: 462-463. 1988. TYPE: Venezuela. Amazonas: Dept. Rio Negro: hilly savanna, Serrania de Vinilla, ca 20 km SW of Mavaca, 2?26'N, 65?20'W, ca. 420 m, Huber 6201 (holotype, VEN, isotypes, MO, NY). Figure 354. Description based on dried material only. Usu- ally terrestrial, sometimes epilithic, to ca. 35 cm tall; internodes short; cataphylls narrowly trian- gular, 5.5 cm long, narrowly acute at apex, drying reddish brown, the uppermost persisting semi-in- tact, promptly weathering to fibers. Leaves erect; petioles 3.2-5.2 cm long, less than 5 mm diam., C-shaped, weakly sulcate with raised margins and an obscure medial rib adaxially, rounded abaxially; sheathing to midway; 5.6-7.3X as long as petiole; geniculum scarcely thicker and drying darker than petiole, ca. 0.5 cm long; blades coriaceous, elliptic to obovate-elliptic, acuminate at apex (the acumen short, with margins thickened and turned upward), acute to obtuse at base, 16-28 cm long, 4.3-9.5 cm wide, broadest at or slightly above the middle, 3-3.6X longer than broad; upper surface incon- spicuously punctate, lower surface „= дагк glandular-punctate (those glands ca. 1.5 mm diam.); both surfaces drying yellow-green, glossy; midrib convexly raised on both surfaces; primary lateral veins 5-8 per side, departing midrib usually at ca. 10? angle from midrib, then spreading at 40? angle, moderately straight or weakly to markedly arcuate to the margin; collective vein lacking or arising from second to last primary lateral vein. /nflores- cences erect, longer than leaves; peduncle 27-39 cm long, 5.6-7.3 х as long as petiole; spathe twist- ed, green, linear-lanceolate, to 6 cm long, 1 cm wide, thickened at base, inserted at 30? angle on peduncle, acuminate at apex (the acumen inrolled), decurrent for 7-10 mm at base, promptly drying and ultimately deciduous or nearly so; prominently and oly еше adaxially, не ригрје-рипс- tate eins) with pale, short raphide cells; stipe 4-6 mm 4 spadix green tinged with maroon, Vli tapered at apex, erect, to 10.5 cm long, 5-6 mm diam.; flowers rhombic, 2-2.3 mm long, usps mm wide, the sides mod- erately straight; 7-9 flowers visible in principal spiral; tepals obscurely pale-punctate; lateral tepals 1.6-1.9 mm wide, the inner margins broadly rounded, the outer margins 2-sided; pistils not at all exserted; stigma linear, 0.6 mm long; stamens held just above tepals in a tight cluster, inclined over and obscuring the pistil; anthers 0.4 mm long, 0.7 mm wide; thecae ovoid, moderately divaricate. Infructescence with spathe withered, bearing ber- ries in the basal fourth only; berries dark wine-red toward apex, paler to whitish in lower part, to 7 mm diam.; seeds 2 per berry, slightly flattened, 3.8 mm long, 2.6 mm diam enrfare and Anthurium vinillense is endemic to Venezuela, known only from the type locality in the Serrania del Vinilla in southern Amazonas. This species is characterized by its small size, short petioles that are sheathed to half their length, glandular-punctate, elliptic to subelliptic, coria- ceous blades which dry with prominently raised reticulate veins on both surfaces, and dark wine- red berries. Anthurium vinillense is closely related to both А. xanthoneurum and A. bonplandii subsp. guay- anum and will perhaps prove to be only a subspe- cies of А. bonplandii. The type locality of Ап- thurium vinillense is a relatively remote and poorly explored area. Anthurium bonplandii subsp. guayanum has been collected in numerous local- ities to the north and south of the Serrania del Vinilla, but plants of subsp. guayanum are much Volume 78, Number 3 1991 Croat 749 Anthurium sect. Pachyneurium larger and never have the reticulate veins promi- nently raise P p NEZUELA. AMAZONAS: Dpt. Río Negro, ca. 20 km W of Mavaca, 420 m, 2?26'N, 65?20'W, Huber 6201 (MO, NY, VEN). Anthurium wagenerianum K. Koch & Bouche, Ind. Sem. Hort. Berol. App. 7. 1853. TYPE: Venezuela. Caracas, Wagener s.n. (Type de- stroyed; cultivated at Vienna; Schott illustra- tions 542-544 serve as lectotype). Figures 335-338. Anthurium bes inp i Bunting, deua: as mr 295. 1986. Venezuela. Cultivated at Jard. Bot. Maracaibo, ly collected by van der Werff in Falcón: Dt amora e of Rio Hueque, Coro- orón Н of Coro, Bunting 7704 (holotype, NY, аин МО, УЕМ). Terrestrial, often epilithic; stem short, 1.5-3.5 cm diam.; roots moderately numerous, ascending to descending, greenish gray, often fuzzy or woolly- pubescent, thick, blunt, to 15 cm long, 2-7 mm cataphylls subcoriaceous, broadly lanceo- late, 3-7 cm long, acute at apex, yellowish green weakly tinged with red, drying brown to reddish brown, usually with visible raphide cells, persisting semi-intact, eventually as coarse linear fibers. Leaves erect to erect-spreading; petioles (4)6-25 cm long, drying (1)4-10 mm diam., subquadran- gular to trapezoidal or D-shaped, flattened to broad- ly sulcate to obtusely V-sulcate adaxially, with the margins sharply to bluntly raised, rounded to 3-ribbed abaxially; geniculum paler and thicker than petiole, becoming fissured transversely with age, 0.5-1.5 cm long; blades subcoriaceous to coriaceous, narrowly to broadly elliptic or oblan- ceolate, acute to obtuse or rounded at apex (the acumen briefly cuspidate), acute to attenuate at base, (8)1 7-70 cm long, (4)6-30 cm wide, broad- est at or above the middle, the margins conspic- diam.; uously wavy, minutely undulate when dried; both surfaces matte to weakly glossy to semiglossy, dark to medium green above, slightly paler below, with linear raphide cells conspicuous on drying (es- pecially below), occasionally with pustules on either surface; midrib flat to weakly angular at base, becoming narrowly acutely raised toward the apex above, prominently 3-ribbed to convexly raised at base, becoming convexly raised toward the apex below; primary lateral veins (3)4-7 per side, de- parting midrib at 30-65? angle, + straight then arcuate close to margin, prominently and narrowly convex above, less narrow and prominent below; interprimary veins few, especially in juvenile leaves, obscure; tertiary veins sunken to obscure above, weakly raised and darker than surface below; col- lective vein arising from near the apex, weakly sunken above, somwhat raised and darker than surface below, 2-10 mm from margin. /nflores- cences erect to spreading, shorter than leaves; pe- duncle 8-54 cm long, 4-6 mm diam., (1)1.5-3x as long as petiole, medium green tinged reddish at base, terete; spathe spreading to reflexed, subco- riaceous to moderately coriaceous, plain green to tinged with red-purple, lanceolate to broadly lan- ceolate, sometimes ovate, 2-6 cm long, 0.5-2 cm wide, broadest at or near the middle, inserted at 80-90? angle on peduncle, acute to acuminate at apex (the acumen inrolled, 2-5 mm long), obtuse to decurrent at base; spadix creamy to green to gray-green at anthesis, becoming red to purple- violet post-anthesis, cylindroid to weakly tapered, sessile or stipitate 1-25 mm, mostly erect, some- times slightly curved, held at 130-180" angle from peduncle, (1.5)4-12.5 cm long, 5-12 mm diam. midway; flowers mostly square to occasionally 4-lobed, (1.2)1.6-3.1 mm in both directions, the sides straight to smoothly sigmoid; 5-10 flowers visible in principal spiral, 5-8 in alternate spiral; tepals matte, very minutely papillate, with pale protuberances when dried; lateral tepals 0.8-1.8 mm wide, the inner margins straight to convex, often broadly so, the outer margins 2-, occasionally 3-sided; pistils early-emergent, semiglossy, densely and minutely papillate, bright green; stigma linear to ellipsoid, 0.2-0.3 mm long; stamens emerging from the base, the laterals preceding the alternates y 6-13 spirals, the 3rd preceding the 4th by 3- 6 spirals, held above the tepals in a circle around the pistil; filaments not exserted when dried; an- thers creamy, sometimes tinged with pink, 0.6– 0.9 mm long, 0.6-1 mm wide; thecae oblong- ellipsoid, not at all or only slightly divaricate; pollen faintly spicy scented, pale yellow to creamy white fading to white. /nfructescence spreading to pen- dent; spathe persisting, green; spadix to 13 cm long, to 2 cm diam., with berries scattered through- out; berries red to purple, oblong to obovoid, acute to mammilliform at apex, 6-8.2 mm long, 3.2- 4.5 mm diam.; pericarp thickened, with numerous raphide cells; seeds 2 per berry, broadly ellipsoid or oblong- беу она 3.2-5 mm long, 1.6- mm thick, attached to carpel wall at radicle and by a thick strand of fibers running midway down the carpel to the basal end of the ruit. mm iam., Anthurium wagenerianum is known from the Cordillera Costal in Venezuela and from Curaçao 750 Annals of the Missouri Botanical Garden in the Netherlands Antilles. It is found on steep, rocky cliffs and dry slopes, rarely reported from primary forest, from sea level to 600 m in tropical moist, tropical dry, and premontane dry forest life zones. This species can be distinguished by the blades, which dry with pale raphide cells conspicuously visible under low magnification and with conspic- uous undulations along the margin, and by the petiole, which is more or less quadrangular, adax- ially flat to broadly sulcate, and abaxially usually squared or 3-ribbed. The spathe and spadix of А. wagenerianum are both quite small, and the spadix is cream-colored to green at anthesis, sometimes with the tepals fainty tinged with purple. Anthurium wagenerianum is deceptively sim- ilar to 4. crassinervium in habit, range, and leaf shape. The above-mentioned blade and petiole characteristics, however, aid in identification of dried specimens. Furthermore, dried abaxial leaf surfaces of А. wagenerianum do not demonstrate the highly light-reflective character nor the readily distinguished concentric epidermal cell pattern of А. crassinervium (see that species for more de- tails). The spathe and spadix of А. wagenerianum are, moreover, quite short and stout in comparison with those of А. crassinervium. Both species have red berries basally attached to the spadix by tepalar fibers at maturity. Anthurium wagenerianum, however, has a drier berry upon rehydration, its seeds being attached to the inner carpel wall at its apical end by a thick strand of fibers running mid- way down the carpel toward the basal end of the hydrated berry by a mucilaginous substance typ- ically found in fruits of most members of sect. Pachyneurium. Another feature separating A. t are the pis- tils which are earlier еван а! i anthesis in 4. wagenerianum than is true of А. crassinervium. wagenerianum from Bunting et al. 11993 is included in А. wage- nerianum with reservation. It is reportedly an ep- iphyte from a wet forest life zone, with an adaxially convex petiole. All three of these characteristics conflict with the norm for lower leaf epidermal characteristics, stubby spadix, А. wagenerianum. 118 and short spathe suggest that из placement in this species is, however, correct. Anthurium huequeense was compared by Bun- ting (1986) with А. cra conspecific with А. wagenerianum and cannot ђе ssinervium, but is in fact distinguished from typical material of the latter. NETHERLANDS ANTILLES. CURACAO: Christoffel Moun- tain, Tuinen 79GR00095 (К). VENEZUELA. b pen Puerto Cabello, 100 m, Curran & Haman 1140 (GH, US); Las Trincheras- El Cambúr, 1.5 km N of ы. оуег Ко Trincheras, N ed SEL, US), 54544 (GH, Valencia, El Cambür, Bunting 2848, 2854 (NY); 100- 300 m, Bunting 13512 (cultivated at MO). DISTRITO FEDERAL: W of Naiguata, 0 m, Gentry & Berry 14735 (MO); Rio Las Caracas, above town of Las Caracas, 100 m, Croat 21620 (МО). FALCÓN: Sierra de San Luis, Piedra de Agua, 600 m, 11?08'N, 69?40'W, Liesner et al. 7634 (MO); Dtto. Acosta, Via Pilancones, 6 km of Mirimire, Ruiz & Rondon 3726 (MY); Dtto. Colina, Rio Ricoa, S of Las Dos Bocas, 200 m, 11?19'N, 69°24'W, Steyer- mark & Gonzales 113632 (МО); Dtto, Zamora, Delta of Rio Hueque, Coro- Morón, 100 km E of Coro, 50 m, Bunting 7704 aped originally collected as van der Werff 3522) (MO, NY, VEN), Croat 71734 (MO). MI- RANDA: Carenero- Chirinena, 2 km NW of Carenero, 0- 5 m, Steyermark & Bunting 102313 (MO); Rio Chu- paquire, new road 1 k i 7 km E of Cupira, 0-150 m, 10°09'N, 65°38'W, Liesner & Gonzalez 11906 (MO, WIS), 11917 (MO); Brion, Laguna Grande, 0-30 m, 10°33'N, 66°04'W, Berry et al. 3740 (MO). zULIA: Dtto. Mara, Rio Guasare, near Destacamento Guasare No. 1 (La Yolanda), abajo del Destacamento, 200-250 m, Bunting et al. 11993 (MO). Anthurium watermaliense hort. ex L. Bailey Nash, Stand. Cycl. Hort. 1:303. 1922. TYPE: Colombia. Not seen, introduced into hor- ticulture via Watermall, Belgium. Neotype: Panama. Bocas del Toro: Changuinola-AI- mirante, Railroad Station Milla 7.5, less than m, Croat 38129 (lectotype MO- 2820846; isoelectotypes, CAS, Е, MEXU, NY, PMA, S, SEL, UC, US). Figures 333, 334. Terrestrial; stem to 25 cm long, (1)1.5-3 cm diam.; roots numerous, descending, velutinous, 3- 4 mm diam.; cataphylls thin, 4.5-14 cm long, obtuse to acute at apex, persisting as coarse linear fibers. Leaves erect-spreading; petioles (12)25- 60(88) cm long, ca. 4-5 mm subterete, flattened to slightly and narrowly sulcate with blunt to rounded margins adaxially, rounded abaxially; geniculum, 1-1.5 ст long; blades moderately co- riaceous, ovate-triangular, long-acuminate at apex, deeply lobed at base, 21-60 cm long, (14)20–40 cm wide; anterior lobe 20-45 cm long; the pos- erior lobes mostly oblong, (7)14-26 cm long, " 5)5-10(16) cm wide, rounded at apex; sinus parabolic to hippocrepiform; both surfaces semi- glossy; midrib convexly raised above and below; basal veins 5-8 pairs, 2nd through 8th coalesced iam., Volume 78, Number 3 1991 Croat Anthurium sect. Pachyneurium 4.5-8 cm, prominulous above and below; posterior rib mostly naked, weakly turned up on outer mar- gin; primary lateral veins 4-8 per side, departing midrib at 40—45° angle, prominulous in grooves above, prominulous below; tertiary veins conspic- uously raised below when dried; collective vein arising from lst or 2nd basal vein, raised above and below, equally as prominent as primary lateral veins when dried, 3-6 mm from margin. /nflores- cences spreading, shorter than leaves; peduncle (12)25-60(75) cm long, 3-7 mm diam., 0.5-1.5 x as long as petiole; spathe erect to reflexed, sub- coriaceous, usually dark violet-purple, sometimes green, lanceolate to ovate-lanceolate, 5-15(21) ст long, (1.5)2-6(8.5) cm wide, inserted at 75? angle on peduncle, long-acuminate at apex, acute to rounded (rarely cordate) and conspicuously decur- rent at base; stipe 0.3-3 cm long; spadix white to greenish, yellow or purple, tapered, 7-19 cm long, 5-10 mm diam. near base, 2-5 mm diam. near apex; flowers rhombic to 4-lobed, 1.5-3.4 mm long, 1.8-3.1 mm wide, the sides straight to jag- gedly sigmoid; 5-12 flowers visible in principal spiral, 6-16 in alternate spiral; tepals matte, punc- tate; lateral tepals 1.5-2.2 margins convex to + straight; pistils raised, pro- truding beyond the stamens at anthesis, glossy, weakly exserted just before stamens emerged, dark green; stigma narrowly ellipsoid, ca. 0.5 mm long, droplets appearing 2-3 days before stamens emerge; stamens emerging rapidly from the base or some- times in a scattered pattern throughout, held erect above the tepals; filaments translucent, exserted, 0.3-0.4 mm long, 0.6 mm wide; anthers creamy white, 0.9 mm long, 0.9-1 mm wide; thecae el- lipsoid, divaricate; pollen white. Infructescence with spadix to 22 cm long, ca. 2 cm diam., bearing berries in the basal portion only; berries yellow to usually orange, ovoid to obovoid, ee beaked at apex, 10-15 mm long, ca. 4– mm wide, the inner This species is known from Costa Rica and Pan- ama, between sea level and 2,400 m; most collec- tions have been made at about 750 m in premon- tane and lower montane rainforest. Populations occur near sea level in tropical moist and premon- tane wet forest. The species is extremely variable, especially in leaf size and shape, and spathe and spadix color. Anthurium watermaliense can be recognized by its ovate-triangular leaf blades with pronounced posterior lobes (a unique feature within the section), its broad, decurrent and frequently dark purple spathe, tapered spadix with exserted stamens, and usually orange berries. is species is most easily confused with spec- imens of Anthurium cotobrusii, which have un- usually long posterior lobes, but the latter species differs in having the pistils shorter than the stamens at anthesis and stamens that do not persist exserted. nthurium watermaliense might also be соп- fused with 4. standleyi. See discussion of that species for distinguishing characters. Although originally reported to have come from Colombia, no wild collections from present-day Co- lombia are known. Since Colombia encompassed Panama at the time of its introduction, it is believed that 4. watermaliense was originally collected in Panama and may never be found in Colombia. In Panama, it has been found no further east than western Veraguas. WITHOUT LOCALITY: Drake 219 (P) Costa Rica: ALAJUELA: Monteverde, Chomogo trail and Refugio El Valle, 1,600 m, Kennedy & Guindon 3799 (MO). can- TAGO: Palo Verde, beyond Quebrada Cangreja, 1,500 m, Luteyn & Wilbur 4376 (MO); SE э! MN = Raiz de Hule Camino, 1,200-1, ‚ Croat (MO, PMA); 4.5 km 5 of bridge on du cde on Lourdes, ca. 8 km S of Cartago by air, Finca El Chaparral, 1,500 m, 9°4.9'N, 83°55'W, Liesner & Judziewicz 14612 (MO); Quebrada Cangreja, 3 km S of Pan American Hwy., NY, RSA, Turrialba * nature trail, 9*54'N, 83°39'W, Liesner et al. 15326 (МО); Tapantí Reserve, 1,400-1,700 m, Goméz 19262 (MO). HEREDIA: Braulio Carrillo Park, Zurqui, 1,700-2,000 m, Goméz 20220 (B, CR, К, МО, RSA); Cerro Zurqui, southern slope, 5 km N of San Luis Norte, 1,800 m, 10%03'N, 84°02'W, Stolze 1548 (Е, MO); Rio San Rafael, Atlantic slope of Volcán Barva, 1,500 m, 10%13'N, 84?05'W Grayum 7021 (MBM, МО); between Rio San Fernando and Rio Sardinal, Atlantic ا‎ of Volcan Barva, 1,850- 1,880 m, 10?12'N, A W, Grayum 7342 (MO). LIMÓN: Puerto Viejo de ue Manza o at Punta Mona, Panama border, 0-50 m, Taylor & Skotak 4435 (DUKE, MO), 4529 (DUKE); range of hills 2 km S of Manzanillo de Talamanca, E of Quebrada pus Me. 190 m, 9*37'N, 3 (МО); Cordillera de Talamanca, Rio Madre de EA Barbilla, Quebrada Canabral, 280-400 m, Grayum et al. 8839 ( MO); Reserva Biol. Hitoy Cerere Valle del Río La Estrella, 400 m, 9?40'15"N, 83?03'20"W, Herrera 2356 (CM, CR, Bs SE of Lago Dabagri, Telire, Goméz et id 23225 (CR, МО); oe 10 km N of Puerto Lim 3m, a 347 (MO); Siquirres, Las Brisas de Pond 300 m, Gomez et al. 23396 (MO, NY, US); Cerro Pirripli, just SE of Puerto Viejo de Talaman nca, n 50 m, Croat 43224 (CM, МО); Rio Sixaola, 10- 752 Annals of the Missouri Botanical Garden 50 m, 9°35'№, 82%53'E, Burger et al. 10444 (Е, MO). SAN JOSÉ-HEREDIA: Cerro Zurqui, southern slope, end of Calle Zurqui, 1,800-2,000 m, Almeda & Nakai 3706 m Rio diy е Cerro de Zurqui, 1,600-1,800 , 10%03'N, 1'W, Burger et al. 9342 (МО). SAN ef Carrillo о. 700 m, Gomez et al. 21168(MEXU, MO). PANAMA. BOCAS DEL TORO: Changuinola- Almirante, Railroad Station at Milla 7.5, less than 1 m, Croat 38129 (CAS, F, IBE, S МЕХ Џ, MO, NY, PMA, RSA, S, SEL, ,US, VDB, W Miss Croat & Porter 16246 (MO); Criollo (just above fon: Vista)- Quebrada Higue- rón, е trail, Kirkbride & Duke 784 (MO); 15 km : of town of Changuinola, Changuinola—1 dam site, 300- 500 m, rds 3146 (МО). c ese i Quebrada Hondo- a Divide, on ар -Chiriquicito Trail, Kirk- МО); Cerro Colorado, 1,200-1,500 m, Croat 33265 (F, MO, SEL, US, VEN), Folsom & Collins 1754, 1764 (MO), Folsom et al. 4840 (MO), Mori & Dressler 7849, 7852 (MO), Sullivan 359 (МО); Cerro Colorado mine area, from Chami station to 9 mi. along rd., 1,100-1,750 m, 8*35'N, Road, Gualaca- Fortuna Dam, 5.9 mi. NW of Los Planes de Hornito, 1,370 m, 8?43'N, 82?15'W, Croat 49898 (MO). vERAGUAS: Santa Fe Region, N of Escuela Agricola Alto Piedra, 700-900 m, Croat 49035 (MO), Croat & Folsom 33958 (CM, F, MO), Folsom 2994 (MO), Mori & Kallunki 2563 (MO); 6-7 km W of Santa Fe, 900 m, Мее 9717 (MO, PMA). Anthurium willifordii Croat, sp. nov. TYPE: Peru. Loreto: Napo River, Explorama Camp, on Río Sucusari, below 200 m (originally col- lected by Jack Williford), Croat 61087 (ho- lotype, MO 3244489; isotypes, B, K, NY, USM). Figures 339, 340. Planta epiphytica; internodia brevia, 1-2 cm diam.; cataphyllum persistens intactum mox deciduum; petiolus quadrangularis ad p interdum subteres, 2- .5 cm longus, 4-5 mm mina oblanceolata aut oblanceolata- elliptica, (14)18- 50 cm longa, 5.5-16 c lata; nervis primariis lateralibus 6-13 utroque; aa culus 2.5-9 cm longus; spatha erecta, atropurpureus, navicularis, 2.3-4 cm longa, 1-2 cm lata; spadix dilute purpureo-violaceum aut griseo-purpureum, cylindroideus, 1.7-4 cm longus, 2-5 mm diam.; baccae magentae. roots nu- merous, dense, spreading to descending, pale to Epiphytic; stem short, 1-2 cm diam.; medium green, smooth, bluntly tapered, moder- ately elongate, 3-4 mm diam.; cataphylls subco- riaceous, broadly triangular, 1.5-5 cm long, acute at apex, pale yellowish green, drying brown, per- sisting intact, soon deciduous. Leaves spreading; petioles 2-4.5 cm long, 4-5 mm diam., erect- spreading to spreading, quadrangular to trapezoi- dal, sometimes subterete, with a medial rib and the margins prominently raised adaxially, sharply 3- 5-ribbed abaxially, the surface slightly pale-speck- led; geniculum becoming reddish tinged, 0.4–0.7 cm long; blades subcoriaceous to moderately co- riaceous, oblanceolate to oblanceolate-elliptic, ob- tuse, minutely apiculate at apex, obtuse to rounded at base, (14)18-50 cm long, 5.5-16.5 cm wide, usually broadest above the middle, the margins sometimes undulate; upper surface matte-velvety, moderately quilted, medium green, lower surface velvety with conspicuous crystalline cells, slightly paler and often tinged with red; both surfaces dry- ing brown to grayish brown; midrib prominently acute at base, becoming narrowly raised toward the apex above, prominently acute and tinged with reddish below; primary lateral veins 6-13 per side, departing midrib at 50—70° angle, straight, flat to weakly raised above, weakly raised and darker than surface and tinged with red below, drying slightly raised above and below; interprimary veins less prominent than primary lateral veins, darker than surface below; tertiary veins slightly darker than surface below, drying weakly raised; reticulate veins obscure; collective vein arising from near the base or in upper third of blade, weakly sunken to flat above, weakly raised and darker than surface be- low, equally as prominent as primary lateral veins, -10 mm from margin. Inflorescences erect; pe- duncle 2.5-9 cm long, 2-3 mm diam., x as long as petioles, green faintly tinged purplish, terete with obscure ridge; spathe erect, subcoria- ceous, dark purple to purplish violet (B & K red- purple 2/5), ovate-elliptic, navicular, 2.3-4.7 ст long, 1-2 cm wide, broadest in the lower third, cuspidate-acuminate at apex, acute and weakly decurrent at base; spadix reddish (B & K red 7/ 10) to faintly purplish violet or grayish purple, very short, cylindroid, erect, 1.7-4.7 cm long, 2-6 mm diam.; flowers rhombic and weakly 4-lobed, 1-1.5 mm long, 1.5-1.7 mm wide, the sides + straight parallel to spiral, jaggedly sigmoid perpendicular to spiral; 8-9(14) flowers visible in principal spiral, 17-21 in alternate spiral; tepals matte, densely and minutely papillate; lateral tepals 0.5-0.9 mm wide, the inner margins rounded to bluntly 3-sided, the outer margins 2—3-sided to shield-shaped and 4-sided; pistils raised, purple, much darker than the tepals; stigma slitlike, 0.2-0.3 mm long; sta- mens emerging from near the middle, lateral sta- mens emerging to % the distance to the apex before alternates emerge at lower 43 of spadix; anthers white, 0.4—0.5 mm long, 0.4-0.5 mm wide, con- tiguous at the surface of the tepals; thecae narrowly ovoid, conspicuously divaricate; pollen yellow, fad- ing to white. /nfructescence with berries scattered throughout; berries obovoid, magenta, rounded at apex. Volume 78, Number 3 1991 roat 753 Anthurium sect. Pachyneurium Anthurium willifordii is endemic to Loreto De- partment in Peru, where it has been collected once in the vicinity of the Explorama Camp on the Rio Sucusari, a tributary of the Rio Napo, at less than 200 m. It was found here as an epiphyte in a tropical moist forest life zone. This species is a very distinctive one, the chief characters being the short, stubby spadix on a short peduncle and the quilted leaf blades, which are matte above, velvety and frequently tinged red below. The petioles are of variable cross-sectional shape: quadrangular to trapezoidal or sometimes subterete; they are always sharply 3-5-ribbed abaxially. o other species is likely to be confused with A. willifordit, and only A. superbum and A. re- flexinervium share the character of quilted, velvety or red-tinged leaf blades. The new species is named in honor of Jack Williford, of Brandon, Florida, who originally col- lected live plants along the Rio Sucusari. PERU. LORETO: Rio Маро, Explorama Camp, on Río Sucusari, below 200 m, originally collected by Jack Wil- liford, Croat 61087 (B, К, MO, NY, USM Anthurium wurdackii Bunting, Acta Bot. Ve- nez. 10: 279. 1975. TYPE: Venezuela. Ama- zonas: Cerro Yapacana, 3?45'N, 66?45'W, 825 m, Steyermark & Bunting 103099 (ho- lotype, VEN; isotypes, MO, MY, NY, US). Figure 341. Description based on dried material only. Ter- restrial or epilithic; stem short, stout, growing hor- izontally, 1.5-4 cm diam.; cataphylls thick, ca. 4– 7 cm long, dark brown, persisting + intact or weathering into fine linear fibers. Leaves erect to spreading; petioles 22-55 cm long, 5-8 mm diam., subterete, somewhat flattened and narrowly sulcate with rounded margins adaxially, rounded abaxially; geniculum 1-1.5 cm long; blades thickly coria- ceous, broadly ovate-elliptic to somewhat broadly lanceolate in larger blades, acute to obtuse at apex, shallowly cordate at base, 28-53 cm long, 10.5- 26 cm wide, broadest at or just below the middle; sinus broadly and shallowly arcuate; upper surface glossy, dark green, lower surface; midrib flat at base; becoming convexly raised toward apex; basal veins 3—4 pairs, prominently arcuate-ascending, usually merging with the margin, at least the first vein extending to near or above the middle of the blade; primary lateral veins 3-4 per side, departing midrib at 40—60° angle, slightly arcuate, steeply ascending, raised above and below; interprimary veins few, less conspicuous than primary lateral veins; tertiary veins sunken above, raised above and below; collective vein arising in the upper half of the blade or absent. Inflorescences longer than leaves; peduncle (44)55-81 cm long, ca. 4-7 mm diam., 1-2x as long as petiole; spathe erect- spreading at anthesis, becoming reflexed to re- curled, subcoriaceous, green tinged with maroon at apex at anthesis, becoming purple, lanceolate, (6.5)8-14.5 cm long, 1.5-2.5 cm wide, decurrent for 1.5-2 cm at base; spadix maroon, stipitate to 7—13 mm, narrowly tapered, 6.5-18 cm long, 6– 7 mm diam. near base, 3-4 mm diam. near apex; flowers rhombic, 2.7 mm long, 2.9 mm wide, the sides straight to weakly sigmoid; 8 flowers visible in principal spiral, 6 in alternate spiral. /nfruc- tescence with spathe persisting; spadix 7-21 cm long, 1.5-2 cm diam.; berries white, maroon at apex, ca. 8 mm diam. Anthurium wurdackii is known only by two collections from Amazonas, Venezuela, on Cerro Yapacana at 825 m in premontane wet forest, and on Cerro Neblina at 780 m in lower montane moist forest. This species is an atypical member of the section and can be recognized by its subterete petioles, coriaceous, broadly ovate-elliptic leaf blades that are shallowly cordate at the base and have three to four pairs of basal veins and steeply ascending, mostly free-ending primary lateral veins. Also char- acteristic are the long peduncle, maroon, sharply tapered spadix, and whitish berries that are maroon at the apex. Anthurium wurdackii is not closely related to any other species in sect. Pachyneurium. In leaf texture, it is most similar to А. bonplandii subsp. bonplandii or subsp. guayanum, but those differ in lacking basal veins and having oblanceolate to elliptic leaf blades that are attenuate to obtuse at the base. VENEZUELA. AMAZONAS: Cerro Yapacana, 825 m, Stey- ermark & Bunting 103099 (MO, MY, NY, US, VEN); Dpt. Río Negro, Cerro dig: Camp IV, 15 km NNE of Pico Phelps, north bra of river in canyon, 780 m, 0%51'N, 65°57'W, pn 16664 (B, MO). Anthurium xanthoneurum Bunting, Phytolo- gia 60(5): 298. 1986. TYPE: Venezuela. Ama- zonas: Dept. Rio Negro, Cerro Aratitiyope, ca. 70 km SSW of Ocamo, 900 m, 2°10'N, 65?34'W, Steyermark et al. 130054 (holo- type, NY; isotypes, MO, VEN). Figure 342. 754 Annals of the Missouri Botanical Garden Description based on dried material only. Epi- lithic; stem short, ca. 2.5 cm diam. Leaves with petioles less than 10 cm long; geniculum thicker than petiole, 2.5 cm long; blades thickly coria- ceous, oblanceolate-elliptic, obtuse at apex, obtuse to narrowly rounded at base, 30-55 cm long, 8.5- 15 cm wide, broadest above the middle, the mar- gins prominently revolute; upper surface semi- glossy and weakly quilted, lower surface paler, matte and with conspicuous brown glandular punc- tations, yellow-green; midrib convexly raised above, more prominently so below; primary lateral veins 8-10 per side, departing midrib at 30-40? angle, weakly arcuate to the margin, convexly raised above in moderately deep grooves, conspicuously raised below; interprimary veins absent; tertiary veins prominulous, especially on the upper surface; col- lective vein absent. /nflorescence with peduncle 50 cm long, ca. 6 mm diam.; spathe unknown; spadix (post anthesis) brownish green, long-tapered, erect, 24 cm long, ca. 8 mm diam. near base, ca. 4 mm diam. near apex, broadest at the base. Anthurium xanthoneurum is known only from the type collection from Cerro Aratitiyope in Ama- zonas, Venezuela, at 990 m in a premontane rain- forest life zone. This species is distinguished by its heavily co- riaceous leaf blades with steeply ascending primary lateral veins, and its slender, long-tapered spadix. Anthurium xanthoneurum is closely related to А. bonplandii and especially resembles subsp. guayanum. That taxon differs, however, in having mostly broader leaf blades with the primary lateral veins ascending at a wider angle. VENEZUELA: AMAZONAS: Dpt. Rio Negro. Cerro Arati- tiyope, 70 km SSW of Ocamo, streams feeding Rio Mani- pitare, 990-1,670 m, 2°10'N, 65?34'W, Steyermark et al. 130054 (NY, VEN). EXCLUDED SPECIES Anthurium aduncum (Vell. Conc.) Schott, Prod. Aroid. 478. 1860. TYPE: Brazil. Illustration in Vellozo (1825). (Fl. Flum. Tab. 124 serves as the type.) This species is based on an illustration (Pothos aduncus) in Vellozo's Flora Fluminensis (1825). Schott (1865) transferred it to Anthurium and included it in his section Pachyneurium. Engler (1905) placed the name among his “Species du- bia." There are only two Pachyneurium species known from southern Brazil, 4. coriaceum G. Don in Sweet and 4. solitarium (Vell. Conc.) Schott, and А. aduncum does not look remotely like either of them. The narrow blades with numerous veins indicate that it is most likely a member of sect. Urospadix. Its exact affinities may never be known, since that group has many similar species. Anthurium agnatum Schott, Oesterr. Bot. Z. 8: 181. 1858. TYPE: Nicaragua. Oersted s.n. Type destroyed. (Field Museum photo 29807 serves as the type. This species, included in sect. Pachyneurium by both Schott (1860) and Engler (1905), was based on a sterile Oersted collection from Nicara- gua. Although reported by Engler to be at Copen- hagen, no such specimen exists there today. This may have been in error, because an Oersted col- lection bearing the name Anthurium agnatum was photographed (Field Museum 29807) at the Vienna herbarium (W), before the latter was destroyed during World War II. Schott a prepared an illustration of th tt illustration 468, Fiche 15:610). The photograph of the spec- imen and the Schott illustration are all that exist toda The leaf blades of Anthurium agnatum are shaped like those of А. spathiphyllum N. E. Br., but have fewer than 10 primary lateral veins (vs. 20—30 for the latter). It may be an unusual spec- imen of А. bradeanum Croat & Grayum, a species similar to A. spathiphyllum, but that species, though possessing fewer primary lateral veins, has ovate-elliptic blades less than five times longer than d. The blades of the Oersted collection are narrowly oblanceolate and over six times longer than broad. The unavailability of a specimen for examination of leaf texture, coloration, and ve- nation makes it impossible to identify it with cer- tainty, so it remains a nomen dubium. broa Anthurium andicola Liebm., Vidensk. Meddel. Dansk Naturahist. Floren. Kiobenhavn 1: 22. 1849. TYPE: Mexico. Veracruz: Santa Maria Alpatlahua, 2,500 m, Liebmann s.n. (lecto- type, Designated as a Pachyneurium by Engler (1905), but rejected due to its lack of involute vernation, this species is a member of sect. Be- lolonchium (A. andicola Alliance); see Croat (1983). Anthurium brownii Masters, Gard. Chron. 11. 744. 1876. TYPE: Colombia. Valle: vic. Bue- naventura, Wallis s.n. (type, К). Volume 78, Number 3 1991 Croat Anthurium sect. Pachyneurium 755 This species was included in sect. Pachyneu- rium by Engler (1905), but it lacks involute ver- nation. It is possibly a member of sect. Belolon- chium. Anthurium cartilagineum (Desf.) Kunth, Enum. : 79. 1879. ТҮРЕ: Venezuela. (Schott Drawings 678-680 serve as the type.) Included in sect. Pachyneurium by Engler (1905), but lacking involute vernation. Section un- known. Anthurium cordatum (Willd.) G. Don in Sweet, Hort. Brit. 633. 1839. ТУРЕ: West Indies, Willdenow Herbarium Cat. Nr. 3101 (type, B). Included in sect. Pachyneurium by Engler (1905), but lacking involute vernation. Section un- known. Anthurium cuspidatum Masters, Gard. Chron. I. 428. 1875. TYPE: Colombia?, Hort. Vietch, Jan. 1875 (K). This species was erroneously included under sect. Pachyneurium in my revision of Anthurium for Mexico and Middle America (Croat, 1983), based on a misidentified plant with involute vernation. It is a member of sect. Polyneurium. Anthurium grandifolium (Jacq.) Kunth, Enum. Pl. 3: 17. . TYPE: Jamaica (erroneously ге- ported as from Venezuela), Jacquin type not seen. (Willendow Herbarium specimen 3102 (B) might serve as the type if the Jacquin specimen is not located.) This species was included in sect. Pachyneu- rium by Engler (1905), but rejected due to the presence of supervolute rather than involute ver- nation. Section unknown. Anthurium hacumense Engl., Bot. Jahrb. Syst. 25. 363. 1898. TYPE: Costa Rica. Puntarenas: of Rio Hacum, near Buenos Aires, 250 m, Ton- duz 6536 (holotype, B; isotypes, BR, CR). Like Anthurium hookeri (see below), А. hacu- mense has glandular-punctate blades. Having over- looked the glandular punctations, Engler no doubt placed the species in Pachyneurium owing to its large size and rosulate habit; it is a typical member of sect. Porphyrochitonium. Anthurium hookeri Kunth, Enum. Pl. 3: 74. 1841. TYPE: Schott Drawing 517 serves as the lec- totype (designated by Mayo 1982). Though treated as а Pachyneurium by both Schott (1860) and Engler (1905), Anthurium hookeri does not belong there. It differs from all species of Pachyneurium by having supervolute vernation and leaf blades with scalariform second- ary venation and glandular punctations on the low- er blade surface. It probably belongs to its own section or is an unusual member of sect. Porphy- rochitonium. Anthurium lilacinum Bunting, Acta Bot. Venez. 10: 273. 1975. TYPE: Venezuela. Carabobo: hs of Rio San Gian, above Los Tan- ques and La Toma, S of Barburata, 750-900 m, Steyermark 95294 (holotype, VEN; iso- type, MY). Anthurium lilacinum was erroneously reported as a member of sect. Pachyneurium by Croat and Lambert (1986). This was based on its rosulate habit and otherwise superficial resemblance to spe- cies of this section. Later studies revealed it to lack involute vernation. It is probably a member of sect. Urospadix. Anthurium longispathum Carriere, Rev. Hort. 498. 1888. TYPE: Guadeloupe (based on cul- tivated collection, no type specimen prepared 2). The species was treated by Епрјег (1905) as a Pachyneurium. It is a synonym of Anthurium grandifolium (Jacq.) Kunth (see that species). Anthurium maximum (Desf.) Engl., Pflanzenr. IV. 23B(Heft 21): 77. 1905. TYPE: No specimen or illustration pertaining to the type is known. Anthurium maximum was described by Des- fontaines in 1829 in Pothos. Schott did not deal with the name at all, but Engler (1879) treated it as a synonym of A. fontanesii in his first treatment of Anthurium. In his 1905 revision, Engler cor- rectly transferred Pothos maxima to Anthurium and treated А. fontanesii as a synonym of А. maximum. Desfontaines’ brief description com- pares the plant with Pothos crassinervia Jacq. from which it was said to differ by its larger blade "broadened toward the apex, its longer, sulcate petiole and its longer spadix." Engler may have seen some preserved material of 4. maximum, perhaps a specimen at the Paris Herbarium pre- 756 Annals of the Missouri Botanical Garden pared by Desfontaines. However, if any such col- lection ever existed, there is no longer any record of it. Engler's description of А. maximum closely matches Schott's original description of А. fonta- nesii, now considered a synonym of 4. crassiner- vium, but the vague nature of А. maximum pre- cludes making any definitive treatment, and it is treated here as a nomen dubium. Anthurium michelii Guillaumin, Bull. Mus. Hist. Nat. (Paris) 31: 263. 1925. TYPE: Panama. Bocas del Toro: hills beyond Fish Creek La- goon, Wedel 2276 (holotype, P). This species was erroneously reported as a Pachyneurium by Croat (1983), and it was later suggested (Croat, 1986) to be a member of sect. Urospadix. However, the latter group is probably restricted to eastern South America and has less conspicuous primary lateral veins. Anthurium michelii is not a Pachyneurium owing to its su- pervolute vernation and is perhaps best placed in a new, unnamed section with 4. decurrens (see discussion in Introduction). Anthurium seleri Engl., Bot. Jahrb. Syst. 25: 459. 1898. TYPE: Guatemala. Huehuetenango, vic. Chacula, Seler 2643 (holotype, B). This species was included in Pachyneurium by Engler (1905), but it lacks involute vernation. It is probably a member of sect. Belolonchium (А. andicola Alliance). Anthurium tricarinatum бойго, Ant. Ecuator. Adiciones 1. 1903. TYPE: Ecuador. Canar: valley of Rio Canar, Rimbach 81 (type lost). Though a well described species, Anthurium tricarinatum cannot be assigned with certainty to any known section. Sodiro compared it with 4 leonianum, a very typical Pachyneurium, but he also said the leaves were black-punctate on the lower surface, a feature most closely associated with sect. Porphyrochitonium. Engler (1905) treated the species as а Pachyneurium, but his description appears to be based on that of Sodiro, and there is no way to be sure that he ever saw a specimen. Unless a specimen or a photograph of the Rimbach specimen becomes available, the spe- cies is best considered a nomen dubium. Anthurium weberbaueri Engl. Pflanzenr. IV. 23B(Heft 21): 81. 1905. ТУРЕ: Peru. Sandia, 2,100-2,300 m, Weberbauer 543 (holotype, B; isotype, C). This species was included in Pachyneurium by Engler (1905), but belongs in sect. Digitinervium by virtue of its scalariform veins and glandular punctations. LITERATURE CITED ADAMS, C. D. 1972. Flowering plants of Jamaica. Univ. West Indies, اا‎ А amaica 1. BAKER, J. G. W. Saunder, Botanicum BERLIN, B. & P Basic Color Terms, Their Universality Evolution. Univ. California Press, Berkeley. BOGNER, J. . Н. Мсогѕом. 1991. Revised classi- of Araceae with dichotomous keys. Willde- wia (in press). Boe G. S. 1965. eT on Mexican Ara- ae. Gentes Herb. 9(4): 3 19 New taxa of PRTA Araceae. Phy- tologia 605) 295. New taxa of Venezuelan Araceae (II). ШО 64: 459-486. . Notes on Araceae. Ann. Missouri Bot. Сага. ze 917-919. Croat, T. B. 1980. does behavior of Anthurium (Araceae). Amer. J. B : 888-904. . Studies in e ceae III: new species of Anthurium from Central America. Selbyana 5(3-4): 315-341 3. A revision of the genus Anthurium (Araceae) of Mexico and Central America. Part I: Middle America. Ann. Missouri Bot. A revision of the genus Anthurium TORTE ) сі Mexico and Central America. Part П: Panama. d m Syst. Bot. Missouri Bot. Gard. 14: 1-204 1987. 13. Anthurium plowmanii Croat, a new species from interior South America. Candollea ^ 1979. The genus Anthurium (Araceae) in Costa Rica. Brenesia 16(Suppl. 1): 1- & С. S. BUNTING. 1979. аиа of Anthurium descriptions. Aroideana .L RT. 1986. The Ar of Vene- zuela. Жайны 9(1-4): а. Скозву, М. ndex Мизсогит— а comput- erized miscological database. Bull. Brit. Bryol. Soc. 48: 25-26. . Маспл. 1986. Tropicos: The Botanical Database at the Missouri lem Garden. Missouri Botanical Garden, St. Lou Dopson, C. Н. € A. H. боль. 1978. Flora of det Palenque Science Center, Los Rios Ecuador. byana 4(1-6): 1-628. ENGLER, А. 1905. Araceae-Pothoideae. In: F = Das Pflanzenreich IV. 23B(Heft 21): 1- FLORES, С. L., L. JIMENEZ, X. MADRIGAL, a Ms NCAYO & Е. TAKAAKI. 1971. Мара de tipos de vegetación de la Repüblica Méxicana. 1:200,000. Secretaria de Recursos Hidraulicos. GRAYUM, M. Н. 1984. Palynology and Phylogeny of the Araceae. Ph.D. Dissertation. Univ. Massachu- setts, Amherst. Volume 78, Number 3 roat 757 Anthurium sect. Pachyneurium HoLDRIDGE, L. R., W. Н. HATHEWAY, T. LIANG & J. А Тоз. 1971. Forest Environments in Tropical Life Zones. Pergamon Press, New York. HowARD, В. А. 1 Nomenclatural notes on the Araceae of the PE Antilles. J. Arnold Arbor. 60: 9. LINNAEUS, C. Species Plantarum 2: 785-1684. 2nd edition. Stockholm. МАТОРА, E . Nuevas plantas de México. Ann. Inst. Biol. Univ. Mexico 32: 149. ScHorr, Н. У. 1860. DONUM Systematis Aroidear- um. Vindobonae [W . Icones Aroideae s mo [Micro- fiche edition.] Inter n Postrassse 14, 6300, 2% a or 4,800 SODIRO, L. 1902. B Ecuatorianos. Anal. Uni Quito 15(111): 38 . 1906. ei SE II Anturios е Anales Univ. Centr. Ecuador 22(156): 2 97. Flora Indiae a cae etc., Anthurium acaule (Jacq.) Schott пена and West Indian ''bird's-nest" Anthuri- ums. Kew Bull. 36(4): 691-719. Ray, T. 5. ment in the Агасеае. Атег 1 аен. Айй, J. Bot. 1986. Growth correlations within the seg- . J. Bot. 73: 993-1001. emoria Explicativa La VELLOZO, J. o. nl. 385- 3 1829 (1825). Diversity of shoot organization in the io de Janeiro Araceae. Amer. J. Bot. 74: 1373-1387. 19 Survey of shoot organization in the 84. 75: 56- APPENDIX 1. Taxonomic status of species included in Schott's grex Pachyneurium (1860).* 55 А. wagenerianum K. Koch & Bouché 96 A. recusatum Schott — A. fendleri Schott 57 A. fendleri Schott 58 А. brachygonatum Schott — A. schlechtendalii Kunth 59 A. spectabile Schot 60 A. huegelii Schott — A. hookeri Kunth 61 А. mexicanum Liebm — A. schlechtendalii Kunth 62 A. rugosum Scho — A. crassinervium (Jacq.) Schott 63 A. crassinervium Schott — A. crassinervium ind ) Schott 64 А. hookeri Kunth Not sect. Pachyneuriu 65 A. affine Schott 66 A. ellipticum K. Koch — A. crassinervium (Jacq.) Schott 67 A. tetragonum ae = A. schlechtendalii Kunt 68 A. egregium Schot = A. crassinervium (Jacq.) Schott 69 A. O Kunth 70 А. acaule Sc — A. acaule (Jacq.) Schott 71 А. a Brongn. ex Schott = А. dombeyanum Brongn. ex Schott 72 A. agnatum Schott Nomen dubium 73 A. solitarium Schott 74 A. aduncum Schott = А. aduncum (Vell. Conc.) Schott section Urospadix? * The numbers are those used by Schott. 1975. Du ecológico de Bolivia, Flora 91. Icon. 9: 121-123. (1835) 758 Annals of the Missouri Botanical Garden APPENDIX 2. Taxonomic status of species included in Engler's sect. Pachyneurium (1905).* 7 А. pendulifolium N. E. Br. 8 А. paraguayense Engl. 9 А. preussii Engl — A. crasinervium (Jacq.) Schott 10 А. joseanum Engl. — A. protensum Schott 11 4. wagenerianum К. Koch $ Bouché 12 А. ellipticum К. Koch & Bouch — A. crassinervium (Jacq.) Schott 13 A. acaule sensu ene (1860) & Engl. (1905) non (Jacq.) Schott — A. crenatum Kunth 14 А. recusatum Schott = А. fendleri Schott 15 А. hookeri Kunth not sect. Pachyneurium 16 A. jenmanii Engl. 17 А. trinitatis Engl. — A. jenmanii Engl. 18 A. tenuispadix Engl. = А. uleanum Engl. 19 А. brachygonatum Schott = А. schlechtendalii Kunth 20 A. buenaventurae Engl. — A. fendleri Schott 21 А. uleanum Engl. 22 A. tarapotense Engl. 23 A. cubense Engl. 24 A. glaziovii Hook.f. — A. solitarium (Vell. Conc.) Schott 25 А. hacumense Engl. sect. Porphyrochitonium 26 A. strictum N. E. Br. — A. oxycarpum Poeppig 27 А. tetragonum Hook. ex Schott — A. schlechtendalii Kunth 28 A. schlechtendalii Kunth 29 А. crassinervium (Jacq.) Schott 30 A. maximum (Desf.) Engl. Nomen dubium (See Excluded Species) 31 А. leonianum Sodiro 32 A. acutifolium Engl. 33 A. rugosum Schott — A. crassinervium (Jacq.) Schott 34 A. affine Schott 34a А. latissimum Engl. 35 A. cymatophyllum К. Koch & Sellow Nomen dubium (See Excluded Species) 36 A. barclayan ngl. 37 A. ernestii Engl 38 4. agnatum t Nomen dubium (See Excluded Species) 39 A. kunthianum Liebm. — A. schlechtendalii Kunt 40 A. nobile Engl = А. solitarium (Vell. Conc.) Schott 41 А. dombeyanum Brongn. — A. dombeyanum Brongn. ex Schott 4la A. rigidissimum Engl. — A. dombeyanum Brongn. 42 A. fortinense Engl. — A. schlechtendalii Kunth 43 A. tricarinatum Sodiro Nomen dubium (See Excluded Species) 44 A. weberbaueri Engl. sect. Digitinervium 45 A. fendleri Schott 46 A. rusticum N. E. Br. ex Engl. = A. paraguayense Engl. 47 А. salviniae Hemsley 48 A. spectabile Schott 49 A. lindmanianum Engl. 50 A. martian Koch & Kolb 51 А. selloum K. Koch 52 А. longispathum Carriere — A. grandifolium rus ) Kunth 53 A. grandifolium (Jacq.) Kunth Not sect. Pachyne 54 A. boucheanum К. = A. ao (Des. Kunth 55 A. liebmannii Schott — A. umbrosum Liebm 56 А. umbrosum Liebm Not sect. Po aeu dst 57 А. арине (Willd.) G. Don Not sect. Pachyneurium 58 А. andicola Liebm. Not sect. Pachyneurium 59 A. cartilagineum (Desf.) Kunth Not sect. Pachyneurium 60 A. brownii Masters Not sect. Pachyneurium 61 A. appunianum Schott — A. cartilagineum (Desf.) Kunth 62 A. seleri Engl. Not sect. Pachyneurium * The numbers are those used by Engler. Those with letter subscripts were treated at the end of the revision. Volume 78, Number 3 1991 Croat Anthurium sect. Pachyneurium 759 APPENDIX З. Species of Р, t in Schott (1860) and Engler (1905) where applicable). Croat revision Schott Engler acutifolium Engl. x r. acutifolium var. herrerae Croat *acutissimum Engl sect. Urospadix Engl. affine Schott x *angustilaminatum Engl. subsp. angustilaminatum *subsp. e Croat anorianum asplundii i Croa m R. Schultes & Maguire bonplandii Bunt ubsp. bo nplandii Bunting ы) cuatrecasii Croat subsp. guayanum (Bunting) Croat bradeanum Croat & Grayum brenesii Croat & R. A. Baker bucayanum Croat cataniapoense Croat caucavallense Croat colonicum K. Krause concolor K. Krause consobrinum Schott coriaceum G. Don in Sweet cotobrusii Croat & R.A. Baker cowanii Croat crassinervium c ) Schott crenatum (L.) Kun cubense Engl. bei ан Croat ombeyanum Brongn. ex Schott l. ernestii En var. ernestii var. oellgaardii Croat eximium * fasciale Sodiro fatoense K. Krause fendleri Schott galactospadix ids glaucospadix Cro guanchezii kis halmoorei Cro hammelii Croat harlingianum Croat * holmnielsenii Croat iramirezae Bunting jenmanii Engl grex Oxycarpium grex Acamptophyllium x sect. Urospadix Engl. sect. Episeiostenium Schott sect. Urospadix Engl x as A. acaule (Jacq.) Schott sensu Schott sect. Epsieiostenium Schott 760 Annals of the Missouri Botanical Garden APPENDIX 3. Continued. Croat revision Schott Engler johnsoniae Croat latissimum Engl. А x x linguifolium Engl. sect. Urospadix Engl. luteynii Croat machetioides Matuda maguirei A. Hawkes manabianum ii manuanum Cro martianum K. E & Kolb х *napeaum Engl. sect. Urospadix Engl. *narinoense Croat ervatum Croat nizandense Matuda *obscurinervium Croat oerstedianum Schott x sect. Urospadix Engl. *ottonis K. Kra oxycarpum Poeppig & Endl. grex Oxycarpium sect. Oxycarpium Schott *oxyphyllum Sodiro sect. Urospadix Engl. achylaminum Croat * palenquense Croat * pallatangense Croat кое е Епр]. х у oroicoanum Croat г. paraguayense релйн lum N. E. Br. x *penningtonil Croat prolatum Croat & R. A. Baker protensum Schott grex Erythropodium sect. Urospadix Engl. subsp. arcuatum Croat subsp. protensum pseudospectabile Croat purpureospathum Croat ranchoanum Engl. sect. Calomystrium Schott reflexinervium Croat remotigeniculatum Croat salvadorense Croat salviniae Hemsley x *santiagoense Croat sarukhanianum Croat & Haager schlechtendalii Kunth x x subsp. jimenezii (Matuda) Croat ш seibertii Croat & R. A. Baker selloum K. Koch x simpsonu Croat Volume 78, Number 3 Croat 761 1991 Anthurium sect. Pachyneurium APPENDIX 3. Continued. Croat revision Schott Engler solitarium (Vell. Conc.) Schott x as A. affine Schott sparreorum Croat spathiphyllum | Е. Вг. sect. Episeiostenium spectabile Schot X x standleyi Croat & R. A. Baker superbum Madison subsp. brentberlinii Croat perbum subsp. su tarapotense Engl х tenaense Сгоа gl. r. nanayense Croat a anum upalaense Croat & R. A. Baker validifolium K. Krause vaupesianum Croat venosum Griseb. sect. Cardiolonchium Schott vinillense Buntin wagenerianum K. Koch & Bouché x х watermaliense hort ex L. Bailey & Nash willifordii Croat wurdackii Bunting xanthoneurum Bunting * Those species marked to the left with an asterisk are members of series Multinervia. All others are members of series Pachyneurium. An x in the column under Schott or Engler indicates that the species was included in their respective systems. Annals of the 762 Missouri Botanical Garden » X OOO < O < O © ра OOOO O O “фе ох asuaodpiun]po asuanj21D2 пашо 714519 UNUDÁDINQ 1152u24Q < О X шпироррлд штиоХот ‘dsqns пзродлауто “dsqns npurjduoq :dsqns npuvjduoq UNPUNIOLISDO UNUDAD]91DQ n$Dui01] “IBA > D]OI1UILD “Ie unaindindo1]D ‘тел uınaındındo11D npunjdsp unupuoun siuadaasnqio -dsqns штуитштууопдир :dsqns штитио ју пир aui umnunssimon 91019149] “TBA O O штуојутор стел штђојуто зотриј 159 M eunuod1y Kengereg 12918 вивих) ‘14 weung вивАпсу e[anzaua A eraog nied dJopenoy etquio[or) eureued BOTY 21507) епаелео seınpuoH өре [H e[euro]en^) ezuog OMXIJA "92992 peorqde12098 шпипгихузра ‘f хамзаау Croat 763 Volume 78, Number 3 1991 Anthurium sect. Pachyneurium оо ох O O ~ O О ~ ~ * O O О ж nupuual тидвјотишјоц штитви110у мађригј. asuaojvf әртәѕој unmunxa np4002]]20 ‘тел 11182112 ‘тел nysausa UNUDAIQUOP xippdsi4n2 ISUIQNI UNIDUILI шт1а19и188012 шти11{05и09 40]02u02 штпотиојоз әѕиә]]рарәттрэә зотриј јод епциодлу Kengei1eq zeg еџето) ‘14 weung виеАп=) е[әп2әиәд enog nəd Iopenoy etquio[or) вшеиеј тогу 21507) ende лес seınpuoH лорвајеб [4 вјешојепо) ezjog оохәр 'penuguo) “py хамзачу Annals of the 764 Missouri Botanical Garden Ox О x xO OO оо ~ О O O xO asuaKnnSpapd ‘тел UNUDOINOLIOI "IBA asuaKonn&pa1nd эзиэвитт ра шпалшәХхо $1u0310 штитлрг1“120 шталгимповдо вортоцоцорш пиХәт] әѕиә]ә10] зотриј 159 M вициэ8 гу Aendeieg вл ви) '14 WeULING BUBAN“) e[anzaua A eaog плод Jopenoy equo) | O вшвива 9 50" Bony F1 en BeIN sempuoH лорвлјеб [4 вјешојепо) anjog 09IXO9]A 'penunuo?) “p хтамааау Croat 765 Volume 78, Number 3 1991 Anthurium sect. Pachyneurium оо O x O O xO ох OOK > оо О x xO iXa]punis 12194195 unungowyo2s ториг1у22]ц28 *dsqns О nzauaunf :dsqns трриг1удојуов О штитио упао ISUIORDIUDS IDIUIAJDS ISUIJOPDA]DS штуотоиог тошол шталдилхајол штироузирл unyiodsosundind apqp122dsopnosd wnsuajoid *dsqns штутаар 'dsqns wnsuajoid un1vjoad штпиргзирла пиоштоја пиојћилиига штђојуприга ватриј 15а M вициэ8 гу Аеп9 еле гае: вивтпх) ‘14 шешип вивќ по) вјопхзизА eaog под | »4 ~ ~ Jopenoy | ж О equipo | О eureueqd тогу ?1s07) епделео звариоН PEARS n е[ешәзепгсу ss T] OMXIJA 'penunuo) ‘p хамзаау Missouri Botanical Garden Annals of the 766 'sIn220 satoads ay} э1эЦм Адипоо 1210 Aue 01 Jaja sx *Áieoo[ э4Аз Əy} 0} 1э}эл SƏPƏ :910N + со oo OO + ЄТ 8с О ж S 11 $5 L £I tc O x 0 зопшоричј 0 9 sadÁ] exe] Jo 4 [810], штатаиоцирх штуојтрура asuanjpdn шпитэ]т ‘тел ISUIÁDUDU “IBA шпиројт әѕиәриәј әѕиәјоар1р] ungiadns ‘dsqns 11u1]42Q1024Q `dsqns unq4adns зотриј 359 M eun ua y Aendeieg [вл еиепгсу ‘1J weung eurÁno) e[gnzaua A enog плод лорепојј equipo?) ешвива во 81500) паев seInpuoH лорелјеб | ејешәјепо) eziog 02IX9]A 'penunuo) `p xIaN3ddVy Volume 78, Number 3 Croat Anthurium sect. Pachyneurium Do a d ~“. qe ^ IGURES 1-4.— 1. Anthurium cubense Engl., Neill 2730, showing involute vernation. — 2. А. hacumense Engl., showing convolute vernation. —3. A. michelii Guill., Panama: Fortuna, Croat & Grayum 60252, showing Pachy- neurium-like habit. — 4. А. harrisii (Graham) Endl., Croat 38340, showing Pachyneurium-like habit. 768 Annals of the Missouri Botanical Garden FIGURES 5-8.— 5. Anthurium hacumense Engl., Croat 33558, showing Pachyneurium-like habit. —6. А. hookeri Kunth, cultivated, Virden s.n., showing convolute vernation and Pachyneurium-like habit. — 7. Anthurium harlin- gianum Croat, Croat 50816, showing erect, debris-trapping roots үө ш by roots). —8. А. asplundii ies cultivated at Selby Bot. Gard. (SEL 82-265), showing new flush of ro Volume 78, Number 3 769 1991 Croat Anthurium sect. Pachyneurium VET дува » O df. { : FiGURES 9-12.—9. Anthurium pendulifolium М. E. Br., cultivated by Burle-Marx, showing semi-intact cataphyll and downward-growing roots. — 10. A. asplundii Croat, Croat 55586, showing dense roots, severed petioles & hook- shaped cataphylls.— 11. А. dombeyanum Schott, Croat 50919 showing straight, lanceolate, intact cataphylls. — 12. А. sparreorum Croat, Croat 57195, cultivated by T. Fennel, Homestead, Florida, showing weathered cataphylls. 770 Annals of the Missouri Botanical Garden FIGURES 13-16.— 13. Anthurium dombeyanum Schott, Croat 58323, showing debris-covered roots.— 14. А barclayanum Engl., Croat 507 12, showing debris-trapping roots. — 15. A. salviniae Hemsley, Croat 32786, showing cucullate cataphyll.— 16. А. harlingianum Croat, cultivated at Selby Bot. Gard., Croat 67410, showing hook-shaped cataphyll. Volume 78, Number 3 771 1991 Croat Anthurium sect. Pachyneurium FIGURES 17-20.— 17. Anthurium loretense Croat, Croat 51226, showing circinate vernation. — 18. А. Croat, Croat 25521, showing mostly intact cataphylls. — 19. А. sparreorum Croat, Croat 38666, showing intact and weathered cataphylls. — 20. А. salviniae Hemsley, Panama: Barro Colorado Island (not collected), showing variation in leaf blade base shape. luteynii 772 Annals of the Missouri Botanical Garden CROSS-SECTIONAL PETIOLE SHAPES IN ANTHURIUM A. Basically terete B. C. D. ! 2 3 4 5 6 7 8 9 10 1 D-shaped or broader than thick 1 2 3 4 5 6 7 U-shaped or thicker than broad ОООО U Markedly angular 1 2 3 4 5 6 7 8 . Markedly ribbed abaxially v REM FIGURE 22. Number Produced FIGURE 21. ien sectional petiole shapes in Ап- thurium. — А. Basically terete: ranging from esulcate (1), to ier ium and spine sulcate (2 & 3), to obtusely and broadly sulcate (4), to narrowly and obtusely sulcate (5), to narrowly and acutely sulcate (6), broadly and acutely sulcate (7), shallowly and broadly sulcate (8), flat adaxially 9), flat adaxially with marginal ribs (10), to flat 2 with marginal and medial ribs .—B. D-shape broader than thick: ranging from flat adeciallr with кзн о medial ribs (6 & 7). — C. U-shaped or thicker than broad: ranging from broadly and obtusely sulcate (1) to narrowly and obtusely sulcate (2), narrowly and acutely sulcate with acute margins (3), shallowly and acutely sulcate (4), sulcate with acute margins (5), to sulcate with shar margins and a medial rib (6). — D. Markedly angular: ranging from ели with acute angles (1), = чи га se angles (2), trapezoidal (3), obtusely triangular (4), acutely triangular (5), acutely osa о and sharply ribbed around the remaining circumference (4), or with one or more ribs on the sulcus (5) O Inflorescences Jan T Y 1 T 1 Ји! Aug Sep Oct Nov Dec Graph illustrating inflorescence and leaf production over 2-year period. Volume 78, Number 3 Croat 773 1991 Anthurium sect. Pachyneurium . Syntype arium acuti foliu Engler к В+ Takes. ar: ser ег © А we * иллә. еам. - = dn " R9 FIGURES 23-26.— 23. Anthurium acutifolium Engl., Pittier 4099. —24-25. A. acutissimum Engl., Madison 2094.— 26. A. affine plex Storr 194 (cultivated at Kew). 774 Annals of the Missouri Botanical Garden FIGURES 27-30.— 27. Anthurium acutifolium AE Croat 59853. — 28-29. А. affine Schott, Hahn s.n. (photo by B. Hahn). — 30. А. anorianum Croat, Croat 567 Моџте 78, Митбег 3 Croat 1991 Anthurium sect. Pachyneurium D FIGURES 31-34.—31. Anthurium angustilaminatum Engl. ssp rait Sodiro s.n. —32. A. angus- tilaminatum Engl. var. cibuserpentis Croat, Steyermark 53764. — 33, 34. A. anorianum C, Croat 56757. Annals of the Missouri Botanical Garden © GURES 35-38. — 35, 36. Anthurium asplundii Croat, Croat 50693. — 37. т ee Croat, Croat 55586. — 38. A. а P Schultes & Maguire var. atropurpureum, Croat 56 Volume 78, Number 3 Croat 1991 Anthurium sect. Pachyneurium i AC omm MOM = CX Sao NN У VN dj ре Su a / - ( FIGURES 39-42. — 39. Anthurium atropurpureum В. Schultes & Maguire var. arenicola Croat, Croat 50303 40, 41. A. atropurpureum R. Schultes & Maguire var. arenicola Croat, cultivated at ET Bot. Gard. (SEL 78. 1146). —42. A. atropurpureum В. Schultes & Maguire var. arenicola Croat, Croat 53698 778 Annals of the Missouri Botanical Garden FIGURES 43-46.— 43. Anthurium s, Cg ii Croat 55608. — 44. A. atropurpureum R. Schultes & Maguire var. atropurpureum, Croat 56927. — A. atropurpureum R. Schultes & Maguire var. arenicola Croat, Croat 53582. —46. A. barclayanum Engl., Vides 50702. Volume 78, Number 3 779 1991 Croat Anthurium sect. Pachyneurium E P ~ 5 0 4 * | а L FIGURES 47-50. — 47-49. Anthurium barclayanum Engl., Croat 50702. —50. A. basirotundum Croat, Plowman 59804. Annals of the Missouri Botanical Garden v tt va >. » ERAR ox Mic [$17421 Li < 055067747 \ FIGURES 51-54.— 51, 52. Anthurium basirotundum Croat, Plowman 59804. — 53. A. bonplandii Bunting subsp. 51,5 bonplandii, Croat 59326. — 54. A. bonplandii Bunting subsp. bonplandii, Croat 59260. Volume 78, Number 3 1991 Croat Anthurium sect. Pachyneurium FIGURES 55-58.— 55. Anthurium a Bunting subsp. bonplandii, Croat 59260. — 56-58. A. bonplandii Bunting subsp. cuatrecasii Croat, Croat 5506 Annals of the Missouri Botanical Garden FicunES 59-62.— Anthurium barclayanum Engl., Thompson 373.—60-62. A. bonplandii Bunting subsp. guayanum (Bunting) ur Croat 54062. FIGURES 63-66.—63. Anthurium bradeanum Croat & Grayum, Croat 35751. —64. А. bradeanum Croat & Miri Croat 364804. — 65. A. brenesii Croat & В. A. Baker, Croat 46923. — 66. A. bucayanum Croat, Croat 615 Volume 78, Number 3 Anthurium sect. Pachyneurium ad Р » р 1 70} Р " / FIGURES 67-70.—67-69. Anthurium ђгепези Croat & R. Croat 50912. A. Baker, Croat 46923. —70. A. bucayanum Croat, Annals of the Missouri Botanical Garden FLORA CUTUCUENSIS FIGURES 71-74.—71-72. Ant isi pig i ada Croat, Croat 61597. —73. A. bushii Croat, Madison et al. 3285. — 74. A. campii Croat, Croat 61 Anthurium sect. Pachyneurium Croat 2 © 5 ua = = d o ~ e Ф ра = Ss <= > ~ S > < © ~ ~ ~ N © Uu ~ © ~ 3 © ~ O = © o = O & Е S o Е 3 ~ ~ 3 X ~ R x © ~ из ~ co ~ | из ~ Y) ш = а) o . E Volume 78, Number 3 786 Annals of the Missouri Botanical Garden FIGURES 79-82.— 79-80. Anthurium сатри Croat, Croat 61562. —81. А. carchiense Croat, Madison 3998. — 82. А. cataniapoense Croat, Croat 59319. Volume 78, Number 3 Croat 1991 Anthurium sect. Pachyneurium FicunEs 83-86.— 83-85. Anthurium cataniapoense Croat, Croat 59319. —86. A. caucavallense Croat, Croat 56729. Annals of the Missouri Botanical Garden X \ 1 Ё, а њ »Á (e L4 FIGURES 87-90.—87. Anthurium caucavallense Croat, Croat 56729. —88. А. colonicum К. Krause, Croat 25831, specimen destroyed.— 89. А. colonicum K. Krause, Croat 33640. —90. А. concolor K. Krause, Croat 33637. Volume 78, Number 3 Croat 789 1991 Anthurium sect. Pachyneurium Буш 91-94.— 9], 92. Anthurium caucavallense Croat, Croat 56729. —93. A. colonicum K. Krause, Croat — 94. А. concolor K. Krause, Croat 3363 Annals of the Missouri Botanical Garden ¡AA fa d FIGURES 95-98,— 95, 96. — b id К. Krause, Croat 33637. —97. А. coriaceum С. Don in Sweet, Croat 53757.— 98. Pu coriaceum C. in Sweet, cultivated by Burle-Marx. Volume 78, Number 3 Croat 791 1991 Anthurium sect. Pachyneurium FicunEs 99-102.— 99, 100. Anthurium Mere Schott, cultivated at Las Cruces Bot. Gard. (# 75-375). — 101. A. consobrinum Schott, Burger & Baker 151.—102. A. coriaceum G. Don in Sweet, cultivated at Prague. Annals of the Missouri Botanical Garden FiGuRES 103-106.— 103. Anthurium cowanii Croat, Cowan & Soderstrom 1758. —104. Anthurium crassi- nervium (Jacq.) Schott, Croat 60621. —105. A. crassinervium (Jacq.) Schott, Croat 383. 38. —106. A. crenatum (L.) Kunth, cultivated at Kew. "d oO $’. PREX Y = й Annals of the Missouri Botanical Garden й ы. Е“ Д = 0 Ф ~ ) ¥ DAT Я ч | х, у и „У Es 21; = FIGURES 115-118.— 115. Anthurium crenatum (L.) Kunth, cultivated at Кем. — 116, 117. А. cubense Engl., Neill 2730. —118. A. curtispadix Croat, Croat 55207. Volume 78, Number 3 Croat 1991 Anthurium sect. Pachyneurium B LÀ i) 4 ^ د‎ що Я РУ" м a | , , * FIGURES 119-122.— 119. Anthurium dombeyanum Brongn. ex Schott, Peru: Machupicchu (photo by T. Plow- man).— 120, 121. А. dombeyanum Brongn. ex Schott, Croat 58366. —122. А. dombeyanum Brongn. ex Schott, Croat 50847. 796 Annals of the Missouri Botanical Garden FIGURES 123-126.— 123. Anthurium curtispadix wy. e 55207. —124. A. ernestii Engl. var. ernestii, Croat 50993. —125. A. eximium Engl., Croat 67622. — 126. А. fasciale Sodiro, cultivated at Munich Bot. Gard. (#83/2905; photo: Bogner). — FIGURES 127-130.— 127. Anthurium dombeyanum Brongn. ex Schott, Croat 58324. — 128. А. ernestii Engl. var. ernestii, Croat 50994. —129. А. ernestii Engl. var. ernestii, Croat 50993. —130. A. ernestii Engl. var. ernestii, Croat 58591. Croat 797 Volume 78, Number 3 1991 Anthurium sect. Pachyneurium - : T? e Ld е ا‎ iv РРР: OM ARAL tah? ud ats O ERA МУ У A У edi 4 N EI 3 + tata A ме Seatac — = * y FIGURES 131-134.—131. Anthurium ernestii Engl. var. oellgaardii Croat, Oellgaard et al. 34801. — A. eximium Engl., Croat 57240.—133. A. fa asciale "= cultivated at Munich Bot. Gard. (#83/2905; pitis Bogner). — 134. А. fatoense K. Krause, Croat 572 Volume 78, Number 3 Croat 1991 Anthurium sect. Pachyneurium „ње "РУЦИ tA oe ae У po Rud AA A \ © FicunES 135-138.— 135. А. a Krause, Croat 49893. — 136. А. fatoense K. Eve Folsom 2949. — 137. A. fendleri Schott, Croat 54612. —138. A. fendleri Schott, cultivated by Bette Waterbury Annals of the Missouri Botanical Garden ‘ a > Wi / "wy AR T. MES q^ 47 А i> ДА y 9 » LI FIGURES 139-142.— 139. Anthurium fendleri Schott, Croat 56578. —140. A. galactospadix Croat, cultivated by Burle-Marx. — 141, 142. А. glaucospadix Croat, Croat 56753. Volume 78, Number 3 Croat 1991 Anthurium sect. Pachyneurium AN N MX 1 FIGURES 143-146.— 143. Anthurium galactospadix Croat, е by Burle-Marx.— 144. А. glaucospadix Croat, Croat 56753. —145. A. glaucospadix Croat, Croat 52000. —146. A. halmoorei aan Croat 45337. FIGURES 147-150.— 147. Anthurium halmoorei Croat, Croat 45337. — 148. A. hammelii Croat, Hammel 6160 (photo: Hammel).— 149. A. harlingianum Croat, Ecuador: Pastaza (not collected). — 150. A. harlingianum Croat, са 49688. Annals of the Missouri Botanical Garden RN SaR 4 М Dl 1 FIGURES 151-154.—151. Anthurium glaucospadix Croat, Croat 57518.—152. A. halmoorei Croat, Croat 45337.—153. A. hammelii Croat, Hammel 6160.—154. A. harlingianum Croat, Croat 49688 Volume 78, Number 3 Croat 803 1991 Anthurium sect. Pachyneurium FicURES 155-158.— 155. Anthurium harlingianum Croat, Croat 50290. —156, 157. A. holmnielsenii Croat, Croat 61640. —158. А. jenmanii Engl., Hay 2827 Annals of the Missouri Botanical Garden 162 # FIGURES 159-162.—159-160. Anthurium holmnielsenii Croat, Croat 50689. —161. А. holmnielsenii Croat, Croat 61640. —162. A. jenmanii Engl., cultivated at Port of Spain Bot. Gard. Volume 78, Number 3 Croat 805 1991 Anthurium sect. Pachyneurium p 163-166.—163, 164. A. johnsoniae Croat, Croat 62835. —165. A. knappiae Croat, Knapp et al. — 166. A. krukovii Croat, Duarte 6958. 806 Annals of the Missouri Botanical Garden и 167-170. — 51 167, 168. Anthurium latissimum Engl., 170. A. latissimum Engl., Croat 52230 Croat 57727. —169. A. latissimum Engl., Croat Volume 78, Number 3 Croat 1991 Anthurium sect. Pachyneurium FicunES 171-174.— 171. Anthurium lennartii Croat, Harling & Andersson 22159. —172, 173. A. leonianum Sodiro, Madison et al. 4960. —174. A. leonianum Sodiro, Croat 38986. 808 Annals of the Missouri Botanical Garden "TN ~ 177 178 FIGURES 175-178.— 175. Anthurium lindmanianum Engl., cultivated at Munich Bot. Gard. — 176. A. lindman- ianum Engl., cultivated by Burle-Marx.—177. A. lindmanianum Engl., Croat 57157, cultivated at Tropic World, San Diego. California. — 178. А. жене Engl., Croat 50697. Volume 78, Number 3 1991 Croat Anthurium sect. Pachyneurium 809 FIGURES 179-182.— 179. Anthurium johnsoniae Croat, Croat 62835. — 180. А. -r Engl., 50697. —181. А. llewelynii Croat, Croat 51092. — 1 82. А. loretense Croat, Croat 5122. Croat 810 Annals of the Missouri Botanical Garden н 183-186.— 183, 184. Anthurium ин Croat, Croat 51092. —185. А. llewelynii Croat, Croat 57212.— 186. А. loretense Croat, Croat 5122 Volume 78, Number 3 1991 FIGURES 187-190.— 187. Anthurium luteynii Croat, Croat 4907 1. —188. A. luteynii Croat, Croat 27169. — 189. A. luteynii Croat, Croat 25521. —190. А. machetioides Matuda, Ramirez 144. 812 Annals of the Missouri Botanical Garden FLORA OF ECUADOR | Ficures 191-194.— 191, 192. Anthurium manabianum Croat, Harling & Andersson 24750.— 193. А. ga- lactospadix Croat, Prance et al. 2924. —194. A. martianum K. Koch & Kolb, cultivated at Munich Bot. Gard. Volume 78, Number 3 Croat 813 1991 Anthurium sect. Pachyneurium MARIO VARGAS CUZCO, PERU а NTARE PEBUVIANAE маном ^erbarium 's lands bosbeheer \ N FIGURES 191A-194A. ies Anthurium maguirei А. Hawkes, LBB 16571.—192A. А. manuanum Croat, A Vargas 17747. —1934. selloum K. Koch, St. Johns, Virgin Is., cultivated ja Conrad то — 1944 martianum K. Koch & E pe ыы at Munich Botanical Garden 814 Annals of the Missouri Botanical Garden FIGURES 195-198. — 195. е luteynii Croat, Croat 27156. — 196. А. martianum К. Koch & Kolb, cultivated at Munich Bot. Gard. — 197. 4. napaeum Engl., Croat 55782. — 198. A. nervatum Croat, Croat 27329. Volume 78, Number 3 Croat 815 1991 Anthurium sect. Pachyneurium 201 ‘ » ARDEN anan IU FIGURES 199-202.—199. Anthurium napaeum Engl., Croat 57035. — 200. А. napaeum E Croat 50876. — 201. A. narinoense Croat, Gentry et al. 55343. — 202. А. nervatum Croat, Croat 48912 816 Annals of the Missouri Botanical Garden 9 FIGURES 203-206. — 203, 204. Anthurium obscurinervium Croat, Croat 50914. — 205. А. oerstedianum Schott, Croat 43419. —206. А. oerstedianum Schott, Croat 35293. Volume 78, Number 3 817 Croat Anthurium sect. Pachyneurium FIGURES 207-210.— 207, 208. Anthurium nizandense Matuda, Croat 45756.— 209, 210. А. ottonis К. Krause, Solomon 8652. Annals of the Missouri Botanical Garden e "^d „+ Зи eS FIGURES 211-214.—211. Anthurium oxycarpum Poeppig, Plowman 7569. —212. А. oxycarpum Poeppig, Plowman 3221.—213. A. oxycarpum Poeppig, Plowman & Kennedy 5699. Ба А. oxyphyllum Sodiro, Croat Volume 78, Number 3 Croat 819 1991 Anthurium sect. Pachyneurium FIGURES 215-218.— 215, 216. Anthurium oxyphyllum Sodiro, Madison s.n., ijo at Selby Gardens. — 217. A. pachylaminum Croat, Croat 58161. —218. A. pachylaminum Croat, Croat 5099 Annals of the Missouri Botanical Garden Volume 78, Number 3 Croat 821 1991 Anthurium sect. Pachyneurium FIGURES 223-226.— 223, 224. Anthurium gt a Croat, Croat т — 225. А. pallatangense Engl., Croat 61545. —226. A. paraguayense Engl. var. paraguayense, Beck 3 e FIGURES 219-222.— 219, 220. Anthurium e Croat, Croat 58161. —221. A. palenquense Croat, Croat 38670. — 222. A. pallatangense Engl., Croat 6154. 822 Annals of the Missouri Botanical Garden ¿qe ип Му 4 . "AS T "m a “> " [229 : SEN | sk ` ~ EX. ~ 3 түз ۳ ús РТ. 1 ر‎ 2 a a b> А и FIGURES 227-230.— 227. Anthurium paraguayense Engl. var. paraguayense, Beck 3296. — 228. А. para- guayense Engl. var. coroicoanum Croat, Croat 51657. —229. A. paraguayense Engl. var. coroicoanum Croat, Croat 51704. — 230. A. pendulifolium N. E. Br., Croat 57196. Volume 78, Number 3 Croat 823 Anthurium sect. Pachyneurium г i DE 231-234.— 231. Anthurium paraguayense Engl. var. paraguayense, Beck 3296.—232. A. - guayense Engl. var. coroicoanum Croat, Croat 51704. — 233. A. pendulifolium М. E. Br., Croat 58610. 1334. А. ам N. Е. Br., Croat 57196. Annals of the Missouri Botanical Garden = MEA wa. i FIGURES 235-238. — 235, 236. Anthurium penningtonii Croat, Croat 49439. — 237, 238. A. plowmanii Croat, Croat 53563. Volume 78, Number 3 825 1991 Croat Anthurium sect. Pachyneurium FIGURES 239-242.— 239. Anthurium plowmanii Croat, Croat 53542,, cultivated by F. Fuchs, Homestead, Florida. — 240. A. plowmanii Croat, Croat 53701. —241. A. pranceanum Croat, Prance et al. 12640. —242. 4A. prolatum Croat & R. A. Baker, Croat 46979. Annals of the Missouri Botanical Garden FIGURES 243-246.— 243, 244. Anthurium mom Croat & R. A. Baker, cultivated at Las Cruces Bot. Gard. — 245. A. prolatum Croat & R. A. Baker, Croat 47087. — 246. A. protensum Schott subsp. protensum, Croat 44501. Volume 78, Number 3 Croat Anthurium sect. Pachyneurium t Җ Dos, FIGURES 247-250.—247, 248. Anthurium protensum Schott subsp. protensum, Costa Rica: а СаггШо National Park (not collected). — 249, 250. А. protensum Schott subsp. arcuatum Croat, Croat 60077 Annals of the Missouri Botanical Garden RES 251-254.— 251. Anthurium db ne Schott subsp. arcuatum Croat, Croat vim — 252. жол Croat, Croat 48674.—253. A. pseudospectabile Croat, Croat 48777. — 254. 4. pseudospectabile Croat, Croat 66592 Volume 78, Number 3 1991 Croat Anthurium sect. Pachyneurium FIGURES 255-258.— 255. Anthurium pendulifolium М. E. Br., Croat 57196.— 256. А. pow Croat, Croat 49439. — 257. A. penningtonii Croat, Croat 58215. — 258. A. plowmanii Croat, Croat 53542 LES 259-262.— 259. Anthurium prolatum Croat & R. A. Baker, cultivated at Las Cruces Bot. Gard. — 260, 61. A. protensum Schott subsp. protensum, Croat 367 14. —262. А. ranchoanum Engl., Croat 36053. Annals of the Missouri Botanical Garden FIGURES 263-266.— 263. Anthurium purpureospathum Croat, Croat 60145. — 264. A. n um Croat, Croat 33139. — 265. A. ranchoanum Engl., Croat 36053. — 266. А. ranchoanum Engl., Croat 47 Volume 78, Number 3 Croat 831 1991 Anthurium sect. Pachyneurium FIGURES 267-270.— 267, 268. Anthurium reflexinervium Croat, Plowman 7585. — 269. А. salvadorense Croat, Croat 42092. — 270. A. salviniae Hemsley, Panama: Barro Colorado Island (not collected). Annals of the Missouri Botanical Garden а дани LLL e». 9 «le a M PU CX ГС OTT A Y. M „ > k i Y "Ts UNA S P" FIGURES 271-274.— 271. Anthurium reflexinervium Croat, Plowman 7585.— 272. A. remotigeniculatum Croat, cultivated at Selby Bot. Gard. (SEL 81-2284). — 273. A. salvadorense Croat, Croat 42169. — 274. A. salviniae Hemsley, Croat 56579. Volume 78, Number 3 1991 Croat Anthurium sect. Pachyneurium FicuRES 275-278.— 275. Anthurium remotigeniculatum Croat, cultivated at Selby Bot. Gard. p 81-2284), Croat 58360. — 276. A. salviniae Hemsley, cultivated at Selby Bot. Gard. —277. А. sarukhan m Croat Haager, Hauser s.n., cultivated at Prague. — 278. А. schlechtendalii Kunth subsp. schlechtendalii, Croat 41525. FIGURES 279-282.— 279, 280. Anthurium schlechtendalii ze subsp. Jimenezii (Matuda) Croat, Croat 45694. — 281. A. schottianum Croat & R. A. Baker, Croat 43247. — 282. A. seibertii Croat & R. A. Baker, Croat 48568. Annals of the Missouri Botanical Garden 1 y wx Es, FIGURES 283-286.— 283. Anthurium santiagoense Croat, Palacios 1492. — sarukhanianum Croat & Haager, Haager s.n., cultivated at ceni — 285. A. s Е ^as subsp. schlechtendalii, Croat 41802. — 286. A. schlechtendalii Kunth subsp. schlechtendalii, Croat 394 Volume 78, Number 3 Croat 835 1991 Anthurium sect. Pachyneurium 5 мађ q ~ FIGURES 287-290.— 287. Anthurium schottianum Croat & | Е. A. Baker, Croat 43247. — 288. А. seibertii Croat R. A. Baker, 10555.—289. А. selloum K. Koch, Britton & Shafer 560.—290. А. simpsonii Croat, Simpson & Schunke 392. 836 Annals of the Missouri Botanical Garden FIGURES 291-294.— 291. Anthurium seibertii Croat & R. A. Baker, Croat 26494. — 292, 293. A. solitarium (Vell. Conc.) Schott, Croat 61894. —294. A. soukupii Croat, Croat 5074 Volume 78, Number 3 1991 " , AMEN ns ` ` zi E. J FIGURES 295-298.— 295. Anthurium solitarium (Vell. Conc.) Schott, Croat 61894. — 296. А. solomonii Croat, Solomon 13000. — 297. А. soukupii Croat, Croat 50747. — 298. А. soukupii Croat, Croat 58298. 838 Annals of the Missouri Botanical Garden M gne 299-302.— 299, 300. Anthurium sparreorum Croat, cultivated at Selby Bot. Gard. gin #15-7 23).— 301, 302. A. sparreorum Croat, Croat 57195, cultivated by T. Fennel, Homestead, Florid Volume 78, Number 3 Croat 839 Anthurium sect. Pachyneurium M ARR FIGURES 303-306.— 303. Anthurium spathiphyllum М. E. Br., Croat 44309. —304. A. spathiphyllum N. Е. Br., Grayum & Schatz 644. —305. A. spectabile Schott, cultivated at Las Cruces Bot. Gard. — 306. А. standleyi Croat & R. A. Baker, Croat 43431. Annals of the Missouri Botanical Garden des URES 307-310. — 307. Anthurium standleyi Croat & К. A. gar Croat 43431. —308. A. superbum Madison superbum, cultivated at Selby Bot. Gard. (SEL #77-61). —309. А. superbum Madison а ан ВА by D. Hull, San Diego, California. — 310. A. superbum subsp. brentberlinii Croat, Berlin 5 Volume 78, Number 3 1991 Croat Anthurium sect. Pachyneurium FIGURES 311-314.— 311. Anthurium w Engl., Croat 58130. — 312. А. tarapotense Engl., Croat 58115. —313, 314. А. tenaense Croat, Croat 49631 842 Annals of the Missouri Botanical Garden 317 — — | FIGURES 315-318.— 315. Anthurium uleanum Engl. var. uleanum, Croat 4966 1. —316, 317. A. uleanum Engl. var. uleanum, Croat 50412. —318. A. uleanum Engl. var. nanayense Croat, Foster 4227. Volume 78, Number 3 843 Croat 1991 Anthurium sect. Pachyneurium FIGURES 319-322.— 319. Anthurium spectabile Schott, Croat 46968. — 320. A. standlevi Croat & R. A. Baker, Croat 43439. —321. A. upalaense Croat & R. A. Baker, Croat 36342. —322. A. validifolium K. Krause, Croat 44445. Annals of the Missouri Botanical Garden FIGURES 323-326. — 323. Anthurium upalaense Croat & К. A. Baker, Croat P 342. — 324. А. upalaense Croat & Baker, Croat 43221. —325, 326. A. validifolium K. Krause, Croat 66204. Volume 78, Number 3 Croat 845 Anthurium sect. Pachyneurium FIGURES 327-330.— 327. Anthurium wonky О К. Krause, Croat 10667. — 328. Anthurium venosum Griseb., cultivated at Havana Botanical Garden (Photo I. Arias).—329. А. vaupesianum Croat, Croat 56795. — 330. А. vaupesianum Croat, Zarucchi & Davis 1185. 846 Annals of the Missouri Botanical Garden FIGURES 331-334.— 331. Anthurium validifolium К. Krause, Croat 44445. —332. A. validifolium К. Krause, Croat 10667.—333, 334. A. watermaliense hort. ex Bailey & Nash, Croat 33958. Volume 78, Number 3 Croat 847 Anthurium sect. Pachyneurium FIGURES 335-338. Anthurium wagenerianum К. Koch & Bouché, Croat 54543. Annals of the Missouri Botanical Garden Antrim Kant a VTL FLORA Of VENEZUELA FIGURES 339-342.— 339, 340. Anthurium willifordii Croat, Croat 61087. т v = Bunting, Stey- ermark & Bunting 103099. —342. A. xanthoneurum Bunting, Steyermark et al. Volume 78, Number 3 Croat 1991 Anthurium sect. Pachyneurium FIGURES 343-346.— 343, 344. Anthurium galactospadix Croat, Croat 62633. — 345, 346. A. solomonii Croat, Solomon & Escobar 12494 (Photo J. Solomon). Annals of the Missouri Botanical Garden FIGURES 347-350.— 347. Anthurium тинн Croat, cultivated by Marilyn and Al Johnson, Miami. — 348. A. vi а Jonker & Jonker, Daniels & Jonker 1241. —349, 350. A. venosum Griseb., cultivated at Dewey Fisk, Mia Volume 78, Number 3 Croat 851 1991 Anthurium sect. Pachyneurium The New York Botenical Garden Avion nr кочу d m Verrirue Cros cepas Sm ta шанагы p we». Thomasi; War "Mese Esperanza» Macacona inca de к errore с. Miramar. ПЫ. «40-5 841610 300-400 а Terrestre; enpádice amarillos ы espata verde. = ч. ч. Thomas У qu. 3 5 2 P. Betella and A. Centurión А = - from the net tun Financial assistance шы saith Noyes Foundat Lon cat GARDEN HERBARIUM (MO Terres rre, € sobre rocas. Eapádico de сеје PR y Syd ре T wide so-marrdn; turon Tale vine «эсе, бирлеге ю- diano, envía serde ^4. O | 13 Да шы FIGURES 351-354. и Паја о уаг. "е Croat, var. nov., Herrera 372.--352. Anthurium atropurpureum var. thomasii Croat, var. no as et al. .— 353. Ат hurium lyon в Ramirez & Laskowsky 310 ud by J. Meyers). — 354. ا‎ ae vinillense Bunting, Huber И 852 Annals of the Missouri Botanical Garden Anthurium Schott 539, 540, 541, 543, 544, 545, 547, 555, 559, 561, 562, 563, 564, 568, 572, 573, 574, 617, 636, 643, aha a — 2 о m > 48, 549, 550, 551, 552, 683, 754, 755, 772 grex Acamptophyllium Schott 759 grex Erythropodium Schott 542, 759 grex Oxycarpium Schott 542, 543, 544, 572, 759, 760 grex Pachyneurium Schott I 573, 757 grex Xialophyllium Schott 5 sect. m Schott њи 544, 559, 562, 621, 756 sect. ен Schott 550, 648, 760 sect. Cardiolonchium Schott 562, 631, 761 sect. Dactylophyllium Schott 562 se sect. Oxycarpium Schott 542, 544, 760 sect. Pachyneurium Schott 539, 540, 541, 542, 543, 544, 545, 546, 547, 548, 549, 550, 551, 552, 555, 556, 557, 558, 559, 560, 561, 562, 563, 564, 565, 566, 567, 569, 571, 572, 573, 574, 597, 629, 631, 634, 643, 648, 652, 657, 660, 666, 670, 679, 684, 688, 689, 703, 712, 719, 728, 748, 750, 753, 754, 755, 756, 757, 758, 759, 762, 767, series Pachyneurium 539, 546, 547, 549, 553, , 965, 573, 761 series anaes Croat 539, 544, 545-546, 548, 52, 553, 555, 564, 565, 567, 571, 573, 596, 599, 622, 649, 662, 683, 684, 687, 689, 692, 695, 696, 702, 718, 731, 761 t. Polyneurium Engl. 539, 542, 755 к. Polyphyllium Engl. 543, 545, 572 sect. Porphyrochitonium Schott 543, 551, 562, 755, sect. Te tras spermium Schott 545 ect. Ре , 680, 754, 755, 756, 757, 759, 760 series Obscureviridia Engl. =, series Flavescentiviridia Engl. 5 sect. Xialop hyllium сој: 543, oed 670 caule Scho tt 636, 75 acaule лын EN 572, 757, 759 acaule var. brevipes Engl. 635 acaule var. portoricensis Kuntze 635 pee rey Engl. 542, 554, 556, 564, 568, 570, , 628, 758, 759, 762, 773, 774 acu Engl. var. acutifolium 563, 594-595, actin var. herrerae Croat 563, 577, 594, 595, 159, , 851 кол Engl. 551, 554, 558, 565, 568, 571, 579, 595-596, 662, 759, 762, 773 acutissimum var. maius Sodiro 595 aduncum Schott 754, aduncum (Vell. Conc. ) Schott 541, 754 affine Schott 541, 549, 550, B 554, 558, 560, 566, 569, 572, 592, 593, 596-598, 757, 758, 759, 761, 762, 773, 774 agnatum Schott 542, 754, 757, 758 agoyanense Sodiro 542 . agoyanense var. eleutheroneuron Sodiro 542, 640 gl. 539, 542, 543, 544, 562, 629, A. A. ~ EN a А. ~ anaa ој А. a mons ъъ BRR BP Ro BRR . atropurpureum R. Schultes albopunctatum Sodiro 696 andicola Liebm. 544, 621, 754, 756, 758 angustilaminatum Engl. 542, 554, 558, 560, 565, 568, 571, 598, 599, 731, 759, 762 angustilaminatum Engl. subsp. angustilaminatum 560, 581, 598-599, 600, 759, 762, 775 angustilaminatum subsp. cibuserpentis Croat 579, 581, 599-600, 759, 762, 775 angustilaminatum var. albidum Sodiro 542, 598 angustilaminatum var. brevipes Sodiro 542, 598 angustilaminatum var. crassum Sodiro 542, 598, 59 a var. gladiatum Sodiro 542, 598 anorianum t 554, 556, 564, 568, 569, 570, 589, 600- 601, 747, 759, 762, 774, 775 appunianum Schott 544, 758 asplundii Croat 565, 568, 570, 588, 589, 601-602, 9, 762, 768, 769 e 718 atropurpureum R. Schultes & Maguire 542, 549, 57, 559, 568, 569, "570, 602, 603, 610, 667, 759, 762 & Maguire var. atropur- pureum 542, 546, 555, 557, 560, 563, 564, 565, 566, 580, 602–604, 759, 762, 776, 778 , 604–606, 607, 667, 762, 777, 778 atropurpureum var. thomasii Croat 565, 220i 580, 62, 589, 603, 605, 606-607, 759, 694, 700, 759, . barclayanum Engl. 542, 547, 549, 550, 551, 555, 558, 560, 562, 563, 565, 568, 571, 587, 589, 592, 602, 607-608, 674, 728, 758, 759, 762, 770, 118, 779, 782 basirotundum Croat 555, 558, 565, 568, 580, 582, 590, 608–609, 759, 762, 779, 780 . bonplandii Bunting 542, 546, 549, 557-558, 568, 569, 570, 603, 609, 610, 611, 613, 615, 652, 657, 664, 668, 672, 748, 750, 754, 759, 762 . bonplandii Bunting subsp. bonplandii 542, 548, 555, 557, 559, 589, 605, 609-612, 615, 672, 680, 762, 780, 781 564, 566, 569, 583, 585, 586, 587, 704, 753, . bonplandii subsp. cuatrecasii Croat 555, 557, 562, 564, 566, 569, 583, 585, 587, Е 591, 592, 603, 610, 613- pens 615, 759, bonplandii subsp. guayanum Croat 54 A 549, 555, 558, 566, 567, 569, 583, 589, 592, 610, 614- 616, 634, 664, 668, 748, 753, 754, 759, 762, 782 bonplandii subsp. rionegrense Bunting 609 boucheanum K. Koch 544, 758 аср cil Schott Ss 721, 757, 758 bradeanum t & Grayum 550, 554, 555, 556, 560, 562, 563, 568, 570, 576, 616-617, 648, ET 8 & `В. А. ВаКег 543, 554, 556, 560, 563, 568, 570, 575, 617-618, 706, 759, 762, 782, 78 brownii се ye 754, t 568, E Qm 618-619, 621, buenaventurae Engl. 651, 695, 758 Volume 78, Number 3 Croat Anthurium sect. Pachyneurium 853 1991 А. bushii Croat 564, 568, 570, 585, 593, 619-620, А. englerianum Bunting 662, 664 759, 762, 784 А. enormispadix Matuda A. is Croat 555, 558, 568, 571, 578, 619, 620- А. ernestii Engl. 542, 549, 550, 557, 563, 564, 565, ‚ 733, 759, 762, 784, 785, 786 568, 569, 571, 640, 642, 643, 644, 654, 669, А. canasas Croat 651 681, 740, 758, 759, 76 А. carchiense Croat 553, 554, 558, 564, 565, 568, A. ernestii Engl. var. ernestii 554, le 566, 584, 587, 571, 581, 621-622, 693, 759, 762, 785, 786 591, 593, 642-646, 759, 763, A. cartilagineum (Desf.) Kunth 544, 755, 758 А. ernestii vi oellgaardii Croat 2 557, 584, 585, A. cataniapoense Croat 550, 566, 569, 570, 583, 585, 87, , 646-647, 759, 763, 798 588, 623-624, 660, 664, 676, 759, 762, 786, A. eximium Engl. 547, 550, 554, 556, 560, 563, 568, 787 570, 575, 647-650, 759, 763, 796, 798 A. caucavallense Croat 549, 555, 564, 568, 571, 578, A. fasciale Sodiro 542, 554, 557, 565, 568, 571, 579, 581, 588, 591, 592, 623-624, 656, 665, 671, 580, 648-650, 684, 719, 731, 759, 763 759, 763, 787, 788, 789 А. fatoense К. Krause 548, 554, 556, 560, 563, 568, A. colonicum K. Krause 544, 550, 554, 556, 559, 560, 577, 627, 650, 759, 763, 798, 562, 563, 564, 568, 570, 574, 624-626, 631, A. ace 1 541, 544, 548, 550, 553, 555, 556, 685, 759, 763, 788, 789 , 963, 564, 566, 567, 568, 570, 585, A. concolor K. Krause 542, 552, 553, 554, 556, 560, 586, We 605, 633, 638, 651, 663, 757, 758, 563, 564, 568, 569, 570, 576, 588, 626-627, 759, 763, 799, 800 744, 759, 763, 788, 789, 790 A. flavescens Роерр. 683 А. consobrinum Schott 542, 544, 545, 548, 551, 552, А. fontanesii Schott = 633, 634, 755, 756 553, 554, 556, 559, 560, 562, 563, 568, 570, A. fortinense Engl. 7 8 576, 594, 617, 627-628, 648, 650, 759, 763, А. galactospadix es 550, 551, 565, 569, 571, 591, 4-655, 759, 763, 800, 801, 812, 849 A. consobrinum var. cuneatissimum Engl. 627 A. giganteum Matuda non Engl. 542, 716 A. pur ms (Willd.) G. Don in Sweet 544. 755, 758 А. glaucescens Kunth 629 А. сопасешт Endlicher 629 А. glaucospadix Croat 549, 551, 554, 562, 563, 564, A. coriaceum G. Don in Sweet 554, 555, 558, 560, 566, 568, 571, 578, 593, 624, 655-656, 671, 566, 569, 572, 581, 628-630, 754, 759, 763, 159, 763, ie 801, 802 90, A. glaucum Schot A. corocoroense Bunting 614, 615 A. glaziovii Hook. f. 728, 758 A. costaricense Engl. 705, 707 A. Em (Jacq.) Kunth 544, 755, 758 A. cotobrusii Croat & В. A. Baker 544, 554, 556, 560, А. guanchezii Bunting 543, 548, iu sor 569, 570, 562, 568, 570, 574, 630-631, 712, 751, 759, 582, 610, 656-657, 662, 759, 763, 793 А. guaiquinimae Bunting 614, 6 А. cowanii Croat 565, 567, 569, 570, 589, 631-632, А. guayanum Bunting 542, 614 759, 763, 792 A. hacumense Engl. 542, 543, 562, 755, 758, 767, А. crassinervium (Jacq.) Schott 541, 554, 556, 560, 563, 564, 566, 567, 568, 570, 571, 588, 592, А. pen vier 543, 554, 555, 556, 560, 562, 563, 593, 632-635, 636, 750, 756, 757, 158, 759, 977, 578, 657, 719, 724, 759, 763, 763, 792, 793 А. crassinervium Schott 757 ~ ~ خخخ ھچ debe crenatum (L.) Kunth 541, 549, 554, 556, 560, 562, 567, 568, 570, 572, 575, 635-637, 721, 758, 759, 763, 792, 793, 794 сиђепзе Engl. 542, 550, 553, 554, 556, 560, 563, 564, 567, 568, 569, 570, 572, 575, 576, 577, 588, 637-639, 758, 759, 763, 767, 793, 194 cuneatissimum (Engl.) Croat 617 curtispadix Croat 549, 551, 568, 571, 591, 639- 640, 759, ee Mies 794, 796 cuspidatum Mas 755 Au Schott 542, 687, 688 atophyllum K. Koch & Sellow 542, 640, 641, ?758 decurrens Poeppig 543, 544, 756 disparile Schot ди 593. 640- 642, 671, 682, 690. 757, 758, 163, 169, ds 771, 795, 796 о Кі22 gregium ru 632, E 757 а ит К. Косћ 7 ellipticum K. Koch & же 632, 633, 634, 758 759, a ~ А. А. А. 4. ки & OAK ui ds . hammelii Croat 563, 564, 568, 570, 575, 658-659, 759, 763, 801, 802 harlingianum Croat 548, 555, 557, 562, 563, 564, 565, 568, 571, 583, 584, 585, 588, 590, 602, 620, 659-661, 676, 759, 763, 768, 770, 801, 2. harrisii (Graham) Endl. 767 ALAN е 554, 558, 565, 568, 571, 579, 61- , 759, 763, 803, 804 E Kunth yn 542, 543, 567, 636, 755, 7571, 768 huegelii Schott 757 huequeense Bunting 749, 750 . iramirezae Bunting 543, 566, 567, 569, 570, 582, 610, 657, 662, 759, 763, 851 jenmanii Engl. 542, 549, 550, 554, 556, 560, 566, 567, 568, 570, 576, 583, 585, 588, 589, 652, 662, 758, 759, 763, 803, 804 jimenezii Matuda 575, 723 johnsoniae Croat 547, 553, 554, 556, 564, 568, 571, 581, 665, 760, 764, 805, 809, 850 joseanum Engl. 542, 706, 758 knappiae Croat 565, 568, 571, 572, 578, 665-666, 690, 760, 764, 805 854 Annals of the Missouri Botanical Garden ~ Lis R ы ~ m PN a ~ ~ hhh‏ ج > a aa ь hanah mon a a à aa a à à anan krukovii Croat 566, 569, 570, 580, 666-667, 704, 720, 758 lanjouwii Јопкег & Jonker 542, 565, 567, 569, 570, 80, 582, 610, 667-668, 760, 764, 850 ig uam Engl. 542, 549, 554, 557, 562, 565, 568, 571, 578, 580, 582, 584, 590, 668, 714, 158, 160, 764, 806 lennartii Croat 554, 558, 565, 568, 571, 578, 669- 670, 680, 760, 764, 807 leonianum Sod. 542, 555, 564, 568, 571, 584, 586, 641, 670-671, 756, 758, 760, 764, 807 liebmannii Schott 544, 758 lilacinum Bunting 544, 562, 755 lindmanianum Engl. 542, 546, 555, 558, 566, 569, 570, 572, 583, 587, 592, 671-673, 758, 760, 764, 808 linguifolium Engl. 542, 546, 565, 568, 571, 581, 89, 608, 673-674, 760, 764, 808, 809 . llewelynii Croat 547, 555, 557, 560, 563, 565, 568, 71, 578, 591, 674-675, 738, 760, 764, 809, . longispathum Carriére 544, 734, 755, 758 . loretense Croat 547, 555, 557, 564, 565, 569, 571, 3, 588, 623, 632, 660, 675-677, 760, 771, 809, 810 764, . luteynii Croat 543, 547, 548, 552, 554, 556, 560, 561, 562, 563, 564, 568, 570, 576, 677, 760, 764, 771, 811, 814 macedoanum А. Hawkes 597 machetioides Matuda 563, 568, 570, 575, 678-679, 160, 764, 811 macrophyllum (Swartz) Schott 727 maguirei А. Hawkes 542, 565, 567, 569, 570, 591, 592, 679-680, 760, 764, 813 manabianum Croat 565, 568, 571, 581, 670, 680, 760, 764, ласа оне C. D. Adam 8 manuanum Croat 565, 568, 571, 591, 680-681, 740, 760, 764, 813 martianum K. Koch & Kolb in K. Koch 542, 567, 569, 579, 581, 681-682, 758, 760, 764, 812, 813, 814 maximum (Desf.) Engl. 542, d 756, 758 mexicanum Liebm. non Engler 720, 757 michelii Guillaumin 543, 756, 767 napaeum Engl. 554, 558, 564, 565, 568, 571, 573, 81, 588, 682-683, 695, 760, 764, 814, 815 narinoense Croat 564, 565, 568, 571, 581, 683, 719, 760, 764, 815 nervatum Croat 543, 554, 556, 560, 562, 564, 568, 570, 574, 625, 631, 684, 760, 764, 814, 815 ан РАМА Matuda 554, 555, 556, 563, 568, 570, 577, 685–686, 724, 760, 764, 817 nobile Engl. 728, 729, 758 obscurinervium Croat 552, 555, 558, 568, 571,579, 622, 686, 693, 760, 764, 816 oerstedianum Schott 542, 548, Vies е 563, 568, 570, 575, 687-688, 760, 764, ottonis К. Krause 542, 565, 566, oe 571, 579, 688-689, 760, 764, 817 . oxycarpum Poeppig in Poepp. & Endl. 542, 544, 555, 557, 560, 563, 564, 565, 566, 569, 571, 666, 689-692, 758, 760, 764, 818 oxyphyllum Sodiro 552, 554, 558, 565, 568, 571, 578, 692-693, 760, 764, 818, 819 A. pachylaminum Croat 546, 555, 557, 565, 568, 569, ~ А. А. А. a ~ ~ м ~ А. ~ А. paraguayense Engl. var . ратавиауепзе . pendulifolium 97, . protensum Schott $ protensum subsp. arc meom. n n а | © 570, 585, 819, 821 586, 590, 693-694, 715, 760, 764, . palenquense Croat 549, 551, 555, 558, 565, 568, 571, 579, 581, 683, 694–695, 760, 764, 821 . pallatangense Engl. 542, 555, 558, 565, 568, 571, 581, 695-697, 702, 760, 764, 82 paraguayense Engl. 542, 549, 557, 560, 562, 563, 565, 566, 568, 571, 697, 758, 760, 764 . paraguayense 554, 557, 569, 586, 587, 590, 591, 593, 697-698, 699, 760, 764, 821, 822, 823 var. coroicoanum Croat 549, 555, 557, 565, 566, | 571, 590, 698-699, 714, 760, 764, 822, N. E “Br. 542, 547, 550, 553, 555, 559, 561, 564, 565, 569, 570, 580, 586, 699- 701, 758, 760, 765, 769, 822, 823, 829 penningtonii Croat 564, 565, 568, 569, 571, 578, 580, 649, 696, 701-702, 731, 760, 765, 824, 8 peripense Engl. 682, 683 plowmanii Croat 543, 549, 550, 554, 557, 560, 563, 565, 566, 568, 569, 572, 582, 586, 590, 592, 675, 702-704, 730, 756, 760, 765, 824, 825, 829 . pranceanum Croat 548, 555, 557, 566, 569, 570, 582, 704-705, 760, 765, 825 . preussii Engl. 632, 633, 758 aker 543, 554, 556, 568, 760, 765, 825, 826, prolatum Croat & 0, 574, 618, 705, 706, 9 . protensum Schott 542, 554, 556, 560, 563, 568, 570, 618, 705, 706, 707, 726, 746, 758, 760, 765 subsp. protensum 556, 560, 575, 708, 760, 765, 826, 827, 829 uatum Croat 556, 560, 564, 575, 658, 706, 707-709, 760, 765, 827, 828 pseudospectabile Croat 543, 546, 547, 550, 554, 556, 560, 563, 564, 568, 569, 570, 575, 577, 709-710, 734, 760, 765, 828 705-707, . purpureospathum Croat 543, 549, 550, 553, 554, 556, 560, 562, 563, 564, 568, 570, 576, 711, 744, 760, 765, 830 ranchoanum Engl. 544, 554, 556, 561, 563, 568, 570, dea 631, 685, 711-713, 734, 760, 765, 829, 8 Nas UM 638, 651, 653, 757, 758 bcc Croat 549, 554, 557, 565, 568, 569, 9. 584, 589, 713-714, 753, 760, 765, 710- 831, remotigeniculatum Croat 548, 549, 555, 557, 563, 565, 568, 570, 584, 714-715, 760, 765, 832, 833 dr Engl. 640, 758 ug rumicifolium Sodiro 6 rusticum N. E. Br. ex Engl. 697, 758 пије пен Сгоа! 543, 550, 554, 556, 560, 563, 568, 569, 570, 577, 715-716, 760, 765, 831, 832 . salviniae Hemsley 542, 546, 550, 553, 554, 556, Volume 78, Number 3 1991 Croat 855 Anthurium sect. Pachyneurium EN a EN 33 P. Croat 564, 565, 568, 571, 579, . sarukhanianum Croat & Haager 550, 553, 563, . schottianum Croat 560, 561, 563, 564, 568, 569, 570, 576, 578, , 626, 716-718, 721, 758, 760, 118-719, 760, 765, 834 578, 719-720, 760, 765, 833, 834 schlechtendalii Kunth 541, 552, 554, 556, 563, 568, 569, 570, 577, 679, 716, 717, 721, 724, 744, 757, 758, 760, 765 schlechtendalii Kunth subsp. schlechtendalii 560, 577, 720-723, 760, 765, 833, 834 . schlechtendalii subsp. jimenezii Croat 556, 560, 563, 575, 658, 686, 719, 720, 723-724, 760, 765, 4 Е. А. Вакег 543, 549, 554, 556, 560, 563, 568, 569, 724-725, 760, 756, 833, 835 544, 570, A. scopulicola Standley & L. 0. Williams 594 ~ aa a TEN seibertii Croat & R. A. Baker 543, 554, 556, 560, 563, 568, 569, 570, 574, 678, 706, 725-727, 746, 760, 765, 833, 835, 836 seleri Engl. 544, 756, 758 selloum K. Koch 542, 567, 568, 572, 574, 727- 728, 758, 760, 765, 813, 835 simpsonii Croat 555, 558, 565, 568, 571, 581, 728, 760, 765, 835 sodiroanum Engl. 670 . solitarium Schott 757 . solitarium (Vell. Conc.) Schott 541, 550, 555, 558, , 569, 572, 588, 590, 597, 728-730, 761, 765, 836, 837 754, 758, ИН "ps 555, 557, 565, 566, 568, 571, 579, 703, ‚ 761, 765, 837, 49 bp. сте 565, 566, 568, 571, 579, 599, 600, 649, 730–732, 761, 765, 836, 837 А. sparreorum Croat 554, 558, 559, 565, 568, 571, bs bs ~ mom ~ ~ ~ a EN м А spathiphyllum 563, 5 582, 588, 619, 621, 671, 732-733, 761, 765, 769, 771, 838 Е. Вг. 542, 550, 552, 553, 733-734, mN. 560, 68, D 575, 617, 648, 754, 761, 765, spectabile Schott 541, 547, 550, 554, 556, 568, 570, 575, 709, 734-735, 757, 758, 765, 839, 843 standleyi Croat & R. A. Baker 543, 544, 548, 554, 556, 560, 561, 562, 563, 568, 569, 735- 736, 751, 761, 765, 839, 840, 843 strictum N. E. Br. 689, 690, 691, 758 шн. (Vell. Conc.) Stellf. 629 0, S x 554, 557, 766 563, 761, . superbum Madison subsp. а 590, 736-731, 761, 766, 840 А Mm subsp. brentberlinii Croat 565, 590, 737, , 761, 766, 840 taraptens Engl. 542, 555, 557, 563, 565, 568, , 584, 591, 593, 609, 675, 737-739, 758, M 166, 841 — Kia 564, 568, 571, 587, 591, 739, 761, PE aes Engl. 739, 740, 741, 758 a a E EN . uleanum var. na А. tessmannii K. ee 342, 739, 741 A. tetragonum A. tetragonum Ho ens ex Schott 717, 720, 758 A. tikalense Lundell 2 A. tricarinatum Sodiro 542, 756, 758 A. trinitatis Engl. 662, 758 A. ulea Engl. 542, 557, 562, 564, 566, 568, 569, 571, 644, 647, 681, 691, 739, 740, 741, 758, 761, 766 uleanum Engl. var. шеапит 554, 557, 584, 586, 590, 591, 592, 739-742, 743, 761, 766, 842 nayense Croat 551, 555, 557, 565, 582, 585, 586, 591, 592, 740, 742-743, 761, 766, 842 . umbrosum Liebm. 544, 758 upalaense Croat & R. A. Baker 543, 547, 552, 554, 556, 560, 562, 563, 568, 570, 576, 577, 626, 711, 743-745, 761, 766, 843, 844 valerii Standley 711 . validifolium K. Krause 552, 554, 556, 560, 563, 568, 570, 576, 745-746, 761, 766, 843, 844, , А. vaupesianum Croat 551, 555, 557, 564, 569, 571, 582, 590, 601, 632, 676, 746–747, 761, 766, 845 А. И Griseb. 544, 554, 556, 567, 568, 572, 574, 47-748, 761, 766, 845, 850 А. ou Standley & L. O. Williams 711 А. vinillense Bunting 543, 566, 567, 569, 570, 578, 589, 593, 748–749, 761, 766, 851 А. wagenerianum K. Koch & Bouché 541, 554, 556, 559, 566, 567, 568, 591, 593, 633, 749-750, 757, 758, 761, 766, 847 А. watermaliense hort. ex Bailey & Nash 548, 560, 568, 514, = 736, 750, por 766, 846 А. w eberba ri Engl. 542, A. wilifordi Croat 547, 554, 557, 565, 569, 571, 580, -753, 761, 766, 848 A. p ПА 542, 554, 558, 566, 567, 569, 570, 580, 582, 662, 753, 761, 766, 848 А. xanthoneurum Bunting 543, 549, 566, 567, 569, 570, 593, 610, 748, 753-754, 761, 766, 848 А. yutajense Bunting 544, 680 Eulaema bomboides 695 Lagenandra Dalz. 539, 544, 548 Mauritia L. 647 Pedicellarum Hotta 541 Philodendron Schott 547 Pothoidium Schott 541 Pothos L. 541, 755 Pothos acaulis Jac cq. Pothos aduncus Vell. Conc Pothos coriacea Graham non m 628 Pothos crassinervia Jacquin 541, 632, 634, 755 , 635 Pothos crenata Linn . 541, 572 Pothos glaucus Schott ex Kunth non Wall. nec Link & Otto ex Steud. 629 Pothos macrophylla pi TAA Pothos maxima Desf. 541, 755 Pothos solitarius Vell. о 541, 728 Pothos subcaulescens Vell. Conc. 629 Appendix 1: Schott's Treatment of Pachyneurium .......... эш 757 Appendix 2: Engler's Treatment of Pachyneurium . 158 Appendix 3: Accepted Pachyneurium Names and Their Treatment by Schott and ата 25199 Appendix 4: Pachyneurium Geographical Checklist | 762 Illustrations Ё 767 Index z= 1 852 dx or Announcing the publication of the latest Thesaurus Dracularum Ry A Lon sid Rode ыды: 8. repa ү бе Саша wanslations by Fite Bar t | acularum is a popular п monograph of the genus Dracula to the mountains млина Dr America, In this Seri јез, each of the 90-odd species will be illustrated by a watercolor painting and accompanied by бам S, discussions and distributions in both English and German. Distribution maps and black and white las ама. are also provided. In Ln eu: format. : coms orders. av 'ailable. The x cipal | six poles, ee $41 outside TM A CUR. S MS оша US) | ————— "Ul tu — ve an п order, bead d waaay “onder i in US. funds, T through а U. a US. nm nl d r tsa id i S No. ar 1 СОМТЕМТ5 А Revision of RO Section Pachyneurium (Araceae) Thomas B. Croat -aaa E | p Abstract | | 539 | * . Methods and Manis | го 540 — Acknowledgments : “History of Section Pockyneuxitun Sectional Relationships within Anthurium ` Morphology of томах Structures `. Growth Patterns Ae Stems FOE Y Roots Cataphylls .. _ Morphology of Reproductive aeons у Inflorescences . _ Peduncles _ Spathes. E Зе е ls BAT } AE b cona a! (кең Structures san € Species Descriptions 3. Excluded Species ... а Cited . Annals of the паш. Pe Volume 78 \ Number 4 X^ Volume 78, Number 4 Annals of the Fall 4994 Missouri Botanical Garden | The Annals, published quarterly, contains papers, primarily in systematic botany, соп. - tributed from the Missouri Botanical Garden, St. Louis. Papers originating outside the | Garden will also be accepted. Authors should write the Editor for information concerning arrangements for publishing in the ANNALS. Instructions to Authors are printed in the back | of the last issue of each volume. | Editorial Committee Marshall R. Crosby Gerrit Davidse Editor, Missouri Botanical Garden | Missouri Botanical Garden | , колде John D. Dwyer | Amy Scheuler Missouri Botanical Garden & | Managing Editor, Saint Louis University Missouri Botanical Garden Prec M Ss Peter Goldblatt Diana Gunter Missouri Botanical Garden Editorial Assistant, Missouri Botanical Garden Dale E. Johnson | : Missouri Botanical Garden Magdalen Lampe | Cirri Moran . Henk van der Werff | Publications Staff Missouri Botanical Garden | А For subscription information contact Department The ANNALS OF THE MISSOURI BOTANICAL al | Eleven, Р.О. Box 299, St. Louis, МО 63166-0299. (ISSN 0026-6493) is published quarterly 7 8 | Subscription price is $75 per volume U.S., $80 Missouri Botanical Garden, 2345 Tower Стоу е | Canada and Mexico, $100 all other countries. Four enue, St. Louis, MO 63110. Second class po | | | | is ili es. issues per volume. Beginning in 1991, recipients of _ paid at St. Louis, MO and additional von ap Dep artment 166-0299. the ANNALS will automatically receive Novon, the PosrMASTER: Send address changes 10 Garden's newly established scientific quarterly. THE Missouri BOTANICAL GARDEN, Eleven, Р.О. Box 299, St. Louis, MO 63 (O Missouri Botanical Garden 1991 О THIS PUBLICATION 1$ PRINTED ON ACID-FREE PAPER. +. ye Volume 78 Annals Number 4 of the Y 1991 Missouri Botanical Garden MONOGRAPH OF THE Robbin C. Moran? NEOTROPICAL FERN GENUS STIGMATOPTERIS (DRYOPTERIDACEAE)! ABSTRACT Stigmatopteris contains 24 species with no varieties or subspecies. It i is distinguished from all other genera of ther Е shared by а! or nearly all species of the genus but also found in other dryopteroid ferns are nonindusiate sori, thin-textured leaves, glabrous laminar tissue, acuminate-serrate pinna, apices, thin, flaccid, tan or orangish scales, erect or short-creeping stems, and minutely pubescent adaxial grooves of the axes. Within the Dryopteridaceae, Stigmatopteris forms a clade with ea eie Olfersia, and Polybotrya, held together by the synapomorphy of pubescent pee xial grooves of the a The genus is entirely neotropical and primarily montane, ranging from the Antilles and southern Mexico to о. Brazil pe conspicuously absent from most of the Amazon | . The Andes are the cen a of diversity for the genus, harboring eleven species, two of which are endemic. The tal mountains of southeastern Brazil are another center, with six species, all of which are endemic. The Greater Antilles have also been important in the diversification of Stigmatopteris because they contain three species, all of which are endemic. All members of the genus grow in wet forests, generally from 200 to 2,000 m RESUMEN Stigmatopteris contiene 24 especies sin subespecies ni variedades. Es distinguido de todos los otros géneros de las dryopteridáceas (sensu stricto) por las numerosas glándulas punctatas, escamas con dientes o cilias numerosas, células glandulares terminando los ápices de los dientes o cilias, y ápices clavatas de las venas que terminan atrás del margen de la lámina. Otras características compartidas por todas o casi todas las especies en el género, pero encontrado ! I thank Alan R. Smith (UC) for helpful discussions on all aspects of je work and Robert С. Stolze еј S first pointing out to me that Stigmatopteris needed a modern о а M. Tryon (then at СН, now SF reviewed the nomenclatural parts of the species treatments and suggested заа ai orge У. Crisci (LP) id assistance in running the HENNIG86 program used in the cladistic мети of the species, and шы Еогего (МО and Warren D. Stevens (MO) facilitated funding for annotation labels and scanning electron microscopy. Gerrit Davidse (MO) and Henk van der Werff (MO) also provided support and encouragement for this project. I thank the curators of the following herbaria for making their specimens available for study: AAU, M, CM, COL, CR, DUKE, F, C, GH, K, L, LIL, MICH, MU, NY, P, PH, PORT, QCNE, SI, UC, 05, 7. 2 Missouri Botanical Garden, Р.О. Box 299, St. Louis, Missouri 63166, A ANN. MISSOURI Вот. GARD. 78: 857-914. 1991. 858 Annals of the Missouri Botanical Garden también en otros helechos dryopteroides, son los soros no indusiados, hojas membranáceas, tejido de la lamina glabro, Sitematopteris hace xiales ntanas costales cas. Las endémicas. Las ree illas han sido importante también en la diversificación de Stigmatopteris, con tres especies, todas de las cuales son endémicas. Todas las especies en el género existen en bosques hümedos, generalmente de 200 a 2,000 m TAXONOMIC HISTORY Stigmatopteris was first proposed by Christen- sen in 1909. Until then, all of its species had been placed in Dryopteris, a genus that comprised a vast, unrelated assemblage of nearly 1,000 species. Four years later, in the first volume of his Dryop- teris monograph, Christensen (1913) changed his mind about the rank of Stigmatopteris, placing it back in Dryopteris as a subgenus, but with the proviso that it could be treated, along with most of his other subgenera, as a distinct genus. Now- adays, pteridologists follow Christensen’s original placement, giving Stigmatopteris full generic sta- tus. Christensen’s (1913) Dryopteris monograph also widened the definition of Stigmatopteris by in- cluding three species in what he informally called the “Peltochlaena group," which differed from true Stigmatopteris by having an indusium and a thick, nonglandular lamina. This group is now con- sidered part of a separate genus, Cyclodium, which was recently monographed by Smith (1986). The species of Cyclodium previously placed in Stig- matopteris are listed in the excluded taxa section. GEOGRAPHY Stigmatopteris is entirely neotropical, extend- ing from the Antilles and the state of Veracruz, Mexico, southward to southeastern Brazil (Fig. 1). The genus is absent from the Amazon basin except along the western portion adjacent to the Andes. It is primarily a montane genus, occurring in wet forests from 200 to 2,000 m, with the highest elevation being 2,500 m recorded for 5. rotundata in Panama. Species diversity and endemism in Stigmatop- teris can be conveniently discussed using the re- gional centers of high endemism and species rich- ness delimited by Tryon (1972) for tropical American ferns (Fig. 2). The region of highest diversity and endemism for the genus is the Andes, containing eleven species and two endemics (S. ichthiosma, and S. pellucidopunctata). Two spe- cies (S. bulbifera and S. pterorachis) are near- endemics, as they also occur outside the Andean Region in the Santa Marta Mountains of northern Colombia. Within the Andean Region, Colombia and Ecuador have more species (ten) than any other country (Tables 1, 2). he coastal mountains of northern Venezuela have two species of Stigmatopteris, one of which is endemic (S. nephrodioides). Although these mountains do not harbor a large number of species, they could be considered a secondary regional cen- ter in Tryon's scheme because they have many endemic species. The region with the next highest number of species is Central American, with seven species, one of which is endemic (S. contracta). With the exception of S. contracta, all of the other species extend southward into the Andean Region, giving the two regions a high floristic affinity. Southeastern Brazil has played a major role in the diversification of Stigmatopteris because it has six species, all of which are endemic. The Greater Antilles have also been important, having three species, all of which are endemic. The Guianan and Mexican centers contain few species and thus have not been important in the diversification of Stigmatopteris. The Mexican Re- gion boasts one endemic (S. chimalapensis), but it is known only from one collection and may even- tually be found outside of the region. MORPHOLOGY Internal glands in the lamina. The glands can usually be seen with a dissecting microscope and, in oblique light, they appear as tiny raised bumps on the leaf surface ‘Fig, 3a, c). They often dry blackish. Perhaps the best way to see the glands Volume 78, Number 4 Moran 859 Stigmatopteris Monograph TROPICAL AMERICA | b NC | ~ ~ т D | xL | Mm | | | 20 7 | | | / У y | | ^a |) | \ _ \ ў 4 h FicurE 1. Distribution of Stigmatopteris. is to look through a hand lens with the lamina illuminated from behind. Viewed this way, the glands appear as pellucid dots in the lamina (Fig. 3b). The glands are more conspicuous in some species than in others. Occasionally, they are so large and abun- dant that they are visible to the naked eye, but in S. opaca, a species with a thick lamina, the glands are not visible. The genus is named for these glands: from the Greek stigmatos, punctate, and pteris, fern. Other genera of dryopteroid ferns also have internal glands, but these are extremely small and visible only by thin-sectioning and viewing under a compound microscope (Widen et al., 1983). The large glands in Stigmatopteris are probably ho- mologous with these smaller ones, but nothing is known about the glands’ structure, development, chemical contents, or role in metabolism. Scales. The scale characteristic that distin- guishes Stigmatopteris from all other genera of dryopteroid ferns is the presence of uniseriate cilia that end in a bulbous, often orange or brownish, glandular cell (Fig. 4). In other dryopteroid ferns, marginal teeth or processes are formed by the projecting end walls of two adjacent cells (Moran, 1987, fig. 7) and are never uniseriate or terminated by a glandular cell. The scales of Stigmatopteris are more uniform than those in other genera of dryopteroid ferns. Although the shape differs, all species of Stig- matopteris have scales that are thin, flaccid, and tan or light orangish. The scales decrease in size from the proximal to distal portions of the lamina until they become uniseriate or only 2-3 cells wide at the base (Figs. species of the genus, and a dissecting microscope 860 Annals of the Missouri Botanical Garden TROPICAL AMERICA о 200 400 $0 00% m \ ЭЖЕ. ый. „Жы ый ыйы \ =====—=—=== o 100 200 300 400 500 600 miles URE 2. Distribution of Stigmatopteris species within regional centers of diversity and endemism for tropical American ferns defined by Tryon (1972). Pr imary centers are indicated with dark shading, secondary center s with light shading. The number on the left is the total number of species, and the number in parentheses indicates how many of those species are endemic. is needed to see them. Their lateral processes are usually terminated by glandular cells. Veins. Unlike most dryopteroid ferns, the veins of Stigmatopteris end behind the margin in clavate apices which are best seen adaxially (Fig. 23a). All other genera of dryopteroid ferns (sensu stricto) have the veins ending at or close to the margin, and the apices are slender, not thickened. ndusia. Stigmatopteris is nonindusiate. Three species have an indusiumlike scale associated with the sorus (Figs. 17b, 18a), but these scales are certainly not homologous with the large, firm, thic textured, circular indusia found in the outgroup genus Cyclodium. Although independently derived in Stigmatopteris, the lack of an indusium is not unique because many other dryopteroid ferns also lack an indusium. Lamina shape and texture. All| species have pinnae with long-acuminate apices that are serrate to the tip. The pinnae are equilaterally symmet- rical, but some of the large species such as S. contracta and S. lechleri may have basal pinnae that are elongated basiscopically. The texture of the lamina is thin and papyraceous, except for that of S. opaca, which is thick and slightly charta- ceous. Hairs. All species of Stigmatopteris lack hairs except in the adaxial grooves of the rachis and costae. These hairs are 0.05-0.2 mm long, 1- or 2-celled, unbranched, and colorless to reddish. They are also found in the same position in Cyclodium, Olfersia, and Polybotrya and are a synapomorphy that holds these three genera and Stigmatopteris on the same clade within the Dryopteridaceae (see evolutionary relationships section). Stigmatopteris heterophlebia and S. bulbifera, however, are ex- Volume 78, Number 4 1991 Moran 861 Stigmatopteris Monograph TABLE l. Distribution of Stigmatopteris by country. Endemics are in boldface. TABLE 2. The species of Stigmatopteris and their distribution by country. Mexico: 3 species, 1 endemic. chimalapensis, longi- кане sordida. Guatemala: 2 species, no endemics. longicaudata, sordida. Honduras: 1 species, not endemic. longicaudat Nicaragua: 2 i min no endemics. heterophlebia, sordida. Costa Rica: ecies, no endemics. contracta, hetero- phlebia, PA longicaudata, sordida. Panama: 6 species, no endemics. contracta, heterophle- bia, killipiana, longicaudata, michaelis, sordida. Cuba: 1 species, 1 endemic. hemiptera. Jamaica: 2 species, 1 endemic. gemmipara, jamaicensis. Наш: 1 species, not endemic. gemmipara. Guadeloupe: 1 species, not endemic. rotundata. Dominica: 1 species, not endemic. rotundata Martinique: 1 species, not endemic. rotundata. St. Lucia: 1 species, not endemic. rotundata St. Vincent: 1 species, not endemic. rotundata. Grenada: 1 species, not endemic. rotundata. Trinidad: 1 species, not endemic. rotundata. French Guiana: 1 species, not endemic. rotundata. Guyana: 1 species, not endemic. longicaudata Venezuela: 5 species, 1 endemic. lechleri, longicaudata, nephrodioides, pterorhachis, rotundata Colombia: 10 species, no endemics. bulbifera, hetero- phlebia, ichthiosma, killipiana, lechleri, longicaudata, michaelis, pterorhachis, pellucidopunctata, sordida. Ecuador: 10 species, no endemics. bulbifera, heterophle- bia, ichthiosma, killipiana, lechleri, longicaudata, mi- chaelis, opaca, pterorhachis, pellucidopunctata, sordi- da. Peru: 5 species, no endemics. n UR lechleri, lon- audata, cup pellucidopunctat Bo livia: 2 species, no endemics. Е pellucido- punctata. Brazil (southeastern): 6 species, 6 endemics. breviner- vis, caudata, heterocarpa, prionites, tyucana, ulei. ceptions to the general condition of hairlessness in the genus because their axes are covered abaxially with capitate-glandular hairs. Sporangia. The sporangia are like those de- scribed by Moran (1987) for Polybotrya, except that the stalks lack a hairlike paraphysis. Petiole anatomy. cross section of the pet- iole shows that the vascular bundles are arranged in a mushroomlike outline, with the base of the mushroom oriented adaxially (Fig. 12a). This is typical of dryopteroid ferns (Moran, 1986, 1987; pers. obs.). Aerophores run down the sides of the petiole and are evident in cross section by a gap in the thickened peripheral collenchyma of the petiole. The petiole bases do not persist or accu- mulate starch after the rest of the leaf has withered. In other words, Stigmatopteris lacks trophopods. 1. S. brevinervis (Fee) R. C. Moran: southeastern Bra- zil. 2. S. bulbifera R. C. Moran: Colombia, Ecuador. 3. S. caudata (Raddi) C. Chr.: southeastern Brazil. 4. S. chimalapensis Mickel & Beitel: southern Mexico. 5. S. contracta (Christ) C. Chr.: Costa Rica, Panama. 6. S. gemmipara C. Chr.: Jamaica, Haiti. 7. S. hemiptera (Maxon) C. Chr.: Cub 8. S. heterocarpa (Fée) Rosenstock: southeastern Bra- zil. 9. S. heterophlebia (Baker) R. C. Moran: Nicaragua, Costa Rica, Panama, Colombia, Ecuador, Peru. 10. S. ichthiosma (Sodiro) C. Chr.: western Ecuador. 11. S. jamaicensis (Desv.) Proctor: Jamaica. 13. 8, ode Lellinger: Costa Rica, Panama, Colom- bia, Ecuador. 13. S. Мея (Мен.) С. Chr.: Colombia, Ecuador 14. S. longicaudata (Liebm) C. Chr.: Mexico, Guate- mala, Honduras, Costa Rica, Panama, French Gui- ana, Venezuela, Colombia, Ecuador, Peru, Bolivia. 15. S. michaelis (Baker) C. Chr.: Panama, Colombia, Ecuador. 16. S. nephrodioides (Klotzsch) C. Chr.: een 17. S. opaca (Baker) C. Chr.: Ecuador, Per 18. S. pellucidopunctata (C. Chr.) C. Chr.: Ecuador, Peru, Bolivia. 19. S. prionites (Kunze) C. Chr.: southeastern Brazil. 20. S. pterorhachis R. C. Moran: Venezuela, Colombia. 21. S. rotundata (Willd.) C. Chr.: Guadeloupe, Domini- ca, Martinique, St. Lucia, St. Vincent, Grenada, Trin- idad, French Guiana, Guyana, Venezuela. 22. S. sordida (Maxon) C. Chr.: Mexico, Guatemala, Nicaragua, Costa Rica, Panama, Colombia, Ecuador. 23. S. tyucana (Raddi) C. Chr.: southeastern Brazil. 24. S. ulei (Christ) Sehnem: southeastern Brazil. Costa Rica, Venezuela, Puppe Spores. I examined the spores from 11 spe- cies of Stigmatopteris with the scanning electron microscope (SEM). All of the spores were typical of those found elsewhere in the Dryopteridaceae, being monolete, having a smooth exospore, and showing a broadly inflated or narrowly folded peri- spore (Fig. 5). The inner layer of the perispore is thin and usually adheres to the smooth exospore; the outer layer is thicker and compact. The perisporal surface on and between the folds was nearly uniform in all 11 species examined (Fig. ). The surface is relatively unornamented, being either smooth (Fig. 5f) or beset with low, diffuse ridges (Fig. 5c, d). Because of the uniformity, the spores do not provide helpful characters in the subgeneric taxonomy of the genus. Ornamentation of the perispore differs in Stig- matopteris from that of the most closely related Annals of the Missouri Botanical Garden sl FIGURE 3. Internal punctate glands in the lamina of Phe i cn (a-c) and rachis-costa juncture showing adaxial awg pubescent with short hairs (d). White bars indicate 1 m a, b. S. bulbifera (Aligaard et al. 35866, AU).—c. S. contracta (Moran 3330, МО). —а. S. heterocarpa (Haerchen 106, B). Volume 78, Number 4 863 1991 Moran Stigmatopteris Monograph шш УХ MN ү} y SE Doni ~ ИЛА. ` (52 CLs ELSA SA 25 SS S N Y 23 222 2: SÍ Я g Т2 7 2 CFLS == Ota WES ERA r2 72 Ó IS 4224 222 == LA = у а 2 у ДЕ 2 1 77) WN, 2, у 1/9 ATA LIES => <<) [7 M У SA) AST Y) ММ), NU SSH к sr, A LT 4) 99/2 4. ер Y Y P, 2 5 Ss EN M M М P А А СА, у n : = ACE AS R.C. MORAN 1979 FIGURE 4. Scales of Stigmatopteris. The scale shown in b is from a petiole base; all others are from the abaxial surface of the lamina and show the transition from uniseriate scales (proscales or microscales) to multiseriate ones. АП are the same magnification as shown in d, except b and the far right scale in а. —а. S. pellucidopunctata (Moran 3624, МО). —b. S. ichthiosma (Moran 3545, MO).—c. S. sordida (Gómez 18641, СК). —а. S. lechleri (Moran 3582, МО). —е. S. opaca (Moran 3576, MO).—f. S. sordida (Moran 3315, MO). Annals of the Missouri Botanical Garden Volume 78, Number 4 1991 Moran 865 Stigmatopteris Monograph dryopteroid genera, i.e., Polybotrya, Cyclodium, and Olfersia. These genera have more ornamen- tation on the perispore. For example, the spores of Polybotrya are typically densely echinate (Mo- ran, 1987; Tryon & Tryon, 1982); those of Cy- clodium are usually finely papillate but can be variously ornamented (Smith, 1986; Tryon & Tryon, 1982); and those of Olfersia are sparsely echinate (Moran, 1986; Tryon & Tryon, 1982). Although these types of spore ornamentation are distinct, it is premature to determine which type is ancestral and which is derived. Such a deter- mination must await a comprehensive comparative study of perisporal ornamentation in the Dryop- навео, Therefore, І have not used spore ог- namentation in the cladogram for Stigmatopteris and related genera (Fig. 6 CHROMOSOME NUMBER Walker (1966), who has made the only chro- mosome count for Stigmatopteris, found that n — 41 in S. jamaicensis (reported as 5. nothochlaena, a synonym). This number relates the genus to the dryopteroid ferns. Walker also noted that meiosis occurs very early in Stigmatopteris during the crozier stage of leaf development. Unfortunately, most workers (in- cluding myself) have collected samples for chro- mosome counts past the crozier stage, at a stage more typical for ferns in general, i.e., when the sori are small and white, but on expanded leaves beyond the crozier stage. This probably explains why only one count exists for the genus. EvoLUTIONARY RELATIONSHIPS Family relationships. That Stigmatopteris is a member of the Dryopteridaceae (sensu stricto of Holttum, 1984) is clearly shown by several char- acteristics which it shares with all genera of that amily: x — 41, more than three vascular bundles in the petiole and these arranged in an omega shape when viewed in cross section (Fig. 12a), and grooves on the adaxial surface of the axes that are decur- rent into those of the next lower order. The Dryop- teridaceae are closely related to the Tectariaceae (sensu stricto of Holttum, 1984), but Stigmatop- teris differs from the latter family by lacking Cteni- CYCLODIUM STIGMATOPTERIS OLFERSIA POLYBOTRYA FIGURE 6. Cladogram of Cyclodium, Stigmatopteris, Olfersia, and Polybotrya. The numbered black bars rep- resent the derived character states given in Table 4. The open bar below Olfersia represents a character state re- versal. See Moran (1986) for the additional apomorphies and synapomorphies pertaining to Olfersia and Polybo- trya. tis hairs (i.e., small catenate, several-celled hairs with the color aggregated in the cross walls), and having sulcate, decurrent axial grooves. Generic relationships. Stigmatopteris most closely resembles Cyclodium (sensu Smith, 1986), another dryopteroid genus, as shown by similarities in leaf cutting and venation and by the laminae, which are hairless except for the grooves of the rachis and costae. In fact, the two genera are so similar in these respects that some pteridologists (e.g., Christensen, 1913, 1920; Tryon & Tryon, 1982) have regarded them as one. They are, how- ever, best kept as distinct. The two genera exhibit 11 morphological differences (Table 3), and this is more than that which occurs between other pairs of dryopteroid genera, such as Arachniodes and Dryopteris, Arachniodes and Maxonia, or Cyr- tomium and Polystichum. If there is to be con- sistency in recognizing dryopteroid genera, at least as regards the number of characters that separate them, then Cyclodium and Stigmatopteris should be maintained as distinct. In the Dryopteridaceae, Stigmatopteris and Cy- clodium form a clade with Polybotrya and Olfersia (Fig. 6, Table 4). This clade is defined by the presence of hairs in the adaxial grooves of the axes Fig. 3d). This character state is considered syn- apomorphic because all other dryopteroid ferns (sensu stricto) are hairless in the adaxial grooves of the axes. This character state would be weak if ~ FIGURE 5. 5046, US).—d. S. tyucana (Brade 5892, МУ). —е. S. opaca (ВИваага 3315, MO).—g. S. rotundata (Cremers 9140, UC) as in Spores of Stigmatopteris. —a, c. S. longicaudata (Mickel 6392, NY).—b. S. lechleri (von Sneidern et al. 35859, AAU).—f. S. sordida (Moran —h. S. caudata (Webb s.n., NY). Spores c-h the same scale 866 Annals Чава BL Garden TABLE 3. Comparison of Stigmatopteris and Cyclodium. Hairs of the lamina: е (except in grooves of the rachis and costae). achis and costae: 0.05-0.2 m Lamina cutting: similar à in many species, cf. S. rotundata and C. Hairs in grooves of the SIMILARITIES mm long, usually blunt-tipped, е опе- [0 few- celled. guianense, ог S. o and C. meniscioides. Venation: similar in free-veined species such as S. nephrodioides and C. pi. d in а ннн НЕ species such as 5. heterophlebia and C. meniscioides. Character DIFFERENCES Stigmatopteris Cyclodium Indusium: Lamina texture: Pinnule Mobi NS Position of vein ending: Shape of vein tips: Internal laminar glands: Scale frequency on axes: Scale color Scale texture: Marginal teeth of scales: Geog Gland at the apex of scale teeth: raphy: absent papyra anadromous basally, catadromous behind margin clavate present common light brown or orangish thin uniseriate present primarily Andean resent chartaceous or coriaceous anadromous throughout at margin slender absent rare or lacking dark brown thick composed of upturned margins from two adjacent cell walls absent primarily Guianan ig: else correlated with it; it could have evolved the clade would be Sx ае The м = that this character evolved only once is supported by the fact that the type of pubescence is the same in all the genera: the hairs are 0.05-0.2 mm long, usually blunt- tipped, colorless, and one- to few-celled (Smith, fig. 10b-e, в, h). Olfersia cervina, the only species of Olfersia, is an excep- tion, having glabrous grooves. Because Olfersia is the sister genus to Polybotrya (Moran, 1986), its lack of hairs in the grooves is best interpreted as a character state reversal (Fig. 6). Some species than once, and i 86; Moran, 1987, of Polybotrya subg. Polybotrya, such as P. os- mundacea an hairs in the grooves (Moran, 1987, fig. 49c) than those found in the grooves of Cyclodium and Stig- matopteris. These longer hairs, however, are oth- erwise similar to the smaller ones and are probably derived from them The clade with Stigmatopteris, Olfersia, and Polybotrya is defined only by one synapomorphy: loss of the indusium (Fig. 6, Table 4). This synapo- morphy is weak and therefore considered tentative because no others correlate with it. Furthermore, the number of times the indusium has been lost in theoides, have much longer TABLE 4. Characters and character states used in the cladistic analysis of Stigmatopteris, Cyclodium, Olfersia, and Polybotyra (Fig. 6). Character state polarities were determined by comparison to other genera in the Dryopteridace- ae, such as Arachniodes, Dryopteris, and Polystichum. Character Ancestral condition Derived condition Pubescence i in adaxial grooves of the axes absent 1. 2. Indus 3. го pam in the lamina 4 . Marginal teeth ог processes of the scales present 5. Glandular cell at apex of scale teeth or absent processes 6. Vein termination at margin 7. Shape of vein tip slender 8. Texture of lamina papyraceous or chartaceous composed of the upturned margins from two adjacent cells present present uniseriate present behind margin clavate membranaceous Volume 78, Number 4 1991 Moran 867 Stigmatopteris Monograph TABLE 5. Data matrix used in the cladistic analysis of 23 species of Stigmatopteris, with Cyclodium used as the outgroup. See Table 6 for the meaning of characters and character states. Characters — Тахоп 5 Cyclodium pterorhachis bulbifera heterophlebia killipiana aca longicaudata heterocarpa rotundata prionites nephrodioides michaelis tyucana chimalapensis jamaicensis hemiptera ichthiosma sordida pellucidopunctata lechleri oo0ororoooroo0oo00o0000000000 o0o0o0oo0o0oo0o0orrOoo0o0o0o00o00o00o00o000 ت ي‎ $ ъ ож жж с +н+:]++-Е|Е | Е < ш tn Co Go бл (o л (n (n (n (n (o Co (o Cn Co Co Co Co Co Co Co Co © ooooooooooooroo0ooroorrrol]» © to № — к — М > мо ке фо دت‎ м دن دن دن‎ > > > > > > > © cOocooco—-ocoo—-ooooooooooooooo|o OoOoococoocoooooooooooooo-—--—oo|a O—-—-—-c—-ococooooooooooooooooc!|oc OoOoocoocoocooocoooooococoo-----o|o fern evolution suggests that it could have easily been lost independently in Stigmatopteris and in the Olfersia- Polybotrya clade. Species relationships. А cladogram of species relationships was dap using the HENNIG86 5 (Farris, 1988; Platnick, plying the implicit enumeration option for calcu- lating trees. Table 5 shows the data matrix used in the analysis, and Table 6 lists the characters and character states used in the data matrix. Stig- matopteris ulei was not included in the data matrix because it appears to be of hybrid origin. Another feature of HENNIG86 was also used: the succes- sive weighting procedure, which calculates weights from the best fits to the most parsimonious trees using rescaled consistencies (rc), which are the products of the character consistency (c) and the character retention index (r). The product is scaled to lie in the range 0-10 (Table 7). The weighting procedure is repeated on successively produced trees until the trees no longer change (Farris, 1988). Cyclodium was used as the outgroup to deter- mine character state polarities in Table 6 (see sec- tion on generic relationships for justification of us- ing Cyclodium as the outgroup). All species of Cyclodium have the ancestral character states giv- en in Table 4 for characters 1, 2, 4, 6, 7, and 8; therefore, it was simple to determine polarity. How- ever, because Cyclodium has both character states for characters 3, 5, and 9 (Table 6), it could not be used to determine character state polarity. For ABLE 6. Characters and character states used in the cladistic analysis of 23 species of Stigmatopteris (the probable hybrid S. ulei not included). The ancestral state equals 0; the derived state (s) equals 1 (or more). The hypothesized evolutionary pathways for all the character states are unilinear. l. i scale: absent n Laine (1). 2. Buds: absent (0); 1 — nt (1). 3. Veins: psa (0); AR otia (1). 4. Distal pinnae bases: short-decurrent (0); long-decur- rent 5. Lamina dissection: 2-pinnate-pinnatifid (0); nearly 2-pinnate (1); pinnae cut 14-54 to the costa (2); pinnae cut 14-34 to the costa (3); entire or nearly so (4). 6. Basal basiscopic lobe of the distal pinnae: adnate to costa (0); adnate to the rachis (1) 7. Hairs on rachis and costa: absent (0); present (1). Petiole scale texture: thin, flaccid (0); thick, rigid (1). 9. Number of pinna pairs: 12-18 (0); 1-12 (1). со 868 Annals of the Missouri Botanical Garden TABLE. de „ахаа попе sakes (c) a retention t fit arsimonious trees roi HENNIG86, which was "ва ји calculate weights. Final weights were often the third round of suc- cessive weighting procedures in HENNIG86. Weights were truncated to integers Con sist- Reten- Range Num- епсу tion Weight Char- of Бег of index index (re x Final acter steps steps (c) (r) 10) weight l 1 1 1 1 10 10 2 1 2 0.5 0.5 2 3 1 1 1 1 10 10 4 1 3 0.3 0.5 | 5 4 4 1 1 10 10 6 1 2 0.5 0 0 7 1 1 1 1 10 10 8 | 2 0.5 0.7 3 9 1 1 1 1 10 10 these three characters, polarity was determined by deferring to Holttum's (1968) hypotheses about the general trends in fern evolution. This method is not entirely satisfactory, but in this case it is the best that could be done with present knowledge. Regarding character 3 (free vs. anastomosing veins), Holttum (1968) believed that the evolu- tionary trend has been from free to anastomosing. Therefore, free veins were assigned as the ancestral character state (Table 6). He also believed that the earliest ferns had highly divided leaves and that the evolutionary trend has been toward less divided leaves. Accordingly, character 5 (leaf division) has the most divided leaf as the ancestral state and the least divided leaf (1-pinnate with entire pinnae) the derived state. Finally, for character 9, Holttum believed that the evolutionary trend has been to develop fewer pinnae; therefore, the smaller num- ber of pinnae was considered as the derived state. HENNIG86 found three equally most-parsi- monious cladograms (Fig. 7), which differed only in the branching patterns for seven of the most derived species—a group that consists primarily of species with anastomosing veins and entire to shallowly lobed pinnae. All of the trees had 17 steps and a consistency index of 0.7. Cladogram 2 (Fig. 7) has two character state sequences. The one shown involves a reversal of character 4 at the node that defines 5. opaca and S. killipiana. The other sequence assumes a parallelism of char- acter state 4.1 at the line leading to 5. longicau- data and at the node which defines S. pterorha- chis, S. bulbifera, and S. heterophlebia. I do not believe that one of these sequences is more likely than the other; both seem equally probable. At the beginning of this study, I thought that species from different regions might form mono- phyletic groups, such as the six species from south- eastern Brazil, or the three species from the Greater Antilles, or the eleven species from the Andes. Although two of the species endemic to the Greater Antilles proved to be sister species, no general correlation exists between the cladogram and the geography of the other species. In other words, no synapomorphy exists that holds together the spe- cies from these regions. The species with anastomosing veins and entire to shallowly lobed pinnae are distinctive and form a monophyletic group (Fig. 7, trees 1, 2, 3). They could be considered a subgenus, much the same way that Soromanes is considered a subgenus of Polybotrya (Moran, 1987). However, S. longi- caudata is intermediate between the group and the rest of the genus, and so I have not formally named the group. Key and Descriptions Stigmatopteris C. Chr., Bot. Tidsskr. 29: 292. 1909. TYPE: Stigmatopteris rotundata (Willd.) C. Chr. Dryopteris subg. пм ји Chr.) C. Chr., Kon- gel. Danske Vidensk. Selsk ., Naturvidensk. Afd. dn 7, 10: 73. 1913 Plants terrestrial; rhizome short-creeping, scaly at the apex, the meristeles surrounded by whitish parenchyma (not darkened sclerenchymatous sheaths); petioles with 4 to many vascular bundles in cross section (Fig. 12a), continuous (not artic- ulate) with the rhizome; lamina 1-ріппаїе to 2-pinnate-pinnatifid, the apex pinnatifid, the tissue glabrous on both surfaces, with punctate glands in the mesophyll (Fig. 3a-c); pinnae linear to паг- rowly oblong, often with a decurrent basal wing, the apices attenuate and serrate; pinnae or seg- ments arranged anadromically in the proximal pin- na pairs, becoming catadromically arranged in the distal pinna pairs; adaxial surface of the rachis and costae sulcate, the sulci decurrent into those of the next lower order, pubescent adaxially within the grooves (Fig. 3d), the hairs 0.05-0.2 mm, 1- or 2-celled, colorless, usually blunt-tipped; abaxial surface of the rachis and costae stramineous or occasionally brown, lacking hairs abaxially or (in two species) glandular-pubescent, scaly, the scales thin, translucent, light brown to orangish, ciliate or denticulate, the cilia or teeth terminating in a Volume 78, Number 4 1991 Moran 869 Stigmatopteris Monograph heterophlebia killipiana с ~ Ф = = 2 5 [7] = o c = ~ о ~ с Ф Ld = $ с = о с 2 с 7(1) с = o a о = о = c x 2 4(1) longicaudata heterocarpa heterocarpa S а 2 = за 2 oC s vo 5 Ф 2 - d 45. = 3 o 2 o a = с = с oO ~ a c = Ф о $ © : = 5 а © ~ 2 2 - = © Ф = 5 Ф = а. - о а c x о a с = 7 (1) = E = o с q о а 7(1) 3 E 4(0) 9 | 4(1) o с 9 heterocarpa nephrodioides chimalapensis tyucana + 5(4) — ли — с |] o c э s Е o c a o 2 о е © б с 9 s ? © = а 2555 o - = E 5 3 <= c 2 Ба EE - = 22 Ф о 2... >. a? a o = = о o О = а = 3 5 | Е о 3 2 E an S 5 | 5 o Е о 5 = 1(1) 5(2) ° = | © o о 2 5(1) 5(3) FIGURE 7. Cladogram of species relationships in Stigmatopteris. Cladograms 1, 2, and 3 attach to dotted far right of the bottom tree. Cross bars indicate apomorphies; the X in cladograms 2 and 3 indicate reversals. darkened, slightly enlarged, glandular cell (Fig. 4); veins either free or inconsistently anastomosing (i.e., mostly anastomosing but with a readily no- ticeable number of free veins), the tips clavate adaxially (hydathodous?), terminating before the margin; sori round or slightly oblong near the cos- tae; indusia absent; sporangial stalks not paraphy- sate; spores monolete, with broadly folded perispore and low, broken ridges (Fig. 5). Chromosome num- ber, x = 41 870 Annals of the Missouri Botanical Garden Кеу TO THE SPECIES OF STIGMATOPTERIS la. d to the species of the Antilles and southeastern Brazil. . Species of the Antilles. 3a. Pinnae crenate or cut e than % to de costa; Lesser Antilles ..... 21. S. rotundata (Willd.) C. Chr. ЗЬ. Pinnae cut 4-% to t sta; Greater Antilles 4a. Pinnae with the Ша o "lobe arius from the rachis e 14а); Cuba cn EE . S. hemiptera (Maxon) С. Chr. 4b. Pinnae PN the basal basiscopic lobe arising from the costa; DR Haiti. 5a. Lamina lacking buds; pinnae 15-23 cm, pairs 22-26; veins 8-11 per segment; sori covere ed by an indusiumlike iub petiole ca. half the length of the lamina; Jamaica ..... UN 11. S. jamaicensis (Desv.) Proctor 5b. Lamina bearing buds; pinnae (4-)9-14 cm, pairs 9-20; veins 5-7(-8) per segment; sori lacking an indusiumlike scale; petiole equaling the length of the SERA Jamaica, Haiti 6. S. gemmipara C. Chr. 2b. Species of parri. Brazil. Pinnae cut У or less to the costa; veins 3-7 per lobe. Та. Pinnae entire to lobed less re V4 to the costa „u 8. S. heterocarpa (Fée) Rosenstock 7b. d lobed 44-14 to the . Sori inframedial; fertile xm 1.1-2.3 ст wide, the sinuses broad, U-shaped .................. 19. S. pui (Kunze) C. Chr. 8b. Sori medial to supramedial; fertile pinnae 2.5-4 cm wide, the sinuses narrow, V-shape 23. S. tyucana (Raddi) C. Chr. 6b. Pinnae cut 14-74 to the costa; veins 5-15 per lobe. 9a. Pinnae cut 24-76 to the costa, the distal ones often with the basal basiscopic lobe adnate and ca. perpendicular to the rachis; (Fig. 10b); veins 9-15 per lobe . 3. S. caudata (Raddi) C. Chr. 9b. Pinnae cut 4-24 to the costa, the distal ones lacking a basal basiscopic lobe adnate to the 10a. Sinuses narrow, V-shaped .......... l. S. i ue (Fée) R. C. Mora Sinuses broad, U-shaped . S. ulei (Christ) Shine lb. Key to all species (including those of the Antilles and southeastern Brazil). lla. Veins anastomosin l2a. Rachis and costae puberulent abaxially. l e pinna pairs 7-10; buds often present in the axils of the = эе distal pinnae these best seen adaxially) 2. S. bulbifera R. C. Moran 13b. Free pinna pairs 1-3(-4); buds absent or (in Mut rarely pres . 5. ennt (Baker) R. C. Moran 12b. Rachis and costae lacking hairs abaxially. 4a. Petiole abaxially atropurpureous, lustrous .... . 12. S. killipiana Lellinger 14b. Petiole abaxially brown, tan, or stramineous, dull. Lamina thin, internal punctate glands easily visible pe magnification); sori Tund: areoles of sterile leaves with 1 or 2 irregular, curved o excurrent veinlets western Venezuela, northern Colombia ل‎ 20. terorhachis R. Moran 15b. Lamina thick, the internal punctate glands not visible (with не round or oblong with a strong tendency to become arcuate; areoles of st ded Ls with 1 straight (or nearly so) excurrent veinlet (Fig. B. е — ELEAN . S. opaca |: (Baker) C. Chr. llb. Veins free or with a few casual anastomoses 16a. Pinnae entire, crenate, or cut less than l4 to the costa . Pinnae confluent in the distal 14-14 of the lamina by a conspicuous, long-decurrent wing (Fig. 21b) 14. S. longicaudata (Liebm.) C. Chr. 17b. Pinnae confluent in the distal 14-14 of the lamina, the a short- — rent. 18a. Pinna pairs 18-32; Lesser Antilles, Guianas, and eastern Venezuela .................................. 1. 5. али (Willd.) C. Chr. 18b. Pinna pairs 12-17; ا‎ a Brazil .......................... 8. S. heterocarpa (Fée) Rosenstock 16b. Pinnae cut % to all the way to the 19a. Pinnae confluent in s distal va 6 of the lamina by a conspicuous, long-decurrent wing. 20a. Pinnae (4.5-)9-11 cm; leaves 0.3-0.7 m; sori pare with multicellular, ark brown hairs S. michaelis (Baker) C. Chr. 20b. Pinnae 14-20 cm; leaves 0.8-1.4 m; sori not рату or (in South America) with filiform scales among the sporangia ............. S. lo ngicaudata (Liebm.) C. Chr. 19b. Pinnae confluent in the distal 14-144 of the uie the pd short-decur Distal pinnae with the basiscopic basal lobe adnate and ca. perpendicular to the rachis (Figs. 10b, 14a), with the costule arising from the ra 22a. || innae 4-7 ст wide; lamina lacking n veins 9- 15 рег segment; south- astern Bra 2il 3. S. caudata (Raddi) C. Chr. 22b. Pinnae 1.5-2.5 cm wide; lamina of small sterile leaves with buds in the axils Volume 78, Number 4 Moran 871 1991 Stigmatopteris Monograph of the distal pinnae; veins 6-10 per segment; Cuba (Е . hemiptera (Maxon) С. Chr. 21b. Distal pinnae with the basal basiscopic lobe к to the costa or only partly adnate to Fs rachis, the costule arising from the 23a. Lamina 2-pinnate-pinnatifid 13. S. lechleri (Mett.) C. Chr. 23b. Lamina 1-pinnate-pinnatifid to nearly 2-pin ostular scales bullate or buds Bone: in v axils of the distal pinnae; amaica, Haiti. 25a. Pinnae 15-23 cm, pairs 22-26; lamina without buds in the axils of the distal pinnae; sori covered by an indusiumlike scale ....... . Jamaicensis (Desv.) Proctor 25b. Pinnae (4-)9-14 cm, pairs 9-20; lamina with buds in the axils of the distal pinnae; sori not covered by an indusiumlike scale . gemmipara C. Chr. 24a. Costular e not bullate; buds absent in the axils of ы distal pinnae; southern Mexico to southeastern Brazil. 26a. = ofi the medial pinnae overlapping the rachis (Figs. 1 2c, e, 3a); Costa Rica, Panama, northern Venezuela. А innae 3-6 cm wide, ou *4 to nearly all the way to the costa; Costa Rica, Panama ..5. 5. contracta (Christ) si Chr. 27b. Pinnae 2-3.5 cm wide, cut 14-4 to the cost a; norther Venezuela ________ 16. S. nephrodioides (Klotzsch) с Chr. 26b. Bases of ds medial pinnae not Е Шы the rachis; souther Mexico to Bolivia, southeaster 28a. Pinnae 8-12 cm; aan Mexico 4. S. chimalapensis Mickel & Beitel 28b. Pinnae 12-40 с 29a. Pinnae 1. 1- 2. 3 ст wide; veins 3-5 per lobe; south- Am 19. S. prionites (Kunze) C. Chr. 29b. Pinnae 2-6 cm wide; veins 4-14 per lobe. 30a. o cut 4-'% to the costa; southeastern а] 23. S. гуисапа (Када) C. Chr. 30b. Pinnae cut 14-25 to the c 31a. i. subtended by an побыце scale g. 17b); western Ecuador .......................... rS 10. S. i ichthiosma (бойго) C. Chr. 3lb. Sori lacking an indusiumlike scale. 32a. Scales of the costae ње уза И lanceolate to linear, mixed with smaller кай ones; Colombia A wa к. и. (С. Chr.) C. Chr. 32b. Scales of the costae and costules ovate to lanceolate, filiform scales absent. 33a. Sori medial; ране Мех- ico to Es OR NOS "T S sordida (Maxon) . Ch 33b. Sori inframedial; south- zl. 34a. Sinuses narrow, shaped |... . 1. 5 brevinervis (Есе) . C. Moran 34b. кө broad, -shaped ________ eee 24. S. po = Chr.) 1. Stigmatopteris brevinervis (Fée) R. C. Mountains, June 1869, Glaziou 3333 pro oran, comb. nov. Phegopteris brevinervis parte (lectotype, here designated, P); Organ Fee, Crypt. Vasc. Brésil 1: 243, t. 77, fig. 2. Mountains, Glaziou 2400 pro parte (P). Fig- 1869. SYNTYPES: Brazil. Rio de Janeiro: Organ ure 8. Ma 872 Annals of the Missouri Botanical Garden 15 cm "T 7 ORA > БРУС es FIGURE 8. Stigmatopteris brevinervis. —a. Abaxial surface of a pinna (Brade 8250, UC).— b. Leaf (Dusén 725a, distal half, NY; proximal half, МО). —с. Fertile pinna (Glaziou 2400, P). Моште 78, Митбег 4 1991 Могап 873 Stigmatopteris Monograph МАР 1. pou bradei Rosenstock, Repert. Spec. Nov. g. 21: 347. 1925. TYPE: Brazil. Sào Paulo: марна, п Serrinha, Brade 8250 (holotype, not seen; isotypes, PH, UC). Rhizome short-creeping; leaves to 1.5 m; pet- iole ca. equaling the lamina, stramineous to brown, scaly at the base, the scales 5-12 x brown, thin, lanceolate; lamina 1-pinnate-pinnati- fid, lanceolate, lacking bulblets, only a few distal- most pinnae with short-decurrent bases; pinnae 16-24 x 2.3-4 cm, 15-20 pairs, cut !$—-25 to the costa, the base sessile or nearly so, the basal pinnae short-stalked; segments 5-6. m wide, separated by narrow, V-shaped sinuses, entire to denticulate at the apex; rachis and costae stra- mineous or occasionally the rachis tan, sparsely scaly, the scales ovate to lanceolate; veins free, not forked, 5-8(-10) per lobe; sori inframedial; indusial scale absent. 1-2 mm, Distribution of Stigmatopteris bulbifera (dots) and 5. brevinervis (diamonds). Additional specimens examined. BRAZIL. MINAS GERAIS: Serra de Babilonia (Minas), 6 Apr. 1868, Glaziou 2400 (P). PARANÁ: Serra do Mar, Volta Grande, 400 m, Dusén 725a (BM, СН, МО); Guaratuba, Dusén 13727 (GH). RIO DE JANEIRO: Or Glaziou 3333 (P). Reitz C 206 (US); Itacorubi, Mri и (US); Morro da Lagoa, Rohr 353 (US). 540 PAULO: muni- cipio Iguape, وا‎ Brade 8250 (PH, UC); munici- pio Iguape, Serrinha, Brade 8343 (US). STATE UNKNOWN no locality, Müller 1038 (US). s brevinervis is endemic to the dud МО юе southeastern Brazil (Map 1); the only recorded elevation is 400 . It closely resembles S. tyucana, the only difference being the degree of cutting. In S. tyucana, the pinnae are cut 14-14 to the costa; in S. ки the pinnae are cut 1—28 to the costa. Stig brevinervis has a larger leaf and more pinnae than does 5. tyucana, which suggests that 5. brevinervis Annals of the Missouri Botanical Garden may be merely a larger, more deeply cut form. Another difference is that most specimens of S. гуисапа have medial to supramedial sori, whereas all those of S. brevinervis have inframedial sori. These two species are very similar and perhaps future study will show that they are conspecific. 2. Stigmatopteris bulbifera R. C. Moran, sp. nov. TYPE: Ecuador. Napo: Cerro Huacamo- yos, on road Baeza-Tena, ca. 34 km from Baeza, 2,000 m, Mllgaard et al. 35866 (holotype, AAU, 5 sheets). Figure 9. Map 1. Folia 1-pinnata, pinnis 7-10 utroque costae latere, gemmis praesentibus in axillis a distalium; rhachis et costae abaxialiter pU pubescentes; venae irre- gulariter anastomosante Rhizome short-creeping, the internodes 0.5-2 cm, the apex scaly, the scales 1-4 mm, lanceolate, dark brown, entire; leaves to 0.9-1.3 m; petioles nearly equaling the lamina, dull, brown, scaly, the scales 5- mm, lanceolate, brown, entire, not shiny adaxially; lamina 1-pinnate, oblong-deltate, buds usually present in the axils of the distal or medial pinnae; free pinna pairs 7—10; basal pinnae not reduced, widest near the middle, stalked, the stalk 2-5 mm; medial pinnae 18-25(-35) x 2.5- 4.5(-6) cm, widest near the middle, the margins crenate-serrate or, in large leaves, cut № to the costa; rachis and costae tan, puberulent and scaly, the hairs less than 0.1 mm, erect, capitate-glan- dular; veins partly free and partly anastomosing; sori discrete, round, lacking an indusial scale. Additional specimens examined. COLOMBIA. CAQUE- A: Putumayo, cerca de Buenosaires, 2,060 m, Hageman 2072 (COL). CHOCO: between Pueblo Rico (Risa- ra alda) е Istmina (Chocó), along Quebrada Antón, 15 km W of Santa Cecilia, 240 m, Croat 70911 (MO, UC). HUILA: Cordillera Oriental, Caquetá side of Huila- Caquetá divide, 20 km SE of Garzón, 2,380 m, Little 9380 (COL, GH, US). MAGDALENA: Sierra Nevada de Santa Marta, finca Los Arroyitos, 1,600-1,700 m, Kirkbride 2377 (COL). QuiNDpío: “Quindio” m 615 (COL). ECUADOR. PICHINCHA: Saloya, road to Las Palmeras, mountain left of road across from Las Гарпаз, 1,850 m, Heinrichs 550 (F, Z Stigmatopteris bulbifera occurs in Colombia and Ecuador (Map 1), where it grows in wet forests from (240-)1,600 to 2,380 m. It resembles 5. heterophlebia in that both are l-pinnate, have anastomosing veins, and are beset with numerous capitate-glandular hairs on the abaxial surface of the axes. The two species are easily distinguished, however, by the characteristics given in the key. Other less constant characteristics are that 5. bul- bifera tends to be a much larger plant, with leaves over one meter tall, and that 5. bulbifera grows at higher elevations than S. heterophlebia. In gen- eral, S. bulbifera has a greater number of free veinlets per pinnae, sometimes up to 90% free. The buds are best seen on the adaxial side of the eaf. 3. Stigmatopteris caudata (Raddi) C. Chr., Bot. Tidsskr. 29: 302. 1909. Polypodium cau- datum Raddi, Opusc. Sci. 3: 288. 1819. Dryopteris caudata (Raddi) C. Chr., Index Filic. 257. 1905. TYPE: Brazil. Rio de Janeiro: Gávea, Tijucas, Raddi s.n. (holotype, FI or PI not seen). Figure 10. Map. 2 ОИ caudatum Kaulf., Enum. Filic. 113. 1824. E: Brazil. Rio de Janeiro: Mertens s.n. (holotype, 17— — destroyed; isotype, C not seen; photos F, СН ex C). Rhizome short-creeping; leaves to 2 m long; petiole 14 to equaling the length of the lamina, brown, scaly, the scales 3-8 x 1-2.5 mm, lan- ceolate, brown, thin, lacking fibrillose scales; lam- ina 1-pinnate-pinnatisect, lanceolate, lacking buds, glands obscure; pinnae 15-32 x 4-7 cm, 15-20 pairs, cut 25—78 to the costa, sessile, the distal ones adnate with the basal basiscopic lobe adnate to the rachis; lobes 3-8 cm wide, separated by broad, open, U-shaped sinuses, serrate on all sides, oc- casionally lobed, the basal basiscopic segments re- duced; rachis and costae tan to brown, scaly, the scales nonbullate, ovate, lanceolate; veins free, 9— 15 per lobe, unbranched or 1-forked with a short acroscopic branch which (in fertile segments) bears the sorus; sori lacking an indusiumlike scale. Additional specimens examine BRAZIL. ESPÍRITO SANTO: Jatiboca, Brade 18228 MEM PARANÁ: bari) o morro 7-astrada US, Z); Serra do 40a (F, G); municipio Paranagua, Pico Torto, encosta oriental, 300 m, Hatschbach 22873 icinio M quecaba, Rio do Custa 26665 (UC). RIO DE JANEIRO: Petropolis “Independencia,” Alston & Lutz 289 (BM); Nouvelle Fribourg, Claussen 133 (G, P); Palmeiras, Preston s.n. (K); Mt. Corcovado, 1815, Cunningham s.n. (BM, NY); Rio Corcovado, Du sén 2538 (МО); Organ Mts., 1837, Gardner 131 (K); Rio de Janeiro, 1822, Gaudichaud s.n. (В, Е, С, К, Р); Rio de Janeiro, Glaziou 2396 (P), 6414 (B, P); Sebas- tianopol, Martius s.n. (B); alto Macahé, Mendonca 1355 By Rio de Janeiro, Miers s.n. (B, BM, K, P); Rio de Janeiro, Mosén 2694 (B); Rio de Janeiro, 1840, Regnell 254a (GH, NY, Р); 1841, Regnell 358 (US); near Rio de Janeiro, Sello s.n. (B, BM); brook trail between Pai- neiras and Jardim Botanico, 100-400 m, 1. B. Smith ~ Volume 78, Number 4 1991 Moran 875 Stigmatopteris Monograph FIGURE 9. b. Lamina (Heinrichs 550, d I Stigmatopteris bulbifera. —a. Juncture of rachis and costa with bud (Øllgaard et al. 35866, AAU).— 7 istal half, B; proximal half, F).—c. Part of fertile pinna (Hageman & Leist 2072, COL). — d. Part of fertile pinna (Ollgaard et al. 35866, AAU). 1391 (GH); near Rio de Janeiro, U.S. Exploring Expe- dition 23 (US); near Rio de Janeiro and Bahia, Webb s.n. (MICH, NY, US). SANTA CATARINA: municipio de Sombrio, Retiro, Araranguá, Reitz 227 (US); municipio de Meleiro, Araranguá, perto da cidade, Reitz C205 (US); municipio de Brusque, Morro de Bateia, 100 m, Reitz 1909 (BM); municipio de Brusque, Ribeiráo do Ouro, 600 m, Reitz 3542 (US); municipio de Ibirama, Horto Flo- restal I.N.P., 400 m, Reitz & Klein 3077 (US); municipio de Florianópolis, Sertao da Lagoa, Rohr 1072 (B, L, NY, Annals of the Missouri Botanical Garden 779 FIGURE 10. c. Basal pinnae (collector unknown, B) US) municipio de Joinville, о 191 (Е); урен da Gloria, Feb. 1884, (Ле 68 (K), 70 © ЗАО PAUL Eje Apr. 1922, cie s.n. (US); Peroupara, Bra de 478(NY, UC); Serra do Mar, Ген шаш 21821(GH); no locality, 1906, Wacket s.n. [Rosenstock, Filices aus- образ или по. 286] (B, GH, P, UC). Stigmatopteris caudata is endemic to south- eastern Brazil (Map 2) where it grows in wet forests from 0 to 600 m. It is the most highly cut species ar sngmatopte ris in southeastern Brazil and has parated by broad, о Stigmatopteris caudata. —a. Venation and sori. — b. Distal third of the lamina (Rohr 1072, US). — U-shaped sinuses (Fig. 10a). In its size and cutting, this species resembles S. ichthiosma, an Ecuador- ian species, but differs by its more deeply serrate or lobed segments, wider sinuses, generally wider pinnae, and lack of an indusiumlike scale. The | scales also differ between the two species: S. caudata has thin scales, whereas at least some of ties on 5. ichthiosma are darkly sclerified adaxially. Although I have not seen the type specimen of this species, I feel confident on the basis of Raddi's Volume 78, Number 4 1991 Moran Stigmatopteris Monograph 877 МАР 2. original description and his later illustration (Raddi, 1825, fig. 39) that the name is here applied cor- rectly. > . Stigmatopteris chimalapensis Mickel 4 Beitel, Mem. New York Bot. Gard. 46: 356, fig. 129E, Е. 1988. TYPE: Mexico. Oaxaca: Municipio Santa Maria Chimalapa, Canada E de la vereda a La Gloria, ca. 8 km SE de Santa Maria, 350 m, Hernández 1282 (ho- lotype, NY). Figure 11. Rhizome erect, short; leaf 70 cm; petiole equal- ing the lamina, brown to stramineous, sparsely scaly, the scales 1-3 mm, ovate to lanceolate, entire, appressed, shiny black abaxially; lamina 1 -pinnate-pinnatifid, lanceolate, bulblets lacking; basal pinnae equilateral, widest at or slightly above the middle, short-stalked, the stalk 2 mm; medial pinnae 8-12 x 1.3-2 cm, equilateral, widest at the base, cut са. У or slightly less to the costae, 14 pairs, sessile; lobes entire or slightly serrate at the apex, the proximal acroscopic ones elongated; rachis brown, costae stramineous, both axes scaly, the scales irregularly ovate to lanceolate, the mar- gins ciliate; veins free, unbranched, 4-6 per lobe; sori lacking an indusial scale. This species is known only from the type col- lection. It resembles S. sordida, with which it may eventually prove conspecific, but differs by its smaller size and less-cut pinnae. Distribution of Stigmatopteris caudata (left) and 5. heterocarpa (right). 15 cm FIGURE 11. Stigmatopteris chimalapensis. — Leaf (Hernández 1282, NY). 878 Annals of the Missouri Botanical Garden сг у її U AA \ CANN \ ЇЇ PIS fT Pa) ^ UN 00 JUN 4 М UN | Y R Morin 1986 FIGURE 12. Sti дечка и contracta (Moran 3187, МО). —а. Cross section of petiole base. —b. Proscales or microscales (uniseriate) and multiseriate scales from the distal portions of the abaxial surface of the lamina.— c. Medial pinnae. — d. Venation ier sori. —е. Segments of the pinna bases overlapping the rachis. — f. Large basal pinna. 5. Stigmatopteris contracta (Christ) C. Chr., GH, NY ex BM); Colombia (Nouvelle Gra- Bot. Tidsskr. 29: 304. 1909. Aspidium cau- nada), Schlim 1683 (P—herb. Christ?, BM). datum var. contractum Christ, Bull. Herb. (This specimen is here referred to S. pellu- Boissier sér. 2, 6: 162. 1906. SYNTYPES: Costa cidopunctata.) Figure 12. Rica. Cartago: Navarro, 1905, 1,100 m, Wercklé s.n. (lectotype, here designated, P; Rhizome up to 10 cm, erect or ascending; leaves isolectotypes, BM, MICH, NY, US; photos, 1-3 m; petiole 12-34 the length of the lamina, Volume 78, Number 4 1991 Moran 879 Stigmatopteris Monograph brown, the scales 1 -3 mm, ovate to nearly circular, opaque, appressed, dull brown adaxially, shiny black abaxially, these sometimes mixed with larger, thin, translucent, flaccid scales; lamina 80-200 x 40- 95 cm, l-pinnate-pinnatifid, lanceolate, bulblets lacking; medial pinnae 20-35 x 3-6 cm, linear, cut % to nearly all the way to the costae, pairs 20-26, the base overlapping the rachis; basal pin- nae 7-15 cm wide, elongated basiscopically, widest near the middle where the basiscopic basal seg- ments are twice as long as the acroscopic ones; segments entire to crenate or serrate, the fertile ones contracted, their sinuses 1-2 times the width of the segments, U-shaped, the basal segments on both sides prolonged and overlapping the rachis; rachis brown; costae stramineous, scaly, the scales 1-2 mm, few, scattered, ovate, entire or nearly so, never fibrillose; veins free, 8-18 in the medial segments of the medial pinnae, those of the fertile segments unbranched, those of the sterile segments branched or unbranched; sori with indusial scale absent or rarely present, the scale ovate-lanceolate, borne on the basiscopic side, stipitate and held above the sporangia. Additional specimens examined. Costa Rica. ALAJUELA: canyon of Río Cariblanco and W slope and summit of ridge between Río Cariblanco and ar Quicuyal, — of cep 840-950 m, A. R. Smith 2 (MO, NY, UC); Monteverde Cloud Forest Reserve, 1,250-1, n m, Burger et al. 10735 (CR, F, NY, US); along road from San Ramón to Bajo Rodríguez, at Río Cataratas, Croat 68108 (MO, UC); Monte Verde slope of ridge along quebrada draining E to Río Cataratitas, ca. 20 km NW of San Ramón, 850 m Smith et al. 2247 (CR, MO, UC); Reserva Foresta) de San Ramon, Rio San Lorenzito, 800-1,000 m, Herrera et al 336, 0, UC). А о: Pejivalle, Тело 8. . El Мипесо, ndley & Torres R. '5106 1 (US), $); Navarro, 1,500 m, Torres 49 (US) without locality, Endress s.n. (K); 10 km SW of Navarro on a winding trail ca. 10 km S of Cartago and 1 km S of Finca Montecristo, 1,400-1,600 m, White & Lucansky 1968150 (DUKE, GH); Reserva de a Gómez 1133 (CR), 2330 (CR, F, GH, MO, NY, US), 18905 (AAU, CR, MO, UC); forest near the od to arque Nacional Tapanti, 1, , NY); along tributary of Quebrada Casa Blanca m, Grayum ridge е Plantanillo, 1,200 (NY, UC, US); Turrialba, Rojas 6. s.n. (CR); Platanillo, 650- 900 m, de la Sota 5238 (US), White & Lucans 1968174 ery HEREDIA: Virgen del Socorro, 1, 000 m, Moran 3166 (MO), 1,500 m, Wagner & Gomez 79027 (CR); ae above the Upper Sarapaqui Val- ley, ca. 1,200 m, Scamman s.n. (GH). LIMON: Siquirres, Stork 2263 (MICH). SAN RAMON: La Hondura, 1,000 m, Nisman S. 98 (GH), Gómez 281 (CR), 1,300-1,700 m, Standley 37834 (US); vicinity of La Hondura, 4 km S to 1 km N on Rte. 220, between Volcán Irazü and Volcán Barba, 1,250-1,500 m, Mickel 2241 (NY), 2249 (NY, UC, US). PANAMA. CHIRIQUÍ: Boquete, Dexter Trail, Corn- man 1091 (F, GH, MICH, MO, UC); vicinity of El Bo- е 1,000-1,500 m, pig 914 (USy valley ba e Rio Piarnasta, са. 5 mi. E of Boquete, 1,600 Killip 5159 (GH, MICH, МО, UC, US). Stigmatopteris contracta grows from 650 to 1,600 m in the montane forests of Costa Rica and Panama. With leaves 1—3 m long, 5. contracta is the largest species in the genus. The segments of the sterile and fertile leaves are subdimorphic, dif- fering in width, cutting, and venation. The name contracta refers to the fertile segments, which are narrower than the sterile ones. This species might be confused with large plants of S. sordida, which also grows in Costa Rica and Panama. Stigmatopteris contracta can be distin- guished by its pinna bases that overlap the rachis, costal scales that are ovate and nearly entire, fertile segments that are contracted, and wide sinuses that are one to two times the width of the segments. А scalelike indusium is present on the type but absent from the other specimens. I chose Wercklé s.n. as the lectotype because it preserves the previous usage of the species name, and it is more widely distributed in a The other syntype cited by Christ, Schlim 1683, here referred to S. pellucidopunctata. 6. Stigmatopteris gemmipara C. Chr., Kongl. Svenska Vetenskapsakad. Handl. ser. 3, 16: 36, t. 9, figs. 5-7. 1937. Dryopteris gem- mipara (C. Chr.) Maxon ex Proctor, Bull. Inst. Jamaica, Sci. Ser. 5: 26. 1953. TYPE: Haiti. Hotte, western group, stony ridge above La Mare Proux, near Torbec, 6 Dec. 1925, Ek- man 5268 (holotype, S? not seen; isotype, IJ not seen). Figure 13 Rhizoine erect to decumbent, compact; leaves to 1 m; petioles equaling the length of the lamina, brown, scaly, the scales 0.5-1 cm, linear-lanceo- late, brown, a few black, sclerotic, opaque; lamina 1-pinnate-pinnatifid, narrowly elongate-triangular to lanceolate, with buds in axils of distal pinnae, the apex attenuate; pinnae (4-)9-14 x (1.5-)2.5- 3.5(-4) cm, 9-20 pairs, widest at or near the base, cut ог more to the costa, the basal basiscopic segment occasionally free, reduced, the base of the proximal pinnae stalked, the stalk 4-6 mm; seg- ments 4-6 mm wide at the middle, serrate, the apex obtuse, the basal basiscopic segment not ad- 880 Annals of the Missouri Botanical Garden FIGURE 13. 15 cm Stigmatopteris gemmipara (Ekman 10113, US).—a. Pinna with bud in axil.—b. Leaf with buds. — c. Fertile pinna, second from the base of b.—d. Petiole scales. nate to the rachis; rachis and costae stramineous to light brown, scaly, the scales 2-3 mm, lanceolate to filiform, nonbullate, spreading, conspicuous, the costular scales nonbullate; veins free, mostly un- branched, 5-7(-8) per segment; sori lacking an indusiumlike scale. Additional specimens | examined. JAMAICA. PORTLAND: ca. 5 mi. SW of Priestman's River on the E slope of John Crow Mts., ca. 500 m, Proctor 4257 (US). HAITI. Massif de la Hotte, western group, Torbec, Morne 10113 (US). Stigmatopteris gemmipara grows only in Ja- maica and Haiti in woods over limestone soils from 500 to 1,200 m. This species could be confused Volume 78, Number 4 1991 Moran 881 Stigmatopteris Monograph FIGURE 14. is adnate to the rachis rather than to the c Stigmatopteris hemiptera. —a. Buds in the п | distal pas note that the basal basiscopic lobe osta (León et al. 10 МУ). — b. Leaf (Ekman 14356, GH).— c. Distal pinna (Biman 14356, GH).—d. Proximal part of a large E con (Wright 1053, US). with 5. jamaicensis, which is endemic to Jamaica, but that species lacks bulblets and has an indu- siumlike scale subtending the sorus. 7. Stigmatopteris hemiptera (Maxon) C. Chr., Index Filic. Suppl. 3: 174. 1934. Dryopteris hemiptera Maxon, Contr. U.S. Natl. Herb. 24: 59. 1934. TYPE: Cuba, Sep. 1859-Jan. 1860, Wright 1053 (holotype, US; isotypes, BM, G, GH, K, MO, fragment B, US). Figure 14 Rhizome decumbent, compact; leaves 45-130 cm long; petioles / to equaling the lamina, stra- mineous to dull brown, sparsely scaly, the scales narrowly triangular, translucent or nearly so, not black and opaque; lamina | -pinnate-pinnatifid, lan- ceolate, buds produced on small sterile leaves in the axils of the distal pinnae, the apex evenly tapered; pinnae 11-18 x 1.5-2.5 cm, 15-20 free pairs, widest below the middle, cut У2—% to the costa, large plants occasionally with a free basal acroscopic pinnule, the base adnate in the distal 14 of the lamina, with the basal basiscopic lobe arising from the rachis rather than the costa; seg- ments 3—4 mm wide, the margins serrate, the apex rounded; rachis and costae stramineous, scaly, the scales scattered, nonbullate; veins 6-10 per seg- ment, mostly unbranched; glands not easily visible; sori nonparaphysate, lacking an indusiumlike scale. Additional specimens examined. CUBA. ORIENTE: Si- erra Maestra, Loma del Gato, Clement 961 (GH, P, UC, US), 1,000 m, 1602 (2); Hioram & Clement 6442 (GH, US); Sierra Maestra, on the divide below Rio Yara and Rio Palmamocha, ca. 1,00 zZ =| un м un e a o o — + =g O "O = c Ùn = =S = [авд D Ф Su Ф = = e دو‎ o л et = e o Aserraldero San Antonio de umbres, region o of La Bayamesa, 1,400-1,500 m, Morton 9569 (US); Loma Annals of the Missouri Botanical Garden FiGURE 15. 15 cm Stigmatopteris heterocarpa (Oliviera & Cordeiro 909, MO).— a. Leaf. —b. Venation and sori. Volume 78, Number 4 1991 Moran 883 Stigmatopteris Monograph del Gato, Ravins 961 (UC). SANTIAGO DE CUBA: Sierra oma del Gato, 1,000 а M о US), 10486 (NY, US); Maestra Ridge, 0 m, Fre. León 11133 (NY, US). Stigmatopteris hemiptera is endemic to the Sierra Maestra mountains of eastern Cuba where it occurs in wet forests from 900 to 1,500 m. A diagnostic characteristic of this species is that the basal basiscopic lobes of the distal pinnae are adnate to the rachis, not the costae (Fig. 14a). Apparently, the buds are only on sterile leaves, unlike the closely related S. gemmipara, which has buds on the sterile and fertile leaves. This observation, how- ever, should be confirmed by field studies, as there are few specimens of each species. p Stigmatopteris heterocarpa (Fée) Rosen- stock, Repert. Spec. Nov. Regni Veg. 21: 347. 1925. Phegopteris heterocarpa Fée, Crypt. Vasc. Brésil 1: 100, t. 30, fig. 2. 1869. TYPE: Brazil. Rio de Janeiro: Rio de Janeiro, Glaziou 2401 (holotype, P; isotypes, K, P). Figure 15 Polypodium carrii Baker in C. Mar end Fl. Bras. 1(2): 597.1870 ер ми сатги (Ва . Chr., Index Filic. 257 mid carrii (Ba ke г) С. Сћ 0 PE: Brazil. Rio asilaris Rosenstock, epert. Veg. 21: 347. 1925. TYPE: ulo: municipio gua pe, Morro das Pedras, Brade 8231 (holotype, 5 not seen; iso- type, US) Rhizosome short-creeping; leaves to 1.5 m, pet- iole ca. equaling the lamina, stramineous to brown, scaly at the base, the scales 5-12 x 1-2 mm, brown, thin, lanceolate; lamina 1-pinnate-pinnati- fid, lanceolate, lacking bulblets, only a few distal- most pinnae with short-decurrent bases; pinnae 14-23 x 1.7-3.2 cm, 12-17 pairs, nearly entire, crenate, or cut % to the costa, the base sessile or nearly so, the basal pinnae short-stalked; lobes entire; rachis and costae stramineous, sparsely scaly, the scales ovate, lanceolate, or oblong; veins free, not forked, 3—5 per lobe; sori medial to inframedial; indusial scale absent. о Addit =“ К imens exam ке, Вима. СН), 200 т, 14680 (Е, С, СН, y municipio Morretes, Estr. da Graciosa, Grota Funda, Hatschbach 10748 (B, LIL), 500 m, 24376 (UC); municipio Tijucas do Sul, Rod. Br 468, Rio Itararé, 900 m, Hatschbach 25939 (NY, UC, 05); municipio Morretes, Jurapé, ao 41955 (UC, Z); municipio Morretes, Estagao rumbi, Kummrow & Hatschbach 2536 (MU, UC, п municipio Morretes, Véu de Noiva, Oliveira & Cor- deiro 909 (MO, U „ Wag- ner "4402 (P). SANTA тее по locality, Pabst 799 (05); ~ чый Biguacü, Reitz C91 9 Brusque, 50 m, Reitz 3141 (US); Horto Ени I.N.P., Ibirama, m, Reitz & Klein 3068 (US); Blumenau, Schenck 494 (B); и پوت ت‎ Sert goa, Rohr 392 (GH, US), 1006 (B, UC, US); Ta ec. 1904, Schmalz s.n. паса Filices austro- brasil. iin no. 106] (B, GH, NY, P, UC); Morro da Cruz, Sehnem 808 (GH); Morro do Antao, 250 m, Seh- nem 3086 (B, US); municipio Brusque, Aza ja, Brusque, 35-135 m, Smith & Reitz 6127 (US); Blumenau, Span- паре! 332 (US); Зао Francisco, Guli, Ule 187 (P). são PAULO: municipio Iguape, Morro das Pedras, 8510 (UC); Peroupara, Brade 8473 (NY, UC), 8474 (NY, UC), 8476 (NY, UC); Ilha do Cardoso, Morro da Сартасао, en di- reccion al Morro dos Tres ud 180-570 m, Rio oe Forero et al. 8793 (COL); Blumenau, Paso nsa, Haerchen 106 PRüsenistock. Filices austro-brasil. exsicc. no. 106] (B, P, UC). Stigmatopteris heterocarpa is endemic to southeastern Brazil, where it grows in wet forests from 50 to 900 m. It resembles S. rotundata, a species of the Lesser Antilles and northern South America, but differs by its fewer pinna pairs. It is sometimes difficult to distinguish this species from S. tyucana, the only difference being that S. tyucana has pinnae more deeply cut and generally broader than does S. heterocarpa. Fieldwork cou- pled with cytological studies may provide further evidence as to whether these two species are truly distinct. Rosenstock's var. basilaris was said to differ from typical S. heterocarpa only by its inframedial sori. This character intergrades with medial sori and is therefore not worthy of distinction. 9. Stigmatopteris heterophlebia (Baker) R. C. Moran, Fieldiana, Bot. n.s. 27: 45. 1991. Polypodium heterophlebium Baker, J. Bot. 22: 363. 1884. Dryopteris heterophlebia (Baker) C. Chr., Index Filic. 270. 1905. TYPE: Costa Rica. Province unknown: Laguna, 305 m, Harrison 59 (holotype, K). Figure 16. ap Rhizome short-creeping, the internodes 0.5-1 cm, the scales mm, ovate to lanceolate, opaque, black, rigid, entire; leaves 45-100 cm; petiole ca. equaling the lamina, dull brown or tan, puberulent, scaly, the scales of two types, the first type 2-5 mm, opaque, appressed, shiny black adaxially, en- 884 Annals of the Missouri Botanical Garden / y FIGURE 16. Stigmatopteris heterophlebia. —a. Leaf (Moran 3613, МО). —b. Sori and irregularly anastomosing ) veins. — с. Decurrent pinna bases (redrawn from Christensen, 1909). tire, ovate-lanceolate, slightly falcate, the second type mm, translucent, spreading, brownish, lanceolate; lamina 1 -pinnate, oblong-deltate to lan- ceolate, buds lacking or rarely (in NW Ecuador) present; free pinna pairs 1-3(-4); basal pinnae usually somewhat reduced, widest near the middle, stalked, the stalk 2-4 mm; medial pinnae (or lobes) 10-16(-18) x 2.5-3.5(-5) cm, widest at or near the base, the margins entire to crenate, the base adnate and decurrent to or nearly to the next pinna pair; rachis and costae tan, light brown, or stra- mineous, puberulent and scaly, the hairs less than А ense, erect, capitate-glandular (when fresh), the scales to 3 mm, lanceolate, not fibrillose; veins part free and part anastomosing; sori discrete or confluent where the veins anastomose, indusial scale absent. Additional specimens examined. NICARAGUA. ZELAYA: Cerro Saslaya, 20 km W of Siuna, 1,100-1,400 m, Neill 1886 (CR), 3826 (CR, MO). Costa RICA. ALAJUE- A: Reserva Forestal de San Ramon, 800-1,200 m, Her- rera Ch. et al. 162 (MO, UC), 316 (AAU, MO, NY, UC, US); canyon of Rio Cariblanco and W slope and summit of ridge between Rio Cariblanco and Quebrada Quicuyal, SW of Cariblanco, 840-950 m, A = Upala, falda sur de Volcán Miravalles, Chacón s.n. (СК); Bajos de Jamaical Reserva de San Ramón, 700-1,000 m, Chacón 1776 (CR); Colonia Virgen del Socorro, Sara- piqui, Ocampo S. 1029 (CR); above San Ramón on road Моште 78, Митбег 4 1991 Могап 885 Stigmatopteris Monograph to Los Angeles Norte, ca. bove San Ramón, 610 m, McAlpin 1313 nue ТЕ. AM Reserva de Monteverde, vertiente Pacifico, 1, 520- m, k C Ramón’s Biological ma Station, Moran 3188 (CR, F, GH, МО); 11 km San Ramón, 1,000 m, Mickel 2955 (NY, UC, US); Univ. de Costa Rica Reserva, Fila Rio Reventazón ur rd Instituto Interamericano of Turrialba, 750 m, Mickel 2622 (NY, US); ca. 22 km E of Turrialba, high ridge above Platanillo, 1,200-1,450 m, Mickel 3447 (МУ); El Мипесо, 1,400 m, Stark 2722 (05), El Muneco, on the Ri 1,400-1,500 m, ет & Torres К. 51069 (US); Tapanti Reserve, 1,400-1,700 m, Gómez 19288 (AAU, MO, UC). GUANACASTE- ALAJUELA: slopes of Miravalles, above Bijagua, ca 0 m, Gómez et al. 19173 (AAU, MO, UC). HE m, Moran agner & Góm = 7902 ја (СЕ); forest between Rio Реје and Río Sardinalito, Atlantic slope of Volcán Barva, 700-750 m, Grayum 6651 (MO, UC); 1 of Rio Sardinal, ca. 0 rayum et al. m, Proctor 32327 > slopes of Cerro yv 9 km N of Sa m, ie: 1634 (CR, Е, UC, US). LIMÓN: vic. of Guápiles, 300-500 m, Standley 37169 (GH, US); along Costa Rica, of Gua о Su Chacón & Parque Nacional Braulio Carrillo, Estación La Mont ura, Rio Zurqui, ca. 1,500-1,600 m, A MO, UC). PROVINCE UNKNOWN: 1901-1905, Wercklé s.n. (NY, US). PANAMA. BOCAS DEL TORO: — Road, near Continental Divide, Fortuna Dam area, 1,000 m, Churchill et al. 4549 (MO, UC); Cerro кэм гаи са. 8.6 mi. W of Сћате, Croat 69137 ( C) along road between Fo of bridge over Fortuna Dam, 3.2 mi. Divide, 700 m, Croat & Grayum 60216 (MO, UC), 60217 (AAU, MO, UC); vic. of Fortuna Dam, below pass on Chiriqui Grande Mas ca. 800 m, McPherson 9730 ме road diy . N of Con imental Divide, 700 m (US); La Fortuna hydroelectric project, 1,300-1, Hammel 2227 (AAU, CR, MO); Fortuna Dam area, N d а Sum | fn. war m ا‎ --- | ~ >? ~ | o 2 a’ 20 | я | E <> = · C [II 4. р | S |o. | | | и) | | 7i | 47-2 ss | Ж 90 8 5 J |70 МАР 3. Distribution of Stigmatopteris heterophle- bia. of reservoir, Quebrada Bonito to E of road, 1,100 m, Churchill 5255 (МО); vic. of Cerro Colorado, above San Felix, 1,450 m, McPherson 12677 (MO); Dtto. Boquete, Fortuna dam site, along trail following Continental Divide, 1,100 m, van der Werff & van Hardeveld 6762 (MO, NY, UC); along road between Fortuna Lake and Chiriqui Grande, 4.5-5 km N of dam over Fortuna Lake, 1,100- 1,135 m, Croat & Grayum 60043 (MO); along road between Gualaca and Fortuna Dam site 7.9 mi. beyond (NW of) Los Planes de Hornito, 1,300 m, Croat 49916 (МО); vic. of Fortuna Dam in valley of Río Chiriqui, along aqueduct trail for water supply to IRHE facilities, 1,100- 1,200 m, Croat 66559 (AAU, MO, NY, UC); valley of the Rio Piarnasta, ca. 5 mi. E of El Boquete, 1,525- 1,600 m, Killip 5387 (Е, UC, US). coa f£: valle de Antón, 1,000 m, Alston 8745 (BM Pie in. Gode: 274 77 (MO, slope above Río Pine azo, 9m nt 700-1,200 m, Croat & raid 1010 (MO). боа. CAUCA: Coteje, Rio Timbi И 8951 (Е, СН, К, 05); Cordillera Oriental, comisaria of Putumayo, Oreto (Indian pueblo) on Rio RD tributary of pos Pu- UC, US tumayo, 450 m, Ewan 16775 (BM, ). CHOCÓ: Municipio de San José del Palmar, hoya del Río Torito, afluente del Rio Habita, 850-950 m, Forero et al. 7407 Pacifico, Rio Cajambre, San Isidro, 5- casas 17336 (US). ECUADOR. ESMERALDAS: banks of Rio Palavi, 0.5 km below Awa encampment, 100 m, Hoover et al. 4304 (MO, UC), 4325 (MO, UC). MORONA-SANTIAGO: along new road Méndez-Morona, 650 m, van der Werff & Gudiño 11125 (MO, UC). маро: Anangu, Parque 886 Annals of the Missouri Botanical Garden Зз om 15 cm FIGURE 17. Sti | ichthiosma. — i and venation. — b. Sori subtended by indusiumlike scale and poivre scales on the UC. d, Sodiro s.n., a Nacional Yasuní, 260-350 m, Ollgaard et al. 39229 (AAU) 12 km SW de Coca, por el camino se llama Los Zorros, 240, Moran 3613 (MO); Estación Experimental de INIAP, San Carlos 6 km al SE de Los Sachas, 250 m, Baker 5934 (NY); at Río Payamino 60 km along Río Payamino W of Coca, 350 m, Holm-Nielsen & Jeppesen 811 (AAU, GH). PicHINCHA: NE of Vicente Maldonado, Reserva de ENDESA, 600 m, Laegaard 51611 (AAU, MO). PERU. LORETO: Santa Rosa, lower Rio Huallaga below ا‎ ca. 135 m, Killip & Smith 28984 (NY, US). Stigmatopteris heterophlebia occurs from Nic- aragua to Peru (Map 3) in wet forests from 100 to 1,900 m. It resembles S. bulbifera in its glan- inna pairs. A further difference is that S. heterophlebia usually lacks buds, whereas S. bulbifera has them on the adaxial surface of the rachis—costae junctures. This characteristic is not completely constant be- cause two specimens of S. heterophlebia from the department of Esmeraldas, Ecuador, have buds. This species is also variable in the scaliness of the axes. Several specimens from Costa Rica and Panama (Hammel & de Nevers 13578, Lellinger & White 1692, Croat & Liesner 1010) are ex- . Sor — c. Medial Ийе of pinna a.—d. Large basal pinna. a-c, Cazelet & Pennington tremely scaly but are typical in all other charac- teristics, and no other feature seems to correlate with the scaliness. Stigmatopteris heterophlebia has been previ- ously called S. alloeoptera (Kunze) C. Chr., a name that cannot be used because it is of uncertain application (see section on names of uncertain ap- plication). 10. Stigmatopteris ichthiosma (Sodiro) C. Chr., Bot. Tidsskr. 29: 302. 1909. Polypo- dium ichthiosmum Sodiro, Recens. Crypt. Vasc. a 59. 1883, as ichtiosmum in So- diro, Crypt. Vasc. Quit. 294. 1893. Dryop- teris ра (бод го) С. Chr., Index Filic. 271. 1905. ТУРЕ: Ecuador, “Atcao, 1,500 m, S. Florencia, etc.” Sodiro s.n. (holotype, Q; isotypes, BM, K, P). Figure 17 Polypodium dentatum Baker, Ann. Bot. (London) 5: 456. 1891, nom. 1 longipetiolata C. nom. nov. for P. de тад “Andes of Ecuador," Sodiro s.n. (holotype, К). Rhizome short-creeping; leaves to 2 m long; Моште 78, Митбег 4 1991 Могап 887 Stigmatopteris Monograph petiole 12 to equaling the length of the lamina, brown, scaly, the scales 4-6 x 2-2.5 mm, ovate, symmetrical, occasionally with darkly sclerified ab- axial surfaces, the base cordate (appearing false- peltate), fibrillose scales absent; lamina 1-pinnate- pinnatifid, lanceolate, lacking bulbils; pinnae 20- 40 x 2.5-6 cm, 15-20 pairs, cut ca. 25 to the costa, sessile or the basal ones short-stalked, the stalk 2-7 mm; segments 4-10 mm wide, serrate at the apex, the sides entire, serrate or rarely lobed, basal basiscopic segment not reduced; rachis and costae stramineous or brown, scaly, the scales non- bullate, ovate, lanceolate, or filiform; veins free, 7-14 per segment, unbranched or 1-forked with a short acroscopic branch which in fertile segments bears the sorus; sori with an indusiumlike scale. Additional specimens examined. ECUADOR. CHIM- M). s.n. (AAU, P, Q); Toachi, Sep. 1874, " Sodiro 48/1 (K); 20 km 5 omingo de los та 300 т Y, UC, US); road Chi- m-Nielsen et al. 24819 (ААО); адшрисипа Biological | Field Station, ca. 5 km 5 of Nanegalito an N of Quito, 1,400-1,700 m, Moran 5250 (MO, ОСМЕ); 2.5 km E of ma к а, ‚200 m, Moran 3545 (МО, 0 СА). VINCE UNKNOWN: near Rio Blanco, 1,800 m, Mille s.n. (P near Angamarca, 12/905, Sodiro s.n. (P); no location, Stübel 97 (B). Stigmatopteris ichthiosma occurs in the west- ern Апдез of Ecuador in wet forests from 300 to 1,800 m. It resembles S. pellucidopunctata in leaf shape and cutting, but differs by its ovate to ceolate-falcate) petiole scales, and the presence of an indusiumlike scale that subtends the sorus (rath- er than none). Furthermore, S. ichthiosma occurs only in the western Andes, whereas 5. pellucido- punctata. occurs only on the eastern side of the ndes. The specific epithet is usually spelled ichtiosma; however, Sodiro's (1883) original spelling is ich- т He later altered it to ichtiosma but did give a reason (Sodiro, 1893). According to “жы (1883), the dried plant smells like fish, thus the specific epithet from the Greek icthys, fish, and osme, odor. I prefer to maintain the original spelling because its derivation is clear and because Article 73.1 of the Code (Greuter et al., 1988) states that the author's original spelling should be maintained. I cannot, however, detect a fishlike odor from the dried plants. Christensen (1909, 1913) listed this species from Cuba and Jamaica; however, the specimens he cited represent two closely related species: S. hemiptera (Maxon) C. Chr. (endemic to Cuba) and S. jamai- censis (Desv.) Proctor (endemic to Jamaica). he other Ecuadorian specimens collected by Sodiro (see specimens examined) might also be types, because he says “etc.” for the type localities, meaning other sites on the western slopes of the Andes. 11. Stigmatopteris jamaicensis (Desv.) Proc- tor, Brit. Fern Gaz. 9: 221. 1965. Cystopteris jamaicensis Desv., Mém. Soc. Linn. Paris 6: 263. 1827. TYPE: “Habitat in Antillis" [un- doubtedly Jamaica], Tussac s.n. (holotype, P; photos, BM, GH, UC, US ex P). Figure 18. Dryopterts nothochlaena Maxon, Contr. U.S. Natl. Herb. 24: 58. 1922. Sti p on) ain о (Мах. enit. Chr., Index Filic. Su . 1934. ТҮРЕ: Jamaica. St. Thomas: n na qi Pass, 610- 823 m, Maxon 1738 Malos, US). Rhizome erect, compact; leaves to 2 m long; petioles са. 4 the length of the lamina, stramineous to brown, glabrate to densely scaly (especially to- ward the base), the scales 1-1.5 cm, spreading, ovate-lanceolate; lamina 1-pinnate-pinnatisect, or fully 2-pinnate at the base, oblong to lanceolate, lacking buds; pinnae 15-23 x 2-4(-5) cm, 22- 26 pairs, widest at or below the middle, cut 76 to nearly to the costa, the base sessile or nearly so; segments 3—7 mm wide, serrate, the apex rounded; rachis and costae stramineous, scaly, the scales ovate to lanceolate, those of the costules bullate; veins free, 8-11 per segment, mostly simple, not forked, scaly, the scales ovate, bullate; sori covered by an indusiumlike scale similar to those on the veins. Additional specimens exam . JAM CLARENDON; Second Breakfast Pate one Twe 610 m, Underwood 1613 (NY, US); Tweedside, Harris NY). ND: Ma- B 1,175-1,500 m, Maxon & Killip 7 US); Seamen's Valley, 150-250 m, Maxon & Killip 7 (F, GH, NY, DO Dollwood, 915 m, Harris 7277 (NY), Jenman s.n. (NY); Hardwar Gap, 1,219 m al. 521 (DUKE, GH, MICH); spur of John Crow a Mil is 450-625 m, Maxon 9307 (NY, US). e Mts., Trafalgar, Perkins 1161 (G, GH). ST. THOMAS: Mansf eld , near Bath, 300-500 m, Maxon 2388 (NY. 05), Mansfield near Bath, 1861, Gilbert s.n. (BM, G 563 (GH, MICH); Cuna Cuna trail, above Mattis River, 888 Annals of the Missouri Botanical Garden 1 cm FIGURE 18. Stigmatopteris jamaicensis (Maxon 1738, US ).—a. Sori subtended by an пре ка dai and bullate ee ovate scales on the costule.—b. Distal two-thirds of the lamina. —c. Abaxial surface of a pin 300-500 m, Maxon & Killip 169 (F, GH, NY, US); Bath to Cuna Cuna Pass, Chrysler 4777 (MICH); near Cuna Cuna Pass, 823 m, l gabon et 2696 (NY); moun- 50 Cuna Cuna Gap and vicinity, 600- GH, US). PARISH UNKNOWN: without locality, Jenman 60 (UC), 62 (UC), 77 (UC); Hart 200 (US); Sherring s.n. (US); slopes above Tweedside, 610-915 m, Maxon 970 (05); lower slopes of Mt. Moses, 610 m, Maxon 1071 (US); Blue Mts., Linden 1683 (P) Stigmatopteris jamaicensis is endemic to Ja- maica where it grows in wet forests, 150-1,200(- 1,500) m. It can be distinguished from all other species of Stigmatopteris by the bullate scales on the costules. Walker (1966) counted the chro- mosomes of this species and found that n = 41. This is the only count known for the genus. 12. Stigmatopteris killipiana Lellinger, Proc. Biol. Soc. Wash. 89: 730, fig. 8. 1977. TYPE: Colombia. Chocó: hillside above Rio Mutata ca. m above its junction with the Rio El Valle, near Alto del Buey, ca. 850 m, Lellin- ger & de la Sota 194 (holotype, US; isotypes, COL, CR, HUA not seen, LP not seen; photos DUKE, F, С, GH, MO, NY, UC ex US). Figure 19. Map 4. Rhizome erect or short-creeping; leaves 70—150 cm; petiole ca. equaling the lamina, abaxially atro- purpureous and lustrous, adaxially stramineous or tan, scaly and with a scurf, the scales of two types, the first type 2-5 mm, opaque, appressed, shiny black abaxially, entire, ovate-lanceolate, slightly falcate, the second type 4-9 mm, spreading, brownish, the scurf ap- pressed, arachnoid, cauducous, light brown or tan; translucent, lanceolate, lamina 1 -pinnate, oblong-deltate to lanceolate, buds lacking, free pinna pairs 5-13; medial pinnae 14– 18(-26) x 3-4.5 cm, widest at or just above the Volume 78, Number 4 1991 Moran 889 Stigmatopteris Monograph uo 9] 2 э —X —— Li Moran 1157 Es ае AC di killipiana. —a. Lea MO). — b. Distal part of the lamina showing the transition from E to nondecurrent pinna bases (Grayum et al. 5495, МО). —с. Basal portion of a basal pinna (Grayum et al. 5495, MO). base, the margins entire to crenate, the base sessile, the upper pinnae with a short decurrent base; basal pinnae ca. equaling the above pinnae, widest at or just below the middle, stalked, the stalk 2-3 mm; rachis atropurpureous at the base, the costae stra- mineous, both axes lacking hairs, scaly, the scales lanceolate to linear, not fibrillose, scattered; veins partly free and partly anastomosing irregularly; sori discrete or confluent with adjacent one where the veins anastomose, an indusial scale lacking. Additional dq examined. COSTA RICA. SAN 105Ё: hills at SW part of Montanas Jamaica, ca. 2.5 km ig of Bijagual de ¡A Carara Reserve, 460-520 , Grayum et al. 5495 (MO). PANAMA. CANAL ZONE: Pipeline pe from Gamboa, Armond 557 |" NW of Gam along “Pipeline Road," McAlpin 1396 (DUKE): n near Frijoles, Killip 2919(US); between Frijoles and Monte Lirio, 30 m, Killip 12147 (GH, US). coc o El Valle de Anton, 800 m, Alston 8803 (BM, M m, Alston 8848 (AAU, BM, CR, F, UC, US); Pargue Nacional Soberania, camino del Oleoducto, Rio Pilon Vásquez 209 (UC), 245 (UC). DARIÉN: Cerro Sapo, ca the W le of Cerro Pirre. 550-760 m, Croat 68882 90 Mar 4. Distribution of Stigmatopteris killipiana. (МО). PANAMÁ: along road to Alto de Pacora, 2 km N of Biological Station, km 56 Moran 3611 (MO, QCA), Evoy 44 (NY), Grayum & Zamora 9375 (MO, UC). PICHINCHA: Hotel Tinalandia, ca. 25 km E of Santo Domingo de Los Colorados, N side of Rio Toachi, 1,000 m, Moran 3562 (MO, QCA) Stigmatopteris killipiana grows in and along streambeds from Costa Rica to Ecuador (Map 4), from 30 to 1,000 m. It differs from the similar species S. pterorhachis and S. heterophlebia by its atropurpureous rachis with abundant scurf and by the greater number of free pinna pairs (5-13) beneath the pinnatifid apex. It further differs from S. heterophlebia by its nonpuberulent axes. 13. Stigmatopteris lechleri (Mett.) C. Chr., anske Vidensk. Selsk. Skr., Na- gopteris lechleri Mett., Fil. Lechler. 2: 25. 1859. TYPE: Peru. Puno: San Gabán, Lechler 2497 (holotype, B; fragments, BM). Figure 20. Map 5 Bommer & Christ, Bull. Herb.‏ — ا Boi 2, 4: 659. 1896 [Bull. Soc. Roy. Bot.‏ opteris cyclocolpa à 2. 1914. TYPE: Costa Rica. : Foréts de Tsaki, Talamanca, e m , Apr. А Tonduz 9480 (holotype, BR not seen; iso- , US, fragment B, i BM ex BR). CON елы, ecuadorensis C. Chr., Kongel. Daria Vidensk. Selsk. Skr., en Е. Afd. ser. 8, 6: 890 Annals of the Missouri Botanical Garden ك Я ў FIGURE 20. pinnae.— d. Bas 29. 1920. TYPE: Ecuador. “Andes of Quito," Sodiro holotype, B; isotypes, BM, P in herb. Christ). Polypodium punctatum Spruce in Hook., Sp. Fil. 4: 262. 862, nom. illeg., non (L.) SW. 1802. die uut prasinum Baker, Syn. Fil. 312. 186 а Danske Vidensk. Selsk Afd. . 7, 10: 79. 1913. LECTOTYPE (designate sd by (зан. 1913): dio San Martín: Tarapoto, Mt. Guayrapurima, Aug. 1856, Spruce 4719 (K; isolectotypes, BM, G, P; photos F, MO ex K). Rhizome erect-ascending; petioles ca. equaling the lamina, brown to tan, scaly, especially toward the base, the scales 8-20 x 1-2(-3) ceolate to linear, brown to blackish, the margins flat or enrolled, entire; lamina 2-pinnate-pinnatifid, 0.6–1.2 x 0.4-0.8 m, deltate or nearly so, bulb- lets absent; pinnae 25-40 x 5-15 cm, widest at the base, not overlapping the rachis; pinnules 4— 10 x 1-2.3 cm, narrowly lanceolate-triangular, sessile to broadly adnate and decurrent, lobe pinnatisect; lobes oblong to rectangular, entire to mm, lan- serrate apically; rachis and costae stramineous or brown, scaly, the scales linear or narrowly trian- gular, tortuous, fibrillose; veins unbranched or rare- ly (in the larger segments) branched, nearly reach- ing the margin; sori lacking an indusial scale. Additional specimens examined. COSTA RICA. LIMÓN: Canton de oe Amubri, margen izquierda del Rio Urén, Loma Sheub, 150 m, Herrera Ch. 2980 (CR, MO). Zona Protectora Ба W side of plateau separating headwaters of N fork of Rio Dantas from headwaters of AN - e с. MORAN еї ан lechleri (Moran 3582, MO). — a. Basal part of pinnule. — b. Petiole scales. — c. Medial al p о Barreal, bis Barbilla drainage, 600-660 m, Grayum et al. 7922 ( C, US). VENEZUELA. PORTUGUESA: Dtto. Sucre, Villa Rosa, 15 km SE de Bis Potrero, 1,400- 1, PORT, UC). COLOMBIA. CALDAS: Santa Cecilia, Cordillera Occidental vertiente Occidental, 800 m, von Sneidern 5046 US). EL VALLE: Digua Valley, Rio Blanco. 500 m, т. 7842 (ВМ). ло. cane SANTIAGO: Cor- да ега de Cutucú, western slopes, n trail from Logrono m, Madison et al. 347 a , ОСМЕ); Cantón Archidona, faldas S del Volcán Sumaco, carretera Hollin- Loreto, km 31, comuna challa Yacu, 1,200 m, Palacios 4030 (MO, QCNE); Moran 5119, MO, QCNE; ig ahi de Orella- na, cerca del Canon de los Monos en Potrero, 250 m Zak & Jaramillo 3635 (MO). PICHINCHA: a Quito- Reserva a ENDESA, Corporación Forestal Juan Manuel Durini, 800 m, bus 0011 (AAU); freq. in silvis de los а Oct. 1883, Sodiro s.n. (SI, UC); crescit in silvis tropicis ad San Miguel, Aug. 1875, Sodiro (fragment, B). PERU. HUÁNUCO: 25 km NE of Tingo María. La Divisoria, 1,700 m, Moran 3696 (MO, USM). Pasco: Provincia Oxapampa, Gran Pajonal, 2-3 km N of Che- quitavo, 1,200 m, D. N. Smith 5082 (MO) Stigmatopteris lechleri occurs from Costa Rica to Venezuela and Peru (Map 5), from 200 to 1,700 m. It is easily identified by its 2-pinnate-pinnatifid lamina, which is the most finely divided in the genus. The petiole scales vary considerably. Spec- imens from most parts of the range have linear or narrowly triangular scales with inrolled margins, but specimens from Colombia have wider lanceolate scales with flat (not inrolled) margins. Since inter- mediate specimens exist and no other character Volume 78, Number 4 Moran 891 1991 Stigmatopteris Monograph | 7 4 eae S | | P SS ча во МАР 5. Distribution of Stigmatopteris lechleri. correlates with the scale differences, the variation does not merit nomenclatural distinction. Christensen (1920) distinguished S. cyclocolpa and S. ecuadorensis from S. lechleri by the degree of pinnule adnation and cutting. However, these characters are variable, even on a single leaf as one goes from the more highly divided base to the less-divided medial regions. Apparently, Christen- sen saw only incomplete type material, which rep- resented the different degrees of cutting. 14. Stigmatopteris longicaudata (Liebm.) C. Chr., Bot. Tidsskr. 29: 300. 1909. Polypo- dium longicaudatum Liebm., Kongel. Vi- densk. Selsk. Skr., Naturvidensk. Afd. ser. 5, 1: 209. 1849. Dryopteris longicaudata (Liebm.) Maxon, Contr. U.S. Natl. Herb. 13: 18. 1909. SYNTYPES: Mexico. Veracruz: “Ва- ranca de Huitamalco,” Liebmann s.n. Шерт. Pl. Mex. 2694, Flora Mex. 737, 738, 739] (lectotype, 737, chosen by A. R. Smith, Fl. Chiapas, Part 2, 213. 1981: C—folio her- barium not seen; isolectotype, K). Figure 21. Map 6. Stigmatopteris palmensis Rosenstock, Repert. Spec. Nov. Regni Veg. 22: 12. 1925. ТУРЕ: Costa Rica. San José: La Palma, Brade & Brade 497 (holotype, S? not seen; isotypes, UC, US) Rhizome erect or creeping, the internodes ca. 1 cm, the scales 1-2 mm, black, rigid, entire, ovate-lanceolate; leaves 0.8-1.4 m; petiole 1⁄2 to equaling the lamina, brown or less often stramin- eous, the scales of two types, the first type 2-5 mm, opaque, appressed, shiny black, entire, ovate- lanceolate, the second type 4-9 mm, translucent, spreading, brownish, lanceolate; lamina 1-pinnate, ovate to subdeltate, lacking bulbils; pinnae mostly 14-20 x 2-2.5 cm, linear to narrowly lanceolate, 12-20 pairs, entire, crenate, or cut less than М(— V5) to the costae; pinnae confluent in the distal 1⁄2- уз of the lamina by the decurrent pinna bases, lower pinnae free, subsessile, or stalked, the stalk 2-5 mm; rachis and costae glabrous to densely scaly, the scales ovate to lanceolate, or (in Vene- zuela, Peru, and Bolivia) linear and fibrillose; veins free, or with a few casual anastomoses, 3-6 per segment; sori lacking an indusiumlike scale, medial to supramedial, filiform scales present among the 892 Annals of the Missouri Botanical Garden шә GI N 0 ми FIGURE 21. Stigmatopteris longicaudata. —a. Leaf (Moran 3206, MO). — b. Medial pinnae showing decurrent ). bases (Liesner & Stannard 16909, NY). — с. Venation and sori (Moran 3206, MO sporangia in plants from Venezuela, Peru, and Bo- Additional specimens examined. MEXICO. CHIAPAS: 10 km above Rayón Mezcalapa along road to Jitotol, Munici- m zd Rayón, 1,700 m, Breedlove & Smith 32672 (F, Н, NY); Monte Bello National Park, Municipio of La Hoi 1,300 m, Breedlove 35128 (F, MO); E of aguna Tzikaw, Monte Bello National Park, Municipio of La “Trinitaria. 1,300 m, Breedlove 35028 (MICH, MO, NY); 1 m ENE of Dos Lagos above Santa Elena, Municipio of La Trinitaria, 1,170 m, Breedlove 56494 ovirosa 500-6 Mickel 6369 (NY); 29 km S of Valle Nacional, 80 km N of Ixtlán de Juárez, Mickel 6392 (МУ); Dtto. Ixtlán, 7 km S of Vista Hermosa, 71 km N of Ixtlán de Juárez on Rte. 175, 1,645 m, Mickel & Pardue 6727 (NY); Dtto. Ixtlán, 2-3 km S of Vista Hermosa, 75-76 of Ixtlán de Juárez on Rte. 175, 1,615 m, mi р Pardue 6543 (NY, UC); road from Quia aca to Tuxtepec km 125-153, Hellwig 491 (NY), 492 (NY), 494 (NY): Ixtlan-Tuxtepec road, 35-39 km S of Valle Nacional, km 96-100, Mickel 1523 (MICH, NY, US); wa ae Sierra de Juarez between Tuxtepec ie Oaxaca, 21.5 m 5 of bridge at Valle Nacional, 5 of La Chiguz boldt, ca. Re m, Hallberg 1724 (NY). P PUEBLA: Limonateno, Municipio de ueytamalco, 1,000 m, Mic Е 484 (МІСН), Ventura A. 484 (NY). VERAC RUZ: 6 kin road N of Paz m iquez, Municipio Yecuatla, 1,200 = et al. 2 (F, NY, Zy Plan de Almansa, Municipio de Ес 600 m, Ventura A. 3306 (CR, NY); Lomas de Santa a Municipio de Yecuatla, Ven- шта A. 3630 (NY); camino Bastonal a Santa Marta, Municipio Catemaco, ттмен PD et al. 5457 (F Calavera, 5 km adelante de Atalen, Municipio de Atzalan, Calzada 5212 (F); 9 km E of Tebanca (9 km E of Lago Catemaco), 980 m, Nee & Schatz 19868 (Е); 12 km S 9 of Misantla, 1,350 m, Bohs et al. 1704 (GH); ca. 8 km Volume 78, Number 4 Moran 893 Stigmatopteris Monograph МАР 6. S of Misantla, ca. 750 m, Conant 792 (GH), qi Santa we Ds Misantla, Hahn 0, 69 (UC). GUATEM MALA. ALTA VERAPAZ: , Aldea San Chilasco, Guzmán-Valdez RUN Hochwald i in Pauy ‚350 m, von iui rue 3865 (GH, US). ect со: vic. of Maxbal, 17 mi. N of Barillas, 1,500 m, Stey- ermark 18744 (F, US» done Negro, 2 mi. E of Las Palmas, 1,600-2,000 m, Steyermark 51710 (F, bis HONDURAS. COMAYAGUA: hills above Varsovia, near tow of Cerro Azul, 900 m, Hazlett 3029 (MO). Costa RICA, ALAJUELA: ca. 20 km N of San Ramon, at the Univ. of San Ramón's Biological Field Station, cloud forest, 1,100 m, Moran 3206 (CR, F, GH, MO); Monteverde Cloud Forest Reserve, 1,520-1,580 m, Dryer 366 (F); Monte Verde Reserve, Pag" slope in valley or Rio Penas Blancas, 900-9 со а al. 14238 (MO). CAR- TAGO: са. 10 es E of Tapanti, Wilbur & Stone 8919 (DUKE); 16 km P 5 ros Stolze 1487 (CR, F, UC); dis ries 18778 (CR, MO, PORT, lverde, Gómez & Wagner 6856 (CR) (one Se ар Reserve, 1,500 m, T Distribution of Stigmatopteris longicaudata, the most widely distributed species in the genus. Croat 36176 (CR, Nes hpa i of Quebrada Casa Blanca Tapanti, 1,3 m & Sleeper 3683 (MO, ы bsy gui Ei Ms mountain from road into Tapanti Res 1 km $ of jet. of Quebrada Salto and Rio Gra on ЈЕ pe 1,500- | ,800 m ‚ А. В. Smith et al. ar the concrete bridge, 1,50 u 6773 (CR, F, GH). HEREDIA: Estación Biológica La Selva, at confluence of Rio Sarapiqui and Rio Puerto Viejo, Atlantic slope, 50-80 m, Grayum et al. 7910 (DUKE, МО); vic. of La Palma, on the road to La Hondura, 1,500- 1,700 m, Maxon & Harvey 8075 (US); Sardinal, 1,600 m, Gómez 2182 (CR, F, GH, NY, UC, US); ar Rio Peje and Rio Ardinalito, pose slope of Volcan Barva, 480-520 m, Grayum 6893 (MO); Zona Protec- tora, N slopes of Volcan E dec Rio Peje and Rio Guacimo, along Quebrada Cantarana, 190 m, Gra- yum & Schatz 3089 (DUKE); Zona Protectora-Q, Canta Rana Magsasay, 400 m, Chacón 852 (CR). SAN JOSE: PE unnamed N fork of Rio Zurqui, 1,500-1,600 m, A. R. Smith et al. 1678 (MO, NY, UC). PANAMA. BOCAS DEL TORO: vic. of Cerro Colorado, on trail along creek 894 Annals of the Missouri Botanical Garden 8. A mi. from Camp Chami, 1,400 m, McPherson 8885 O, UC); Cerro Colorado, ca. 8.6 mi. W of Chamé, Croat 69108 (MO, NY, PMA, UC, 05); along road N m N of Continental Divide, 700 m, Croat & Grayum 60215 (MO). CHIRIQUÍ: vic. of Cerro Colorado mine above San Felix, 1,450 m, McPherson 12035 (MO); above Boquete, aid iiu to Pate Macho on Continental Divide, 1,850- m, McPherson 12310E (MO); Canassa, 2,500 m, Killip 5444 (US). cock: El Valle, González 14 (МО); La Mesa, above El Valle de Antón, ca. 2 km W of Cerro dan 900 m, ir 37492 (MO, US); upper Caldera ershed, between “Сатр 1” and the divide, Holcomb's val е Е Res 1,650-1,925 m, Maxon 5647 (NY, 05); Holcomb's Trail, above El Boquete, 1,675- > 800 m, ae 5344 (GH, ee uas: Cerro Tute, m N of Santa Fé, 900-1,100 m, Moran 4048 (MO. STRI, Padi along Rio Dos a . 12 km beyond Santa Fé, 450 m, Croat 25818, (MO), 27736 (MO); 11 m from © ка Agricola Alto Piedra on road to Calo- vebora, 450 m, Croat 27541 (MO); 5 mi. W of Escuela Agricola Alto Piedra, 800-1,200 m, Croat 22998 (MO, NY); area of Santa Fé, road past Es ys Agricola Alto Piedra, 700 m, Churchill. et al. 5984 (M O, UC). FRENCH GUIANA. Tumuc Humac, à mi-chemin entre le Mitaraka Sud et le Toukouchipann, 480 m, de Granville 1291 (Z), ac, au du Mitaraka Sud, 400 m, de Granville 1434 (NY, P, Z). VENEZUELA. CARABOBO: las deras arriba de las cabeceras de Rio San Gián, E de , Steyermark AMAZONAS: Depto. Atures, lower forested E slope of 1,760 m peak, 8 km NW of settlement of Yutuje, 3 km Río Coro-Coro ‚ W of Serrani a ro Dougal et al. 3811 (MO, UC). EL VALLE: Cordillera Oc. cidental, vertiente occidental, Hoya del Río Digua, a Е La Cristalina, 1, 000-1, F, US); Cordillera с vertiente occidental, Hoya del Río Digua, lado izquierdo, Piedra de Moler, 900-1,180 m, Cuatrecasa s 15122 (Е, US). ECUADOR. PICHINCHA: NE of Vicente Maldonado, Reserva de ENDESA, 600 m, а 52431 (AAU, MO). CARCHI: wet plateau u above San Marcos de los Coai- єл Ё © а!. 57515 (АА tambo, Cosnipata Valley, Rio Tono, first foothill ridge on road N of Patria, 800 m, Wachter et al. 200 (F). HUANUCO: Tingo María, 700 m, Tryon & Tryon 5330 (BM, F, GH). JUNÍN: E of Quimiri Bridge, near La Merced, 800-1,300 m, Killip & Smith 23906 (NY, US); ridge E of Tingo María, 625-1,100 m, Allard 22558 (US). MADRE DE DIOS: Province Мапа, Cerro de Pantiacolla, Rio Palotoa, 10-15 km NNW of Shintuya, transect to ridge top, 700- 1,000 m, Foster et al. 10908 (F); Parque Nacional del Мапа, Cocha Cashu Biological Station, Foster P-84-91 (F). SAN MARTÍN: Mt. Guayrapurima, near Tarapoto, Spruce 4012 (BM, G, K, P). ucAYALI: Province Coronel Portillo, Sinchono, cerca de la Divisoria, 1,500 m, шс 861 (СН, USM); Sinchono, entre Tingo Мапа y к 1,500 m, Aguilar си плана МО COCHABAMB ince Carrasco, Antahuacana, es a Espirit 750 m, Buchtien 64 no 216 Рту cubi ns confluencia del por el camino a Suapi, al lado del Rio Khusillani, ca. 1 km antes la comunidad Khusillani, 1,500 m, Solomon 18469 (MO). Stigmatopteris longicaudata occurs from Mex- ico to French Guiana and Bolivia (Map 6), 220– 2,100(-2,500) m, and is the most widely distrib- uted species in the genus. Its most distinctive fea- tures are the pinnae, which are less lobed than most other species and, at least in the distal part of the lamina, have long-decurrent bases (Fig. 21b). The specimens from Peru and Bolivia differ from the rest by the linear and fibrillose scales (vs. ovate- lanceolate) on their axes and veins. No other dif- ference correlates with scale type. 15. S haelis (Baker) C. Chr., Bot. "Tidssk. 29: 300. 1909. Polypodium mi- chaelis Baker, J. Bot. 6: 164. 1877. TYPE: Ecuador. Pichincha: near Santo Domingo de los Colorados, Sodiro 2615 (holotype, К; iso- types, P, Q, SI). Figure 22. Map 7 Polypodium sylvicola Baker, J. Bot. 19: 205. 1881. teris sylvicola (Baker) C. Chr., Index Filic. 5. TYPE: Colombia. Antioquia: 1,000 m Kalbreyer 1807 (holotype, K). Rhizome short-creeping; leaves 30-70 cm; pet- ioles 25 to equaling the lamina, the base dull brown, becoming stramineous distally, scaly, the scales opaque, black, appressed, or membranous, tan, and spreading; leaves 0.3-0.7 m lon nate-pinnatifid, winged for 28-34 the length, nar- rowly deltate to lanceolate, the apex evenly atten- uate; pinnae (4.5-)9-11 x (1-)1.8-2.2 cm, pairs 17-20, widest at or slightly above the middle, lobed ca. halfway to the costa, the base of the distal pinnae adnate and decurrent, the base of the basal pinnae stalked, the stalk 1-3 mm; rachis and costae stramineous, scaly, the scales scattered; less than ; lamina 1-рш- 0.4 mm, filiform, the larger ones with several cilia at their base; veins free, unbranched, (3-)4-6(- 7) per segment; sori lacking an indusiumlike scale, paraphysate, the paraphyses ca. equaling the spo- rangia, uniseriate, multicellular, dark brown, spreading, somewhat tortuous. Volume 78, Number 4 1991 Moran Stigmatopteris Monograph 895 шо Gl 7 Yj Y 77 Е CN GAl NN а oper 22. Stigmatopteris michaelis. —a. Le ‚ b, McDonough et al. 371, MO. c Additional io ise examined. PANAMA. PANAMA 15 km from its beginning, 300 m, Ha uie , US). COMARCA DE SAN (MO). CHA: in silvis tropical prope 1882, Sodiro s.n ; NE of Vicente Maldonado, Re- serva de ENDESA, 600 m, Laegaard 52224, (AAU). Stigmatopteris michaelis occurs from Panama to Ecuador (Map 7) on steep clay banks along streams, from 300 to 1,000 m. It is distinctive by the combination of long-decurrent pinna bases, pin- nae lobed about halfway to the costa, and paraphy- sate sori (Fig. 22). The paraphyses are easily over- looked because they resemble stalks of the af. —b. Base of medial pinnae. — c. Paraphysate sorus. —d. Medial , d, Sodiro s.n., sporangia, but they can be distinguished by their arker color, slightly tortuous habit, and one-celled (not three-celled) apex. In general, S. michaelis has a smaller leaf than other species in the genus. The derivation of the specific epithet is unknown. 16. Sti hrodioides (Klotzsch) C. Chr., Bot. Tidsskr. 29: 299. 1909. Polypo- dium nephrodioides Klotzsch, Linnaea 20: 384. 1847. TYPE: Venezuela, Karsten 9, Coll. II (holotype, B). Figure 23. Map 8. Rhizome short, ascending; leaves to 2 m; petioles ca. l^ to equaling the lamina, tan to brown, the scales 3-13 х 1-2 mm, lanceolate, entire; lamina 1 -pinnate-pinnatifid, lanceolate, bulblets lacking; Annals of the Missouri Botanical Garden m o So e . C ^ ч ^ ' ojo $ o | o ° : "d o MUN LJ З К ` - o Sn LJ ^ i o 7 " d К L4 n " o -7 4 o 2 o 80 i. "md MaP 7. Distribution of Stigmatopteris michaelis. pinnae 18-25 x 2.1-3.5 cm, cut 4—2 to the costa, or rarely more in enlarged basal pinnae, the basal ones stalked, the distal ones gradually be- coming sessile, then adnate, the lobe of the medial pinnae often overlapping the rachis, the basal acroscopic lobe often prolonged; asal basiscopic rachis and costae moderately scaly, the scales lan- ceolate; lobes rounded, entire; veins 4-8 per lobe, unbranched; sori lacking an indusial scale. itional specimens examined. VENEZUELA. en and Ocumare de la Costa, Henri Pittier National Park, 3 k . Cordillera Interior, entre El Pauji y El Socorro, hacia la e Consejo, 1,350-1,400 m, Steyermark & Stoddart 118173-A (GH, VEN). CARABOBO: ca. 18- 20 km S of Puerto Cabello, ca. 9-14 km S of San Esteban on old trail to Valencia, 0-4 km S of colonial bridge, 350-700 m, Liesner & Medina 13739 (MO, UC, VEN); arriba de las cabeceras del Rio San Gián, E de Los Tan- ques, S de Borburata, Steyermark & Steyermark 95406-А (F, GH, NY, US, VEN). DISTRITO FEDERAL: old road be- tween Portachuelo and Penita (Petaquire and Carayaca), between Colonia Tovar-Junquito road and Hacienda El Limon, 6-8 mi. below junction of Junquito-Colonia Tovar road, mE & Nevling 95914 (GH, NY, VEN). FALCÓN: Sierra de San Luis, arriba de Sta. Maria, 1,200 m, van der Werff 3488 (MO, UC). MIRANDA: Dtto. Paéz, Fila La Tigra, Quebrada San Juan, 18 km SO de Cüpira, 650 m, Ortega & González 369 (PORT, MO, UC). YARACUY: nis road from Salom to Candelaria, 8-8.5 km above Salom Central Plaza, 1,260-1,290 m, Croat 60765 (MO, NY, UC); Dtto. Bruzual, Montana де Maria Lionza, Quebrada Quibayo, 250-1,000 m, Steyermark et al. 125086 (UC); Cabeceras de | Quebrada Amparo, El Amparo hacia Candelaria, a 7-8 km N de Salom, 1,220-1,250 m, Steyermark & Espinoza 111174 (NY, VEN); Dtto. Nirgua, El Amparo, 7-11 km N de Salom, 1,200-1,300 m, Steyermark & Espinoza 111493 (NY, UC). s nephrodioides is endemic to НЕ. central Venezuela (Мар 8), where it occurs in wet forests from 650 to 1,450 m. A good char- acter to identify this species is the basal basiscopic lobe of the medial pinnae that overlaps the rachis. (Fig. 23a). The only other species of Stigmatop- teris with pinna bases that overlap the rachis is 5. contracta, a Costa Rican species that differs great- ly in leaf cutting (Fig. 12). The name S. alloeoptera (Kunze) C. Chr. may apply to S. nephrodioides, but I have not seen the type and therefore cannot be certain (see section on names of uncertain ap- plication). 17. Stigmatopteris opaca (Baker) C. Chr., Kongel. Danske Vidensk. Selsk. Skr., Na- turvidensk. Afd. ser. 7, 10: 78. 1913. Menis- cium opacum Baker, J. Bot. 15: 166. 1877. Phegopteris opaca (Baker) Christ, Farnkr. Erde 270. 1897, illegit., non (Sprengel) Mett. 1858. Dryopteris christii C. Chr., Index Filic. 257. 1905, a nom nov. for Meniscium opa- cum with the same type, non Dryopteris opa- ca (Don) C. Chr. 1905. TYPE: Ecuador. Sodiro 54/3 (holotype, K; isotypes, P, UC; photos, BM, GH, US ex K). Figure 24. Map 8 Polypodium oligophlebium Baker in Hook. & Baker, Syn. Fil. . 1874, hom. illeg., non Kunze 1850. Phegopteris oligophlebia Salomon, Nomencl. 276. 1883, nom. РАК oligophlebium type. e ا‎ , Index Filie. 283. 1905, nov. for Polypodium oligophlebium Baker, with e same type. e Peru. San Martín: Tarapoto, Spruce 4653 (holotype, K). Polypodium coalescens Baker, J. Bot. 15: 164. 1877. ryopteris coalescens (Baker) C. Chr., Index Filic. 258. 1905. TYPE: Ecuador. Pichincha: near San Miguel de los Colorados, Aug. 1875, Sodiro 54/2 (holotype, K). Rhizome erect, the internodes approximate, scaly at the apex, the scales 1-3 mm, lanceolate, opaque, brown, rigid, entire; leaves to | m; petiole Y2 to equaling the lamina, dull brown on both surfaces, scaly, the scales of two types, the smaller type 5 mm, opaque, appressed, shiny black abaxially, Могап 897 Stigmatopteris Monograph Volume 78, Number 4 1991 FIGURE 23. pinnae showing the basiscopic | Medial pinnae. entire, ovate-lanceolate, slightly falcate, the larger type 4-9 -4 mm, translucent, spreading, brownish, ovate to lanceolate; lamina l-pinnate, opaque, the internal glands not visible, oblong- deltate to lanceolate, bulblets lacking; free pinna pairs 4-10; medial pinnae 12-27 3-5 ет, widest at or just above the base, the margins entire to crenate, the base sessile, the upper pinnae with a short decurrent base; basal pinnae ca. equaling the suprabasal pinnae, widest near the middle, ses- sile; rachis brown, the costae stramineous, both axes lacking capitate-glandular hairs, scaly, the scales lanceolate to linear or amorphous; veins reg- Stigmatopteris nephrodioides, Funck & Schlim 411 (P; lamina apex, GH). obe overlapping the rachis.—b. Distal half of the lamina.—c. Lobes of pinna.—d. 15 cm Fir ihn 7 1 1 / р PC MORAN 1989 i] E —a. Base of medial ularly to irregularly anastomosing, those of the sterile leaf with a single, often straight, excurrent vein projecting from the apex where the two lateral veins join; sori variable in shape, round, oblong, or arcuate, lacking an indusial scale. Additional specimens examined. ECUADOR. COTOPAXI: Quevedo-Latacunga road, km 46 from Queve- do, 600 m, Holm-Nielsen et al. 2906 (AAU, NY, UC). long road Mendez-Morona, 650 m, van der Werf & Gudiño 11169 (AAU, Е, МО, NY, UC). маро: Cordillera Guacamayos, on road Baeza- Tena, ca. 34 km from Baeza, са. 2,000 m, Mllgaard et al. 35859 (AAU); vic. of Jaguar Hotel, Rio Napo, Proctor M NA=SANTIAGO: 898 Annals of the Missouri Botanical Garden воҳ МАР 8. triangles). 387 12 (МУ); ca. 10 km SE of Coca, 250 m, Harling et al. 19737 (UC); camino entre Puyo-Tena, casi 10 km S de Rio Маро, 500 m, Moran 3576 (MO, ОСА). PASTAZA: 2 km W of Yuralpa, 440 m, Holm-Nielsen & Jeppesen 992 (AAU, МУ); Pastaza Canton, 20 km al sur de la población de curary, Zak & Espinoza 5287 (MO, QCNE). PICHINCHA: near San Miguel, Sodiro s.n., Jan 1882 (P, US); Peripa flumine, 300-400 m, а Е 174 (Р); Toachi, Sodiro s.n. (SI; МЕ of Vicente Maldo, Reserva de ENESA, 60 (AAU, MO). ZAMORA-CHINCHIPE: 1,250 m, Ollgaard et al. 90435 (AAU, MO). pteris opaca occurs in Ecuador and Peru kie 8), from 250 to 2,000 m. It is extremely variable in venation and shape of the sori. Although Distribution of Stigmatopteris opaca (dots), 5. pterorhachis (diamonds), and 5. nephrodioides (open the veins are always anastomosing, they may do so irregularly (as in the other species with anas- tomosing veins) or they may do so regularly, with the veinlets straight and producing a single excur- rent veinlet where they unite. This latter type of venation is similar to that found in Thelypteris subg. Meniscium. The sori vary from round to (in the type) arcuate, but some intermediate specimens have shapes between these extremes (Fig. 24b, c, e This species may be distinguished from 5. het- erophlebia by its nonpubescent axes and greater number of pinna pairs. It can be distinguished from Volume 78, Number 4 |. Moran 899 1991 Stigmatopteris Monograph 1 cm FIGURE 24. Stigmatopteris opaca. —a. Leaf (van der Werff & Gudiño 11169, MO). —b. Mixed round and arcuate sori (Moran 3576, МО). — c. Arcuate sori only (Sodiro s.n., К). — 4. Petiole scales (van der Werff & Gudiño 11169, MO).—e. Venation of a sterile pinna, showing single excurrent veinlet from vein junctures (Holm-Nielson et al. 2906, AAU). Annals of the Missouri Botanical Garden "= ڪڪ >= 7 RG AA e ==> === ho >” C W ~ АА, NL. ©? Volume 78, Number 4 1991 Moran 901 Stigmatopteris Monograph S. killipiana by its brown (not atropurpureous) petiole and rachis, thicker lamina, and tendency for the sori to become arcuate. The lamina is thick- er and the glands less visible than in the other species of Stigmatopteris. 18. Stigmatopteris pellucidopunctata (C. Chr.) C. Chr., Bot. Tidsskr. 29: 304. 1909. Polypodium macrophyllum Hook., Sp. Fil. 4: 241. 1862, hom. illeg., non (Blume) Mett. 1856. Phegopteris macrophylla (Hook.) J. Smith, Hist. Fil. 233 Mett. 1859. Dryopteris pellucidopunctata C. Chr., Index Filic. 283. 1905 for P. macrophyllum Hook. TYPE: Peru. San Martin: Mt. Guayrapurima, Aug. 1856, Spruce 4720 (holotype, K; isotypes, BM, G, P, US; photos, Е, GH, MO, NY, UC ex BM). Figure 25. Map 9 5, hom. illeg., non , nom. nov. Rhizome erect or short-creeping; leaves to 2 m long; petiole / to equaling the length of the lamina, brown, scaly, the largest scales 4-6 x 2-2.5 mm, lanceolate, falcate, occasionally with darkly scleri- ed abaxial surfaces, the base cordate (appearing false peltate), with smaller fibrillose scales present, ish; lamina l-pinnate-pinnatifid, lanceolate, lacking buds; pinnae 20-40 x 2.5-6 cm, 15-20 pairs, cut са. 2% to the costa, sessile or the basal ones short-stalked, the stalk 2-7 mm; пеп 4—10 mm wide, serrate at the apex, the sid я serrate or rarely lobed, the basal basiscopic seg- ment not reduced; rachis and costae stramineous or brown, scaly, the larger scales nonbullate, lan- ceolate to linear, mixed with smaller fibrillose ones; veins free, 7-14 per segment, l-forked with a short acroscopic branch that in fertile segments bears the sorus; sori lacking an indusiumlike scale. unbranched or Additional spec vip examined. COLOMBIA. ANTIOQUIA: Municipio San Carlos m E of San Carlos, Lake Punchina, Quebradas та Уша y El Macho, Вгат et al. 1679 (MO). SANTANDER: ca. 25 km E of Puerto Wilches, E side of Magdalena Valley, 70-130 m, Elias 8 (05); vic. of Barranca Bermeja, Magdalena Valley, between Schlim 1683 (BM). ECUADOR. MORONA-SANTIAGO: along new road Mendez-Morona, 650 m, van der Werff & Gudino 11170 (MO, QCNE, UC). PASTAZA: Rio Pastaza, between Destacamento Chiriboga and Apachi Entza, 285 10 - =й й M ` Ñ : К хае JJ x o M; : N » `Җ. Mp МАР 9. Distribution of Stigmatopteris pellucidopunc- tata. m, Øllgaard et al. + (AAU). NAPO: 27 km SE of Coca, Pozo Auca 4, 200 m, Moran 3624 (MO, Q, QCA) Punt. AMAZONAS: б de Bagua, aa Tambillo above Cascadas de Mayasi 5 25 m, Wurdack 2002 (GH, US, USM). HUÁNUCO: Fundo Chela, Sinchono, Агиш lar 927 (USM). MADR 10S: Provincia Мапа, Atalaya, Provincia Oxapampa, Paujil, near Puerto Bermudez, León et al. 311 (USM). BOLIVIA. LA PAZ: Provincia Sud Yungas, Chulumani, 107 km NNE pasando Asunta, Alto Charia sobre el Rio San José afluente, Rio Boopi, 900 m, Beck 8500 Stigmatopteris pellucidopunctata grows from Colombia to Bolivia (Map 9), from 100 to 900 m. It is closely related to 5. sordida and S. ichthiosma but may be distinguished by linear to fibrillose scales on the costae and costules (Fig. 25a). The stiff, adaxially sclerified, appressed petiole scales of S. pellucidopunctata are of slightly different shape than those of S. ichthiosma, being lanceolate and falcate rather than ovate and symmetrical. In Ecuador, S. pellucidopunctata occurs only on the eastern side of the Andes, whereas S. ichthiosma occurs only on the western side. — FIGURE 25. from the abaxial surface of the lamina. — b. Leaf.— Stigmatopteris ПА — a. Proscales ог microscales B and multiseriate scales . Basal portion e with of medial pinna.— d. Rachis and costa linear-lanceolate scales.— e. Petiole ide a, b, d, e, "Moran 3624, MO. c, van pus 7 erf & Gudino 11170, MO. 902 Annals of the Missouri Botanical Garden ^ c MORAN 1137 ЈЕ FIGURE 26. Stigmatopteris prionites. —a. Leaf (Blanchet 2483, ВМ). —b. Venation and sori (Martius 305, B). — c. Fertile pinna (Martius 305, B). Volume 78, Number 4 991 Moran Stigmatopteris Monograph 903 МАР 10. Distribution of Stigmatopteris rotundata (do 19. Stigmatopteris prionites (Kunze) C. Chr., Tiddskr. 29: 298. 1909. Dryopteris prionites (Kunze) C. Chr., Index Filic. 286. 1905. Polypodium prionites Kunze, Flora Beibl. 1: 29. 1839. TYPE: Brazil, from three states, see discussion below: Martius 305 (lec- totype, here designated, B; isolectotypes, BM, BR not seen, С, К, L, MO, NY; photo, BM ex BR). Figure 26. Map 10. Stigmatopteris prionites var. denticulata (Fée) C. Chr., Bot. Tidsskr. 29: 299. 1909. re denticu- lata Fée, Crypt. Vasc. Brésil 1: 32, fig. 2. 69. TYPE: Brazil. Rio de ane Зи de Santa ts) and 5. prionites (diamonds). Anna, Glaziou 2400 pro parte (holotype, P; isotype, P). Rhizome short-creeping; leaves to 1.5 m; petiole ca. equaling the lamina, stramineous to brown, scaly at the base, the scales 4-8 x 1-2 mm, brown, thin, ceolate; lamina 1-pinnate-pinnati- fid, lanceolate, lacking buds, only a few distalmost pinnae with short-decurrent bases; pinnae 11-20 X 1.1-2.3 cm, cut 14-14 to the costa, pairs 19- 26, the base sessile or nearly so; segments entire to denticulate at the apex, separated by broad, U-shaped sinuses; rachis and costae stramineous, 904 Annals of the Missouri Botanical Garden sparsely scaly, the scales ovate to lanceolate; veins free, not forked, 3-5 per lobe; sori inframedial; indusial scale absent. Additional specimens examined. BRAZIL. BAHIA: no locality, Luschnath 120 (В), 1836, Blanchet 2483 (BM, С, P). ESPÍRITO SANTO: Jatiboca, Brade 18405 (С, GH, S GERAIS: Juiz de Fora, Rio do Peixe, Kr rueger 17 107 (UC) Vicosa, pease Riberro, Agricul- ] а (UC); Distrito b (F, GH, K, NY, P, S, Z). STATE UNKNOWN: Me Parahyba, 1886, Galdi а, s.n. Ub Brasilia, Riedel s.n. (US); no locality, Glocker 32 (BM, US). Stigmatopteris те is endemic to south- eastern Brazil ( 10) where it grows in wet forests from 500 to о 740 m. It can be distinguished from all other Sti pteris in Brazil by the combination of narrow (1.1-2.2 cm) pinnae cut up to halfway to the costa, broad, obtuse lobes with wide sinuses, few (3—5) veins per lobe, and infra- medial sori. The fertile pinnae are slightly narrower than those of the sterile. Variety denticulata differs by having more deeply lobed pinnae with the lobes denticulate at the apex. Because this character intergrades with plants having entire lobes, the variety is here placed in synonymy. The type collection has a varied provenance. Although none of the specimens have a more spe- cific locality than "Brazil," the original description by Kunze states that Martius 305 comes from the provinces of Bahia (Porto Seguro), Espirito Santo, and Rio de Janeiro. Therefore, I chose the sheet at Berlin (B) as the lectotype. 20. Stigmatopteris pterorhachis R. C. Mo- ran, sp. nov. TYPE: Colombia. Magdalena: Santa Marta, Sierra del Libano, 1,800 m Smith 1045 (holotype, MO; Toe, ВМ. DUKE, Е, GH, MICH, NY, P, UC, US). Map 8. Folia 1-pinnata, pinnis 4-7 utroque costae latere, gem- mis absentibus; rhachis costaeque abaxialiter glabrae; ve- nae irregulariter anastomasantes Rhizome short-creeping, the internodes 0.5-2 cm, scaly at the apex, the scales 1-3 mm, ovate to lanceolate, opaque, brown, rigid, entire; leaves 50-110 cm; petiole ca. equaling the lamina, dull brown or tan on both surfaces, lacking hairs, scaly, the scales 4-9 mm, translucent, spreading, brown- ish, lanceolate; lamina 1-pinnate, oblong-deltate, bulblets lacking; free pinna pairs (3-)4- 7; basal pinnae usually somewhat reduced, widest near the middle, stalked, e stalk 2-4 mm; medial pinnae (10-)15-20 x 2.5-3.5 cm, widest at or near the base, the margins entire to deeply crenate; distal pinnae with long-decurrent bases; rachis and costae tan, light brown, or stramineous, lacking hairs, scaly, the scales to 1 mm, 1-2 cells wide, fila- mentous, usually with a long central filament and several shorter ones branched from the point of attachment; veins part free and part anastomosing, the areoles with 1 or 2 curved, irregular or forked excurrent veinlets; sori discrete, indusial scale ab- sent. Additional specimens examined. VENEZUELA. LARA: Dtto. Moran, 1,500 m, Rivero et al. 1602 (PORT, UC); Dtto. Menez, between Cubiro and Escalera, 10-15 SW of Cubiro, 1,600-2,000 m, Steyermark et al. 110227 (Е). PORTUGUESA: Dtto. Sucre, Los Paramitos, а 20 km por aire al SO de Biscucuy, 1,000-1,500 m, Ortega et al. 1822 (MO, PORT); Dtto. Guanare, ESE of Paraiso de Chabasquén, along road to Cordoba, ca. 27 minutes from Chabasquén, 1,500 m, А. К. Smith 1031 (AAU, CR, MO, PORT, UC, 2). COLOMBIA. MAGDALENA: Sierra Nevada, Rio Macha, Schlim 856 (K, P); Sierra Nevada Santa Marta, finca Los Arroyitos, 1,600-1,700 m, Kirkbride 2377 (NY, UC). NORTE DE SANTANDER: Ocana, Schlim 596 (BM, C, K) a c Stigmatopteris pterorhachis occurs in north- western Venezuela and northern Colombia (Map 8) where it grows primarily in cloud forests from 1,000 to 1,800 m. It greatly resembles S. heterophlebia, but differs by the absence of capitate glandular hairs on the axes, the presence of 1- or wide, filamentous, basally branched scales on the costae, and the greater number of pinna pairs. -cells- These characters correlate on all specimens except Rivero et al. 1602, which is moderately to sparsely glandular on the abaxial surface of the rachis, but lacks glands on that surface of the costae (unlike S. heterophlebia). The specimen is typical of S. pterorhachis in the other two characters, which separate it from S. heterophlebia. The name 5. alloeoptera (Kunze) C. Chr. has been previously misapplied to this species and to S. heterophlebia (see section on names of uncer- tain application). 21. Stigmatopteris rotundata (Willd.) C. Chr., Bot. Tidsskr. 29: 297. 1909. Aspidium ro- tundatum Willd., Sp. Pl. ed. 4, 5: 247. 1810. Dryopteris rotundata (Willd.) C. Chr., Index Filic. 289. 1905. TYPE: Plumier, Traité Foug. Amer. 29, t. 38. 1705, illustrating a plant from Martinique. Figure 27. Map 10. Mp Ji AR RP A Kaulf., Enum. Filic. 108. Ha 1 е аа (Kaulf.) Те а аи 866. YPES: Martinique, Sieber 350 B. MO, ТА "BM. F, G, GH, MO, NY, P, UC ex 1); Martinique and Montserrat, Ryan s.n. (P-herb. Vahl). Volume 78, Number 4 Moran 905 1991 Stigmatopteris Monograph 15 cm | R Moran 1987 7 AND 2; ^ ул 24 ДУЛУ = RES | EU | 3 ст FicunE 27. Sti ا‎ as rotundata. —a. Venation and sori (Morton 6184, US).—b. Leaf (Morton 6184, US).— c. Costal scales (Alston 494, NY).—d. Costal scales о unknown, K-H1058/87).— e. Costal scales (Duss 4074, NY).—f. Fertile pinna (collector unknown, K-H1058 Stigmatopteris rotundata var. trinidadensis C. Chr., 6259 (US) [these two numbers equal the Botanic Bot. Tidsskr. 29: 297. 1909. SYNTYPES: Trinidad. Garden Herbarium of Trinidad nos. 337 and 1236, Fendler 98 (B, BM, GH, K, MO, UC, US); Cabueres respectively] р | де Заггаро, 11 Моу. 1863, Hart 3707 (US, МУ), Nephrodium imrayanum Hook., Sp. Fil. 4: 86, t. 242. 906 Annals of the Missouri Botanical Garden 1862. Aspidium imrayanum (Hook.) Fée, Mém. Foug. 11: 79. 1866. Dryopteris imrayana (Hook.) Domin, € Kral. Ceské Spoleén. Nauk,-Tr. Mat.- Prír. N.R., 21: 191. 1929. TYPE: Dominica. Imray E b K; isotype, BM; photos, GH, US ex M). Rhizome short-creeping, the scales 2-4 mm, brown, entire, narrowly lanceolate; leaves to 2 m; petiole Yê to equaling the lamina, brown to tan, the scales all of one type, 5-12 x 1-2 mm, lanceolate, translucent, never blackish or thickened, entire, brown; lamina l-pinnate-pinnatifid, oblong to ob- long-lanceolate, lacking bulbils; pinnae 11-26 x 1.5-2.5(-3) cm, -32 pairs, linear or rarel narrowly lanceolate, crenate or cut less than Уз(– V4) to the costa; upper pinnae adnate to the rachis and confluent only in the distal 4—8 of the lamina, the lower pinnae subsessile or stalked, the stalk to mm; rachis and costae scaly, the scales ovate to lanceolate, filiform, or roundish to ovate and ir- regularly lobed; veins free, 3-4 per segment; sori lacking an indusiumlike scale, inframedial to su- pramedial, those on the basal veinlets often at the apex of the vein. Additional specimens examined. GUADELOUPE. Morne du Calababbaye, Belanger 821 (G); no locality, Duss 80 (05); Bon du Marne de la Madeleine, Trois Rivières, Duss 729 (P); Bois de la Calebasse, 450-900 m, inis E 56 1 (NY, P), Bois de la case ‘аде де Vanchelet, 0 m, Duss 4074 (NY, Roche, Questel 2905 (US); Gran Etang, 350 m, 1104 (US); ravine ат. 680 m, Stehlé 1463 (US); Bains Jaunes, 1,050 m, Stehlé 1887 (US). DOMINICA. P); Sylvania estate, 4 UC, US); eue Hodge 104 (GH, US); Lisdara, 470 m, Hodge 237 H, NY, US); 1 mi. E of Laudat, ca. 665 m, Hodge A Hodge 1972 (GH); no locality, Imray 39(В); St. Раш Parish, 0.5 mi. S of Pont Casse, Lellinger 362 (MICH, US); St. Paul Parish, Lellinger 367 (DUKE, GH, US); St. Joseph Parish, S of Tiperie, N side of Morne Couronne, Lellinger 507 (05); Mt. Diablotin, Lloyd 902 along trail to Te m Wilbur 7882 n TINIQUE. No locality, Belanger 821 (BM Lorrain, Duss 1 ; 1 Py St. p Stehlé 3303 (US); Colson, 560 m, Stehlé JS); Riviére ur. кщ m, Stehlé 4883 (UC, US); Bois de la Piro dee A m, Stehlé 6725 : 500 7238 (US); vallée du Lorrain, 650 m, Stehlé 84 36 (MICH, US). Sr. Lucia. Mt. Gimie, Box 1966 (BM). Sr. . No locality, Checkley 7 (B); Charlotte pres Martine), Cooley 8325 (GH); no locality, Eggers 6869 (F, US); Ta Chateau- belair River, 300-4 а 5373 (СН, 05); пррег valley of Richmond River, 330-540 m, Morton 6184 ‚ К); Bois du (05); Eua locality, H. H. Smith & G. W. Smith 659 (B, BM а GRENADA. No locality, Eggers 6164 (F, UC, 5 St. Mark Parish, NW slope of Mt. St. Catherine, Proctor 17244 (MICH, US); no ps Oct.-May 1890- 91, Sherring s.n. (BM, К, US). TRINIDAD. Mount То- cuche, Britton et al. 1361 (F, NY. US); Morne Bleu, N via Santa Cruz, Broadway 6123 (BM, МО); Fendler 98 (B, BM, С, GH, K, MO, NY, UC, US); Morne Bleu, road near summit, Johnston 303 (BM); Morne Bleu, Arima- РМ а — d 104 (BM) Asa Wright Nature Cen of Arima, Mickel 9412 (NY, UC); no locality, dines 214 (MO). FRENCH GUIANA. Montagne de l'Inini, zone prd 150 m, Cremers 9113 pd 9140 (UC); Mt. Galbao, secteur Est, 3?36'N, 53?17'W, 600 m, de Granville 8873 (UC); Montis Dia- id in 1847, Leprieur 270 (P). Guyana. Ereugfaru, aiewac River, along the Santa Maria trail, Alston 494 K, NY) Potaro River, 1899, Jenman s.n. (NY). VENEZUELA. BOLÍVAR: región de Urimán, 760 m, Bernardi 938 (PORT); Uei- а ۰ Luepa and Cerro Vena- то, 1,100-1,300 m, Ste dei & Nilsson 314 (NY, VEN). SUCRE: península de Par , W de Cerro Humo, 700-750 m, Steyermark & Rabe 96115 (NY, VEN). BE Stigmatopteris d the type of the ge- o 1,300 m Antilles, Guianas, and eastern лије Proctor (1977) listed this species from Montserrat and Kra- mer (1978) listed it from Suriname (Versteeg 315), but I have not seen a specimen from those coun- nus, occurs from n the Lesser tries. The scales in S. rotundata are extremely vari- able. Most plants from the Lesser Antilles have filiform to lanceolate scales (Fig. 27d), but plants from Guadeloupe and (often) Dominica have 2-4- mm-long, broadly ovate, flaccid scales (Fig. 27e). Plants from South America have scales that are ess than 1 mm long, roundish to ovate, and ir- regularly lobed (Fig. 27c). Because no other char- acteristics other than geography correlate with scale type, I feel it is best only to mention this variation — 2) rather than formally name it. 22. Stigmatopteris sordida (Maxon) С. Chr., Index Filic. Suppl. 3: 175. 1934. Dryopteris sordida Maxon, Contr. U.S. Natl. Herb. 24: 60. 1922. TYPE: Guatemala. Alta Verapaz: near Cubilquitz, 350 m, von Túrckheim 1432 (holotype, US; isotype, US). Figure 28. Map LL S litoralis Rosenstock, Repert. Spec. Nov. g. 22: 12. 1925. TYPE: Costa TA Limón: Finca Gebr. Hundrisser, Brade & Brade 353 (ho- lotype, S? not seen). Rhizome short, ascending; leaves to 1.3 m; pet- ioles Y2 to about equaling the lamina, stramineous, tan or brown, the scales 3-10 x 1-2 mm, ovate Volume 78, Number 4 1991 Moran Stigmatopteris Monograph ү \ ЈЕ x | 1 ст IGURE 28. Е Stigmatopteris sordida. —a. Leaf. —b. Venation and sori.— c. Leaf.—d. Venation and sori Moran 3607, MO. c, d, Moran 3115, MO. 908 Annals of the Missouri Botanical Garden -afls 190 ъч nd q E EN ` ~ x TON \ оо > EE q ° ERES МАР 11. Distribution of Stigmatopteris sordida. to lanceolate, or lanceolate with an attenuate apex, entire, dark brown or rarely (in Colombia) yellow- brown; lamina 1 -pinnate-pinnatifid, ovate, lanceo- late, or narrowly long-triangular, buds lacking, only the distalmost pinnae with short-decurrent bases; pinnae 15-28 x 2-4 cm, usually cut between уг— 34 of the way to the costa or sometimes more in the enlarged basal pinnae, pairs 16-25, the basal ones stalked, those above gradually becoming ses- sile then adnate, the basal basiscopic lobe not over- lapping the rachis (although sometimes adnate to it), the basal acroscopic lobe often prolonged; lobes rounded, entire or serrate at the apex; rachis and costae stramineous or the rachis brown, or sparsely scaly, the scales ovate to lanceolate; veins generally 7—10 per lobe, unbranched or rarely with a short acroscopic branch; sori medial, lacking an indu- siumlike scale. Additional specimens examined. MEXICO, CHIAPAS: Finca Mexiquito, Purpus 6870 (BM, F, GH, MO, NY UC, US); no locality, Purpus 7256 (GH). OAXACA: Био. Ixtlán, 29 km S of Valle Nacional, 80 km N of Ixtlán de Juárez, trail E of Rte. 175 at a Vista gi toward Ladú, 500-850 m, Mickel 6403 (МУ, Ut МАГА. SUCHITEP бом pe | slope of Mn Ati- Mocá, Guata 0 m, dd h & Wilson 346 (US). HONDURAS. COMA ҮА JA: E 'arsovia,' 24 km SE del lago de Yojoa. "600 m, Claros 7 73 (MO). - NICARAGUA. ZELAYA: Costado sur del Cerro La Pimienta y N del Cerro Hormiguero, a Orilla del Caño Hormiguero, 800-900 m, Grijalva 290 (MO). COSTA RICA. ALAJUELA: near Rio San Rafael, 2 km W of La Marina, Llanura de San Carlos, 550 m, Molina R. et al. 17321 (CR, NY); NW of Zarcero, 850 m, Croat 43598 (CR, MO); W of Marina, 500 m, Burger & Stolze 5055 (CR, F, MO), 5057 (CR, F, GH, NY, US); between San Lorenzo and Los Angeles de San Ramón, above Río San Lorenzo, 620 m, Burger & Antonio 11179 (CR, Е); Rio Chiquito, ca. 40 km on road to Upala, 800 m, Gómez 18641 ( UC). cARTAGO: E of Turrialba, 900 m, Lellinger & White 1428 (CR, MO, US). HEREDIA: Finca La Selva, Puerto Viejo de Sarapiqui, ca. 60-80 m, Grayum 2641 (UC); Cerro Sardinal, ca. 2-2.5 km N of С cs La Martita), 80- ide m, А. R. ied et a 1813 1 d Ко Guac yu Station, 100 m, E 11495 (MO), 12382 (CR, MO). LIMÓN: Dist. Talamanca- Canton, ca. 3 km SW of Suretka, near RECOPE test "dri site, 70 m, Moran < 31 15 (CR, F, 45 Mic А o Colorado Island, Kenoyer 34 (US); o Island, Knight s.n. (US); Parque ЕЯ Soberania, camino del Oleoducto, и 188 (М О, 05). COCLÉ: road to Coclesito, logging camp, 12 mi. from Llano d Churchill et y 4089 (MO): El Valle a a 1,000 m, Alston 8744 (BM). COLOMBIA. ANTIOQUIA: Mu- nicipio de Zaragoza, е де =н аш шагда 500- Volume 78, Number 4 1991 Moran 909 Stigmatopteris Monograph 650 m, Soejarto & Villa 2685 (COL, GH). ECUADOR. LOS RÍOS: Rio Palenque Science Center, 56 km vedo, Morar, 3607 (MO, ОСА, QCNE), Grayum & Zamora 9376 (MO). PICHINCHA: NE of Vicente Maldona- do, Reserva de ЕМРЕЗА, 600 m, Laegaard 51610 (AAU). Stigmatopteris sordida occurs from southern Mexico to Ecuador (Map 11) from 70 to 1,450 m. Maxon (1922), in his original description of this species, stated that the ultimate segments are never serrate, but the type has some serrations on the sterile segments. Specimens that have been col- lected since 1922, from Central and South Amer- ica, show that the segment margins vary from entire to serrate. The specific name is derived from the Latin word sordidus, meaning dirty-looking, dingy, or soiled. I am uncertain what this refers to. Although I did not see the type, 5. litoralis is put here in synonymy because of the original de- scription and because the type was collected from the Atlantic lowlands of Costa Rica where S5. sor- dida is common. Stigmatopteris longicaudata also grows in the Atlantic lowlands but can be excluded from consideration because the original description of S. litoralis stated that the pinnae are cut % to the costa and only the distalmost pinnae are short decurrent— characteristics unlike 5. longicauda- ta. 23. Stigmatopteris tyucana (Raddi) C. Chr., ot. Tidsskr. 29: 298. 1909. күч tyucanum Raddi, Opusc. Sci. 3: 288. 1819. Dryopteris tijuccana (Raddi) C. Chr., iis Filic. 298. 1905. TYPE: Brazil. Rio de Janeiro: Rio de Janeiro, Сама, Mt. Пјиса, Raddi s.n. (holotype, FI or PI not seen). Figure 29. pu Wc yneuron Fée, Crypt. Vasc. Brésil 1: 133. хс : Brazil. Bahia: Blanchet s.n. (P? not en). Bons tenuis Fée, Crypt. Vasc. Brésil 1: 99, t fig. 1. 1869. TYPE: Brazil. Rio de Janeiro: Rio ү ае Glaziou 964 (holotype, P). Rhizome short-creeping; leaves to 1.2 mm; pet- iole ca. equaling the lamina, stramineous to brown, scaly at the base, the scales 5-12 x 1-2 mm, brown, thin, lanceolate; lamina 1 -pinnate-pinnati- fid, lanceolate, lacking bulblets, only a few distal- most pinnae with short-decurrent bases; pinnae 12-20 x 2.5-4 cm, 12-15 pairs, cut М-М to the costa, the base sessile or nearly so, the basal pinnae short-stalked; segments entire; rachis and costae stramineous, sparsely scaly, the scales ovate to lanceolate; veins free, not forked, 4-7 per lobe; sori medial to supramedial; indusial scale absent. Additional specimens examined. BRAZIL. RIO DE JANEIRO: Colonia Leopoldiana, 1836, Blanchet 2483 (BM); Terezópolis, Шпапћо, 450 m, Brade 9972 (GH, UC); Teresópolis, Brade 9977 (NY); Rio de Janeiro, Gaudi- chaud s.n. (С); Petrópolis, Cascatinha, Glaziou 7245 (B, K, P, US); no locality, Martius s.n. (B). S&o PAULO: Alto da Serra, Brade 5892 (NY, UC); Serra do Mar, 300 m, Wacket 50 (NY, UC); Alto da Serra, Parque Cajurü da estacion biologica, Wacket 21149 (BM, NY, UC), Lue- derwaldt 21149 (GH, UC). srATE UNKNOWN: near Rio de Janeiro and Bahia, Webb s.n. (MICH); no locality, Riedel s.n. (B, GH) Stigmatopteris tyucana is endemic to south- eastern Brazil where it grows in wet forests from 300 to 400 m. The specific epithet is usually spelled tijuccana. I prefer to spell it tyucana because that was Raddi's original spelling, and the Code (Greuter et al., 1988) recommends the use of the author's original spelling (when not in error). ough I have not seen the type, 1 am con- fident from Raddi's excellent illustration (1825, t. 37) and description that the name is here applied correctly. 24. Stigmatopteris ulei (Christ) Sehnem, Fl. Illust. Catar. 1ASPI: 100, est. 25, fig. 2. 1979. MAC diis i a Bull. Herb. Boissier sér. 2, 34 2. Dryopteris ulei (Christ) C. Chr., judex dde 299. 1905. TYPE: Brazil. Santa Catarina: municipio de Sáo Fran- cisco do Sul, Serra de Laranjeira, Ule 70 pro parte (holotype, P? not seen). Lamina 1 -pinnate-pinnatifid, apparently lacking buds; pinnae cut ca. 2 to the costa, the distal ones (presumably) lacking a basal basiscopic lobe adnate to the rachis; the lobes nearly entire, separated by broad, U-shaped sinuses; veins free, 5-8(-10) per lobe; sori inframedial. I have the type number from P of Ule 70, but the locality given on the sheet (im schattigen Urwald der Peninsula da Gloria, Feb. 1884) is different from that given in Christ's original description. The specimen is Stigmatopteris caudata. The only specimen that I have seen of 5. ulei is the pho- tograph of a medial pinna in Sehnem (1979). Be- cause І have not seen a herbarium specimen, the above description is incomplete. Stigmatopteris ulei 18 included in the key and the species descrip- tions, rather than the uncertain taxa section, to call attention to this poorly known species and because, on the basis of Sehnem's photograph, it appears distinct. Christensen (1913) thought that this species was probably a form of S. prionites, but S. ulei is intermediate in cutting— perhaps a hybrid?— be- 910 Annals of the Missouri Botanical Garden a \ шә Gl FIGURE 29. Stigmatopteris tyucana. —a. Leaf (redrawn from Када, 1825, t. 37; scale approximate).—b. Veins and sori.—c. Medial pinna (Brade 9977, NY).—d. Basal pinna (Brade 9977, NY). tween S. caudata and S. brevinervis. The inter- mediate cutting is evident in Sehnem’s (1979) pho- tograph, which has 5. ulei placed between 5. caudata and S. brevinervis. Sehnem (1979) cited the following specimen for S. ulei: Brazil. Parana: Paranagua, Morro de Tabaquara, terricola na mata pluvial de encosta de morro, 50-100 m, Hatsch- bach & Imaguire 16361. Obviously, it is difficult to assess whether S. ulei represents a hybrid with- out having specimens for study. EXCLUDED TAXA Dryopteris subg. Stigmatopteris, group Peltochlaena Fée ex C. Chr., Kongel. Danske Vidensk. Selsk. Volume 78, Number 4 1 Moran 911 Stigmatopteris Monograph Skr., Naturvidensk. Afd. ser. 7, 10: 74. 1913. = Cyclodium C. Presl (ide Smith, 1986). Stigmatopteris clypeata (Maxon & C. Morton) i Rd Proc. Biol. Soc. Wash. 89: 730. 1977. teris clypeata (Maxon & C. Morton) OMM ‘(Ade Kramer, Acta Bot. Neer]. 18: 141. dagmatoptens guianensis (Klotzsch) C. Chr., Index Filic. uppl. = Cyclodium guianense (Klotzsch) А. R. Smith (фе Smith, 1986). Ра e (Willd.) Kramer in Kramer & selaar, depen Nederl. Akad. We- ense. Pro oc. e 71: 521. 1968. = Cyclodium menis- es (C. быш `В. Smith (fide Smith, 1986). Жы paludosa (C. Morton) R. Тгуоп 4 A. on, Rhodora 83: 136. 1981. = Сан" те- niscioides wt paludosum (C. Morton) A. R. Smith (fide Smith, 1986). Stigmatopteris sancti- ни (Ноок.) С. Chr., Index Filic. Suppl. 3: 17 = Cylcodium guianense 6). TU varians (Fée) Alston, Kew Bull. 1932 | = Cyclodium varians (Fée) А. R. Smith (fide Smith 1986). NAMES OF UNCERTAIN APPLICATION Stigmatopteris alloeoptera (Kunze) C. Chr., Bot. idsskr. 29: 300. 1909. Polypodium allodopterum Kunze, Linnaea 25: 748. 1835. TYPE: Venezuela. Distrito Federal: near Galipán, 1,200-1,500 m, Sep., Wagner 460 (LZ — destroyed). This name does not represent S. heterophlebia or S. pterorhachis (previously treated as one), to which it was applied by all previous pteridologists. There are three reasons why this former application cannot stand. First, Kunze described the pinnae of Polypodium alloeopterum as deeply incised-cre- nate, yet the pinnae of the former two species are only entire to deeply crenate. Second, Kunze lik- ened the leaf of S. alloeoptera to that shown in Raddi's plate of S. tyucana; yet S. tyucana lacks the broad, decurrent, pinna bases that are distinc- tive for S. heterophlebia and S. alloeoptera. In addition, Raddi's plate clearly shows a leaf with deeply lobed pinnae and free veins. Third, 5. het- erophlebia and S. pterorhachis have never been found in the coastal mountains around Caracas where the type was collected—this despite much collecting activity in that area. y species of Stigmatopteris—and a common one—-which occurs around Caracas is 5. nephrodioides. This species has deeply lobed pin- nae, free veins, and is like the leaf pictured in Raddi's plate. I strongly suspect, but cannot prove, that P. alloeoptera applies to 5. nephrodioides, the later name. І do not, however, want to replace the well-known name S. nephrodioides with S. alloeoptera because I have not seen the type of S. alloeoptera and because S. nephrodioides has been widely used. ани perforata Fée, Мет. Foug. 5: 248. 185 PE: Country and collector unknown (P — im Mougeo t) Christensen (1920), who app tly saw the type, said that this name applied to a species of Stig- mens received from Paris (P), I did not see any specimens with this name. LITERATURE CITED CHRISTENSEN, С. 1909. On Stigmatopteris, a new nus of ferns with a review of its species. Bot. Tidsskr. 29: 291-304. 1913. A monograph of the genus Dryopteris. Part 1. The tropical American pinnatifid-bipinnatifid species. Kongel. Danske е, = Skr., Na- turvidensk Afd. ser. 7, 10 A bep NS ed m Dryopteris. Part. 2. The tropical American bipinnate-decompoun Danske о Selsk. Skr., Na- Farris, J. 5. 1988. HENNICS6. Version 1.5 [photo- copy of E се accompanying program]. GREU ITER, W. et al. 1988. International Code of Bo- an са Nomenclature. Regnum Veg. 118: 1-328. O R. 968. Flora of Malaya, E 2. Ferns. db Printing Office, Singapor . 1984 udies of Pub genera allied t to | Pee. taria, 1. А commentary on recent schemes of clas- sification. Fern Gaz. 12: 313-319. KRAMER, K. U. 1978. The pteridophytes of s Uitgaven Natuurw. Studiekring Suriname Ned. 1922. Studies of tropical о . 7. Contr. 0.5. Natl. Herb. 24: 33-63. Moran, В. C. 1986. The wp fern un Ol- fersia. Amer. Fern J. 76: 78. — ——. 1987. Monograph o of ay д ЖК fern ge- nus Polybotrya о Bull. Illinois Nat. ist. Surv. 34: PrATNICK, М. І. 1989. An empirical comparison of я parsimony programs, II. Cladistics 5: 145 PROCTOR, GR 1977. d In: R. A. Howard (editor), Flora of the Lesser Antilles 2: 1-414. Каро, С. 18 antarum quee iensum no et A m Pars. 1. Tipographia Aloisii Bs Flor. um A. 979. Aspidiáceas. In: К. dira (editor), Flora Illustrada Catarinense, I ASPI. 1: 1-356 =. R. 1986. Revision of the neotropical fern us Cyclodium. Amer. Fern Эш А. 1883. Recensio cryptogammarum vascu- larum iis Quitensis. Quito 1893. Машин vasculares Quitensis. Quit TRYON, R. M. 1972. Endemic areas and geographic speciation in tropical American ferns. Biotropica 4: 121-131. 912 Annals of th Missouri Boo Garden А. Е. Tryon. 1982. Ferns and Allied Plants with Special Reference to Tropical America. Spring- er-Verlag, New York. . 1966. A cytotaxonomic survey of the desea НА E d id Trans. Roy. Soc. Edin- burgh 66: тие; E “|, de EA & D. M. BRITTON. 1983. ‚ On in sin in Bot. ue 20: 407- 417. INDEX TO COLLECTORS’ NUMBERS This monograph is based on the specimens cited below, which consist of about 1,500 herbarium sheets from ap- proximately 530 collections in 25 herbaria (see acknowl- gments for the list of herbaria). Types are in boldface. The abbreviations in parentheses refer to the following species of Stigma topteris, which are arranged alphabet- ically in the taxonomic treatment. brev = brevinervis; bulb = ae caud = caudata; chim = chimalapensis; cont = acta; gemm = gem- mipara; hemip = hemiptera; = = ћегегос arpa; hp = heterophlebia; ic — ichthio ma ; jam = jamaicensis; kil illi = i; long = = longicaudata; mi nephro od opaca = opaca; = pellucidopunctata; pri = prionites; pter = vieron chis; ro = rotundata; sord = sordida; tyu = tyucana. Aguilar 861 (long); 865 (long); 927 (pp). Allard 22558 (long). Alston 494 (го); 8744 (sord); 8745 (hp); 8803 (kil); 8848 (kil). Alston & Lutz 289 (caud). Antonio 2050 (hp). Armond 557 (kil). Aymard et 5 мое (lech). Baker 5934 (hp). Barringer & Gómez-Laurito 2583 (hp). Barrington 410 (long). Beck 1625 gos 8500 (pp). А 821 (го). Bernardi 938 (ro). Blanchet 2483 (tyu, pri). Bohs et al. 1704 (long). Box 1966 (ro). Brade 5892 үне 8231 (het); 8250 (brev); 8340 (brev); 8343 (brev); 8473 (het); 8474 (het); 8476 (het); 8478 (caud); 8510 (het); 9972 (tyu); 9977 (tyu); 18228 (caud); 18405 i & Brad 3 (sord); 497 (long). Brant et pp). Breedlove 35028 (long); 35128 (long); pA & Smith 32672 (long). Britton Пак ау 5355 | 6123 T г Burger & Stolze 5055 (sord); 5057 (sord). Burger et al. 10735 (cont). Calzada 5212 (long). aba 1020 (long). Cazelet & n 319 (sord); 852 (long); a 1624 (hp). Checkley 7 (ro). Chrysler 4777 (jam). Churchill 5255 (hp). Churchill et al. 4089 (sord); 4549 (hp); 5984 (long). Claros 73 (sord). Clausen 133 (caud). Clement 633 (hemip); (hemip); 945 (hemip); 949 (hemip); 961 (hemip). Conant 792 (long). Cooley 8325 (ro). Cornman 914 (cont); 1091 (cont); 1180 (hp); 1210 (hp); 1284 (hp). Cremers 9113 (ro); 9140 (ro). Croat 22998 (long); 25818 (long); 27477 (hp); 27541 (long); 27736 (long); 36176 (long); 37492 (long); 43598 (sord); 48097 (long); 49916 (hp); 54475 (ne); 60576 (ne); 60765 (ne); 66559 (hp); 68108 (caud); 68154 (sord); 68882 (kil); 69108 (long); 69137 (hp); 70911 (bulb). Croat & Grayum 60043 (hp); 60177 (long); 60215 (long); 60216 (hp); 60217 (hp). Croat & Liesner 1010 (hp). Crosby et al. 521 (jam). Cuatrecasas 15122 (long); 15240 (long); 17336 (hp). de Granville 1291 (long); 1434 (long); 8873 (ro). de ~ la Sota 5238 bua | & Saito 1232 (caud). Dryer 366 (long); 396 (hp). Dusén 725a (brev); 2538 (caud); 10244 (het); deles 13727 (brev); 14115 (het); 14680 (het); 14687 (caud). Duss 80 (ro); 729 (ro); 1561 (ro); 4074 (го). Eggers 905 (ro); 6164 (ro); 6869 (ro). Ekman 5268 (ретт); 7424 (gemm); 10113 (gemm); 14356 (hemip). Elias 8 (pp). Evoy 44 (kil). Ewan 16775 (hp). Fendler 98 (ro); 198 (ne). Finch 169 (long), Forero et al. 2309 (hp); 7407 (hp); 8793 (het). Foster P-84-91 (long). Foster & Wachter 7437 (pp). Foster et al. 10908 (long). Franco et al. 1148 (kil). Funck & Schlim 411 (ne). Gardner 131 (caud). Glaziou 964 (туи); 2396 (caud); 400 (het, brev, pri; 2401 (het); 3333 (brev); 6414 (caud); 7245 (tyu). Glocker 32 (pri). Gómez 281 (cont); 987 (hp); 1133 (cont); 2182 (long); 2330 (cont); 3522 (hp); 18641 (sord); 18778 (long); 19288 (hp). Gómez & Wagner 6856 (long); 79027 (cont). о et al. 19173 (cont); Dua ыр» 6893 (long). Grayum & Jacobs 3598 (sord). G m & Schatz 3089 (long). Grayum & Sleeper 3683 биш dien а eon & Zamora 9375 (kil); 9376 (sord). Grayum et al. (hp); 5043 (hp); 5495 (kil); 1910 longi 7922 mis Grijalva 290 (sord). Guz- mán-Valdez 892 aerchen 10 6 (he к Hagemann & Leist 2072 (bulb). Hahn 37 (ro); 352 (long). Hallberg 1724 (long). Hammel 2227 (hp); 4858 (kil); 11495 (sord); 12382 (sord). Ham- de Nevers 1 (mi). Hammel & Zuchowski 13880 ( 14238 (long); 14827 (kil). Harling et al. 19737 а Наггіѕ 7277 (јат); 7419 (ја Н 200 (јат); 337 (го); 1236 (го `; 3 Hatch & Wilson 346 (sord). Hatschbach 10748 (het); 22873 (caud); 24376 (het); 24739 (caud); 25939 (het); 41955 (het). Hatschbach & Scherer 26665 (caud). Haught 1931 (pp). Hazlett 3029 (long). Heinrichs 550 (bulb). Hellwig 491 (long); 492 (long); 494 (long). Hernández 1282 (chim). Herrera Ch. et al. 162 (hp); 316 (hp); 336 (cont); 2175 (sord); 2980 (lech). Hioran & Clement 6442 992 (opaca). Holm-Nielsen et al. 2906 (opaca); 24819 (ic). Hoover et al. үлү E 4325 (hp). Imray 17 (ro); 39 Jansson 740a caudi, Jaman 60 (jam). Joao & Rohr 392 (het). Johnston 303 (ro Kalbreyer 1807 (mi). Каг sten 9 (ne). Kenoyer 34 (sord). Killip 2919 (kil); 5159 (cont); 5344 (long); 5387 (hp) 5444 [еза 12147 (kil). Kilip & Smith 23906 (ок 28984 (hp). Kirkbride 2377 т Krueger 17107 (pri). Kummrow & Hatschbach 2536 (het). Laegaard 52431 (long); 51610 (sord); 51611 (hp); 51614 (opaca); 52224 (mi). Lancaster 836 (cont); 893 (hp). Lechler 2497 (lech). Le Gallos 2965 (ro). Lehmann 8951 (hp). Lellinger 362 (ro); 367 (ro); 507 (ro). Lellinger & de la Sota 194 (kil). Lellinger & White 1412 (hp) иза ү 1 al. 311 (pp). lan 10486 (hemip) 11133 115 (ro). Liebmann 737 (lon (hemip). L'Herminier р); 738 (long); 739 (long); 8 e). Li na 16909 (long). Lloyd 902 (ro). Luederwaldt 21149 snis 21821 (caud). Luschnath 120 (pri). Volume 78, Number 4 1991 Moran 913 Stigmatopteris Monograph MacDougal et al. 3811 (long). Madison et al. 3472 (lech). Maguire et al. 36864 (long). Marie 516 (ro). Mar- tius 305 (pri. Maxon 790 (jam); 1071 (jam); 1738 (jam); 2388 (jam); 5647 (long); 8898 (jam); 8863 (jam); 9158 (jam); 9307 (jam). Maxon & Harvey 8075 (long). Maxon & Killip 7 (jam); 169 (jam); 716 (jam). McAlpin 1043 (long); 1313 (hp); 1326 (hp); 1396 (kil). McCallum 104 (го). McDonough et al. 371 (mi). McPherson 6774 (hp; 8885 (long); 9730 (hp) 12035 (long); 12310E (long); 12677 (hp). Mendonca 1355 (caud). Mexia 4893- a (pri); 5492 (pri). Mickel 484 (long); 975 (long); 990 Lo 6369 бу 6392. (long); 6403 (sord); 9412 (ro). e 6543 (long); 6727 (long). Molina R. 132 1 (so rd). Moran 3165 (hp); 3166 (cont); 3187 (dont 3188 (hp); 3206 (long); 3315 (sord); 3330 (cont); 545 (іс); 3562 (kil); 3576 (opaca); 3582 (lech); 3607 ( 4048 (long); 4130 (hp); 5119 (lech); 5250 (ic). Morton 5373 (ro); 6184 (ro); 9569 (hemip). Mosén 2694 (caud). Moya 0011 (lech). Müller 1038 (brev Nee & Schatz 19868 (long). Nee et adi 26382 (long). Neill 1886 (hp); 3826 (hp). Nisman S. 98 (cont). Ocampo S. 1029 (hp); 1771 (sord). Oliveira & Cordeiro 909 (het). Vllgaard et al. 35212 (pp); 35859 (ораса); 35866 (bulb); 39229 "ra idea о Ortega & González 369 rtega et al. 2 (pter). Pabst 799 (he:). Palacios 4030 (lech). Mira) 1161 (jam); 1613 (jam). Pittier 12172 (ne). Prestoe 214 (ro). Proctor 4257 (gemm); 17244 (ro); 20345 (го); 32327 (hp) 38712 Mi Purpus 7256 (sord). Questal 2 Ravins x vo ik Regnell 254a (caud); 358 (caud). Reitz C205 (caud); C206 (brev); 227 (caud); 276 (brev); C919 (het); 1909 (caud); 3141 (het); 3542 (cuad). Reitz & Klein 3068 (het); 3077 (caud). Rivero et al. 1602 (pter) Rohr 352 (brev) 392 (het) 1006 (het) 1072 (caud). Rovirosa 322 (long). Scamman 7667 (hp). Schenck es | (het); 794 (pri). Schlim 596 (pter 856 (pter); 1683 of cont). Schmalz 106 (het); 191 (caud). Sehnem 808 (het); 3086 (het). Sieber 350 (ro). Skutch 145 (jam). Smith, A. R. 1031 (pter); 1692 (hp); 2121 (hp). Smith, A. R. et al. 1678 (long); 1813 (sord); 1862 (cont); 1876 (hp); 2194 (long), 2235 ЭРА cc (cont). Smith, H. H. 1045 (pter). Smith, H. H Smith 659 (ro). Smith, L. B. 1391 (caud). pa th, L. B. & Reitz 6127 (het). Sodiro 48/1 (ic); 54/2 (opaca); 54/3 (opaca); 2615 (mi). Soejarto & Villa 2685 (sord). Solomon 18469 (long.) Spannagel ais (he ng); 4653 (opaca); 3109 ens 37834 (cont) Stan & (long); 51710 (long); 110227 (pter); 125086 (ne). Stolze 1487 (long); 1634 (hp). es 2263 (cont); 2658 (cont); 2722 (hp). Stübel 797 Tonduz 94.80 (lech). Torres 49 (cont). Tryon & Tryon 5330 (long). Underwood 1315 (jam); 1613 (jam); 2036 (jam); 2696 ~ jam). Ule | (caud); 70 (caud & ulei); 187 (het). U.S. Expl. Exped. 23 (caud Valerio 295 (hp). van am Werff 3488 (ne); 7010 (mi). n der Werff & Gudino 11125 (hp); 11169 (opaca); 11170 (pp). van der Werff & van Hardeveld 6762 (hp). van der Werff et al. 9566 (mi). Vasquez 188 (sord); 209 (kil); 245 (kil). Ventura A. 484 (long); 3306 (long); 3603 long). Verleysen 174 (opaca). von Sneidern 5046 (lech). von Türckheim 1432 (sord); 3865 (long); 111691 (long). acket 50 (tyu); 286 (caud); 21149 (tyu). Wagner 460 (pter); 1402 (het). Wagner & Gómez 79021a (hp). Wachter et al. 200 (long). Watt 82 (jam). Webster 13329 го); 13566 (ro); 13588 (ro). White & Lucansky 1968150 cont) 1968174 (cont). Wilbur 7882 (ro). Wilbur & Stone 8919 (long). Wilson & Murray 563 (jam). Wright 1053 (hemip). Wurdack 2002 (pp). Zak & Espinoza 5287 (opaca). Pak & Jaramillo 3635 (lech). ~ INDEX TO NAMES ted names are set in roman type; new taxa and uiua ык appear in boldface; all other names are italicized. In parentheses following an italicized name is the accepted tax Aspidium Sw rado Fée (tyucana) caudatum var. аи нем Christ (contracta) k.) Fée (rotundata) (Hoo m Willd. (rotundata) r jamaicensis Desv. (jamaicensis) Dryopteris Adan g. Stigmatopterie (C. Chr.) C. Chr. i excluded coalescens (Ba ker) C ook.) Пени (rotundata) longicaudata (Liebm.) Maxon (long ata) pellucidopunctata C. Chr. о prionites (Kunze) C. Chr. (prioniodes) gemmipara (C. Chr.) Maxon ex Proctor (gemmi- C. Chr. (hemiptera) C. Chr. (heterophlebia) (ichthiosma) longipetiolata C. Chr. (i chlaena Maxon (jamaicensis) paucinervata C. Chr. (opaca prasina (Baker) C. Chr. (lechleri) rotundata C. Chr. (rotundata) sordida Maxon (sordida sylvicola (Baker) C. Chr. (michaelis) tijuccana C. Chr. (tyucana ulei (Christ) C. Chr. (ulei) imrayanum Hook. (rotundata) Phegopteris Fée brevinervis Fée (brevinervis) flavopunctata (Kaulf.) Fée (rotundata) 914 Annals of the Missouri Botanical Garden риба pn (prionites) heterocarpa Fée (heterocarpa) le chleri Mett. о macrophylla (Hook.) J. Smith (pellucidopunctata) oligophlebia Salmon (opaca) opaca (Baker) Christ (opaca) longicaudatum Liebm. (longicaudata) macrophyllum Hook. (pellucidopunctata) michaelis Baker (michaelis) nephrodioides Klotzsch (nephrodioides) oligophlebium Baker in Hook. & Baker (opaca) prasinum Baker (lechleri) prionites Kunze (prionites) punctatum Spruce in Hook. (lechleri) sylvicolum Baker (michaelis) tyucanum Raddi (t na) Stigmatopteris C. Chr. alloéoptera (Kunze) C. Chr. uncertain carrii (Baker) C. Chr. (err caudata (Raddi) C. Chr. chimalapensis Mickel & Beite clypeata (Maxon & C. ‘Morten Lellinger excluded contracta (Christ) C. C cyclocolpa (Christ) С. Gir koui ecuadorensis C. Chr. (lechleri) gemmipara C. Chr. guianensis (Klotzsch) Chr. excluded hemiptera (Maxon) heterocarpa Jess ris Rosenstock е t . Chr litoralis Rosenstock (cordial longicaudata Liebm. meniscioides Kramer in Kramer 4 van Donselaar exclude michaelis (Baker nephrodioides (Klotzsch) C nothochlaena (Maxon) C. м la opaca (Baker) C. Chr. palmensis Rosenstock Јуре aa paludosa (C. Morton) R. Tryon « A. Tryon excluded pellucidopunctata (C. Chr yc. Chr. prasina (Baker) C. Chr. (lechleri) pe (Kunze) C. Chr. ar. denticulata (Fée) - Chr. (prionites) Mor var. trinidadensis C. (ЋЕ (rotundata) sancti-gabrieli (Hook. А C. Chr. excluded sordida (Maxon) C. Chr m varians (Fée) Alston excluded ESTUDIOS MORFOLOGICOS EN EL SUBGENERO DICHANTHELIUM DE PANICUM (POACEAE: PANICOIDAE: PANICEAE), CON ESPECIAL REFERENCIA A PANICUM SABULORUM' Osvaldo Morrone? y Fernando Zuloaga? RESUMEN El subgénero Dichanthelium fue establecido por Hitchcock & Chase (1910) en su tratamiento de las especies as anicum. Estos aut a езїїеше еп е caracteres anteriormente mencionados, рага foló en diferentes poblaciones de Pani mor fologi anatomic roseta bd de hojas e y anchas; en primavera las canas son simples, con una inflorescencia terminal casmógama. Lueg „nuevas blecer la variación icha nthelium, habiéndose m., cana d pope iu ён ders el patrón de crecimiento, casmogamia y vet dr presencia de due anatomía y relaciones de la especie con especies norte у sudamericanas del géne ABSTRACT Subgenus о was established within Panicum by Hitchcock & Chase (1910) in their treatment of the e 5 orth American species of the short, broad deine d in spring the culms are simple, w arise at the node cleistogamous inflorescences. In species о г ап nthelium. Morphological and anatomical studies of different populations of Panicum nus. These authors defined Dichanthelium as having an overwintering rosette of with a terminal, chasmogamous inflorescence. Later, branch culms the vernal culms, forming loose to rather dense fascicles of reduced leaves and branches with the present treatment, variation of these characters was analyzed in South American sabulorum American species of subg. Dichanthelium, were carried out, examining the growth patterns, anatomy, and Dcos of this species with other American species of the subgenus. El subgénero Dichanthelium А. Hitchc. et Chase fue establecido por Hitchcock & Chase (1910) en su tratamiento de las especies norteamericanas del género Panicum. Estos autores definieron a Di. chanthelium, sobre la base de especies de América del Norte, por poseer en la temporada invernal una roseta basal de hojas cortas y anchas; en la primavera las canas son simples y llevan una in- florescencia terminal. Luego, numerosas ramifi- caciones nacen en los nudos inferiores de dicha саћа, dando lugar a densos fasciculos de hojas reducidas e inflorescencias menores, disminuyendo con el tiempo el tamaño de las ramificaciones, hojas e inflorescencias (Fig. 1). Hitchcock & Chase (1910) indicaron que las inflorescencias terminales lleva- ban espiguillas casmógamas, que raramente pro- ducian semillas viables, y las inflorescencias axi- lares llevaban espiguillas con flores cleistógamas, con una alta producción de semillas viables. Hitchcock & Chase (1910) mencionaron ciertas excepciones dentro del subgénero еп los caracteres anteriormente descritos, como por ejemplo la au- sencia de roseta invernal en especies del Grupo Depauperata y de roseta y forma otonal, es decir ' Deseamos expresar nuestro agradecimiento a Gerrit Davidse por la lectura crítica del manuscrito. El trabajo de campo en Colombia y Venezuela fue realizado con un subsidio de la National Geographic Society, #3964-88. Zuloaga desea expresar su agradecimiento a la Fundación Guggenheim por la beca otorgada para desarrollar sus investigaciones en el Missouri Botanical Garden, donde la presente revisión ha sido concluida. Ejemplares de herbario de Panicum sabulorum fueron examinados de los siguientes herbarios: BAA, CTES, G, HB, K, LIL, M, MO, NY, P, R, RB, SI, < ‚ М. 3 Саш de Botánica Darwinion, Casilla de Correo 22, San Isidro (1642), Argentina. Zuloaga está actualmente asociado al Missouri Botanical Garden. ANN. MISSOURI Bor. GARD. 78: 915-927. 1991. 916 Annals of the Missouri Botanical Garden FIGURA 1. invernal.— orma primave rend Е dicun ral.—C. Esquema del desarrollo E una especie norteamericana del subgénero Dichanthelium. — А. For a otonal. Circulos negros: inflorescencias en floración E nte en ао Circulos con linea interrumpida: inflorescencias sin espiguillas (ya as. . Círculos negros aídas). El tamaño de los círculos indica el orden de floración de las panoj sin cañas fasciculadas claramente desarrolladas en Panicum viscidellum Scribner. Estos autores tam- bien indicaron que “there is an intermediate stage of branching, in which the plants do not show the characteristic vernal nor autumnal habit. Vernal culms are sometimes produced on plants during the branched condition, because of renewal of ac- tivity, due to increased moisture, excess of nutri- ment, injury, or other causes” (Hitchcock & Chase, 1910: 3 Gould (1974) elevó Dichanthelium a la cate- goria de género, y otros autores, como Clark & Could (1975), Brown & Smith (1975), Gould & Clark (1978), y Gould (1980), apoyaron dicha segregación en base a los caracteres exomorfoló- gicos previamente descritos, tipo fotosintético y ornamentación del antecio superior. Todos los autores anteriormente mencionados basaron sus revisiones en especies de América del Norte de Dichanthelium, indicando que este taxon estaba representado por pocas especies en América del Sur. Más recientemente Lelong (1984), Clayton & Renvoize (1986), Zuloaga (1987) y Webster (1988) consideraron a Dichanthelium como un subgenero de Panicum, senalando que no existen caracteres suficientes para mantener a Dichanthe- lium como un género aparte. Durante una revision de las especies centro y sudamericanas de Dichanthelium (Zuloaga et al., еп prep.), se comprobó que este subgénero está ricamente representado en el área, con cerca de 40 especies; además, estas especies tienen un ma- yor grado de variación morfológica en relación a las especies norteamericanas. Consecuentemente, se ha tratado de establecer en el presente trabajo la relación existente entre las especies norte y sudamericanas de Dichanthelium, considerando la validez de los caracteres morfológicos empleados hasta el momento para su segregación a nivel ge- nérico o infragenérico. Se seleccionó, dentro de los taxa sudamericanos a P. sabulorum Lam., especie con la distribución mas austral, y consecuente- mente con mayores adaptaciones a climas más frios. Estudios anatómicos y exomorfológicos fue- ron realizados en diferentes poblaciones de P. sa- bulorum, analizando el patrón de crecimiento ve- Моште 78, Митбег 4 1991 Morrone 8 Zuloaga Estudios Morfológicos en Panicum getativo y florifero, cleistogamia y casmogamia, presencia de glandulas, ornamentacion del antecio superior y caracteres histofoliares. MATERIALES Y MÉTODOS Especimenes de Panicum sabulorum fueron co- leccionados en las provincias de Buenos Aires, En- tre Rios, Corrientes y Misiones (Argentina), y en учн: El NEED fue herborizado y los ejem- pl SI. Observaciones de las plan- tas vivas fueron realizadas en su ambiente natural entre 1987 y 1990. Plantas vivas fueron colec- cionadas y mantenidas bajo cultivo en el inver- náculo del Instituto Darwinion. Las fotografias de campo fueron tomadas con una cámara Asahi Serm usando una pelicula Ko- dak Panatomic, ASA 3 Los materiales para i morfológicos y ana- tómicos, provenientes de las poblaciones estudia- das, fueron fijados en Etanol 70% (ver lista de especimenes estudiados). Para los estudios histofoliares se empleó material fijado o proveniente de materiales de herbario re- cuperados y rehidratados con Contrad 70 (Schmid & Turner, 1977). Los cortes transversales fueron obtenidos a mano alzada, previamente tratados con FH 5% para rernover la silice. Las secciones fueron coloreadas con safranina-fast green al 1% y mon- tadas en gelatira glicerina. Las epidermis fueron preparadas siguiendo el método propuesto por Metcalfe (1960) Para la detección de células siliceas y suberosas se empleo respectivamente Fenol liquido (Metcalfe, 1960) y Sudán III (Sass, 1940). Amiloplastos fue- ron coloreados con I,KI (Sass, 1940). Para la de- tección de cristales se siguió el método de Sánchez & Casabona (1981). Cortes seriados de glándulas fueron obtenidos con un micrótomo rotativo. El material fue des- hidratado en ura serie de alcoholes ascendentes e incluido en Paramat. Las secciones, de 8-12 um de espesor, fueron coloreadas con safranina~fast green y montadas en bálsamo de Canadá. El liquido secretado por las glándulas fue colec- tado por medio de un tubo micropilar y la presencia de azücares testeada mediante la técnica colori- y ien propuesta por Dubois et al. (1956). a la detección de aceites se empleó Sudán IV erc 1962). Las canas, hojas y estructuras reproductivas fueron estudiadas con un microscopio estereoscó- pico Wild M5 y las preparaciones histológicas con un microscopio fotónico Wild M20 Las microfotografias de glándulas y hojas fueron tomadas en un microscopio fotónico Zeiss. о. де P. sabulorum empleados еп ч езїи- dios anatómicos: ARGENTINA. ears AIRES: Punta Lara, ] ш Escobar, Paraná ne 4544 (SI ; Escobar, al. 3228, 3231 (SI). ga & Morrone 3857 (MO, SI); Colonia Ауш, Zuloaga & Morrone 3852 (MO, SI); de Calabacilla a Nueva chi Zuloaga & Morrone 3846, 3847 (51); de Ceibas а Médanos, 7 km de Medanos, Zuloaga & Morrone 3840 (SI); ruta 12, desvio a Holt, e El Cuartillo, Zuloaga 3864 (MO, SI). MISIONES: n Ignacio, casa de Horacio Quiroga, Zuloaga et al. 3191 (SI; INTA, Cuartel Rio Mirum ruta nac. 14, aprox. 11 km de San Мыке Giberti & Davina 210 (SI). BRASIL. RIO GRANDE DO SUL: Jaguarao, 8 km W de Jaguarao, Allem & Vieira 1868 (Si) Pelotas, Swallen 7137 (US); Santa Victoria do Palmar, Swallen 7396 (US). RIO DE JANEIRO: Parque Nacional Itatiaia, Rio Campo Belo, Zuloaga et al. 2361 (SI). SANTA CATARINA: 4 km S of Campo Alegre, Smith & Klein 7334 (US). SAO PAULO: Sao Paulo, Parque do Estado, Sendulsky 1278 (SI). URUGUAY. CANELONES: Ruta Interbalnearia, km 51, Arro- yo Solis Chico, Dubcovsky 834, 835, 836, 837 (SI). ~ HÁBITAT Panicum sabulorum posee una gran adaptabi- lidad para ocupar hábitats muy diversos, creciendo desde lugares húmedos y sombrios, en bordes o en el interior de bosques y selvas, hasta lugares más abiertos y soleados, en suelos usualmente arenosos y secos, estando presente también en suelos ro- cosos; a menudo es posible encontrar especimenes en areas perturbadas. La morfología de los especimenes está correla- cionada con el hábitat: plantas de lugares húmedos y sombrios son delicadas, apoyantes en la vege- tación o decumbentes; generalmente poseen hojas, inflorescencias y espiguillas glabras (Fig. 2C). El tamaño de los ejemplares varia desde pequeños hasta otros que llegan a 80 cm de alto, con cañas apoyantes en la vegetación. Las poblaciones de lugares abiertos y soleados, y que crecen común- mente en suelos arenosos, son plantas cespitosas, con hojas rigidas y pilosas, e inflorescencias y es- piguillas pilosas (Fig. 2D), variando el tamaño desde especimenes pequeños a mayores, densamente ra- mificados (Fig. 2 En condiciones de invernáculo las plantas pro- venientes de lugares abiertos y arenosos, con en- trenudos muy cortos, producen durante la prima- vera cañas apoyantes, con entrenudos largos, hasta 918 Annals of the Missouri Botanical Garden Ficura 2. Fotografias de campo de P. о — А. Plántulas naciendo en eed — B. Innovaciones apareciendo en la base de la Fade pu te el verano (flecha). — C. Canas densamente ramificadas y nuevas inno- vaciones, con hojas más anchas, en la base de la Boden durante el verano. — D. Planta de lugares arenosos, abiertos. — D Canas simples con inflorescencias terminales formadas durante el otoño. — Е. Inflorescencias terminales ru canas n hojas rigidas. — C. Inflorescencia terminal (flecha) e inflorescencia axilar. — H. Planta ramificada con inflorescencia le (flecha) e inflorescencias axilares Volume 78, Number 4 1991 Morrone & Zuloa 919 Estudios Morfológicos en Panicum o que la misma planta continüa el ciclo el siguiente ano. El tamario de los circulos indica el orden de floración de las panojas de 15 cm de largo, semejantes a los especimenes que crecen en lugares ћитедоз y sombrios. MORFOLOGÍA, DESARROLLO Y RAMIFICACIÓN (Fics. 2, El estudio de las poblaciones citadas durante los ültimos cuatro anos ha permitido observar un in- cremento en la complejidad de los ejemplares a medida que los mismos se desarrollan. En la pri- mavera germinan nuevas plantas, las que desarro- llan canas simples y erectas, con entrenudos largos y hojas anchas, cordadas en la base (Fig. Luego de formadas las primeras 7 u 8 hojas el crecimiento vegetativo cesa y se forma una inflo- rescencia terminal multiflora, la que emerge de un largo pedünculo (Fig. Con posterioridad a la floración de la panoja terminal, la caña principal comienza a ramificarse en los nudos superiores, observándose canas axi- lares a fines de la primavera y comienzos del ve- rano (Fig. 2C). Estas ramificaciones tienen 3 o 4 hojas reducidas, menores que las presentes origi- nalmente en la cana principal, y terminan en in- florescencias axilares paucifloras. En este momento del desarrollo las inflorescencias axilares pueden coexistir con las inflorescencias terminales de la cana principal, o las inflorescencias terminales pue- den encontrarse sin espiguillas (Fig. 3 La ramificacion de las canas continüa baut todo el verano, con inflorescencias laterales nu- merosas, existiendo una consecuente reducción de los entrenudos, hojas, e inflorescencias, lo que da a la planta un aspecto fasciculado. Asimismo, nue- vas innovaciones aparecen en la base de la planta, siendo estas similares a las innovaciones prima- verales, con canas simples con hojas cordadas y anchas (Fig. 3 | final del verano y comienzo del otono continua la profusa ramificación en las caras, habiendo des- aparecido en este punto las inflorescencias termi- nales primarias y las hojas basales anchas de dichas canas primarias (Fig. 1C, 3E, láminas e inflores- cencias faltantes indicadas con trazo discontinuo). Las canas simples formadas en la fase anterior (Fig. 3D) florecen dando una inflorescencia terminal, para posteriormente comenzar a ramificarse en los nudos superiores (Fig. 3 920 Annals of the Missouri Botanical Garden FiGURA 4. Panicum sabulorum. А. Hábito.--B. Ligula.—C. Espiguilla, vista ventral. — D. пе qr h vista dorsal.— E. Pálea inferior. — Е. Antecio dd. vista dorsal.—G. Antecio superior, vista ventral. —H. superior con lodiculas, cariopsis, estigma y anteras. — I. Cariopsis, visto del lado del embrión. — J. Cariopsis visto Bd lado del hilo (de Zuloaga et al. 3191, SI). Las plantas mantienen durante el invierno la siguiente primavera, las canas que permanecieron mayor parte de la porción aérea, con las canas en la planta continüan ramificándose, formando profusamente ramificadas, no formándose durante pequeñas inflorescencias en el extremo de сада este lapso nuevas innovaciones. Al comienzo dela nueva ramificación (Fig. 3F). A la vez, nuevas Volume 78, Number 4 1991 Morrone & Zuloaga Estudios Мет МЕ еп Ратсит innovaciones aparecen еп la planta, con cañas sim- ples y hojas cordadas, anchas, lo que implica un nuevo comienzo del ciclo (Fig. 3 Debido a la profusa ramificación de las сайаз, las mismas son decumbentes y generalmente apo- yantes en la vegetación. Plantas provenientes de poblaciones de Entre Rios y Buenos Aires preservadas en invernáculo, mantuvieron durante el invierno una activa pro- ducción de ramificaciones laterales y muy fre- cuentemente de inflorescencias axilares. Esta ac- tividad de los meristemas está directamente relacionada con la protección de las plantas a la rigurosidad del clima durante el invierno. ESPIGUILLA (Fic. 4) La espiguilla posee una gluma inferior 3-5-ner- via, que alcanza 44 o poco más del largo de la espiguilla, gluma superior y lemma inferior subigua- les y 7-9-nervias, pálea inferior presente, flor in- ferior ausente y antecio superior endurecido, pa- piloso y con el ápice de la lemma cortamente чиа se hallaron espiguillas trifloras en el material estudiado, con flor inferior neutra, con su correspondiente lemma, una flor intermedia hermafrodita con lemma y pálea endurecidas y una flor superior tambien hermafrodita con lemma y pálea endurecidas. Espiguillas trifloras han sido ci- tadas previamente en el género Panicum (Palacios, 1968; Zuloaga & Sendulsky, 1988) y en el género Lasiacis (Davidse, 1978) En espiguillas maduras se observó un eleosoma en la porción basal de la espiguilla, en la región periférica de la raquilla entre la inserción de la gluma inferior y la lemma inferior, hallándose acei- te en las células. La presencia de eleosoma en gramineas ha sido generalmente correlacionada con la mirmecofilia (Berg, 1985; Davidse, 1987). El antecio superior en P. sabulorum es endu- recido y biconvexo, liso y con papilas simples dis- tribuidas en hileras longitudinales regulares, pu- diendo observarse pelos y estomas en el apice de lemma y pálea respectivamente. CASMOGAMIA Y CLEISTOGAMIA (TABLA 1) Las espiguillas que se disponen en las inflores- cencias terminales y axilares son isomorfas, ho- mógamas, no existiendo diferencias morfológicas entre las flores cleistógamas y casmógamas. Asi, no se observaron diferencias en las anteras de flores casmógamas y cleistógamas, variando entre 0.3 a 1 mm de largo. El análisis del material de P. sabulorum mostró que no existe una correlación entre inflorescencias LA l. Porcentaje de espiguillas con cariopsis y flores cleistógamas y casmógamas en inflorescencias ter- minales y axilares. Total % Inflorescencia terminal (noviembre) Espiguillas con cariopsis 639 93.55 Espiguillas sin cariopsis 44 6.45 Flores casmógamas 489 99.59 Flores cleistógamas 2 0.41 Inflorescencias axilares (enero) Espiguillas con cariopsis 86.9 Espiguillas sin cariopsis 15 13.1 Flores casmógamas 100 86.9 Flores cleistógamas 15 13.1 terminales y flores casmógamas e inflorescencias laterales, o axilares, y flores cleistogamas. Espi- guillas con flores casmógamas son frecuentes tanto en las inflorescencias terminales como en las la- terales; espiguillas con flores cleistógamas son es- casas y se pueden hallar en ambos tipos de inflo- rescencias y en cualquier momento del periodo de floración. Cuando presentes las espiguillas con flo- res cleistogamas se encuentran en inflorescencias que llevan también espiguillas con flores сазто- mas. El estudio de 683 espiguillas maduras en inflo- rescencias terminales permitió confirmar que el 93.5% d sobre 491 espiguillas de inflorescencias terminales e las mismas forman cariopsis. Además, producidas en noviembre se estableció que el 99.59% de las espiguillas son casmógamas. Observaciones realizadas en 115 espiguillas pro- venientes de inflorescencias axilares permitió de- terminar que el 86.975 producen cariopsis y un mismo porcentaje corresponde a espiguillas cas- mógamas. La variación en la presencia de flores cleistó- gamas en las inflorescencias terminales y axilares de P. sabulorum posiblemente esté condicionada por factores ambientales, por lo que la cleistogamia podria ser clasificada como facultativa (Connor, GLÁNDULAS (Fic. 5) Nectarios extraflorales han sido citados ocasio- nalmente para la familia Poaceae, como por ejem- plo en trabajos de Bowden (1971), Davidse (1988), Fahn (1979), Mattei & Tropea (1908, 1909), Nicora (1941), Reeder & Toolin (1989), y Linder & Ellis (1990). Pelos glandulares multicelulares han sido mencionados en especies de la sección Clavelligera Stapf de Panicum (Kabuye & Wood, y glándulas crateriformes para la lemma inferior de especies de la sección Stolonifera (A. 922 Annals of the Missouri Botanical Garden y Y Uim * ay he P a > FicURA 5. А-С, Е, С. Fotomicrografias de glándulas de % sabulorum. — A. Eje de la inflorescencia. — B. Glándula en vaina. — C. Glándula en eje de la inflorescencia. — E. Cor te longitudinal de glándula en eje de la inflorescencia. — С. Glándula en eje de la panoja — D. Fotomicrografia de glándula de la vaina de P. surrectum. F, H. Fotomicrografias de glándulas de P. adenorhachis. —F. Glándula en vaina.—H. Glándula E lamina (A-C, E, С, de Zuloaga y Morrone 3544, SI; D, de Chase 9555, US; F, H, de Mori et al. 12451a, Volume 78, Number 4 Morrone & Zuloaga 923 1991 Estudios Morfológicos en Panicum con fenol-ácido sulfúrico (Dubois et al., 1956), Hitchc. & Chase) Pilger por Zuloaga & Sendulsk y (1988 Glándulas multicelulares secretoras fueron ob- servadas sobre pedúnculos, ramas de la inflores- cencia (Fig. 5A, С, E, С), pedicelos, у en entre- nudos o vainas (Fig. 5B), de numerosos especi- menes de P. sabulorum. La presencia de dichas glándulas es marcadamente variable, hallándose una gradación desde ejemplares con abundantes glandulas a otros con una menor densidad, o au- sentes. Las glándulas son mas frecuentes en las inflo- rescencias y su presencia es más notoria al co- mienzo de la primavera sobre las inflorescencias terminales, siendo raras, о ausentes, en las inflo- rescencias que se forman durante el verano y el toño. Las glándulas son sésiles, crateriformes, y de forma anular o elipsoide, midiendo de 0.15-0.3 mm de largo por 0.1-0.16 Sobre la superficie glandular es posible observar a trasluz mm de ancho. pequeñas gotitas brillantes; la mayor producción glandular se da por la mañana y al atardecer, decreciendo a medida que el organo envejece. El tejido secretor se halla ubicado en la depresión de la glándula. En vista paradermal las células secretoras son cuadradas en contorno, más peque- ñas que las células epidérmicas vecinas, de paredes anticlinales lisas, rectas y engrosadas, semejando el tejido a una rejilla (Fig. 5B, C; Fig. 6). Las células epidermicas que rodean al tejido glandular son de contorno rectangular y se hallan realzadas en relación al resto de la epidermis. El tejido glandular es unistratificado en corte transversal y está restringido a la epidermis (Fig. SE). Está compuesto por un número variable de células prismáticas, alargadas transversalmente, en empalizada y sin espacios intercelulares. Las células tienen aproximadamente 25 um de largo en corte transversal y longitudinal y poseen paredes del- gadas a excepción de las paredes radiales, las que tienen un conspicuo espesamiento junto a la pared tangencial externa. El citoplasma es denso y gra- nuloso y se tiñe intensamente, con un gran nucleo junto a la pared tangencial interna. La cutícula es delgada y fuertemente cutinizada en las células epidérmicas que la rodean (Fig. Las glándulas carecen de tejido vascular, y el tangulares en corte longitudinal e isodiamétricas en corte mn de paredes lisas y con pocos o sin cloroplast 1 análisis cod ЕН del liquido secretado por las glandulas multicelulares, a traves de un ensayo permitió detectar la presencia de azücares. Esto permite inferir que las glándulas funcionan como nectarios extraflorales. La función de los nectarios extraflorales en P. sabulorum no es conocida. Bentley (1976), Elias & Gelband (1976, 1977), sostienen que los nec- tarios extraflorales atraerian a las hormigas, pro- tegiendo a las plantas del ataque de herbivoros. Davidse (1988) citó que las glándulas presentes en Pentastichis podrian ser parte de un mecanismo de defensa contra los herbivoros. En las poblaciones aqui analizadas de P. sabulorum no se observó interacción entre hormigas y plantas con nectarios extraflorales. Los nectarios presentes en P. sabulorum difieren de los pelos glandulares citados para la sección Clavelligera por Kabuye & Wood (1969) por ser estos ültimos pedicelados. А su vez, son aparen- temente similares a los citados para espiguillas de especies de la sección Stolonifera por Zuloaga & Sendulsky (1988), no habiendo estos ültimos au- tores estudiado en detalle dichas glándulas. Glándulas similares a las presentes еп Р. sa- bulorum han sido detectadas, dentro del subgénero Dichanthelium, en especies norte y sudamerica- nas, como por ejemplo P. adenorhachis Zuloaga & Morrone (Fig. 5F, H), P. stigmosum Trin., P. sphaerocarpon Elliott, P. strigosum Muhl., P. sco- parium Lam. y P. surrectum Chase ex Zuloaga & Morrone (Fig. 5D). CARACTERES HISTOFOLIARES (Fic. 6) Epidermis abaxial. Zonas costales angostas de 3-6 hileras de células de ancho; cuerpos siliceos halteriformes. Zonas intercostales conspicuas, de 10-14 hileras de células; células largas cercanas a las zonas costales rectangulares, tres veces más largas que anchas, con paredes anticlinales longi- tudinales onduladas; células largas de la región intercostal media cuadradas a poligonales, con pa- redes sinuosas a casi paralelas. Células intercostales cortas rectangulares, alargadas transversalmente y con paredes sinuosas, solitarias o en pares, sepa- radas por una o más celulas largas. Aparatos es- tomáticos con células subsidiarias angostamente triangulares, dispuestos en 2 hileras a cada lado de la zona costal, de 28.5-49.5 um de largo por 20-26 um de ancho, alternando con células in- terestomáticas; células interestomáticas con pare- des anticlinales transversales cóncavas. Micropelos bicelulares, de 4 .9 um de largo, célula basal e paredes paralelas, de tamano igual o menor que la célula distal; célula distal fusiforme, aguda y con paredes más delgadas que las de la célula basal. 924 Annals of the Missouri Botanical Garden FIGURA 6. e P. acuminatum. —E. Corte transversal. Zuloaga & Morrone 3840, SI; E, F de Zuloaga 2490, Macropelos presentes o ausentes, cuando presen- tes en las zonas intercostales, unicelulares y de paredes gruesas, agudos y con la base bulbosa hundida en el mesófilo. Aguijones no vistos. Ра- pilas ausentes. Lámina en corte transversal. Transcorte pla- no de 170-360 um de espesor. Costillas conspicuas en la cara adaxial, dispuestas sobre todos los haces vasculares, surcos adaxiales poco marcados № del Ж; у iii, aa мак иу is Ag etr a тысы о и Д "T A-D. Anatomía de P. sabulorum. — A, C. Corte transversal. — B, D. Epidermis abaxial. E, F. Anatomía —F. Epidermis abaxial (А, B de Zuloaga & Morrone 4544, SI; C, D de SI). ancho de la lámina; costillas y surcos inconspicuos en la cara abaxial. Nervio medio inconspicuo, con un haz vascular primario indistinguible estructu- ralmente de los laterales; 2-5 haces vasculares de segundo orden entre haces vasculares de primer orden contiguos; haces vasculares contiguos se- parados por 5-8 células del mesofilo, a una dis- tancia de 195-265 um entre haces contiguos. На- ces vasculares de primer orden trabados, de contorno circular, rodeados por dos vainas; vaina Volume 78, Number 4 1991 Morrone & Zuloaga 925 9 Estudios Morfológicos en Panicum externa parenquimática completa, formada por 15- 18 células globosas, vacias o con escasos cloro- plastos no especializados; extensiones de la vaina Tl presentes hacia ambas epidermis, —3 columnas de células; vaina interna mes- tomática completa. Haces vasculares de segundo orden trabados, de contorno poligonal; vaina ex- terna parenquimática completa, formada por 6- 10 células globosas; extensiones de la vaina pa- renquimática en una o dos columnas, cada columna con 3-7 células hacia la cara adaxial, y con 2-4 células hacia la cara abaxial. Clorénquima ите- gularmente radiado alrededor de los haces vascu- lares, células tabulares, raquimorfas, espacios ае- reos intercelulares numerosos. Células buliformes en abanico, dispuestas en grupos regulares de 3- 5 células en los surcos adaxiales. Células epider- micas abaxiales, opuestas a las células buliformes, globosas, de mayor tamano que las adyacentes a la zona intercostal. Amiloplastos presentes en las células clorenquimáticas. OBSERVACIONES Se ha observado una correlación entre los ca- racteres exomorfológicos y el hábitat de las plantas con los caracteres histofoliares. Ejemplares que crecen en ambientes hümedos y sombrios, con es- piguillas pequenas y glabras, usualmente poseen láminas delgadas, de 1 m de espesor, estomas pequeños de 28.5-44.1 um de largo, y epidermis con macropelos aislados a frecuentes. Las plantas que habitan en lugares abiertos, are- nosos y drenados, y que generalmente tienen es- piguillas mayores y pilosas, poseen láminas mas gruesas, de 220-360 шт de espesor, estomas ma- yores, de 33.70-49.5 um de largo, y macropelos abundantes Micropelos 3-celulares fueron observados en el ejemplar Giberti y Davina 210. Estos micropelos son fusiformes, con la célula basal de paredes mas engrosadas que la célula media y distal. Pelos mul- ticelulares fueron previamente reportados para la tribu Paniceae en Paspalum (Türpe, 1966) y Pan- icum validum Mez (Zuloaga et al., 1989). DISCUSIÓN El estudio raorfologico de P. sabulorum у es- pecies afines de América del Sur, como por ejemplo P. superatum, P. stigmosum, y P. surrectum ha confirmado la variación existente en los caracteres tradicionalmente utilizados para distinguir a Di- chanthelium a nivel genérico o infragenérico. ROSETA BASAL La roseta basal, propia de las especies norteame- ricanas, no se observa en P. sabulorum o en las restantes especies de Dichanthelium que crecen en México, Mesoamérica y América del Sur. Asi por ejemplo en P. sphaerocarpon, P. acuminatum Sw., y P. aciculare Desv. ex Poiret, especies co- munes en Атепса del Norte y con una morfologia similar al resto de las especies del subgénero que crecen en el área, no se forma una roseta basal en las plantas que habitan en Meso y Sudamérica. Una excepción se da en especimenes de P. aci- culare coleccionados en Cuba y de P. scoparium de la Repüblica Dominicana, en los que se observa una roseta basal; en este caso se podria sugerir que las especies se introdujeron en las Indias Oc- cidentales desde los Estados Unidos y mantienen, en determinados casos, especializaciones como el poseer roseta basa RAMIFICACIONES El patrón de ramificación presente en P. sa- bulorum ых, con el que se halla еп las del género, como asi tam- s en las restantes especies sudamericanas de Dichantheliu abe Fo la persistencia de las canas du- rante la temporada invernal, y la posterior for- mación de nuevas ramificaciones, e inflorescencias, a partir de las mismas durante la primavera si- guiente. La ausencia de roseta basal y permanencia de las canas, con una eventual producción de espi- guillas, puede explicarse por la ausencia de in- viernos rigurosos tanto en Mesoamérica como en América del Sur. DIMORFISMO FOLIAR El dimorfismo foliar, con hojas de tamano di- ferente en las canas simples y ramificadas, sólo se halla presente en especies de distribución boreal y en P. sabulorum, siendo esta ültima la especie del subgénero Dichanthelium con distribución mas austral en America del Sur. En el resto de las especies de Dichanthelium las canas principales se encuentran ramificadas, llevando las ramifica- ciones hojas de tamano similar a las de la cana principal. Durante el presente estudio se comparó el pa- trón de ramificación y dimorfismo foliar con las restantes especies de Dichanthelium que crecen en América del Sur, en particular con aquellas que 926 Annals of the Missouri Botanical Garden habitan desde América del Norte hasta Colombia y Venezuela. Se analizó material, en vivo y de herbario, de P. acuminatum, P. aciculare y P. sphaerocarpon, proveniente de Colombia y Ve- nezuela; este material no presenta ningün tipo de dimorfismo foliar ni roseta basal, caracteres que si se hallan presentes en A imenes de estas es- pecies de América del Nort CASMOGAMIA Y CLEISTOGAMIA El porcentaje de flores cleistógamas en las in- florescencias es muy bajo o nulo en las especies sudamericanas de Dichanthelium y mayor en las especies que crecen en América del Norte. Esta variación ha sido corroborada en la presente con- tribución mediante el estudio de especies norteame- ricanas de Dichanthelium que llegan hasta Ате- rica del Sur, como por ejemplo P. aciculare, P. acuminatum y P. sphaerocarpon. En el material analizado de estos taxa, coleccionado en Colombia y Venezuela, se hallaron pocas flores cleistógamas. La presencia o ausencia de flores cleistógamas está discant relacionada, tal como ocurre con los caracteres vegetativos, con la adaptación de las plantas a condiciones ambientales desfavorables. ANATOMÍA FOLIAR El análisis de la estructura histofoliar de P. sa- bulorum, y especies de la sección Dichanthelium, permitió concluir que esta sección posee una serie de caracteres histofoliares comunes, siendo todas sus especies no-Kranz, con clorénquima irregular- mente radiado alrededor de los haces vasculares, con yacentes, usualmente de 5 a 8 células; células de la vaina parenquimática de los haces vasculares más de 4 células entre haces vasculares ad- sin cloroplastos especializados, presencia de almi- dón en las células clorenquimáticas, y conspicuas extensiones parenquimátic as de la vaina externa ha acia has enidermis estomas con cellas subsidiarias triangulares a cu- En la epidermis se observan puliformes, y micropelos 2-celulares, fusiformes, con la célula basal igual o mayor que la distal, y célula distal atenuada o aguda. ORNAMENTACIÓN DEL ANTECIO SUPERIOR El antecio superior de P. sabulorum comparte las caracteristicas presentes en el resto de especies del subgénero, con papilas en la superficie y lemma comünmente apiculada o mucronada y con estomas hacia su ápice. CONCLUSIONES El estudio efectuado en poblaciones de P. sa- bulorum, y en otras especies del subgénero Di- chanthelium, seccion Dichanthelium, de Meso- атепса (Zuloaga et al., en prep.) permite concluir que los caracteres tradicionalmente utilizados para distinguir a Dichanthelium no se mantienen cons- tantes dentro del зиђрепего. Dichos caracteres, presencia de roseta basal, dimorfismo foliar y cas- mogamia y cleistogamia, representarian una estra- tegia evolutiva de las plantas a climas más rigu- rosos. El hecho de que las especies que crecen en América del Norte posean caracteres adaptativos más especializados, y la existencia de un alto пи- mero de especies diploides del sübgenero en Ame- rica del Sur (Dubcovsky & permite sugerir que este grupo tiene, contraria- mente a lo sustentado anteriormente por Hitchcock & Chase (1910), Brown (1948) y Gould & Shaw (1983), su centro de origen en America del Sur, desde donde radio hacia América del Norte, adqui- riendo formas morfologicas evolucionadas, distri- buyéndose ampliamente al adaptarse a climas con una estacion invernal marcada. Los resultados del presente trabajo nos permiten concluir que no existen caracteres suficientes para uloaga, en prep.) mantener a Dichanthelium como un репего se- parado de Panicum. Dentro genero Dichanthelium se distingue de los restantes subgéneros por una suma de caracteres, incluyendo de Panicum el sub- (además de la presencia o ausencia de roseta basal, dimorfismo y cleistogamia, ya citados previamente) el que poseen por lo general inflorescencias laxas, difusas, con espiguillas usualmente elipsoides a obo- voides, gluma superior y lemma inferior usualmente 7-9-nervias, antecio superior papiloso, con papilas simples distribuidas regularmente, cortamente mu- cronado o apiculado en el ápice, y tipo fotosintético y anatómico no-Kranz, con extensiones de la vaina parenquimática hacia ambas epidermis. LITERATURA CITADA BENTLEY, B. 1976. Plants Е extrafloral nectaries and the associated ant community: interhabitat dif- indeque in the reduction of herbivore damage. Ecolo- gy 57: 815- i Spikelet structure in Panicum aus- onomic and ecological impli- ustral. J. Bot. 33: 579-583. BOWDEN, B. 1971. Studies on Andropogon ga- и Kunth. VI. The leaf nectaries of Andropogon anus var. bisquamulatus (Hochst.) Hackel (Gra- J. Linn. Soc. 64: 77-80. 1948. A cytological study in the Gra- mineae. Amer. J. Bot. 35: 382-396. & B. М. SMITH. 1975. The genus Dichan- Volume 78, Number 4 1991 Morrone & Zuloaga Estudios Morfológicos en Panicum 927 — Gramineae). Bull. Torrey Bot. Club 102: EUM -13. du. C. A. & Е. У. Соло. 1975. Some epidermal characteristics of Paleas of ro Panicum and Echinochloa. Amer. J. Bot. 62: 743-748. T is ы & S. A. RENVOIZE. 1986. Genera ‚ Grasses of the World. Kew Bull. Addi- ul Geis XIII. London. Connor, H. E. 1979. Breeding systems in the grasses: a survey. Zealand J. Bot. 17: 547-574 DavipsE, С. A systematic study of the genus Lasiacis те Paniceae). Ann. Missouri Bot. Gard. 65: 1133-1254. 19 Fruit dispersal in the Poaceae. Pp. ` 143-155 in T. oderstrom, K. ilu, C. S. Campbell & M. E. Barkworth fedi. Grass Sys- tematics nis Evolution. Smithsonian Inst. Washingto Press, A revision of the genus Prionanthium (Poaceae: Arundineae). Bothalia 18: 143-153 DuBors, M., К. A. GILLES, J. К. HAMILTON, P. А. REBERS & А. Е. 5мин. 1956. Colorimetric method for determination of sugars and related substances. Analytical Chem. 28: 350-356. Erias, T. S. & Н. GELBAND. 1976. Morphology and anatomy of floral and extrafloral nectaries in Campsis ни Amer. J. Bot. 63: 1349-1353. 1977. Morphology, anatomy, and relationship of extrafloral nectaries and hydathodes in two species o atiens пе Bot. mp Gaz. Ded cum 138: 206-212. FAHN, A. ecretory Tissues in Plants. Academic Press, London, New York & San Francisco. GOULD, Е. W. . Nomenclatural changes in the Poaceae. Brittonia 26: 59-60. 1980. The Mexican species of Dichanthelium Brittonia 32: 353-364. 1978. Dichanthelium (Poaceae in the Unitec States and Canada. Ann. Missouri Bot. Gard. 65: 1088-1132. оркен}. — 1983. Grass Systematics. HiTCHCOCK, А. S. & A. 1910. The North Ame- rican он of Panicum. Contr. U.S. Natl. Herb. 1 «eie W. A. 1962. Botanical Histochemistry. Prin- ciples and РА W.H. Fr КАВОУЕ, C. Н. =. & D. Woon. of NS gland hairs in the Gramineae. Bot. Linn. Soc. 62: 69-770. LELONG, M. 1984. New combinations for Panicum sub- genus Panicum and subgenus Dichanthelium (Poa- ceae) of the southeastern United States. Brittonia 36: 262-2 LiNpER, Н. P. & К. P. ELis. 1990. A revision of Pentaschistis di dev Poaceae). Contr. from the Bolus Herb. 12: 1-124. Mattel, С. E. & C. TRoPEA. 1908. Graminacee prov- viste di nettarii A dd Boll. Reale Orto Bot. Giardino Colon. Palermo 7: 113-117. 1909. Ricerche e studi sul genere e ign in 1 rapporto ai nettarii estranuziali. Contr. Biol. Veg. 4: 205- МР С. R. 960. Anatom of the Monocotyle- s. I. Gramineae. Clarendon Press, Oxfor Ком, E. G. 1941. Contri ibución al estudio histológico de | gm epidérmicas de algunas especies de stis. Darwiniana 5: 316-321. 1968 Era PALACIOS, R P. A Variaciones en sek iba er = lg Bol. Soc. Argent. Bot. 12: 38-43. hee: J. В. & Г. J. TooLin. 1989. s on Pap- pophorum (Gramineae: Pappophoreae). on Bot. 58 14: 349-358. SANCHEZ, E. & M. I. САЗАВОМА. 1981. Presencia de cristales en hojas de Paniceae (Gramineae). Bol. Soc Argent. Bot. 20: 83-90. Sass, J. E. 1940. Elements of Botanical Microtechni- que. McGraw-Hill, New York & London. SCHMID, В. & M. D 1977. Contrad 70, an effective softener of herbarium material for anato- mical study. Taxon 26: 551-552 Tiree, А. M. 1966. Histotaxonomia de las especies argentinas del género Paspalum. Lilloa 32: 35-299, 97-498. WEBSTER, R. D. 19 Genera of the North American Paniceae (Piatssi Panicoideae). Syst. Bot. 13(4): -609. ZULOAGA, F. O. 1987. Systematics of the New World ecies of Panicum (Poaceae 306 in T. & M.E. Barkworth (editors), Craii ae and Evolution. Smithsonian Inst. Press, Washington, D.C. T. SENDULSKY. 1988. А revision of Panicum subgenus Phanopyrum section Stolonifera (Poa- ceae: Paniceae). Ann. Missouri Bot. Gard. 75: 429- 455. E 4 J. DuBCOVSKY. Exo- E а. апа cytological studies in Panicum validum (Poaceae: Panicoideae: Paniceae): its systematic position within the genus. Syst. Bot. 14: 220-230 А REVISION OF RADDIELLA (POACEAE: BAMBUSOIDEAE: OLYREAE)' Fernando O. Zuloaga? and Emmet J. Judziewicz? ABSTRACT Raddiella is a genus of seven species of low- to mid-elevation, small herbaceous bamboos (Poaceae: Bambusoideae: Olyreae) that occurs m Panama and Trinidad to Bolivia and Brazil. Raddiella esenbeckii is widespread in savannas and cerrados, but its s x congeners are rare, e , R. minima, with the many occur only on wet rocks in the spray zone below waterfalls. mallest leaves of any bambusoid grass, and R. lunata, with lunate es female florets, are described This first revision of Raddiella includes a key to the recognized species, descriptions, and distribution maps, as important interspecific differences well as | волио electron micrographs of the female florets, which reveal taxonomically Along with the Cuban genus Mniochloa Chase, members of Raddiella Swallen (Poaceae: Bam- busoideae: Olyreae) are the smallest bambusoid sses in the wor e genus was described (Swallen, 1948) on the basis of three species, Rad- diella nana (Doell) Swallen (= R. esenbeckii (Steu- del) Calderón & Soderstrom), the type species; К. truncata Swallen (= Parodiolyra lateralis (Nees) Soderstrom & Zuloaga); and R. malmeana (E. Ekman) Swallen. Swallen characterized the genus as monoecious with the male and female spikelets in different inflorescences; the pedicels of female spikelets not conspicuously thickened upwards; fe- male spikelets with the glumes 3-nerved, and the floret indurate, white, and smooth. Soderstrom (1965) described three new species of Raddiella, R. kaieteurana, R. maipuriensis, and R. potaroensis, and included a key to the six species in the genus. Later, based on priority, Cal- derón & Soderstrom (1980) considered К. esen- beckii as the valid name for the type species of the genus and transferred Panicum molliculmum Swallen to Raddiella (as R. molliculma). Examination of more herbarium material since Soderstrom's summary of the genus has revealed two distinctive new species of Raddiella, and a scanning electron microscope study of female flo- rets has proven useful in clarifying species limits, warranting the following revision of the genus. MATERIAL AND METHODS Classical alpha taxonomic methods were used to study Raddiella. In addition, scanning electron micrographs were made of the female florets of a few populations of each of the species except К. minima, using the same procedures and equipment as described by Soderstrom & Zuloaga (1989). The vouchers for this study are marked with an asterisk (*) in the lists of specimens examined in the taxonomic treatment. MORPHOLOGY Species of Raddiella are remarkable in the Bambusoideae for their very small size. Most spe- cies appear to be annuals and represent the only annual bambusoids, excluding oryzoids. Many are obligate phreatophytes, growing only on wet rocks n the constantly moist spray zone of riversides below waterfalls; exceptions are the perennials К. esenbeckii and possibly R. minima, which often grow in dry savannas and cerrados. Raddiella езепђески and npa R. minima are also unique i nyctinasty, or sleep move- ments, he blades folding upwards at night or under water stress. The blades are firm and conspicuously asymmetrical in R. esenbeckii and R. minima, thinly membranous and only slightly asymmetrical in the other species. int e genu ' We are grateful to Vladimiro Dudas and Bruno Manara for the illustrations, to Bruno Manara for assistance with T the Latin diagnoses, and to се Davidse, Lynn Clark, Edga rdo Romero, ary Sangrey, and the late Thomas R. Soderstrom for help in various way that allowed him to of P erry, editor of the Flora Mis Venezuelan 2 Instituto de Botánica Darwinion, Cas es gratitude to the Guggenheim Foundation for a fellows ship EUM work on this d at the Мон Бели Garden. Figure 13A-G is used with permission ana a de Correo 22, 1642 San Isidro, Argentina * Department of Botany, Birge Hall, Univer: of Wisconsin, Madison, Wisconsin 53706, U.S.A. ANN. Missouni Bor. GARD. 78: 928-941. 1991. Volume 78, Number 4 1991 Zuloaga & Judziewicz 929 Revision of Raddiella Species of Raddiella bear terminal and nu- merous axillary inflorescences; both types are con- tracted, few-flowered, and barely exserted from the leaf sheaths. The several inflorescences arising from the uppermost node each consist of several (rarely one as in R. minima) early-deciduous male spike- lets on short, filiform pedicels. Several axillary in- florescences occur at each node, and the disposition of spikelets in these is more variable. Typically, the entire inflorescence is female, with spikelets borne on short, filiform pedicels. However, at a given node, inflorescences bearing 1-2 terminal female spikelets as well as 1-several short-pedi- celled male spikelets positioned below them can be found. Exceptions to these patterns include R. доча, in which both terminal and axillary re long-peduncled and each consists of 1-2 terminal male and 1-2 subterminal female spikelets; and К. lunata, with racemose-appearing inflorescences that may be up to 12-flowered, but in which the disposition of the sexes in the single, overly mature specimen is not entirely clear. The spikelets of Raddiella are fairly typical of those of other olyroid genera. In the female spike- lets, the glumes vary from persistent in R. esen- beckii and R. minima, to deciduous together with the floret in the other species. The glumes аге 3- , and there is a characteristic indurate, thickened, interglumal internode present. The fe- male floret morphology is ellipsoidal in all species (Fig. 13) except for R. lunata, which has curious, lunate (crescent-shaped in profile) florets that taper to an acute point at both ends (Fig. 14). Floret epidermal texture, however, ranges from smooth to strongly papillose and is discussed in the section on anatomy below The male пјене are hyaline and early decid- uous, and their relative size with respect to the 5-nerved female spikelets is apparently of some taxonomic value. The caryopses of most species are ovoid-ellip- soid, and the hilum is central punctiform or short- elliptical, a type not present in other bambusoid grasses (Calderón & Soderstrom, 1980). However, in R. esenbeckii (but not its putative sister species R. minima) the hilum is short-linear. ANATOMY A scanning electron microscope study of the mature female florets proved useful in clarifying species limits in Raddiella, just as in Panicum (Zuloaga, 1987) and Olyra (Soderstrom & Zuloa- ga, 1989). Most species of Raddiella had distinc- tive lemmatal epidermal features that are described below. Common to all species was the presence near the base and/or apex of the palea of circular 10-15 um diam., present on the female paleas of various species of Olyra L. and Parodiolyra Soderstrom & Zuloaga (Soderstrom & Zuloaga, 1989). These excavations are apparently caused by the deflation of certain cells when т. in the vacuum of the scanning excavations, similar to those electron microsc Raddiella Macc em (Figs. 1, 2). Female lem- ma completely smoot Raddiella ЕР (Figs. 3, 4). Both lemma and palea covered throughout with simple papillae 4—5 um diam. Raddiella lunata (Figs. 9, 10). Both lemma and palea with cells with a single large, central, compound papilla 9-11 um diam., bearing on its raised margins 3—5 smaller papillae 3—4 um diam. Raddiella malmeana (not illustrated). Lemma completely smooth (but available floret slightly im- mature). Raddiella minima. Not included in scanning electron microscope study; in an optical micro- scope, the lemma appears to be completely smooth, as in К. esenbeckii. Raddiella molliculma (Figs. 5, 6). Both lemma and palea with the central portion of each cell enlarged and protruding (ca. 30 x 20 um), and with 10-12 papillae, 3-5 um diam., present along the undulating margins. Raddiella potaroensis (Figs. 7, 8). Both lemma and palea with cells with the central portion of each cell enlarged, 15-18 um diam., surmounted by 7- 10 smaller papillae 3-4 um diam. RELATIONSHIPS The possible affinities of Raddiella with several other putatively related genera of the Olyreae are summarized in Table 1. It will be noted that the type species, R. esenbeckii, has several characters in common with Parodiolyra lateralis (Nees) So- derstrom & Zuloaga rather than with the phreato- phytic species of Raddiella (with the possible ex- ception of the poorly known R. minima): perennial habit; firmly membranous leaves; smooth female florets; and a short-linear rather than punctiform hilum. Arguing for the retention of the phreatophy- tic species within Raddiella is the presence of the following feature in common between them and А. esenbeckii: few-flowered panicles with the sexes usually separate, the males terminal and females axillary. Moreover, R. esenbeckii and P. lateralis have quite different chromosome numbers (n — 10 and 18 (Davidse & Pohl, 1978), respectively), and Annals of the Missouri Botanical Garden FIGURES 2 6. Scanning ici modu of female florets of Raddiella species. — 1, 2. R. esenbeckii (Pires 51123).—3. К. kaieteurana (Egler 1244).—4. К. kaieteurana E ei & dé dads 23089). —5, 6. R. mol- liculma maia & Schultes 616). Scale i 50 um for 1, 3, 5; 10 um for 2, the latter has leaf blades that are not nyctinastic and female spikelets with indurate, strongly clasp- ing glumes that become black at maturity. TAXONOMIC TREATMENT Asterisks (*) indicate vouchers used in this study. Raddiella Swallen in Maguire et al., Bull. Torrey Bot. Club 75: 89. 1948. ТУРЕ: Raddiella nana (Doell) Swallen [= К. esenbeckii (Steu- del) Calderón & Soderstrom], Smithsonian Contr. Bot. 44: 21. 1980 Plants small, monoecious, either tufted peren- nials or delicate annuals, often forming low mats. Volume 78, Number 4 1991 Zuloaga & Judziewicz Revision of Raddiella Ficu (Maguire & Fanshawe 32277).—9, RES 7-10. Leaves small, pseudopetiolate; blades elliptical to ovate-triangular, often strongly asymmetrical and with an apiculate apex, exhibiting sleep movements or not. Inflorescences small, few-flowered, 1-sev- eral from both terminal and axillary nodes, barely exserted from the leaf sheaths, 2-6(- 12) не per inflorescence. Terminal inflorescences male an female or strictly female; axillary inflorescences either all fe- emale and male spikelets; female spikelets minutely cupulate at apices. Spikelets 1 -flowered, the broad female spikelets shorter than more commonly strictly male, rarely male or wit to as long as the narrow male spikelets. Female spikelets falling entire from pedicels or in one spe- cies the glumes persistent and the floret falling, ovate-lanceolate; glumes subequal, membranous, as long as the spikelet, acute, 3-5-nerved, the lateral nerves often obscure; floret elliptical to ovoid, acute, glabrous, thinly coriaceous, stramineous or becoming dark when mature, the margins of the lemma inrolled over the edges of the palea; style 1, stigmas 2. Caryopsis ovoid to globose, brownish tan; embryo small, basal; hilum punctiform or short- Scanning electron micrographs of female florets of species of Raddiella. — 7, 8. R. potaroensis 10. R. lunata (Rondon s.n.). Scale bar 50 um for 7, 9; 10 um for 8, 10. linear. Male spikelets borne on filiform pedicels, hyaline, glabrous, early deciduous, linear to lan- ceolate; glumes absent; lemma 3-nerved; palea 2-nerved; stamens 3 Distribution and ecology. Seven species ranging from Panama and Trinidad to Bolivia and Brazil (to Зао Paulo); savannas, forest margins, or among wet rocks near rivers and waterfalls, from near sea level to 1,5 Raddiella has two centers of diversity: (1) the Guayana Highlands, with four species, two of them endemic (including the Colombian species К. mol- liculma at the far western edge of the Guayana Highlands formation); and (2) the western Brazilian Planalto from Rondónia to northern Mato Grosso and southern Pará, with five species including three endemics. More new species may be expected from oth areas. KEY TO THE SPECIES OF R ADDIELLA la. Female florets smooth 2a. Leaf blades elliptical, + symmetrical, acute at both base and apex; female glumes de- Annals of the Missouri Botanical Garden TABLE 1. Comparison of Raddiella with related genera. Raddiella Parodiolyra Parodiolyra Raddiella (phreatophytic Olyra lateralis (2 species) esenbeckii species) Caespitose Caespitose to — Sprawling Caespitose Sprawling perennials sprawling perennials perennials annuals perennials Nyctinasty - = - + eT Leaf blade asymmetry - (+) - + = Leaf blade texture Firm Firm Firm Firm Delicate Axillary inflore- scences well — —, + —, + + + developed Inflorescence size Large Moderate Moderate Small Small Sexes in separate (+) — + + inflorescences Female spikelet c (6 f f f f pedicels clavate or iform Female spikelet falling — (+) + + — + entire Female spikelet: = + + + + thickened internode between glumes Female floret Variable Smooth Smooth, viscid Smooth Papillose or texture smooth Hilum Fully linear Short-linear Short-linear Short-linear Punctiform А (са. 1/2 length) (са. 1/2 length) length) Chromosome п= 7, 10, 11, 18 Unknown 10 Unknown 20 ciduous along with the floret 2b. Leaf blades ovate-triangular, axymme tri- cal, the ba se truncate, the apex acute or apiculate; female gumes persistent, only the flor et deciduo R. malmeana 3a. Female eu 1.9-2.7 mm long; male spikelets 3-5 mm long, (1-)2- 4 per inflorescence; leaf blades 9-22 , 4-11 mm wide; caryopsis hilum short-linear; wide- . R. Ба w c — . Ф dE Bg [7] n =. =: "a -& 2. 0 n — > E =] 5. = 99 3 E Ф m wide; caryopsis with the um punctiform; rare, Pará, Brazil 5. minima lb. Female florets minutely papillose with simple or f урана papillae all over surface a male florets binate тенин shaped) i in gle compound papilla in the middle of each cell; rare, Rondônia, Brazil nuc ds 4b. Female florets straight and lanceolate in profile, terete, not beaked; female floret . lunata epidermis without a compound papilla in the middle of each cell 5a. Inflorescences long-exserted, some pe- duncles at least 5 cm lo ng; female florets with papillae along the cell mar- gins; rare, Caquetá, Colombia .............. сл c molliculma . Inflorescences with bases included in leaf sheaths, or short-exserted on pe- duncles less than 1 c male florets with 1 -several papilla per cell; Venezuela, Guianas, Bra с c eaf blades simple papillae per cell; male spikelets 2.7-4.2 mm lon T razi with the lower sur- (асе hispid to _strigose through- m long; fem .4-2 mm 1 simple papilla е. iih male spikelets 4-5.8 m 2. R ee . Leaf blades with the lower su face glabrous or with short, iab. rid hairs only on the veins; female . potaroensis Volume 78, Number 4 1991 Zuloaga & Judziewicz 933 Revision of Raddiella 1. Raddiella esenbeckii (Steudel) Calderón & Soderstrom, Smithsonian Contr. Bot. 44: 21. anicum esenbeckii Steudel, Syn. Pl. Glumac. 1: 1854; based on Panicum lat- erale var. 3 Nees, Agrost. Brasil. 213. 1829. Basis in part of Olyra nana Doell in C. Mar- tius, Fl. Bras. 2(2): 329. 1877, nom. superfl. Raddia nana (Doell) Chase, Proc. Biol. Soc. Wash. 21: 185. 1908. Raddiella nana ix Swallen, Bull. Torrey Bot. Club 75: 8 TYPE: Brazil. Amazonas: Rio Negro, a vee tius s.n. (holotype, M? not seen; the same specimen designated as a lectotype of O. nana 2, not seen). Figures 1, 2, 13A-D Densely tufted perennials, with up to 60 un- branched culms per clump. Culms sprawling, de- cumbent to geniculate, the erect portions 8-40 cm tall, wiry; internodes cylindrical, glabrous to sparse- ly pilose with short, retrorse hairs; nodes thickened, densely pilose with whitish retrorse hairs. Leaves in complements of 7-15, nyctinastic, the blades folding upwards at night or under water stress; sheaths striate, glabrous to short-pilose, densely so toward the summit, one margin membranous, the other short-ciliate, short-auricled, the auricle mem- branous on one side; ligules ca. 0.3 mm long, membranous, apically short-ciliate; pseudopetioles ca. 0.5 cm long, densely pilose; blades 9-22 mm long, 4-11 mm wide, ovate-triangular, asymmet- rical, truncate at the base, acute and short-apic- ulate at the apex, flat, firmly membranous, either glabrous, or short-pilose near the base on the upper surface, to densely puberulent on both surfaces, the lower margins ciliate, the lower surface purplish in some specimens. Inflorescences terminal and axillary; terminal inflorescences 3-5, with male spikelets only, on cylindrical, glabrous peduncles 1-2 cm long, each inflorescence with (1—)2—4 male spikelets on short, glabrous pedicels; axillary inflo- rescences both female and male, the female inflo- rescences on short-pilose peduncles 5-8 mm long, each bearing 2 short-pedicelled female spikelets. Female spikelets 1.9-2.7 mm long, wide, ovoid; glumes firmly membranous, greenish becoming blackish, sparsely to densely short-pilose with spreading hairs, 3-nerved, persistent, the in- ternode between the glumes thickened, the lower glume acute, slightly shorter than the acuminate upper glume; floret 1.6-2 mm long, ellipsoid, co- riaceous, smooth, shiny, glabrous, whitish becom- ing dark, deciduous. Caryopsis 1-1.2 mm long, 0.7-0.8 mm wide, long-ellipsoid, dorsally com- pressed; hilum short-linear, ca. / the length of the caryopsis, extending from near the center to near the base; embryo са. У, the length of the caryopsis. Male spikelets borne on minutely cupuliform pedi- cels, 3-5 mm long, ca. 1 mm wide, lanceolate, acuminate, greenish to purplish, hyaline, glabrous to sparsely hispid with short hairs; glumes occa- sionally present, to 3.7 mm long, linear-lanceolate, sparsely hispid; lemma as long as spikelet, lanceo- late, acuminate, glabrous to sparsely hispid with short hairs; stamens with anthers 1.3-3 mm long. Distribution and ecology. Panama and northern South America (including Trinidad) to Bolivia and central Brazil (Fig. 1 1); from 0 to 1,500 m in semishaded or more often dryish, open places: gallery forests, woodlands, savannas, and cerrados, often on rocky or sandy soil. Chromosome number. 10 (Davidse & Pohl, 1972) п = Representative specimens examined. PANAMA. PAN- AMA: 7 mi. N of Cerro Azul, 2,600 ft., 13 Nov. 1965, Blum et al. 1789 (МО); Cerro Campana , 900 m, 21 Jan. 1984, van der Werff & Herrera 6156 (MO); hills NE of Hacienda La Joya, 50-300 m, 9 Dec. 1934, Dodge et al. 16882 (MO). HERRERA: N slope of Cerro Alto Higo, 2,400 ft., 5 Aug. 1978, Hammel 4202 (MO). seges AMAZONAS: Bella Vista, Rio Igará- Paraná, 8 Sep. 1973, Sastre 2152 (US). META: La Macarena, Rio riis Sabanas de Arenisca, Jan. – Маг. 1959, García-Barriga & Jaramillo 17089 (US). vauPÉs: Río Guainía, Caño del Caribe, between Isla del Venado and San Jose, 2?45'N, 67*50'W, 2 Nov. 1952, Schultes et al. 18262 (US); Rio Guainia basin, Río Naquieni, vicinity of Cerro Monachi, June 1948, Schultes & López 10115 (US); Rio Kan- anari, Cerro Isibukuri, 28 Oct. 1951, Schultes & Cabrera 14452 (US). VENEZUELA. AMAZONAS: Piedra Arauicaua, Río Yatua, 27 Sep. 1957, Maguire et al. 41608 (US), а d 1959, ас = Adderley 43457 (US); Rio uma, occasional on moss- vera boulders, up- river q Playa ries 14 "Nov. 1950, Maguire et a 29500 (US); banks of Rio Cuao, above Cuao Creek, one 1948, Maguire & Politi 27376 (US); Rio урне . 1 km above the confluence with Rio Atabapo, 6-19 July 1969, Bunting et al. 3658 (US); Cerro Duida, Rio Cunucunuma, Culebra Creek, 1,500-1,600 m, 21 Nov. BOLÍVAR: cercanias del pariente en iie lado sur del Río Ichün, tributario del Rio Paragua, debajo del Salto Maria Espuma, 4?46'N, 63?18'W, 30 Dec. 1961, Steyermark 90470 (US); Dist. Heres, San Salvador de Paúl, 6°02'N, 62°53'W, 2 Dec. 1982, Davidse & Huber 23128 (US). SUCRE: 8 km S de Santa Fé, 10?16'N, 64?24'W, 230 m, Davidse & González 19069 (MO, US); carretera Guanta- Los Altos, Tamayo 2146 (US). TÁCHIRA: Dist. Uribante, Siberia to Pregonero, 8°55'N, 71?40'W, 1,300 m July 1983, van der Werff & González 5365 (MO). TRINI- DAD. Aripo Savanna along the Cumuto road, 3 Nov. 1929, Broadway 7685 (MO), 5 Aug. 1970, Davidse 2550 US). Guyana. Kaieteur savanna, 5 Sep. 1937, Sandwith 1375 (US), 11 Feb. 1962, Cowan & Soder- rom 1807 (05); Waranama ranch, Harrison & Рег- эбе 1069 (ВЕС, К); Lama Dam, Jenman 5971 (ВЕС). Е 934 Annals of the Missouri Botanical Garden | о 200 400 600 800 !000km | А t pop onm Q 100 200 300 400 500 600 miles | Prepared by Hendrik R. Rypkema FIGURE 11. Distribution of Raddiella Grote Zwiebelzwamp, Lanjouw & Lindeman 872 (NY, U); Sipaliwini savanna, Oldenburger et al. 14 1 (U). FRENCH GUIANA. 10680 (CAY, US). BRAZIL. AMAPA: Matapi, Pires & Silva 4797 (US); Igarapé do Paia, km 110 on road to Amapa, 20 July 1962, Pires & Cavalcante 52233 (MO, NY, US). AMAZONAS: З km E of Borba near the Rio Madeira, 4°23'S, 59°35'W, 22 June 1983, НШ 12753 (Е, MO, US); Rio Jauari, affluent of Rio Aracá, 0°49' 3, 63°20'W, July 1985, Huber et al. 107 10 (МО); Rio Urubú, Serra da Lua, 6 June 1968, Prance et al. 5005 (MO); Manaus, Rio Cuieiras, Igarapé Cachoeira, 19 Dec. 1961, К drigues & Wilson 3191 (05); km 38 on road from © f | | ~ M и Humaitá to Pórto Velho, 6 May 1982, Teixeira et al. 260 (MO). BAHIA: Serra do Sincorá, 16 km N of Barra da Estiva on the Paraguagú road, 13?20'S, 41%20'W, 31 Jan. 1974, H Rio Bartolomeu, 5 May 1980, Heringer et al. 4626 (US); Chapada da Contagem, Par- que Municipal do Gama, ca. 25 km W of Brasilia, 1,100 m, 3 Feb. 1968, Irwin et al. 19474 (US), 700-1,000 ca. km S of Alto do Paraiso, 22 et al. 24921 (MO, US); S Serra Dourada, gallery forest 20 km E of Formosa, 13?45'S, 48*50'W, 18 Мау 1956, Dawson 14956 (US); vicinity of Goiabeira, n- ápolis and Goiás, 23 Mar. 1930, Chase 11510 (MO, US); k ca. 3 km E of Alto Paraiso on road to Nova Roma, Volume 78, Number 4 1991 Zuloaga & Judziewicz Revision of Raddiella 935 Mar. 1973, Anderson 6604 (MO, US). а Редга Caida, 35 km М of Carolina, 7%08'S, 47%25' pr 1983, Taylor et al. E-1240 (US). МАТО CROSSO: ca. 5 m М of Barra do Garças, S face of mountain, 7 May 1973, dun 9885 (05); km 762, Santarém- Cuiabá road, 9935'5, 54°55’W, Amaral et al. 806 (Е, MO). MINAS GERAIS: Vaccaria, Serra do Cipó, Apr. 1925, Chase 9257 (МО, US); Metallurgica, Serra de Ouro Branco, E Dec. 1929, Chase 10290 (MO, iti m E of Diamantina, 790 m 4 27993 (MO, US); Serra do Espinhago, 3.5 km by road SW of Rio Jequiti and Mendanha, 14 Apr. 1973, Anderson 8922 (US). PARA: 7 km SE of Vigía, 0%55'S, 48°04'W, 31 Mar. 1980, Davidse et al. 17689 (МО); Rio Arua, Black 49-844 (US). RONDONIA: Abuna, Rio Madeira, 11 July 1968, Prance et al. 5887 (MO); Pórto Velho, caminho para o km 8, 26 May 1952, Pay a Cordeiro 52-14526 (US). SAO PAULO: Sao José pos, 16 Dec. 1909, Lofgren 4680 (US). to BENI: Prov. Vaca Diez, 15 km W de Guayamerim, camino a Riberalta, en campos con rocas aflorantes, 17 Apr. 1979, Krapovickas & Schinini 35062 (US This species is variable in stature. Alone among members of the genus, the leaf blades exhibit sleep ques folding upwards at night or under wa- ter stre The pas ally similar Parodiolyra lateralis differs in its sprawling habit, larger, more rigid blades lacking sleep movements, and larger, more open mixed panicles with mature female spikelets with blackish, indurate, 3-7-nerved glumes that strongly invest the floret. № Raddiella kaieteurana Soderstrom, Mem. New York Bot. Gard. 12(3): 6. 1965. TYPE. Guyana: Kaieteur Plateau, on moist cliff faces along trail from Kaieteur Plateau to Tukeit, 1,100 ft., 4 Feb. 1962, R.S. Cowan & T.R. Soderstrom 1742. baone US*; isotypes, GH, K, NY, P). Figures 3, 4, 13E-G. Delicate mat-forming herbs of indefinite dura- tion. Culms weak, decumbent, rooting and branch- ing at the lower nodes, the flowering portion 5(- 20) cm long, many-noded; internodes compressed, glabrous. Leaves with the sheaths striate, longer than the interncdes, greenish to purplish, with long hairs near the apex, otherwise sparsely pilose to glabrescent; ligules membranous, 0.3-0.5 mm long; pseudopetioles 0.5 mm long, hirtellous; blades 8- 17 mm long, 3-8 mm wide, flat, delicately mem- branous, densely hispid to strigose on both surfaces with long and short hairs, the base slightly asym- metrical, obtuse, and subcordate, the apex acute and slightly apiculate, the margins ciliate through- out or only toward the base of the blade. Inflores- cences terminal and axillary from the upper nodes, with bases included in leaf sheaths or short-exserted on peduncles less than 1 cm long; terminal inflo- rescences 2-6, mostly male, each bearing 1-2 spikelets, or occasionally female spikelets present; axillary inflorescences female, exserted up to 3 mm on filiform peduncles and bearing 1—3 female spike- lets on divergent pedicels 1.5-5 mm long. Female spikelets 1.4-2 mm long, 0.6-0.8 mm wide, ellip- soid, falling entire; glumes ovate-elliptical, acuminate, subequal, prominently nerved, sparsely short-pilose with spreading hairs to uncommonly glabrous, greenish, the lower glume 3-nerved, sep- arated from the 3-5-nerved upper glume by a short internode; floret 0.8-1.7 mm long, 0.5- mm wide, narrowly ellipsoid, dorsally compressed, mi- nutely papillose, whitish becoming brownish at ma- turity; lemma evidently 3—5-nerved. Caryopsis 0.6- .8 mm long, 0.4-0.7 mm wide, broadly ellipsoid; hilum oblong, placed in the middle of the caryopsis and extending slightly toward the base; embryo № or less the length of the caryopsis. Male spikelets borne on pedicels 0.5-2.3 mm long, 4-5.8 mm long, 0.6–0.8 mm wide, fusiform, hyaline, sparsely short-pilose toward the apex; glumes occasionally present, to 4 mm long, linear-lanceolate; lemma 3-5-nerved; stamens with anthers 2-2.8 mm long. Distribution and ecology. Venezuela (Boli- var), Guyana, Surinam, and northern Brazil (Pará); on damp cliff faces or in the spray of waterfalls, from 100 to 810 m (Fig. 12). ional spec imens examined. VENEZUELA. BOLÍVAR: Te is Ата y-tepui, S side about 1 km from SW corn с ерш 5°54 М, 62*15'W, 550-810 m, 26 Apr. 1986, Liesner & Holst 20369 (MO). GUYANA: semishaded rock face in forest along trail from Kaieteur E Falls to Тикен, about 1 km Maguire & Fanshawe 23087 (M Feb. 1962, Cowan & Soderstrom пау im 1988, Hahn et al. 4163 (BRG, US); anno de Jenman 127 (US). SURINAM.: wert base of N escarpment, under ж 390 m, 11 Aug. 1944, Maguire 24397 (MO, N U); NW side, Geyskes 998 (U). BRAZIL. PARA: | ol (affluent of Rio Tapajos), Cachoeira Kereputiá, in a hole on sandstone rocks, 3 Mar. 1960, Egler 1244 US*). ~ As presently conceived, this species is variable. In Egler 1244 the female spikelets are only 0.9- 1 mm long and the glumes are covered with short hairs, but scanning electron micrographs revealed that female floret texture is similar to that of typical R. kaieteurana. An unusual feature of Cowan & Soderstrom 1918 is the presence of terminal fe- male inflorescences with two spikelets. Liesner & Holst 20369 differs from typical R. kaieteurana in its short-pilose to glabrous leaf blades and com- pletely glabrous female glumes. 936 Annals of the Missouri Botanical Garden ! р 1 E m u J Rad iella kaieteura a XR. lunata h Ш R. malmeana An R. minima | lliculma о 200 400 600 800 !000km 0) ===> O 100 200 300 400 500 600 miles Prepared by Hendrik R. Rypkema | FiGURE 12. Distribution of Raddiella kaieteurana, В. lunata, R. malmeana, R. minima, К. molliculma, and R. potaroensis. es . Raddiella lunata Zuloaga & Judziewicz, sp. Delicate, sprawling, mat-forming annuals. Culms nov. TYPE. Brazil. Rondónia: Serra dos Pacáas decumbent, rooting and branching at the lower Novos, Cab. do Cautário, planta umbrophila, nodes, the erect portions 7-15 cm tall, many- na entrada das lapas, firmas, predas, etc., Mar. | noded, unbranched; internodes cylindrical, hollow, 1917, C. Rondón s.n. [as J.G. Kuhlmann glabrous; nodes thickened, with short, retrorse, 1863] (holotype, RB; isotypes, K, MO, NY, whitish hairs. Leaves with the sheaths striate, gla- SI, SP, US*). Figures 9, 10, 14. brous, slightly inflated, the margins membranous, the auricles membranous, glabrous, the summit Ab omnibus conspecificis differt: spicula, flosculo et truncate; ligules not noted; pseudopetioles ca. 0.1 caryopside lunatis abrupte rostratis, necnon flosculo pa- MM long, brownish, short-pilose; blades 10-15 mm pilla composita in medio cuiusque cellulae praedito. long, 2.5-3.5 mm wide, lanceolate, delicately Volume 78, Number 4 991 Zuloaga & Judziewicz Revision of Raddiella FIGURE 13. А-О. Raddiella esenbeckii (Huber 2126). — A. Habit. —B. Female spikelet. — С. Female floret. — D. Male spikelet. E-G. Raddiella kaieteurana (Liesner 20414). —E. Нађи. —F. Female spikelet. —G. Female floret. E addiella molliculma (Grassl 10024). —H. Female spikelet, dorsal view. — I. Female spikelet, ventral view. J-L. Raddiella potaroensis (Cowan & Soderstrom 2162).—J. Female spikelet, dorsal view. — К. Female spikelet, ventral view. — L. Female floret. membranous, glabrous to sparingly short-pilose on both surfaces (especially on the veins below), the base slightly asymmetrical, acute at the apex, the margins scabrous. Inflorescences both terminal and axillary, with 1-several inflorescences produced at each node, short-exserted on glabrous, filiform pe- duncles up to 10 mm long; individual inflorescence a contracted, racemelike panicle 5-10 mm long, the rachis slightly sinuous, glabrous, alternately bearing 2-12 spikelets on short pedicels 0.5-1. mm long, or occasionally with a lower branch up to 3 mm long bearing 2 spikelets; terminal spikelet female, the lateral ones male or occasionally a few females. Female spikelets ca. 1 mm long, obovoid, lunate in profile, falling entire; glumes delicately membranous, short-hispid, 3-nerved, the lower glume not covering the apex of the floret, the upper glume as long as the floret, flat; floret 0.8-1 mm long, 0.5-0.6 mm wide, obovoid, dorsally com- pressed, concave on the ventral side, convex on the dorsal side, acute at the base, rather abruptly beaked at the apex, coriaceous, whitish becoming brown at maturity, strongly papillate; lemma con- spicuously grooved on the dorsal surface, slightly winged, the margins covering the edges of the palea with wide margins. Caryopsis 0.6-0.8 mm long, 0.5-0.6 mm wide, ovoid, lunate, of the same gen- eral morphology as the floret, slightly winged; hilum 938 Annals of the Missouri Botanical Garden rounded at the middle portion of the caryopsis; embryo / the length of the caryopsis. Male spike- lets 1.7-2.7 mm long, lanceolate-elliptical, acute, hyaline, glabrous; lemma 3-nerved; stamen with anthers 1-1.5 mm long. Distribution and ecology. Known only from the type collection made in the Serra dos Pacaas Novos in Rondónia, Brazil (Fig. 12). This species differs from all other congeners in its lunate (crescent-shaped, hence the specific ep- ithet) female floret that tapers rather abruptly to both ends, and the surface of the floret, which has — А. Habit. — B. Detail of a node. — C. Leaf. — —F. Lower glume.—G. Upp FIGURE 14. nd e ва s.n.).— D. оос . Male spikelet. F-J. Female spikelet er glume. H-J. Floret. —H. Dorsal view. — I. Ventral view. =; RA view. К, L: Caryopsis. — К. Dorsal view. —L. Ventral view. a single compound papilla on each cell. The form of the female spikelets could be associated with dispersion by water. In the mature plants at hand, the culms disar- ticulate readily at the nodes. Whether this repre- sents an adaptation for dispersal or is merely in- dicative of senescence is unclear. 4. Raddiella malmeana (E. Ekman) Swallen, Bull. Torrey Bot. Club 75: 89. 1948. Olyra malmeana E. Ekman, Ark. Bot. 10(17): 21. 1911. Raddia malmeana (E. Ekman) A. Volume 78, Number 4 1991 Zuloaga & Judziewicz 939 Revision of Raddiella Hitchc., Contr. U.S. Natl. Herb. 22: 505. 1922. TYPE. Brazil. Mato Grosso: Santa Ana da Chapada, Buriti, in proruptis rupis ad ca- taractam junta cum Sphagni, 16 June 1894, Malme (Regnell Expedition I) 1684 (ho- lotype, S n.v.; isotype, US*) Annuals. Culms decumbent and rooting at the lower nodes, the erect portions to 20 cm tall (ac- cording to original description), many-noded, un- branched; internodes glabrous. Leaves with the sheaths longer than the internodes, striate, gla- brous, the margins membranous, the auricles mem- branous; ligules ca. 0.3 mm long, membranous; pseudopetioles ca. 0.5 mm long, narrow, brownish, glabrous; blades 5-10 mm long, 2-4 mm wide, elliptical, delicately membranous, flat, sparingly short-pilose on both surfaces or with scabrid hairs on the veins only on the lower surface, + sym- metrical, acute at both base and apex, the margins scaberulous. Inflorescences terminal and axillary from the uppermost nodes, partially included in the sheaths; terminal inflorescences bearing male spikelets only; axillary inflorescences bearing fe- wide, narrowly ovoid, falling entire; glumes acu- minate, short-hispid with long, thickened hairs mainly on the nerves, the lower glume 3-nerved, the upper glume 3-5-nerved; floret ca. 1.5 mm long, narrowly cvoid (but none fully mature), whit- ish, smooth, shiny. Caryopsis not seen. Male spike- lets 1.7-2 mm long, ca. 0.4 mm wide, lanceolate, hyaline, glabrous, shiny; lemma 3-nerved; stamens with anthers ca. 1 mm lon Distribution and ecology. Endemic to wet rocks near streams at low elevations in Mato Grosso 2 and Pará, Brazil (Fig. 1 Additional specimens examined. BRAZIL. MATO GROSSO: Utiarity, Rio Papagaio, margin of waterfall, May 1918, Kuhlmann 1866 (US). РАКА: estrada Santarém-Cuiabá, 877 km de Cuiabá, vale de solo pedregoso, embaixo da Cachoeira da Luz do Rio Curua, 8°45'$, 54257" М, 350 m, 2 May 1983, Silva 196 (MO) This species is related to R. potaroensis and R. kaieteurana, but differs in its smooth female flo- rets, and in female and male spikelets of similar size. 5. Raddiella minima Judziewicz & Zuloaga, sp. nov. TYPE. Brazil. Para: Municipio de Itai- tuba, km 771 estrada Santarém—Cuiaba, próximo a divisao Pará-Mato Grosso, mata de cerrado solo pedregoso, vegetação rupestre, 9°35'S, 54°35'W, 22 Apr. 1983, LL. Ama- ral, N. Silva, O.P. Monteiro, J. Lima, L. Brako, W.D. Reese & M. Dibben 883 (ho- lotype, INPA; isotypes, MO, NY). A Raddiella esenbeckii differt lamina minore (4-6 x 2.7-3.3 mm) et spiculis mascula et femina minoribus (1- 1.3 mm longis), caryopside hilo punctiformi centrali prae- dit Ita. Tiny, straggling herbs of indefinite. duration. Culms to 6 cm long, filiform, geniculate, multi- branched, internodes purple, glabrous; nodes with a few very short retrorse hairs. Leaves in loose complements of 3—5; sheaths shorter than the in- ternodes, usually shorter than the blades, glabrous to short-pilose, 5-nerved, slightly inflated, the up- per margins ciliate, abruptly truncate at the short- ciliate summit; ligule membranous, ca. 0.2 mm long; pseudopetiole 0.1—0.2 mm long, glabrous to short-pilose; blades 4-6 mm long, 2.7-3.3 mm wide, ovate-triangular, asymmetrical, glabrous, the base truncate, the apex acute and apiculate, the margins scaberulous, the abaxial surface occasion- ally purplish. Inflorescences terminal and axillary, terminal inflorescences 2-6, ca. 10 mm long, fi- liform, each terminated by a single male spikelet; axillary inflorescences 1 per each axil, ca. 3 mm long, partly included in the leaf sheath, each of 2(-3) female spikelets on short, filiform pedicels. Female spikelets 1-1.4 mm long; glumes lanceo- late-ovate, acuminate, 3-nerved, glabrous to short- hispid, persistent; floret 0.9-1.2 mm long, ovoid, acute, dorsally compressed, smooth, shiny, white becoming dark, deciduous; lemma 3-nerved. Cary- opsis 0.7 mm long, 0.6 mm wide, ovoid-globose; hilum punctiform, central; embryo small, basal. Male spikelets ca. 1.3 mm long, the lemma soon deciduous, elliptical, hyaline, glabrous; anthers ca. 0.6 mm lon Distribution and ecology. Known only from the type collection, made in campo rupestre near the borders of the states of Mato Grosso and Pará, Brazil (Fig. 12). Raddiella minima is related to R. esenbeckii, but differs in its conspicuously smaller leaf blades and male and female spikelets; caryopsis with punc- tiform, central hilum; and solitary male spikelet per inflorescence. И is not certain whether the leaves exhibit sleep movements as in К. esenbeckii. Along with the Cuban endemic Mniochloa pul- chella (Griseb.) Chase, this species is one of the world's smallest bambusoid grasses, hence its spe- cific epithet. 940 Annals of the Missouri Botanical Garden 6. Raddiella molliculma (Swallen) Calderón & Soderstrom, Smithsonian Contr. Bot. 44: 22. 1980. Panicum molliculmum Swallen in R. Schultes, Bot. Mus. Leafl. 16: 57. 1953. TYPE: Colombia. Caquetá (as Vaupés"): Cerro de El Castillo, Rio Apaporis, moist, shaded sandstone ledge with mosses and Selaginella ѕрр., 16 Jan. 1942, Gutiérrez & Schultes 616 (holotype, US*; isotype, COL). Figures 5, 6, 13H, I. Delicate annuals. Culms decumbent, sprawling, geniculate, the erect portions 5-10 cm tall, many- noded, unbranched; internodes glabrous; nodes compressed, short-pilose with retrorse hairs. Leaves puberulent throughout with spreading, short-pilose hairs; sheaths striate, shorter than the internodes, the upper margins short-ciliate, the summit pilose; ligules membranous-ciliate; pseudopetioles ca. 0.6 mm long, brownish, short-pilose; blades 7-12 mm long, 3-4.5 mm wide, ovate-lanceolate, flat, deli- cately membranous, slightly asymmetrical, the base truncate, the apex acuminate, the margins sca- berulous. abundantly produced from the uppermost nodes (6–15 from the terminal node), long-exserted on glabrous, filiform peduncles 5 cm long; individual inflorescences 2-5 mm long, few-flowered, with 1— 2 terminal, short-pedicelled male spikelets and 1— 2 subterminal female spikelets or occasionally with Inflorescences terminal and axillary, 1 or more male spikelets below; pedicels up to 1 mm long, glabrous. Female spikelets 1.3-1.5 mm long, 0.6-0.7 mm wide, narrowly ellipsoid, falling entire at maturity; glumes lanceolate-ovate, acu- minate, subequal, greenish, short-hirsute, 3-nerved, the lateral nerves inconspicuous; floret ca. 1.3 mm long, 0.6-0.7 mm wide, ellipsoid, whitish, shiny, papillose. Caryopsis 0.8 mm long, 0.6 mm wide, ellipsoid, brownish; hilum oblong in the middle por- tion of the caryopsis; embryo small, basal. Male spikelets 1.3-2.2 mm long, 0.5-0.6 narrowly elliptical, acuminate, hyaline, short-his- pid; lemma 3-nerved; stamens with anthers 0.9- mm wide, 1.3 mm long, purple. Distribution and ecology. Restricted to the Cerro de El Castillo in Caquetá, Colombia (Fig. 12), at elevations of 240-300 m on moist, shaded, sandstone cliffs. Additional specimens examin Mi COLOMBIA. CAQUETÁ: Upper Río Apaporis basin, Cer Е Castillo, 27 July 1943, Schultes 5668a (US), Grassl 10024 (US). 7. Raddiella potaroensis Soderstrom, Mem. New York Bot. Gard. 12(3): 6. 1965. ТҮРЕ. Guyana. Kaieteur Plateau, bottom of Potaro Gorge near Kaieteur Falls to W rim of splash- basin of Falls, 700 ft., 13 Mar. 1962, R. S. Cowan & T. R. Soderstrom 2162 (holotype, US*; isotypes, К, NY). Figures 7, 8, 13J-L. Raddiella maipuriensis Soderstrom, Mem. New York ard. 12(3): 7. 1965. ТУРЕ: Guyana. Imbai- nadai Savannas, Upper Mazaruni River, abundant in spray of Maipuri Falls, Karaurieng River, 1,250 m, 25 Oct. 1951, B. Maguire & D. B. Fanshawe 32277 (holotype, US*; isotypes, NY, U). Annuals. Culms weak, decumbent, branching and rooting at the lower nodes, the flowering por- tions 5-20 cm tall; internodes cylindrical, glabrous; nodes thickened. Leaves with the sheaths longer than the internodes, striate, glabrous, the margins membranous, the apex somewhat truncate, one side of the summit forming an auricle; ligules 0.2-0.4 mm long, membranous; pseudopetioles ca. 0.5 mm long, glabrous; blades 9-17 mm long, 3-7 mm wide, elliptical, flat, glabrous or with short, scabrid hairs only on the veins, slightly acute and asym- metrical at both ends. Inflorescences mainly axil- lary (one terminal inflorescence in the type of R. maipuriensis), 2-4 produced from each node, with the bases included in the leaf sheaths or short- exserted on peduncles less than 1 cm long, 2-3 spikelets per inflorescence, either all female or the terminal female and 1—2 male spikelets below; ped- icels of male and female spikelets 1-3 mm long, filiform, glabrous. Female spikelets 1.3-2.3 mm long, ellipsoid to ovoid; glumes subequal, acute to acuminate, sparsely hispid with spreading, strag- gling hairs, weakly to strongly 3-5(-7)-nerved; floret 1-1.3 mm long, 0.7-0.8 mm wide, ellipsoid, acute, firmly coriaceous, shiny, whitish becoming olivaceous at maturity, papillose; lemma 3-nerved. Caryopsis ca. 0.9 mm long, 0.7 mm wide, slightly obovoid; hilum punctiform in the middle portion of the caryopsis; embryo ca. М the length of the caryopsis. Male spikelets 2.7-4.2 mm long, 0.6– 0.7 mm wide, lanceolate, glabrous; lemma 3-nerved; stamens with anthers 1.7-2 mm long. Distribution and ecology. Apparently en- demic to the Pakaraima Plateau of western Guyana and adjacent Bolivar, Venezuela (two sterile col- lections); on wet rocks near and in the spray zone of waterfalls, from 200-1,250 m (Fig. 12). Addi P кан specimens exami ned. | VENEZUELA. BOLÍVAR: Sierra de Lema, in pow at base of uppermost m of headwaters of Rio anan at base of high andstone N ing bluffs, Seal eos of Rio ۰ 80 km SW of El Dorado, Steyermark 89568 (US); L Volume 78, Number 4 1991 Zuloaga & Judziewicz Revision of Raddiella Gran Sabana, S of El Dorado along road to Santa Elena, ca. km 202, Salto Camá, 1,000 m, Davidse 4872 (MO). GUYANA. Kaieteur Falls, spray zone below falls, Apr. 1988, Hahn et al. 4635 (BRG, US). Judziewicz (1991) did not find enough differ- ences between К. maipuriensis and К. potaroen- sis to separate them as species and therefore placed the first name in the synonymy of the latter. It may be noted that the surfaces of the female florets of both taxa are identical under the scanning elec- tron microscope. The type specimen of К. mai puriensis is larger than the rest of the material studied of this species, with culms ca. 20 cm tall, leaf blades 2 cm long and 0.7 cm wide, and female spikelets 1.8-2.3 mm long with acuminate, strong- ly nerved glumes. Two sterile Venezuelan collec- tions, Davidse et al. 4872 and Steyermark 89568, are tentatively referred to this species based on their nearly glabrous leaf blades. EXCLUDED SPECIES Raddiella truncata Swallen, Bull. Torrey Bot. Club 75: 89. 1948 [= Parodiolyra lateralis (Nees) Soderstrom & Zuloaga)]. LITERATURE CITED CALDERÓN, C. E. & T. R. SopERsTROM. 1980. The genera of Bambusoideae (Poaceae) of the American continent: keys and comments. Smithsonian Contr. 1- Bot. 44: DAVIDSE, C. & R. W. Pour. 1972. eges num- bers, meiotic behavior and notes on asses from ig jw and the West indie а -145 J. Bot. 50: 8. Chromosome numbers of Weiss American speed = ае). Ann. Mis- uri Bot. Gard. 65: Jude E. J. l. UR 87. Poaceae. In A. órts-van Rijn (editor), of the Guianas, oeltz Scientific Publica- по апу. нон. T. R. 5. Gramineae. In: B. Maguire, e Botany of the inm Highlands, Part VI. Mem. New York Bot. Gard. 12(3): 5-7. . O. ZuLoaca. 1989. A revision of the genus Olyra and the new segregate genus Parodi olyra (Poaceae: Bambusoidae: Olyreae). Smithsonian Contr. Bot. 69: 1-79. SWALLEN, J. R. Gramineae. /n: B. Maguire et al., Plant Explorations in Guiana in 1944, Chiefly to the Tafelberg and the т Plateau, I. Bull. Tor rey Bot. Club 75: 56-115. ZULOAGA, F. О. 1987. A LN of the New World species of Panicum (Poaceae: Paniceae). Pp. 287- 306 in T. R. Soderstrom et al. (editors), Grass Sys- tematics and Evolution. Smithsonian Inst. Press, Washington, D.C Peter Goldblatt? and Neil Snow? SYSTEMATICS AND CHROMOSOME CYTOLOGY OF ELEUTHERINE HERBERT (IRIDACEAE)! ABSTRACT The New World genus Eleutherine (Iridaceae, tribe Tigridieae) is treated as comprising two species that are В by a bulbous rootstock, а large subapica | flowers. The more common Е. bulbosa (syn. E. iy to the West Indies and is widely cultivated for its medicinal properties. Less well known is Е. latifolia (here including Е. citriodora), with centers in northern Central America and subtropical South America. Both species have a basic chromosome number of x — all other genera of the tribe have x = 7. The common form of E. bulbos result of a Н inversion in one of the long chromosomes, and it is sexually sterile, unlike Е. latifolia, which is self -compatible and a us. The sys uu и nomenclature of the two species are dealt with in detail, and the medicinal o of E bulbosa are outlin Eleutherine Herbert (1843), a member of the New World tribe Tigridieae of Iridaceae, comprises low-growing bulbous plants with pleated lanceolate leaves and small, white, evening-blooming flowers. Eleutherine bulbosa (Miller) Urban (syn. Е. pli- cata (Sw.) Klatt), which extends across South America from Bolivia and southeastern Brazil to the West Indies, has been known since the sev- enteenth century (Baker, 1892). Currently, Ra- venna (1984a) admits two more species, E. lati- die (Standley & L. O. Williams) Ravenna of rthern Central America and subtropical South еы and E. citriodora (Ravenna) Ravenna, from northern Argentina. We recognize only Е. latifolia and suggest that E. citriodora is conspe- cific with E. latifolia. Two more species have been described from Indochina (Gagnepain, 1934), both conspecific with E. bulbosa. They are sometimes erroneously thought to be native there and are still occasionally recognized (Schultes, 1990). he two species of Eleutherine are alike in their nearly identical small, white, stellate flowers with free, ascending stamens and relatively long, filiform style arms that extend between the anthers. Veg- etatively they share two unusual features, both synapomorphies: a large foliage leaf inserted at the apex of the vegetative part of the stem; and dark red, rather than brown, bulb scales. In other veg- etative features, Eleutherine accords well with Ti- ios in having a true bulb and plicate foliage two important synapomorphies for the tribe c 1990). The common cultivated strain of E. bulbosa is sexually sterile for reasons outlined ine It comprises part of the pharmacopeia of several America Indian tribes and may be a clone that has been dispersed by humans from a source in the eastern Апдеап foothills of Peru, where the most fertile of the few capsulate specimens occur. In this paper we describe the chromosome cy- tology of Eleutherine and provide evidence that although both species appear to have an unspe- cialized morphology, the genus is likely to be a derived genus of Tigridieae. А key to the species, complete nomenclature, and descriptions of the species are provided in the systematics section. MATERIALS AND METHODS Collections of Eleutherine were obtained from wild sources (Table 1) and cultivated at the Missouri Botanical Garden. For cytological study, root tips were pretreated in aqueous m-bromonaphthalene for 4 hours and then fixed in 3 : 1 absolute ethanol ' Support for this study by grants ВУК 85-00148 and ВУК 89-06300 from the U.S. National Science Foundation is gratefully acknowledged. We thank Walter H. Lewis for making wild-collected plants and his ethnobotanical database ан to u A. Krukoff Curator of African Botany, Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166, 5. a Biology Department, Washington University, St. Louis, Missouri 63130, U.S.A. ANN. Missouni Вот. GARD. 78: 942-949. 1991. Volume 78, Nurnber 4 1991 Goldblatt & Snow 943 Systematics and Chromosome Cytology of Eleutherine E 1. Chromosome numbers and collection data for species of Eleutherine studied. Voucher specimens are located at the Missouri Botanical Garden (MO). iploid D Species number Collection data E. latifolia 12 Argentina. Tucumán: Castillo sub Goldblatt 9072 12 Mexico. San Luis Potosi: Tama- pe PU sub Gold- blatt 9 Peru. Loreto: n Hermosa and vicinity, Lewis et al. 10624 E. bulbosa 12 glacial acetic acid. After hydrolysis in 10% HCI for 6—7 minutes at ca. 65?C, root tips were squashed in FLP orcein (Jackson, 1972). Karyotypes were drawn with a camera lucida. Notes on reproductive biology were based on observation of fruit set in greenhouse plants that were not accessible to potential insect pollinators. OBSERVATIONS Two populations of Eleutherine latifolia, one from Argentina and one from Mexico, are self- fertile, seed-producing plants with a diploid chro- mosome number of 2n — 12. They have an iden- tical karyotype (Fig. 1A, C) of one long subtelo- centric chromosome pair 5-6 шт long (including secondary constrictions), and five much shorter, submetacentic to acrocentric pairs 1.5-2 um long. Large satellites are located on the very small short arms of the two long chromosomes. The satellites are ca. 1 um long and much exceed the short arms of the long chromosomes. Although easily over- looked in contracted metaphase preparations, they are a striking characteristic of the karyotype. In prophase, the satellites are easily mistaken for a A FIGURE 1. 5 um.) pair of small chromosomes. The thin secondary constrictions are difficult to see and are sometimes stretched by manipulation during preparation of squashes such that the satellites lie some distance from the arms to which they are attached. Plants of Eleutherine bulbosa from Peru, ap- parently sexually sterile, have a karyotype similar to the above except that the pair of long chro- mosomes do not match. The long arms and satellites differ in length (Fig. 1B). The total length of the chromosomes is, as far as we can estimate by linear measurement of camera-lucida-drawn chromo- somes, close to, if not the same as the F. latifolia populations. Thus, the longer satellite is ca. 2. um long, and the rest of the chromosome including the secondary constriction is ca. 3.5 um long. The chromosome with the short satellite is ca. 6 um long, including the satellite, which is ca. 1 um long, and corresponds in length and morphology to the long chromosomes of Е. latifolia. DISCUSSION CYTOGENETICS The chromosomes of Eleutherine latifolia ac- cord well with those of Е. features of the karyotype, x = 6 with extreme bulbosa in the basic bimodality and a strongly acrocentric long chro- mosome pair to which are attached large satellites. The karyotype was briefly described by Guerra (1988), based on unpublished observations of A. Kenton, but a detailed description and illustration have not been published until now. The basic features of the karyotype of Eleutheri- ne bulbosa (also reported in the literature as Е. plicata) are already known (Zaman et al., 1985; Cherian & Kuriachan, 1988—-based on plants cul- tivated in India; Martinez & De Azkue, 1987; Guerra, 1988— based on plants of South American Karyotypes of Eleutherine latifolia from Mexico (A) and Argentina (C), and Е. bulbosa (B). (Scale Annals of the Missouri Botanical Garden origin). Eleutherine guatemalensis Standley, which has 2n = 28 (Ravenna, 1965), is now Cobana guatemalensis (Standley) Ravenna. The genus Cobana is probably not closely related to Eleuthe- rine and belongs in Tigridiinae rather than Cipu- rinae, where Eleutherine seems well placed (Gold- blatt, 1982 The earliest chromosome counts in the literature for Eleutherine bulbosa are those of Sharma & Talukdar (1959) and Rao (1969), both papers re- porting 2n — 14 and a karyotype of two long pairs of chromosomes, one of which is metacentric. In the same paper, Sharma & Talukdar (1959) re- ported 2n = 12 (and 14) in Cipura paludosa Aublet; their drawings of the chromosomes of the Eleutherine closely match published karyotypes for Cipura paludosa (e.g., Goldblatt, 1982), whereas the karyotypes of their cytotypes of C. paludosa correspond with Eleutherine here, with some minor variation. The plants studied by Rao (1969) as Eleutherine were presumably also misidentified, and their count of 2n — 14 must be ignored. (Cipura, especially C. paludosa, is understandably confused with Eleutherine, but the two genera are probably not immediately related.) Subsequently, Goldblatt (1982) reported 2n — 12 and a structurally heterozygous karyotype in (1985) also found 2n = 12 in E. bulbosa, and they reported Eleutherine bulbosa. Zaman et al. regular bivalent formation at meiosis and the pres- ence of an inversion bridge with a fragment at meiosis I. Based on this, they postulated that a pericentric inversion in one of the two long chro- mosomes was responsible for the long chromosome heteromorphism. А fragment associated with a mei- otic bridge is not compatible with a pericentric inversion (cf. Stebbins, 1971), but Zaman et al.'s interpretation of the meiotic figures, not illustrated, may not have been correct. А study by Guerra (1988) finally demonstrated convincingly the pres- ence of a pericentric inversion in one of the long chromosomes of E. bulbosa, and he also reported a small tandem duplication of part of the short arm of the chromosome with the inversion (and longer satellite), rendering this chromosome slightly longer than its homologue. Guerra's conclusions make the comments of Cherian & Kuriachan (1988) on the nature of the chromosomal restructuring in E. bul- bosa redundant and need not be repeated here. Despite their refutation of the presence of a peri- centric inversion in the species, the meiotic con- figurations illustrated in their paper show the bridg- es formed by the long chromosomes without fragments, consistent with heterozygosity for a pericentric inversi It is sl certain that the ancestral karyo- type in Eleutherine had x = 6 and was structurally homozygous, as described here in E. latifolia and likely to occur in sexually reproducing E. bulbosa, not yet examined cytologically. The mismatched long chromosome pair in the cultivated strain of E. bulbosa has a pericentric inversion, which is responsible for its sexual sterility; this accounts for the absence of capsules on most herbarium material of E. bulbosa and in cultivated plants. Fertile pop- ulations and at least partially fertile individuals do, however, exist. А collection from Peru (Schunke 971, NY, F) comprises several individuals, all bear- ing numerous capsules. Other collections comprise at least one plant with a few developed capsules (Duss 3836 from Guadelupe, US; Costa 164 from Pará, Brazil, F; Faris 475 from Santo Domingo, US). The capsules are small and globose and differ from those of E. latifolia, which are oblong to cylindric and always substantially larger. The basic chromosome number for the entire Tigridieae is probably x = 7 (Goldblatt, 1982, 1990; Kenton & Rudall, 1987), and there are few exceptions in the tribe. Only Gelasine elongata (Graham) Ravenna (syn. G. azurea Herbert fide Ravenna, 1984b) also has n = 6 (2n = 12) (Kenton & Rudall, 1987). The chromosome number is known in only one other Gelasine, G. uruguay- 7 (Ravenna, 1984c). The karyotype of G. elongata has a long acrocentric chromosome pair with conspicuous satellites larger than the short arms (misinterpreted by Goldblatt, 1982, as a pair of small chromosomes), and in this respect resembles Eleutherine. However, G. elon- gata differs in having a second relatively long chromosome pair. The karyotype of G. elongata is structurally heterozygous, and the chromosomes form a ring of 12 at meiosis (Kenton & Rudall, 1987). Karyotypic and cytogenetic differences be- tween Eleutherine and Gelasine elongata make it seem unlikely that the two share an immediate common ancestor. The n — 6 in G. elongata most likely originated from ancestors within the genus with x — 7. The only other member of Tigridieae that is known to diverge from the basic x — 7 is Cypella s m n — 5, but other species of the genus hav which is the basic number for Cypella (Goldblatt, 1982; Kenton & Heywood, 1984). ensis, which has n = SYSTEMATICS AND PHYLOGENY The two species that we propose to recognize in Eleutherine have virtually identical flowers, but Volume 78, Number 4 1991 Goldblatt & Snow 945 Systematics and Chromosome Cytology of Eleutherine FIGURE 2. Morphology of Eleutherine bulbosa (A) and E. latifolia (B). Habits x 0.5, flower and fruiting branch of E. bulbosa full size. they are readily distinguished by differences in their ). In Е. latifolia, generally a short plant, the large apical cauline leaf subtends one or sometimes two subsessile to shortly pedunculate rhipidia (spathe-enclosed flow- er clusters), the basic inflorescence type in Irida- eae (Fig. 2B). In Е. bulbosa the cauline leaf subtends a few pedunculate rhipidia and a second ary axis that terminates in (3-)6-12 rhipidia in umbellate arrangement; thus the rhipidia are ar- inflorescence structure (Fig. 2). ranged in two series, axillary and terminal (Fig. 2A). The flowers of Eleutherine (Fig. 3A) appear to be relatively simple in structure and comparable with those of more primitive genera of Iridaceae. It is tempting to interpret the genus as primitive within Tigridieae, in which a more complex flower structure is the rule. However, Mariceae, sister tribe of Tigridieae, are also characterized by a complex flower structure (Fig. 3B) almost identical with that of Tigridieae (Goldblatt, 1990; Goldblatt & Henrich, 1991), in which the tepals are clawed, the inner tepals have a zone of nectariferous tissue in a fold of the tepal limb, and the thickened style branches are terminally bifurcate and have a sub- Details of the flowers and stamens and FIGURE 3. style of Eleutherine (A) and of the putative ancestral type rimezia steyermarkii (B). Whole flowers eparated stamens and styles and style branches much enlarged apical transverse stigma. It seems more parsimo- nious to interpret the flower of Eleutherine as derived by a process of reduction in complexity associated with self-compatibility and autogamy, a erived type of reproductive biology in Tigridieae that is basically self-incompatible (Goldblatt, pers. obs.). Similar examples of the simplification of an ancestrally complex flower structure are known within Iridaceae in tribe Irideae, notably in subtribe Homeriinae (Goldblatt, 1979, 1981, 1986). Here, flowers with clawed tepals and elaborate petaloid, apically bifurcate style arms with transverse stig- mas, are believed to have been secondarily reduced in several lineages, leading in one genus, Roggev- eldia Goldblatt, to species with stellate flowers with simple filiform style branches that alternate with the stamens. SYSTEMATICS еа Herbert, Edwards's Bot. Reg. 29: t. 7 (1843). TYPE: Е. plicata Herbert (= Е. 2s (Miller) Urban). Galatea Salisb., Trans. Hort. Soc. 1: 310 (1812), nom. Small, seasonal perennials with a fleshy bulb with reddish tunics. Leaves few, plicate, both basal and cauline, narrowly lanceolate. Stem terete, com- prising one long internode, with a large cauline leaf at the apex, subtending the inflorescence. Inflo- rescences rhipidia, these solitary or several and then arranged in a pseudopanicle; rhipidia (spathe- enclosed umbels) several-flowered, stalked or + sessile; spathes green, subequal, relatively short. Flower pedicellate, subtended by membranous 946 Annals of the Missouri Botanical Garden bracts as long as the spathes; perianth white, stel- late; tepals free, subequal, spreading from the base, without nectaries; filaments free; anthers diverging. Ovary obovoid; style short and dividing opposite the upper half of the filaments, the branches un- divided, filiform, extending between the anthers, stigmatic apically. Capsules globose to oblong-cy- lindric, truncate; seeds angular. Basic chromosome number x — Eleutherine consists of two species distributed in Mexico, West Indies, Central and South Amer- ica; cultivated in Africa and Asia, especially the Philippines and Indochina, and now naturalized there. KEY TO THE SPECIES la. Plants with several rhipidia per flowering stem, duncles borne both in the axils of the ondary pu rarely setting fruit, but capsules globose, . E. bulbosa 1b. Plants ah T or rarely 2 rhipidia either + sessile or borne on short peduncles inserted in the axil of the cauline leaf; iid setting 2m and capsules 12-20 mm lon . E. latifolia 1. Eleutherine bulbosa (Miller) Urban, Re- pert. Spec. Nov. Regni Veg. 15: 305 (1915). Basionym: Sisyrinchium bulbosum Miller, Gard. Dict. ed. 8 (1768). Galatea bulbosa (Miller) Britton, Brooklyn Bot. Gard. Mem. 1: 37 (1918) nom. inval., certain, possibly figure in Burman, Pl. Amer. Fasc. 35 & t. 46 f. 2 (1757 as 1755), but not specifically cited by Miller [perhaps a neo- type should be chosen]. [Miller may have had living plants but cited Tournefort, Inst. Rei Herb. 387 (1700), who in turn cited Magnol, Hort. Reg. Monsp. (1697), neither of which has any illustration of the species nor refers to one]. Figure gen. inval. TYPE: un- Ixia americana Aublet, Hist. Pl. Guiana 1: 33 (1775). Galat i ан latifolium Sw., Prodr. 17 (1788) nom. Meg uperfl. Moraea plicata Sw., Fl. Ind. Occid. (1797) nom. illeg. superfl. Манса с (Sw. us Gawler, Curtis's Bot. Mag. 18: t Si (1803). Gal. atea vespertina Salisb., leg Hor ‚1:31 (1812) nom. inval., gen. inval. С же plicata (Sw.) Griseb., Fl. Brit. W. NT рм 589 (1864). Eleuthe- rine (as Eleutherina) plicata (Sw.) Klatt in Martius, Fl. Bras. 3(1): 514 & t. 64, f. 2 (1871). TYPE ees in Burman, Pl. Amer. Fasc. 35 & t. 46 f. 2 (1757 as 1755). Bleutherine anomala Herbert, Edward’s Bot. Reg. 29: (1843). Galatea us (Herbert) Kuntze, мали ids Pl. 2: 701 (1891) nom. inval., gen. val. TYPE: illustration in Bot. Reg. 29: t. 57. Eleutherine од Gagnepain, Bull. Soc. Bot. 81: 66 & f. 67. 1934 et in Fl. Gen. de l'Indo-Chine 6: 676. 1934. TYPE: Vietnam (Annam, Quang-Tri), id 12279 (P) (a form flowering without basal leaves). Site ane longifolia Gagnepain, Bull. Soc. Bot. France 81: 66. 4 et in Fl. Gen. de l'Indo- Chine 6 : 677. 4. TYPE: Vietnam (Annam, ? Ние), Jacquet 611 (P) (described from vegetative plants). Sisyrinchium americanum, radice phaenicea, foliis pli- catis & nervosis etc., Magnol, Hort. Reg. Monsp. 185 (1697). Bermudiana palmaefolio, radice bulbosa Tournefort, Inst. Rei Herb. 1: 388 m Sisyrinchium palmifolium sensu Lam., Encyc. Meth. 1: 408 (1783-85) et sensu Cavanilles, Diss. 6: 348 & se rill, Philipp. J. Sci. 7: 233 (1912) non sensu Lin- naeus). Plants 15-75 cm high, usually sexually sterile and lacking capsules; other characters as for the genus except the following. Inflorescence of several stalked rhipidia, the peduncles borne both in the axil of the cauline leaf and in umbellate fashion on a secondary axis; spathes 11.5-15(-20) mm long; peduncles (12-)20-55 cm long. Tepals 10-18 mm long. Filaments 2-2.5 mm long; anthers 3.5-4.5 mm long. Capsules irregularly globose, showing the outline of the seeds, (2.5-)4-6.8 mm long, seeds rarely more than 3 per locule; seeds 1.8-2.7 mm long, irregularly globose with a concave chalazal end, reddish brown to black (? with age), strongly rugose. Chromosome number 2n — 12. Eleutherine bulbosa is widespread, extending from southeastern Brazil and Bolivia to Venezuela and the West Indies (Fig. 4). The original habitat is uncertain. The most consistently fertile plants we have seen are from a population in the eastern Andean foothills of Peru, but occasional plants with capsules are known from eastern Brazil (Pará), Trinidad, Guadeloupe, and Santo Domingo. The branched inflorescence and, when present, the small, more or less globose capsules distinguish the species from its close relative Е. latifolia. In its extended and branched inflorescence, we assume E. bulbosa is the less specialized species. The flowers of the two are essentially identical. The several specimens sampled cytologically (see discussion of cytology), all sexually sterile, have the two long chromosomes heteromorphic and het- erozygous for a pericentric inversion and a tandem duplication, which accounts for the prevalent re- Volume 78, Number 4 1991 Goldblatt & Snow Systematics and Chromosome Cytology of Eleutherine TROPICAL AMEFICA FIGURE 4. productive sterility. Chromosomes have not been examined in апу fertile plants. Nomenclatural history. The systematics of this relatively uniform species is remarkably complex. Although first described by Phillip Miller in 1768 in the genus Sisyrinchium, his epithet was over- looked by Aublet (in 1775), who named the plant Ixia americana, and by Swartz, who gave it two epithets, Sisyrinchium latifolium in 1788 and Mo- raea plicata in 1797. All three species were based on the same type. Herbert's (1843) Eleutherine anomala was based on plants grown in Britain. Herbert was aware of the four earlier epithets for the species and regarded E. anomala as distinct from them. Herbert provided a generic name for plants that had until 1843 been assigned to no less than five genera, if Salisbury's nomen nudum for Distribution of Eleutherine latifolia (triangles) and E. bulbosa (circles). the genus, Galatea, is included. Curiously, Herbert designated Marica plicata as the type of Eleuthe- rine but did not make the combination E. plicata often attributed to him. Adding to the complexity is Lamarck's use of Linnaeus's Sisyrinchium pal. mifolium for E. bulbosa. Sisyrinchium palmifoli- um must be based on a specimen collected by Ardouin in Brazil (a species of Sisyrinchium still known by this name) rather than on the figure of E. bulbosa in Burman's Plantae Americanae that Linnaeus cited with a query, in the protologue of S. palmifolium. Lamarck's use of the epithet pal- mifolium for E. bulbosa was followed by Jacquin (1790), who referred the species to the African Moraea, and Merrill (1912), among others. In 1934, Gagnepain described two more Eleu- therine species from Indochina that are obviously forms of Eleutherine bulbosa: E. subaphylla, which 948 Annals of the Missouri Botanical Garden lacks foliage leaves; and Е. longifolia, which lacks flowers. Specimens matching both are routinely encountered in herbarium material from the Amer- icas, sometimes from populations that include plants of more normal appearance. Eleutherine bulbosa is an important element of American Indian pharmacopeia and is one of few neotropical Iridaceae with known medicinal uses (Schultes, 1990). It is often cultivated in gardens maintained by Indian tribes (W. Lewis, pers. comm.) and is also fairly widely grown in the tropics as an ornamental. It is used as а vermifuge in Ecuador (Vickers & Plowman, 1984) and is widely docu- mented in the literature and in herbaria as a treat- ment for bloody diarrhea (Ayala, 1984), haemor- rhagia and open wounds (Lewis et al. 9966, 10624; Costa 164; Peters & Padoch 137; Hahn & Tred- well 125), and even as a contraceptive (Schunke 97 1). Eleutherine bulbosa is now naturalized in the Philippines and Indochina and cultivated in Kwazulu, South Africa (A. Hutchings, pers. comm.), where it is also used medicinally. Quisumbing (1951) recorded the use of E. palmifolia (i.e., E. bulbosa) as a diuretic in the Philippines. It is also believed have magical and hallucinatory properties. Schultes (1990) has provided information about chemistry and possible active principles, so far hardly investigated. 2. Eleutherine latifolia (Standley & Williams) Ravenna, Phytologia 56: 195 (1984). Cipura Неа Standley & Williams, Ceiba 1: 75 (19 TYPE: Honduras, El Zamorano, ан of the Escuela Agricola Panameri- cana, 14 Aug. 1947, Molina 498 (holotype, EAP). Figure 2B. Eleutherine citriodora (Ravenna) Ravenna, Phytologia (1984), syn. nov. Eleutherine d ee citriodora Ravenna, Bol. Soc. Argent. н (1965). TYPE: Argentina, in pratis u urbi Tuc mán, Feb. есы Vervoorst s.n. (holotype, Herb. ho 285 n Plants (6-)1 2-20 cm high, other characters as for the genus except the following. Inflorescence of one, rarely two, subsessile or stalked rhipidia borne in the axil of the cauline leaf; peduncle 10- 50(-80) cm long; spathes 15-20 mm long. Tepals 12-14 mm long, 6-8 mm wide. Filaments 2.5-3 mm long; anthers 3 mm long. Style ca. 2 mm long, the branches ca. 4 mm long. Capsules obovoid to oblong-cylindric, 10-28 mm long, with numerous seeds per locule; seeds 2.1-2.3 mm long, conical, the sides somewhat flattened, lightly rugose, dull greenish yellow, dark brown at the chalazal end. Chromosome number 2n — 12 Eleutherine latifolia is widespread but scattered in the Neotropics and subtropics, with centers in northern Central America and Mexico and in south- ern South America, northern Argentina, | ау (Fig. 4), and according to Ravenna (1984а), Bo- livia, but also recorded in Venezuela. The species is strongly autogamous. Because of the inflores- cence that comprises one or occasionally two rhi- pidia, compared with the several in pseudopanicu- late arrangement in E. bulbosa, we consider Е. latifolia to be more specialized. Until now, Eleutherine latifolia has been re- garded as a Central American- Mexican endemic. However, the differences between this and the southern South American Ё. citriodora seem to us too trivial 10 merit continued separation. Her- barium specimens cannot as a rule be distinguished, although Central American plants are often stur- dier. In cultivated samples of single populations from Mexico and from Argentina we noted that the Mexican plants have firmly erect stems, clear green foliage, and flowers lightly acrid-smelling, slightly smaller, and opening earlier in the evenings than the plants from Argentina. The latter have stems that seldom remain erect as the capsules develop, the leaves have a slightly bluish tinge and the flowers, faintly sweet-scented, open a full hour later that the Mexican plants, 6:00-6:30 P.M., and last about an hour later, till nightfall, 8:00-8:30 P.M. We have no information on interpopulational variation (flowers are rarely preserved in dry spec- imens and floral odor is never recorded) in either of the two main centers of the species nor in plants from Venezuela. Thus, at least on available infor- mation, we recommend including E. citriodora in Е. latifolia. The only realistic way to tell the two apart is to know their geographic origin. The dif- ferences enumerated above are cryptic and can only be seen in live plants: even then we do not know how consistent they are across populations. Experience in other lridaceae is that flower size, odor, and phenology will often vary to some extent in widely distributed speci en Ravenna (а Eleutherine citrio- dora as E. bulbosa subsp. citriodora, he distin- guished it from subspecies bulbosa by its somewhat larger rhipidial spathes crowded in the axil of the cauline leaf; the oblanceolate, more acute, tepals; the flowers citrus-scented; and the stamens often suberect. This distinction is the same as that be- tween the Central American E. latifolia and E. bulbosa, except for the sweet floral scent in so- called E. citriodora from Argentina. We have not which is in the seen the type of E. citriodora, private collection of P. Ravenna and is unavailable Volume 78, Number 4 1991 Goldblatt & Snow Systematics and Chromosome Cytology of Eleutherine 949 for study. Our interpretation of the taxon is there- fore based on the diagnosis and the locality infor- mation. The type collection was made near the town of Tucumán, northern Argentina, where plants corresponding to E. latifolia are well known, whereas E. bulbosa does not occur in Argentina. Unlike the closely related E. bulbosa, there are no recorded human uses of E. latifolia. LITERATURE CITED AYALA, F. 1984. Dini on some disi and medic- inal plants of Am an Per Prance & . А. Kalunki (edirs) Ethnobotany i in the Neotrop- s. Adv. Econ. Bot. m J. G. 1892. "Handbook of the Irideae. George Bell & Sons, London CHERIAN, M. & P. I. KURIACHAN. 1988. Cytology of Eleutherine bulbosa: a reinvestigation with special reference to the nature and possible mode of origin E bo heteromorphic chromosomes within the ge- e. Caryologia 41: 741-47. шыл. F. 1934. Iridacées, Amaryllidacees et Lilia- cées nouvelles d'Asie. Bull. Soc. Bot. France 81: 66- т P. 1979. Roggeveldia, а new genus of thern African lridaceae-lrideae. Ann. Missouri ns Gard. 66: 839-844. 1981. Systematics and biology of deis (Iridaceae), Ann. Missouri Bot. Gard. 68: 413-503. 19 Chromosome cytology in I to — systematics of neotropical lridaceae. Syst. Bot -198. . The Moraeas of Southern Africa. Ann. Kirstenbosch Bot. Gard. 14 . Phylogeny and classification of Irida- ceae. Ann. Missouri Bot. Gard. 77: 607-627 & J. E. HENRICH. 1991. Calydorea Herbert (Iridaceae- Tisridicach: notes on this New World ge- nus and reduction to synonymy of Salpingostylis, Cardiostigma, Itysa, and Catila. Ann. Missouri Bot. Gard. 78: 504-511 Guerra, M. pos SANTOS. 1988. Mitotic and meiotic analysis of a pericentric inversion associated with a M duplication 5 Eleutherine bulbosa. Chro- n osoma (Berl.) 97: «Pu W. a on anomala. Edward’s Bot. Reg. 2%: t. 57 Jackson, R. 1972. Chromosomal evolution in Haplo- pappus gracilis: a centric transposition race. Evo- lution 27: 243-2 Jacquin, N. 1790. Moraea к in Icones Plan- tarum Rariorum 2(fasc. 7): t. KENTON, A. C. EYWOOD ЕЯ Cytological studies in South Aerie апак, Pl. Syst. Evol. 146: 87-104. & P. Кораџ. 1987. An unusual case of com- plex heterozygosity in Gelasine azurea egi cun and its implications for reproductive biology. Evol Tend Pl. 1: 95-103. MARTINEZ А. & D. DE AZKUE. 1987. Variación del a no de Botánica, Medellín, Colombia, 1986] MERRILL, E. D. Iridaceae. /n: Nomenclatural and systematic notes on the flora of Manila. Philipp. J. Sel. 7: 233. тш E. 1951. Medicinal Plants of the Philip- s. Philippines Dept. Agric. Nat. Resources Techn. Bull. 16. Manila. 5. Notas sobre аа, II. Bol. Soc. Argent. Bot. 10: 311-322. 1984a. Notes on lridaceae. VI. Phytologia 56: 193-195. A previous valid specific тш for Gelasine azurea (Iridaceae). Phytologia 56: 1984c. The delimitation of Gelasine do ceae), and С. uruguaiensis sp. nov Nordic J. Bot. 4: 347-350. ScHULTES, R. A. 1990. The Healing Forest. Dioscorides Press, Portland, Oregon SHARMA, А. К. & C. TALUKDAR. 1959. Cyto-taxonomic studies on some members of the Iridaceae with special reference to the structural heterozygosity of Cipura paludosa Aubl. The Nucleus 11: 63- STEBBINS, C. L. . Chromosomal Evolution i in High- er Plants. Edward Arnold, London VICKERS, W. Т. & T. PLOWMAN. 1984. Useful _ of the Siona and Secoya Indians of eastern Ecuador Fieldiana, Bot. new ser. 15. ZAMAN, M. A., В. BEGUM & С. R. SAHA. 1985. Karyo- morphology and chromosome behaviour of a form of Eleutherine plicata Herb. (Iridaceae) with an ab- . from Uruguay. e sion. Caryologia 38: 207-211 Peter Goldblatt? John C. Manning,? and Amina Bari’ SULCUS AND OPERCULUM STRUCTURE IN THE POLLEN GRAINS OF IRIDACEAE SUBFAMILY IXIOIDEAE' ABSTRACT Ixioideae, the largest and most derived of the four subfamilies of Iridaceae, is distinct in the family in having specialized pollen grains with tectate-perforate and scabrate sculpturing. This is in contrast to the prevailing tectate- reticulate exine in the rest of the family and its putative allies. In addition, most species have — grains with a 2-banded operculum, a pi that is probably synapomorphic for the subfamily. Exceptions among taxa with monosulcate grains having either a 1-banded or disorganized operculum are seen as most probably derived. Important exceptions to this pattern are Mic ranthus, which has zonasulcate grains with reticulate exine; one of seven species of Thereianthus, 5, s, which also has a reticulate exine except o n the operculum; Zygotritonia, which has 3-aperturate inoperculate дүнк nd Savannosiphon ich the grains ры an irregular distribution of the exine and may be described as inap чайнам but functionally OE REAR The origin of the ا‎ and the significance of the variation within Ixioideae is considered in relation to the systematics of Ixioidea Ixioideae, the largest of the four subfamilies of Iridaceae, are generally considered to have геја- tively uniform pollen grain morphology (Schulze, 1970, 1971; Goldblatt, 1990), and pollen has been accorded scant attention in systematic investiga- tions of the 28 genera and ca. 860 species that now considered to comprise the subfamily (Goldblatt, 1991). Like most petaloid monocots, pollen grains of Iridaceae are usually sulcate, and the exine is tectate-columellate. A reticulate sur- face pattern is probably basic for Iridaceae (Gold- blatt, 1990; J. Walker, pers. comm.) and is the predominant or exclusive condition in Isophysi- doideae, Nivenioideae, and lridoideae, three of the four subfamilies recognized by Goldblatt (1990). However, in Ixioideae, most species of most genera i a pu (micropunctate) and scabrate (spi- late) e a synapomorphy for the subfamily (Goldblatt, 1990, 1991). Within Ixioideae, Syrin- godea and Crocus have an imperforate scabrate exine, while in the relatively specialized Micran- thus the exine is reticulate (Erdtman, 1952; Schulze, 1971). In Syringodea the grains are inaperturate and in Crocus inaperturate, spiraperturate, or poly- rugoidate, clearly derived states for these two very specialized genera. The grains of Micranthus have been described as syncolpate by Schulze (1970). A preliminary study of the pollen grains of Mi- cranthus and the putatively closely related Therei- anthus (Goldblatt, 1989а) to compare in detail the exine patterning revealed two unusual features. In Thereianthus there are discrete bands of exine in the sulcus, clearly separate from the rest of the exine. This structure, which accords with a broad definition of an operculum, is not present in Mi- cranthus. 'The grains of the latter are zonasulcate (perhaps equivalent to Schulze's term syncolpate) in all three species. Independent study of the bi- ology of North African species of Romulea (Bari, in prep.) revealed the presence of opercula in these species as well. These observations prompted us to examine the pollen grain aperture in a selection of species of all genera of Ixioideae. The size of Ixioi- deae precludes an examination of every species, but our results suggest that although the aperture is variable in the subfamily, its structure is constant within most genera so that a more thorough survey seems unnecessary except perhaps in the course ' Support from United States National Science Foundation grant BSR 89-06300 i is ан acknowledged. We d also thank John Skvarla and Annick Le Thomas for advice with v rious aspects of t udy. . Krukoff Curator of African Botany, Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166- 0299, U.S.A. * National Botanic Gardens, Kirstenbosch, P. Bag X7, Claremont 7735, South Afric ' Laboratoire de Biologie Vegetale, morphologie, E.P.H.E, 16 rue Buffon, Paris 75005, Fra ANN. Missouni Bor. GARD. 78: 950-961. Muséum National : на Naturelle, 61 rue Buffon and Laboratoire Phyto- 1991. Volume 78, Number 4 1991 Goldblatt et al. Sulcus and Operculum Structure in Iridaceae Pollen of detailed systematic research. We have included observations of exine patterning and grain size as well as sulcus structure in our results. MATERIALS AND METHODS For examination of the sulcus, 80 species in 26 of the 28 genera (Tables 1, 2) of Ixioideae were examined (Crocus and Syringodea, with imper- forate grains, were not included in the survey). Information extracted from the literature for an- other five species (Radelescu, 1970a, b) is also included in Table Pollen grains from either fresh or herbarium material were placed on glass slides, rinsed briefly with 70% ethanol, and then mounted in Calberla's fluid (Ogden et al., 1974), which effectively re- expanded collapsed grains and stained the exine dark pink while leaving the rest of the grain un- stained. Pollen was taken from herbarium speci- mens collected no more than 20 years ago when- ever possible, as grains from older specimens were found to re-exgand poorly, if at all. Grains were examined under the light microscope and measured within 48 hours of preparation for the following characters: aperture condition and presence and form of the operculum (Table 1); polar and equa- torial diameter, sulcus width and length, operculum (if present) width and length, and interopercular distance when the operculum was 2-banded (Table 2). Measurements (Table 2) represent the mean of 5-10 grains. Descriptive terminology used follows Walker & Doyle (1975 ). F 1 under tl scanning electron iie (SEM), the following method was fol- lowed. Fresh grains were fixed in FAA, and dried grains from herbarium specimens were rehydrated by warming in 10% KOH for ca. 10 minutes. The grains from either source were then dehydrated directly in ethanol and critical-point dried before mounting and coating. Specimens were observed at an accelerating voltage of 15-20 kV OBSERVATIONS APERTURE All genera examined (Table 1) except Micran- thus, Savannosiphon, and Zygotritonia have monosulcate pollen grains (Figs. 1A, B, 2-5). The sulcus is relatively broad and when expanded is usually as wide as the grain or nearly so (Table 2). Syringodea (10 spp.) and Crocus (80 spp.), which have specialized, inaperturate and inaperturate, spiraperturate, and polyrugoidate grains, respec- tively (Erdtman, 1952; Schulze, 1971; de Vos, 1974a; Mathew, 1982), were not included in our study. Zygotritonia has grains with three elliptic ap- ertures and thus is evidently trisulcate (Fig. 1C). The apertures are equidistant from one another and run almost the entire length of the ellipsoid grains. The more or less spherical grains of the mono- typic tropical African Savannosiphon have an ir- regular distribution of exine (Fig. 1D). The overall impression is of random distribution of irregularly shaped stained (exinous) patches of varying size and unstained (nonexinous) wall material. The grains are evidently inaperturate in morphological terms but probably functionally omniaperturate. Since grains of plants from two separate populations were examined, the possibility that our sampling was not representative seems unlikely. n Micranthus the grains of all three species are evidently zonasulcate (Figs. 6, 8). The aperture is ca. 18-20 um wide and forms a continuous band around the grain. We could not determine the orientation of the aperture but assume that it is polar rather than equatorial. There is a fair amount of scattered, granular exine material within the sulcus, usually concentrated close to the sulcus margins (Fig. OPERCULUM In most of the genera with sulcate grains (Table 1) there are two narrow, linear to narrowly elliptic bands of exine within the sulcus (Figs. 1A, 2). By definition, this discrete exinous structure within the sulcus constitutes an operculum (cf. Kremp, 1965; Faegri & Iversen, 1964). The opercular exine ap- pears to have the same microstructure as the rest of the grain (Figs. 2-5). These more or less parallel bands, occasionally joined at their ends, constitute an unusual structure that apparently does not occur in Iridaceae outside Ixioideae. Included among the species listed in Table 1, we note from figu provided by Radelescu (1970a, b) that the following also have а 2-banded operculum: Crocosmia Х cro- cosmiiflora (Lemaire ex Morren) N.E. Br., Gladio- lus illyricus Koch, G. vandermerwei (L. Bolus) Goldblatt & de Vos, G. imbricatus L., Sparaxis tricolor (Schneev.) Ker, and Tritoniopsis pauci- flora (Jacq.) G. Lewis. Exceptions to the above pattern among Ixioideae with sulcate grains are as follows. All four species of [xia examined and also /. Lewis (Radelescu, 1970a), Anomatheca viridis and A. fistulosa (but not the other three species of the genus), five of the seven species of Thereianthus, urpureorosea G. 952 Annals of the Missouri Botanical Garden TaBLE 1. Pollen grain sulcus and operculum type for Ixioideae, including data from Radelescu (1970а, b). Collection information for the species examined here is included in the table. Taxa are arranged alphabetically by genus and species, following the most current taxonomy for the subfamily (Goldblatt, 1991). m/s — monosulcate; 2/8 = перика t/c = Bo s/u = structure uncertain (see text); 2/b = two-banded; 1/b = one-banded; d/o — disorganized; a — absent. All vouchers are housed at the Missouri Botanical Garden Herbarium (MO) unless stated оте Total species per genus is indicated in parentheses after the genus name. Aper- Taxon ture Operculum Collection data Anomatheca (5) fistulosa (Klatt) Goldblatt m/s 1/b Goldblatt 2359 grandiflora Baker m/s 2/b 9 laxa (Thunb.) Goldblatt m/s 2/b x hort. no voucher verrucosa (Vogel) Goldblatt m/s 2/b Goldblatt 2933 viridis (Aiton) Goldblatt m/s l/b Goldblatt 3640, 7190 Babiana (62) dregei Baker m/s 2/b Lavranos 22231 stricta (Aiton) Ker m/s 2/b Goldblatt 2322 Chasmanthe (3) aethiopica (L.) N.E. Br. m/s 2/b Goldblatt 1716 Crocosmia (8) aurea (Pappe ex Hook.) Planchon m/s 2/b Phillips 4670b x crocosmiiflora (Lem. ex Morren) m/s 2/b Radelescu (19702) N.E. Br. Devia (1) xeromorpha Goldblatt & Manning m/s 2/b Snijman & Manning 1194 Dierama (44) a N.E. Br. m/s 2/b van der Zeyde 63 ii (N.E. Br.) Hilliard m/s 2/b Mauve & Venter 1 оет (L.f.) Baker m/s 2/b Dahlstrand 2896 Duthieastrum (1) linifolium (Baker) de Vos m/s 2/b Acocks 24202 (K) m/s 2/b Acocks 460 (BOL) Freesia (11) occidentalis L. Bolus m/s 2/b Goldblatt 4069 refracta (Jacq.) Klatt m/s 2/b Goldblatt 2841 Geissorhiza (82) schinzii (С. чш Goldblatt m/s 2/b Goldblatt 3008 scillaris А. m/s 2/b Walters 2616 Gladiolus (215) abyssinicus (Brongn. ex Lemaire) m/s 2/b Lavranos & Collonette 20357 Goldblatt & de V callianthus Marais m/s 2/b Keeley s.n. communis L. m/s 2/b Martin Cacao & Fernandez (1990) decoratus Baker m/s 2/b Richards 18743 (P) dehnianus Merxm. m/s 2/b Chase 8016 (P) dalenii Geel m/s 2/b Ryding s.n. (UPS) erectiflorus Baker m/s 2/b La Croix 1034 grantii Baker m/s 2/b Polhill & Paulo 1218 (P) gunnisii (Rendle) Marais m/s 2/b Bally & Melville 15706 illyricus Koch m/s 2/b Radelescu (1970b), Martin Cacao & Fernández (1990) imbricatus L. m/s 2/b Radelescu (1970b) italicus Miller m/s 2/b Martin Cacao & Fernández (1990) Volume 78, Number 4 1991 Goldblatt et al. 953 Sulcus and Operculum Structure in Iridaceae Pollen TABLE |. Continued. Aper Taxon ture Operculum Collection data melleri Baker m/s 2/b Goldblatt et al. 8135 muenzneri Har m/s 2/b Pawek 10432 ea (L. Bolus) Goldblatt m/s 2/b Radelescu (1970b) & de Vos Hesperantha (ca. 65) humilis Baker m/s 2/b Goldblatt 6090 radiata (Jacq.) Ker m/s 2/b Goldblatt 6994 Ixia (45) brevituba G. Lewis m/s 1/b Goldblatt & edo 8645 polystachya L. m/s 1/b Goldblatt 3 purpureorosea Lewi m/s 1/b Radelescu Cr rapunculoides мен m/s 1/b van Berkel 140 scillaris L. m/s 1/b Goldblatt 2567 Lapeirousia (36) arenicola Schltr. m/s 2/b Compton s.n. coerulea Schinz m/s 2/b Lavranos 22698 erythrantha Baker m/s 2/b Goldblatt 7515 neglecta Goldblatt m/s 2/b Goldblatt & Manning 9022 Melasphaerula (1) ramosa (L.) N.E. Br. m/s 2/b Low 943 Micranthus (3) alopecuroides (L.) Ecklon 2/5 а Barker 4903 (NBG) junceus (Baker) М.Е. Br. z/s a Winkler 166 (NBG) tubulosus (Burm.f.) N.E. Br. z/s a Goldblatt 8711 Pillansia (1) templemannii L. Bolus m/s 2/b Goldblatt 7907 Radinosiphon (1-2) leptostachya (Baker) N.E. Br. m/s 2/b Goldblatt 5933, 6864 cf. leptostachya m/s d/o Goldblatt s.n. Romulea (90) bifrons Pau m/s 2/b-d/o Moret & Bari 12-08 bulbocodium (L.) Sebast. & Mauri m/s 2/b (joined) Moret & Bari 313-89 va ioica Batt m/s 2/b Moret & Bari 413-89 camerooniana Baker m/s 2/b Gereau & al. 2828 clusiana (Lange) Nyman m/s 2/b Martin Cacao 7 و‎ (1990) columniae Sebast. & Mauri m/s d/o Moret & Bar engleri m/s 1/b-2/b (joined) Moret & Bari n Я is ligustica Parl m/s 2/b Moret & Bari 52-88 major Schousko m/s 2/b (joined) Moret & Bari 743-89 monadelpha (Sweet) Baker m/s 2/b Goldblatt 4036 ramiflora Ten. m/s 2/b Martín Cacao & Fernández (1990) rollii Parl. m/s d/o Moret & Bari 443-89 rosea (L.) Ecklon m/s 2/b Williams 2514 stenopetala Bég. m/s 2/b Moret & Bari 14-18 Savannosiphon (1) euryphylla (Harms) Goldblatt s/u s/u Marais Draper 14 (К); Pawek 7972 Schizostylis (1) coccinea Backh. & Harv. m/s 2/b Gibbs Russel 3016 954 Annals of the Missouri Botanical Garden TABLE 1. Continued. | Арег Тахоп ture Operculum Collection data Sparaxis s^ bulbifera (L.) Ker m/s 2/b Williams 430 Ne Goldblatt m/s 2/b Goldblatt 6162 galeata Ker m/s 2/b Goldblatt 72344 grandiflora (Delaroche) Ker m/s 2/b Goldblatt 6270 tricolor (Schneev.) Ker m/s 2/b Radelescu (1970a) Thereianthus (7) bracteolatus (Lam.) G. Lewis m/s 1/b Grant е "sce 9032 ixioides Lewis m/s 1/b Pillans juncifolius (Baker) G. Lewis m/s 2/b dius one Jackson s.n. longicollis (Schltr.) С. Lewis m/s l/b Compton 12413 (NBC) minutus (Klatt) G. Lewis m/s 1/b Esterhuysen 35830; Goldblatt 5221 racemosus (Klatt) G. Lewis m/s 2/b(-1/b) Esterhuysen 16620 (BOL) spicatus (L.) G. Lewis m/s b Goldblatt 5396, 9032 Tritonia (28) dubia Ecklon ex Klatt m/s 2/b Goldblatt 4944 pallida Ker m/s 2/b Goldblatt 4167 Tritoniopsis (22) pauciflora (Jacq.) G. Lewis m/s 2/b Radelescu (1970a) pulchella G. Lewis m/s 2/b Goldblatt 6839 Watsonia (52) aletroides (Burm.f.) Ker m/s 2/b Goldblatt 2858 borbonica (Pourret) Goldblatt m/s 2/b Goldblatt 7918 marginata (L.f.) Ker m/s 2/b Snijman 972 Zygotritonia (4) bongensis (Pax) Mildbr. t/c a Fay 5544 nyassana Mildbr. t/c a Faden & Faden 74/117 and one North African species of the predominantly southern African Romulea have grains with a sin- us narrowly elliptical band of exine in the center e sulcus (Fig. 1B). This resembles the oper- nd of some genera of Tecophilaeaceae (Simp- son, 1985). In А. viridis, Thereianthus minutus (Fig. 4), and R. engleri the opercular band of some grains lacks exine in the median center, suggesting the incomplete fusion of two separate bands. In one sample of Radinosiphon leptostachya the sulcus has exine fragments evenly scattered over its entire surface, while in three North African species of Romulea the operculum is sometimes partly (К. bifrons) to completely disorganized (К. columnae, R. rollii). In species with a 2-banded operculum, the bands are separated from one another and from the sulcus margins by the pore membrane that remains un- stained in our treatment. There may be scattered, small exine fragments concentrated in the midline of the unstained areas of the sulcus (Figs. 2-5). The opercular bands are two-thirds to nearly as long as the sulcus and converge (rarely appearing fused) at their ends; thus, the interopercular space is widest in the center of the sulcus. The opercular bands range from 3.6-12.5 um wide and are con- sistent in width within species and sometimes also within genera, but are fairly variable in length even within species. GRAIN SHAPE AND SIZE r grains of most species are typically broadly ellipsoid and slightly longer in equatorial than polar diameter, thus prolate (Table 2), but in a number of genera the grains are more or less globose. Grains of Micranthus and Savannosiphon, in particular, are spherical when fully expanded. Size is variable among species within genera as well as between genera. The smallest grains recorded are those of Ixia scillaris (32 X 36 um) and the largest those of Savannosiphon (106 x 106 ит) and Micran- Моште 78, Митбег 4 1991 Goldblatt et al. Sulcus and Operculum Structure in Iridaceae Pollen 955 thus tubulosus (90 x 90 um). Variation in size within genera may be extensive, notably in Gladio- lus (Table 2), but so far we have found no taxo- nomic significance to size per se at the generic level, except that /xia (Table 2) seems consistently to have the smallest grains. The genus stands out in the subfamily in this feature. EXINE SCULPTURING The exine of most genera of Ixioideae is tectate- columellate and usually has a perforate, sparsely scabrate tectum (described as punctitegillate or mi- cropunctate and granulate, spinulate or verrucate). Our study did not focus on this aspect of the pollen because it has been extensively documented by Schulze (1970, 1971) using light and transmission. M electron microscopy and reported under the SE for Duthieastrum (= Duthiella) (de Vos, 1974b), Homoglossum (now Gladiolus) (de Vos, 1976), Tritonia (de Vos, 1982), and Pillansia (Goldblatt & Stein, 1988). Exceptions are the three species of Micranthus (Figs. 6–8) that have a reticulate tectum (Erdtman, 1952; Schulze, 1970, 1971), which Thereianthus racemosus has as well. The operculum of the latter, however, has & perforate exine. Syringodea and Crocus have an imperforate but sparsely scabrate tectum (Schulze, 1971; de Vos, 1974a). Descriptions of the exine structure of pollen grains of Thereianthus and Micranthus, not before examined under the SEM, are as follows. Thereianthus (Figs. 2-4). Ехіпе tectate-colu- mellate, tectum perforate (except 7. racemosus), the perforations small (ca. 0.25-0.55 x the width of the intervening tectum, 7. juncifolius, T. lon- gicollis), (ca. 0.75-1 х the width of intervening tectum, 7. minutus, especially on the proximal pole); finely scabrate; terminating abruptly at the sulcus (T. minutus) or broken into scabrate patches (especially T. juncifolius). In T. racemosus, tectum reticulate, lumina 2-3 X width of the walls, be- coming microreticulate at the sulcus margins with the walls as wide as the lumina; exine of the oper- culum perforate (perforations ca. 0.75-1 x the width of the intervening tectum). Thus, except for T. racemosus, the exine of Thereianthus is typical of Ixioideae. Micranthus (Figs. 6-9). Ехіпе tectate-columel- late, reticulate entirely or becoming microreticulate to perforate (Fig. 9) toward the margins of the sulcus (especially M. tubulosus), and forming iso- lated scabrate islands near the sulcus margins, the muri finely scabrate, lumina 1-3 x the width of the muri (M. alopecuroides), 1.3-4 x (M. jun- ceus), 0.75-22.2 x (M. tubulosus), floor of the lumina containing pilae. DISCUSSION PHYLOGENETIC SIGNIFICANCE AND ORIGIN OF THE OPERCULUM Operculate pollen has been mentioned as oc- curring in Ixioideae by Schulze (1970, 1971), who did not indicate which taxa had an operculum nor its specific form. However, figures published by Radelescu (1970a, b) show an operculum in the sulcus of a few species, drawn as somewhat diffuse exinous material. Referring to the operculum as *colpus ornamentation,’ Radelescu did not de- scribe its structure in detail nor did she include more than a few of the genera possessing an oper- culum in her survey. A 2-banded operculum is reported by Straka & Friedrich (1984) in Gladio- lus bojeri (as Geissorhiza bojeri) and appears in the SEM micrograph of a pollen grain of a Gladio- lus cultivar in Kress & Stone (1982). Martin Cacao and Fernandez (1990) describe a 2-banded oper- culum in species of Romulea and Gladiolus that occur in Spain. The absence of any reference to an operculum (cf. Zavada, 1983) from other stud- ies of the pollen morphology of Iridaceae is prob- ably attributable to the examination of either ace- tolyzed or unexpanded grains. The occurrence of an operculum in all 23 genera of Ixioideae that have monosulcate pollen grains (the basic aperture type for the subfamily), includ- ing members of all three tribes, makes it all but certain that an operculum is a synapomorphy for Ixioideae. The operculum is present in the 2-band- ed form in Pillansia (the only genus of Pillansieae), in three of the five genera of Watsonieae sensu Goldblatt (1989a, 1990), and in 18 of the 19 genera of Ixieae that have monosulcate pollen grains. Although for morphopalynological reasons (A. Le Thomas, pers. comm.) it may be preferable to see the 2-banded operculum as having arisen from а 1-banded operculum, the pattern in extant Ixioi- deae suggests that a 2-banded operculum is an- cestral in the subfamily, and this is congruent with what is known about relationships within Ixioideae (Goldblatt, 1990, 1991; Rudall & Goldblatt, in press). The 1-banded operculum of /xia and some species of Romulea, Anomatheca (all Ixieae) and Thereianthus (Watsonieae) must therefore be re- garded as derived. The occurrence of the 1-banded operculum is scattered in the subfamily and none 956 Annals of the Missouri Botanical Garden ABLE 2. Pollen grain characteristics of selected Ixioideae (voucher information as in Table 1). Measurements for interopercular distance are only for grains with a double-banded operculum; those with a single operculum are scored n/a (nonapplicable) for the character. Measurements were taken only from normal-looking grains and are the mean of 5-10 measurements for each feature. Equatorial diameters are measured across the length of the sulcus. All measurements in microns. Grain diameter Sulcus Operculum Her opercular Taxon Pole Equator Width Length Width Length distance Anomatheca fistulosa 57 58 43 >58 4.8 43 п/а grandiflora 66 76 48 >76 8.4 67 7.2 60 64 48 > 64 7.2 63 9.6 verrucosa 53 61 34 >61 4.8 48 8.4 viridis 76 87 58 >87 12.5/6.3 60 п/а/2.4 51 59 34 >59 12.1 39 п/а Babiana dregei 63 67 43 >67 6.6 51 12.1 stricta 76 89 48 >89 7.2 58 10.8 Chasmanthe aethiopica 52 59 34 53 8.4 48 9.6 Crocosmia aurea 43 57 48 >57 4.8 48 12.1 Пета xeromorpha 46 52 34 >52 3.6 35 8.4 [легата cooperi 46 53 20 >53 4.8 39 6.0 mossil 48 51 41 >51 6.0 41 9.6 репашит 68 65 60 >65 10.4 58 16.9 Duthieastrum linifolium 55 75 51 >75 7.9 70 12.1 Freesia refracta 51 58 36 >58 6.5 41 10.8 Geissorhiza schinzit 89 96 65 >96 8.5 43 19.3 scillaris 44 48 36 >48 4.8 35 11 Gladiolus abyssinicus 60 65 43 >65 8.4 58 8.4 callianthus 75 82 43 >82 8.4 68 12.1 decoratus 59 60 3l >66 5.1 44 8.8 lehnianu 37 40 31 >40 6.6 42 5.9 erectiflorus 60 71 53 >71 9.6 51 9.6 g 50 53 47 >53 5.9 47 7.4 muenzneri 43 48 40 >48 7.2 27 7.2 47 58 34 >58 6.0 48 7.2 Hesperantha humilis 91 51 43 >51 4.5 38 12.5 radiata 48 48 21 >48 4.1 36 6.0 Ixia revit 32 36 24 >36 4.8 27 п/а polystachya 34 38 23 >38 9.6 36 п/а rapunculoides 37 41 27 >41 8.4 36 п/а scillaris 30 41 27 >41 6.6 39 п/а Volume 78, Number 4 Goldblatt et al. 957 1991 Sulcus and Operculum Structure in Iridaceae Pollen TABLE 2. Continued. Grain diameter Sulcus Operculum шне opercular Taxon Pole Equator Width Length Width Length distance Lapeirousia arenicola 67 72 48 = (2 6.0 60 15.7 coerulea 55 55 36 >55 4..8 36 9.1 erythrantha 51 65 38 >65 3.6 53 UL Melasphaerula ramosa 55 61 34 >61 4.9 53 TZ Micranthus tubulosus 90 90 18-20 n/a n/a n/a n/a Pillansia templemannii 63 67 65 >67 9.6 60 9.6 Radinosiphon leptostachya 64 71 36 >71 6.2 71 8.2 Romulea camerooniana 55 60 48 >60 6.0 41 8.4 monadelpha 43 62 24 >62 5.3 43 9.6 гозеа 46 63 25 >63 4.8 43 1.2 Savannosiphon euryphylla 106 106 n/a n/a n/a n/a n/a Schizostylis coccinea 48 95 30 >55 7.2 43 4.8 Sparaxis bulbifera 70 77 48 >78 8.9 60 12.7 grandiflora 58 67 46 67 6 65 15.7 Thereianthus juncifolius 65 70 41 65 6.0 97 7.2 79 76 43 >76 6.6 58 9.6 minutus 55 70 53 >70 19.9 58 п/а racemosus 43 58 51 >58 4.8 43 8.4 spicatus 68 71 55 >71 12.1 65 n/a Tritonia pallida 59 70 36 >70 6.0 53 8.4 Tritoniopsis pulchella 54 59 53 +59 9.6 53 15.6 Watsonia aletroides 63 71 58 >71 12.5 60 16.9 borbonica 87 101 58 >101 8.9 60 10.4 marginata 63 69 68 >69 8.4 55 18.1 Zygotritonia bongensis 63 70 17 >70 п/а п/а п/а пуаззапа 48 57 12 >57 п/а п/а п/а of the genera in which it occurs is especially prim- itive. Romulea is probably a particularly special- ized genus. Moreover, the closest relatives of gen- The single opercular band in Anomatheca fis- era with a 1-banded operculum have a 2-banded tulosa and the single or nearly single band in А. operculum. viridis are noteworthy, for these two species are STATUS OF TAXA WITH l-BANDED OPERCULA 958 Annals и A A Garden t [s " "i ey. ue Lt O! ~ TN. «27. oe tere i eru MA E >>” FiGURE 1. А. Anomatheca laxa, — ara йм а single banded operculum.—C. Zygo Savannosiphon panies 8 ле grain. For duc of grains see Table ~ пота bongensis, unusual members of the genus and are evidentiy taxonomically isolated from the other species of Anomatheca and the closely related Freesia (Gold- blatt, 1972). The two species seem derived in the genus, 4. fistulosa having a long perianth tube and flowers solitary on the branches, and 4. viridis having a green perianth and autogamous repro- duction. The possibility that these two species are united by the 1-banded operculum, or are mis- placed in Anomatheca, should be considered in future studies of this genus. It seems reasonable to regard the 1-banded operculum in 4. fistulosa and А. viridis as derived within the genus on the grounds explained above and because of their own specialized features. The infrageneric variation in the opercular mor- Ж О mas ot EE Expanded pollen grains of Ixioideae i in polar and equatorial view p for D), showing sulcus types. — — В. 1 xia brevituba, monosulcate grain with triaperturate grain, the от without opercula.— phology in Thereianthus probably also has taxo- nomic significance. Thereianthus juncifolius, which has a 2-banded operculum, is isolated in the genus because of its lax inflorescence, but the possibility that it is misplaced is remote. The other species of Thereianthus, T. racemosus, which has a 2-banded operculum, may be primitive in the genus. It stands out morphologically in having small flowers with a short perianth tube (a plesiomorphic character in the genus). It should be considered less specialized than the long-tubed and often zygomorphic-flow- ered members of the genus, all of which have a 1-banded operculum. Thus, in this genus the hy- pothesis that the 2-banded operculum is the basic state also seems likely. In the largely кын m African genus Romulea Моште 78, Митбег 4 1991 Goldblatt et a 959 Sulcus and IE Structure in Iridaceae Pollen FIGURES 2-5. Aperture detail їп th templem == median lacking e» Sulcate T Len of аб — 2. Pillansia templemannit with 2-banded oper . Ther anthus minutus, sulcate grain with 1-banded Wilh i with the ана Be ава grain showing operculum (one of the two а has exfoliated) 3). and scattered exine nes lying on the aperture membrane. Scale bar = 20 um (2, 4, 5; = 10 u (whose close relatives are also southern African), only a few North African species are so far known to have a l-banded or disorganized operculum, although pollen of few of the southern African species have been examined. A 1-banded opercu- lum is thus presumably derived in the genus, but as far as we know there are no morphological or other reasons to view the species with a 1-bande operculum as specialized or derived from ancestors with a 2-banded operculum. The only genus that seems consistently to have а 1-Бапде4 operculum is /xia, and this is presum- ably an apomorphy for the genus. Lewis (1962) regarded /xia as most closely related to Dierama. Dierama, which has a 2-banded operculum, is ap- parently less specialized in its evergreen habit but derived in its pendulous flowers and possibly also in its unusual leaf anatomy (Rudall & Goldblatt, in press; Goldblatt, in prep.). SIGNIFICANCE OF THE ZONASULCATE APERTURE AND RETICULATE EXINE IN MICRANTHUS Zonasulcate grains о еру described as syn- colpate by Schulze, 1970) a are almost certainly both deis for Micran- thus, e a reticulate exine vidently a fairly derived genus of Watson- e (Goldblatt, 1989a icranthus is re. correctly placed in Ixioideae despite its reticulate or largely reticulate exine, which corresponds to the patterning in other subfamilies of Iridaceae. In this connection, the transition from reticulate to microreticulate to per- forate exine toward the sulcus margins, noted in preparation of M. alopecuroides and M. tubulosus, seems noteworthy. In Iridoideae with reticulate ex- ine structure, perforate exine does not occur ad- jacent to the aperture (e.g., Goldblatt & Manning, 1989). The reticulate exine of Micranthus thus 960 Annals of the Missouri Botanical Garden FIGURES 6-9. fragments in rom microreticulate to perforate exine in M. tubulosus. Scale bar — 20 um (6, 8); = n the aperture.—7. M. appears best regarded as a reversal to the basic state for the family. Likewise, we consider the reticulate exine of Thereianthus racemosus a specialization for the species. The possibility that this character is shared with Micranthus must be seriously considered. Th two genera are closely related (Goldblatt, 1989a), and the suggestion that they may be congeneric seems supported by this shared exine character. However, the unusual zonasulcate aperture of Mi- cranthus seems to isolate it effectively from Therei- anthus, despite their shared karyology, unusual leaf insertion, and similar corm, morphology (Goldblatt, 1991). The absence of intermediates between the sul- cate-operculate and zonasulcate apertures in the subfamily makes exact interpretation of the ap- erture of Micranthus impossible. The aperture may have evolved through an extension of the sulcus fruit, and seed Pollen grains and exine detail of Micranthus.— 6. M. dq at whole grains showing exine junceus, detail of exine sculpture. Det . M. tubulosus. —9. ail of transition 5 um (7); = 10 um (9) around the grain through the distal pole and the loss of the operculum. Alternatively, the aperture may have been pushed into an equatorial position by the enlargement of the operculum to match the size of the other part of the exine (thus zonisul- culate). SIGNIFICANCE OF TRI-APERTURATE AND POLYRUGOIDATE GRAINS The trisulcate grains of Zygotritonia are re- markable for Iridaceae and probably for the mono- cots. We speculate that the tri-aperturate condition may have evolved from an ancestor with a double- banded operculum by the enlargement of the oper- cular bands, and the displacement of the membrane areas away from the polar part of the grain toward the equator. Zygotritonia is a specialized genus (Goldblatt, 1989b) and must be derived from stock Volume 78, Number 4 1991 Goldblatt et al. Sulcus and Operculum Structure in Iridaceae Pollen that had a sulcus with a double-banded operculum. Whatever the phylogeny of this unusual aperture type, it is clear that Zygotritonia is additionally isolated in the subfamily by its specialized pollen grains. The remarkable pollen grains of Savannosiphon reinforce our appreciation that the genus, segre- gated from Lapeirousia by Goldblatt & Marais (1979), is quite distinct from the palynologically uniform (as far as is known) Lapeirousia. LITERATURE CITED ON G. 1952. Pollen Morphology and Plant Tax my. Angiosperms. Almqvist & Wiksell, Stock: olm. FAEGRI, К. & J. IVERSEN. 1964. Textbook of Pollen Analysis, 2nd revised edition. Hafner, New Yor GOLDBLATT, P. ] A revision of the genera Lapei- rousia Pourret and Anomatheca Ker in the winter rainfall region of South Africa. Contr. Bolus Herb. 4: 1-111. 1989a. The southern African genus Watsonia а беа: Ann. Kirstenbosch Bot. Gard. 19. 1989b. Revision of the tropical African genus Zygotritonia Tna Bull. Mus. Natl. Hist. Nat., ér., Adansonia 11: 199-212 о and classification of Irida- ceae. Ann. Missouri Bot. Gard. 77: 607-627 19 An overview of the systematics, phy- logeny and biology of the African Iridaceae. Contr. Bolus Herb. 13: 1-74. . MaNNING. 1989. Pollen morphology of сє shrubby Iridaceae, Nivenia, Klattia, and Wit- зета. Ann. Missouri Bot. Gard. 76: 1103-1108. “8 W. MARAIS. 1979. Savannosiphon gen. nov., а segregate of Lapeirousia (lridaceae- Ixioi- deae). Ann. Missouri Bot. Gard. 66: 845-850. & . STEIN. 1988. Pollen morphology of Pillancin L. we аза Ann. Missouri Bot. 1965. Morphologic Lun of dp ir Univ. Arizona Press, Tuc KRESS, J. . E. STONE. 1982. Nature > of the s por wi in зі онат ащ with special reference to the pollen grains of Canna and Heliconia. Grana 21: 129-148. Lewis, С. J. 1962. South African Iridaceae. The genus xia. J. S. African Bot. 28: 45-195. Martin Cacao, M. & I. FERNANDEZ. Contri- ución al estudio palinológico de la familia Tridaceae en Andalucia ыр-чыр (excepto el género Iris L.). Lagascalia 15: MATHEW, B. а The Crick: B. T. Batsford, Lon- on. OGDEN, E. С., С. S. Raynor, J. У. Ha р. М. E: "s. 1974. Manual of Sampling Airborne Pollen. r Press, Londo Hon D. 1970. ` Recherches morphopalynol- giques sur les е d'Iridaceae. Lucr. Сгад. Bot. Bucuresti 1968: 311-35 ——. 19705. Re Sel да wy aon ig ыы sur les espéces d'Iridaceae de la flore Rou . Lucr. Сга i Корал, P. . GOLDBLATT. ss. Leaf зге: and phylogeny of Ixioideae о Bot. J. L So SCHULZE, У. 1970. Beitrage zur peer as Iridaceae-Ixioideae. Wiss. Z. Fr uer ard Jena, Math.-Naturwiss. Reihe. 19: Beitrage zur ee der Iridaceae und ihre Bedeutung fiir die Taxonomie. Feddes Repert. 82: 101-124. Simpson, M. G. 1985. Pollen ultrastructure of the Te- cophilaeaceae. Grana 24: 77-92. SrRAKA, Н. & B. FRIEDRICH. 1984. Palynologia Mad- agassica et Mascarenica. amily 45. Iridaceae. Trop : 13-1 Die Suid- Afrikaanse genus 5у- . African. Bot. 40: 201-254. Duthiella, ^n nuwe prs van die Iridaceae. J. S. African Bot. 40: 301-309. 976. Die Suid-Afrikaanse species van Homo- glossum. J. S. African Bot. 42: 09. 1982. The African genus Triton (байыд Part 1. Ј. S. African Bot. 48: -163. WALKER, J. У. & J. A. DOYLE. ases of angio- sperm phylogeny: palynology. Ann. Missouri Bot. Gard. 62: 664-723. ТАМАРА, М. 83. Comparative morphology of mon cot pollen and evolutionary trends of apertures "m wall structures. Bot. Rev. (Lancaster) 49: 331-379. nia Ker-Gawler 105 SYSTEMATIC PLACEMENT OF THE PLATANACEAE IN THE HAMAMELIDAE! and Robert N. Schwarzwalder, Jr.? David L. Dilcher? ABSTRACT A cladistic analysis of the basal Hamamelidae based upon 92 characters is presented. The results support a close phylogenetic relationship between the subfamily Altingioideae (Hamamelidaceae) and Platanaceae but suggest that present hypot the five subfamilies of heses concerning the derivation of the Hamamelidae from the Hamamelidaceae using Platanaceae as an outgroup yields a phylogeny that corresponds to the Magnoliidae need revision. Analysis of fossil record far better than does current theory. As a result of this study and recent paleobotanical and phytochemical evidence, it is suggested that Hamamelidae are at least as primitive as Magnoliidae. Platanaceae are a monogeneric family of trees with a fossil history that extends to the Albian, Lower Cretaceous (Kutuzkina, 1974; Hickey & Doyle, 1977; Dilcher & Eriksen, 1983; Upchurch, 1984; Crane et al., 1986; Schwarzwalder, 1986). Cronquist (1981) considered Hamamelidales, which include Platanaceae, as an order of subclass Hama- melidae ancestral to all other orders within the subclass except the Trochodendrales. According to this view, Hamamelidales evolved from a Magnoliid or Magnoliid-derived taxon on a line of evolution that first spawned Trochodendrales. Recognition of Platanaceae as a well established angiosperm family in the Early Cretaceous raises questions concerning the current systematic placement of this family. A diverse body of evidence supports a close systematic relationship between Platanaceae and Hamamelidaceae (Tippo, 1938; Jay, 1968; Takh- tajan, 1969; Hickey & Wolfe, 1975; Cronquist, 1981; Zavada & Dilcher, 1986). Wolfe (1973) suggested a relationship between Platanaceae and Eupteleaceae based upon foliar characteristics. Morphological similarities between Platanaceae and Hamamelidaceae—especially floral resem- liance. However, there is no agreement concerning the relative advanced or primitive nature of the two families, and thus the extent of their relation- ship remains unresolved. Tippo (1938) felt that Platanaceae are “clearly more advanced than the Hamamelidaceae anatomically.” Cronquist (1981) considered Platanaceae to have been derived ear- lier than Hamamelidaceae from a common stock of flowering plants. He commented that the flowers of Platanaceae are the more primitive of the two, but stated that “neither family can be derived from the other." PROCEDURES AND METHODS A cladistic analysis of the basal Hamamelidae— Eupteleaceae, Platanaceae, and the five subfamilies of Hamamelidaceae— was conducted using Swof- ford's PAUP (Phylogenetic Analysis Using Parsi- mony) program (Swofford, 1984) on Indiana Univ- ersity's mainframe computer. PAUP options were chosen to consider all possible trees on the basis of maximum parsimony. Character state polarities were established by outgroup analysis using Tetra- centraceae and Trochodendraceae as outgroups. Ninety-two characters are presented for each of the nine taxa in Table 1. Character states are coded for presence or absence for each of the nine taxa surveyed. Presence of a character state in any species of a given family or subfamily resulted in that taxon being scored as positive for the trait. The data used in our analysis were gleaned from the literature or generate chwarzwalder (1986). The 92 characters include aspects of sec- ondary chemical compounds, wood anatomy, epi- dermal characteristics, pollen structure, foliar mor- phology, petiole anatomy, seed structure, floral morphology and pollination adaptation, inflores- cence type, and fruit type. In our analysis the five subfamilies of the Hama- ! We ac с the assistance | Professor Ted Barkley and NSF grants 79-10720 and 85-16657 to D. Dilcher. of Biology, Indiana rsity, Bloomington, Indiana 47405, U.S „А. Current address: Engineering/ ныш үтен 308 qeu ни Library, University of Michigan, Ann Arbor, Michigan 48109-1185, à Bie yartments of Biology and Geology, Indiana University, Bloomington, Indiana 47405, U.S rida, Gainesville, U.S.A. Florida Museum of Natural History, University of Flo „А. Current address: Florida 32611, ANN. MISSOURI Вот. GARD. 78: 962-969. 1991. Volume 78, Number 4 1991 Schwarzwalder & Dilcher 963 Systematic Placement of Platanaceae melidaceae, Altingioideae, Disanthoideae, Exbuck- landioideae, Hamamelidoideae, and Rhodoleioide- ae, were treated as taxonomic equivalents to the Platanaceae, Eupteleaceae, Tetracentraceae, and Trochodendraceae. This subdivision was performed to insure a meaningful analysis of these taxa. The five subfamilies of the Hamamelidaceae are phe- notypically quite distinct. Unlike the other families of the basal Hamamelidae, which display only mi- nor genetic and phenotypic variation, the Hama- melidaceae, with 26 genera and more than 100 species (Cronquist, 1981), are extremely hetero- geneous. If the family is analyzed as one taxon, the large degree of phenotypic variability encom- passed by Hamamelidaceae tends to obscure in- trafamilial and interfamilial relationships. Numer- ous authors have recognized the distinctness of the hamamelidaceous subfamilies and have recom- mended the elevation of one or more of them to familial status (Blume, 1928; Hayne, 1830; Lind- ley, 1836; Wilson, 1905; Nakai, 1943; Chang, 1959, 1964; Skvortsova, 1960; Willis, 1966; Takhtajan, 1969; Melikian, 1971, 1973а, b; Rao, 1974; Rao & Bhupal, 1974; Wolfe, 1973; Dahl- gren, 1977; Jha, 1977). These five subfamilies are considered as sister groups to Platanaceae and Eupteleaceae. Tetra- centraceae and Trochodendraceae, together con- sidered by Cronquist (1981) as members of the most primitive order within the Hamamelidae, were used as separate outgroups to establish character state polarities. Establishing character state polar- ities using the two families phenotypically closest to the Magnoliidae allows us to reconstruct Hama- melid evolution in relation to current evolutionary thought (Cronquist, 1981). By using Tetracentra- ceae and Trochodendraceae as outgroups we test the hypothesis of a Magnoliid-derived Hamameli- dae. Cercidiphvllaceae and Myrothamnaceae are not considered in this analysis, because we feel that evidence suggests neither family as a hypothetical sister group to Platanaceae In order to determine the stability of relation- ships within the cladogram generated from this study, four test analyses were performed. In each test, 5% random error was introduced into the original data set, and PAUP was used to generate a cladogram(s). Using cladistics, one formulates the most direct course of evolution possible to hypoth- esize the evolutionary relationships of a given set of taxa. Evolutionary noise, i.e., homoplasy, occurs in a data set proportional to the extent that mosaic evolution occurs in a group of related taxa. Ho- moplasy is an expected component of any data set and tends to introduce ambiguity into cladistic re- sults. The experimental addition of random error into this data set can be considered as extra ho- moplasy. Well- resolved relationships should be sta- relationships. Since cladistics formulates the evo- lutionary interrelationships of all taxa considered in an analysis, an error in the data set of one change in any taxon has ramifications for all taxa. Rohlf & Sokal's (1981) statistical tables fur- nished random numbers. Three-digit random num- bers were used to designate character states to be changed. The first digit of the number specified the taxon (1-9), and the last two digits specified the character (1—92). As all data are in presence- absence form rather than having three or more character states, characters could be reversed with- out ambiguity. Four tests of the initial cladogram (Fig. 1) were conducted. In each test 41 character states, 5% of the total 828 character states (92 characters for each of nine taxa), were reversed. A listing of the character states reversed in each of the four tests is presented in Table 2. Figure 2A-D illustrates the results of these tests. Consen- sus trees were calculated where more than one tree was determined to possess maximum parsimony. Swofford's CONTREE program, which uses the methods of Adams (1972) and Rohlf (1982), was used in these analyses. RESULTS AND DISCUSSION The cladistic analysis resulted in one tree of a length of 157 steps (Fig. 1). In this cladogram, Platanaceae are placed within Hamamelidaceae sensu lato as a sister group to subfamily Altin- gioideae. Platanaceae and subfamily Altingioideae share 12 derived character states: the presence of cyanidin (character 5), glandular trichomes (char- acter 17), actinodromous to palinactinodromous primary pattern of venation (character 42), the union of intercostal tertiary veins to secondaries (character 47), altingioid marginal venation (char- acter 53), the presence of several vascular traces in petioles (characters 56 and 57), orthotropous ovules (character 60), small seeds (character 64), hard testas (character Ga), minute petals е 74), апа р t ). The position of Eupteleaceae on the калыша per the large synapomorphy list separating this family from Platanaceae and Hamamelidaceae suggest that Eupteleaceae and Platanaceae are more distantly related than is indicated by similarities of foliar features alone. The cladogram suggests that Platanaceae orig- 964 Annals of the Missouri Botanical Garden TABLE 1. Characters and character states for A. Disanthoideae, B. Exbucklandioideae, C. Hamamelidoideae, D. Altingioideae, E. Rhodoleioideae, F. Platanaceae, G. Eupteleaceae, H. Tetracentraceae, I. Trochodendraceae. Su- perscripts: 1. Saupe (1981), 2. Seigler (1981), 3. Dahlgren et al. (1981), 4. Jay (1968), 5. Harborne (1967), 6. Gornall et al. (1979), 7. Metcalfe & Chalk (1965), 8. Tang (1943), 9. personal observation, 10. Bailey & Nast (1945), 11. Zavada & Dilcher (1986), 12. Praglowski (1974), 13. Bogle & hiis (1980), 14. Hickey & Wolfe (1975), 15. Wolfe (1973), 16. Nast & Bailey (1946), 17. Endress (1977), 18. Nast & Bailey (1945), 19. Boothroyd (1930), 20. Bogle (1970), 21. Rao (1974), 22. Schwarzwalder & 8 (1981), 23. Harms (1930). (9 = not applicable; ? = unknown value A B C D E сон 1 1. Cyanogens tyrosine-derived' — = — — — + - — + 2. Monoterpenes iridoid** = — Е + m = = = = 3. Ellagic acid** d eo de Б Ж э > ш о „ 4. Myricetin'** — + + + + + - - $ 5. Cyanidin* — - + ? + = = = 6. Flavonols acylated^ IM ж ш EP = mm oa" 7. Leucodelphinidine* -= ? + + ? + - ? ? 8. Leucocyanidine' + 2 + + 2 + + 2 2 9. Vessels present'* + + + + + + + ES 10. Wood rays heterocellular’ + + + + + = + + + 11. Axial parenchyma banded’ — — + = = + = = + 12. Wood fiber length = 1,600 um: + + + + + + + + 13. Tyloses present’ - - + + = + + о @ 14. Vessel elements > 800 um'* + + + + + — + @ о 15. Number bars/perforation plate > 20** = + + + + = + @ @ 16. Perforation plates simple'^ = - Е — = + - @ @ 17. Trichomes glandular’? — - — + = Ф - € — 18. Trichomes stellate’? — = + = + = = ER 19. Trichomes branched’? = - — - = + = = = 20. Trichomes peltate E + = = = == ET — Secretory cells" + + + $ d: = + + РЕ 22. Crystals solitary" = > + - - = $ 4 23. Crystals clustered" — + — + + + = + = 24. Stomates bus о - — + - = + = + + 25. Stomates paracytic" + — + + = + = = = 26. Stomates aa — — — _ - 4 + = = 27. Stomates cyclocytic? — + - - + = = = Е 28. Epidermis papillate: - - + = d = + = — 29. Pollen штсођраге!! >!» + + + = + + + + + ЗО. Pollen forate!!!» — — + + — = = = = 31. Pollen prolate!!!» + + + – + + + + + 32. Pollen spherical''!*'* HE + + uc ж 33. Pollen oblate!!! = = + = = 8 ra = + 34. Endexine present'™!? + = + + - + + + + 35. Endexine thick at aperture'''* + — - - + س‎ + + + 36. Endexine not through aperture!" — — + + — + = = - 37. Pollen psilate'''?'* + + + — + + – + + 38. Pollen scabrate''!*'* - - – - Е — + = = 39. Pollen гиршаге! >> — + + + = = Еа = - 40. Leaves opposite” — — + = = = = — — 41. Primary veins pinnate”'* — — + + + + + = + 42. Primary veins oe Е = = $ + — + Е 43. Primary veins acrodromous + + — — — — = = = 44. Secondary veins ии = — + + – + + - = 45. Secondary veins camptodromous”'*'* + + — + + + - + + 46. Tertiary veins strongly percurrent” - — + E — + + = = 7. Intercostal tertiaries joining secondaries? = - = + — $ + - — 48. Tertiary veins forming chevrons’ + + + + + + + + — 49. Teeth chloranthoid”'* - — — - - = — + + Моште 78, Митбег 4 1991 Schwarzwalder & Dilcher 965 Systematic Placement of Platanaceae TABLE 1. Continued. 50. Teeth platanoid?'* 51. Secondary veins decurrent?'* 52. Teeth admedially oriented”'* 53. Altingioid marginal venation? 54. Teeth hamamelidoid? 55. Nodes unilacunar*'* 56. Number petiole traces = 1*'!* 57. Number petiole traces > 1*'* 58. Petiole nia ses axillary bud" 59. Stipules present? pre 60. Ovules anatropous!'*!*.1>.20 O00 ею WN × $5835 Ф 5 ЕЕ 5 g < © o c Е 5 5 ® з E Ф s 5 e e n Ф eu = N ss 65. Subchalazal area vascularized'*"*?' . Tegmen two- Е 67. Tegmen persistent in mature seed” А ot bu ndis ending at chalaza'**! sta hard, conspicuous'*! 10. Сев fusediete.rsao 71. Carpel wall open'*!*192 72. Flowers imperfect'*!*92» 73. Flowers perfect!9 920 74. Petals minute or absent!*!*92o22 с an oO © O œ 75. Sepals minute or absent!*!?-19.20.22 76. Carpels sessile!*!*:92» 77. Stamens valvate!$! 7.19.1920 78. Оущез > 119171920 79. А philous'®"7"8 80. Entomophilous'®!”"'* 81. Epigvnous!*! 19.20.23 82. Hypogynous'*15192023 83. Perigynous!*!*192023 84. Flowers solitary or іп pairs'*!7!5192 85. Flowers in spikes!*!7.1*-1>20 86. Inflorescence a суше or cymose raceme'*' 5^ 87. Flowers in hrads!*!?.18.19.20 88. Flowers in pseudanthia! 1715192021 89. Fruit a follicle or (осетите 17,18,19,20,21 0. Fruit ca psule ip 1617,18,19,20,2 91. Fruit an асћепе! 6! 8192021 92. Fruit a samara!?!”.18.19.20,21 o - = = = + = = = + == = a + — E — = — == = + = rem = — — => 2 = = — — = + m — + + + = + = = + + - = = + => + + - +4 + + + + + + = + = + + + = + = + + + + + + + + + = LL = — = = = = = - + + == = - = + + + - =- — + о - + + + + - = = = = — — + + - =- = = = = - + + + + + + + = = + + + + + + + + = = = + + + = + = = = = + + + + + = = = + - = = o. - + -—- + + - + + + = + = = = + =- + = + = + + + ОА О + + + - + = + = + + = + + + + + + + = + + = E + -— — — = = > — + + + = + + + + + = + = + = + = + + = + = = + + + + = + + = + -= LLL XUL LLLA + = + = = = + = LA - = + + = = = = + - + + + -= + =- WX >- uut - =m — + x — a — - = + + + + + + + -= - - - — > — — + — — — — == = > — == + — = inated late in a series of evolutionary events that gave rise to the subfamilies of Hamamelidaceae. The position of Disanthoideae in the cladogram is corroborated by Cronquist (1968, 1981) and Takhtajan (1969); however, this is not unexpected since their placement of Disanthoideae is based upon their similarity to the Magnoliidae. However, the positions of Hamamelidoideae, Altingioideae, and Platanaceae as the most recently derived taxa of the group do not agree with the fossil record. Members of the Platanaceae are first known from fossil leaves and reproductive material from the Albian, Lower Cretaceous (Kutuzkina, 1974; Hick- ey & Doyle, 1977; Dilcher & Eriksen, 1983; Upchurch, 1984; Crane et al., 1986; Schwar- zwalder, 1986). The first megafossil records of Altingioideae are leaves from the genus Liquid- in the Cenomanian, lowermost U Cretaceous, (Newberry, 1898; Chesters et al., 1967) and Upper Cretaceous (Brown, 1933; En- ambar L. 966 Annals of th Missouri Him Garden o) H— НАМА 3 © &ь со ‹о б Y 0 Јело 7 42} ALTI 47 7 60 [— PLAT 64 @ . Abbreviations above are as follows: ТЕТК = Tetracentraceae, TROC = Trochodendraceae, EUPT = Eupteleaceae, DISA — Disanthoideae, RHOD — Rhodo- leioideae, EXBU = Exbucklandioideae, HAMA = Hama- liai is, ALTI = Altingioideae, and PLAT = Plat- anaceae. dress in Cronquist, 1981). megafossils from the Hamamelidoideae are foliar fossils from the lower Paleocene (Mathiesen, 1932; Hollick, 1936; Koch, 1963; Wolte, 1966, 1973). The early fossil records of these putatively ad- vanced taxa and the lack of early fossil represen- tation of the other taxa in the cladogram suggest that the analysis may not be valid. The discrepancy between these results and the fossil record prompted us to test the validity of our initial analysis. Four tests were performed by in- troducing 5% error into the data set and conducting cladistic analyses on each with the PAUP software. Results are presented in Figure 2. The relationship between Altingioideae and Plat- anaceae, and the position of Eupteleaceae are pre- served in the four consensus trees in Figure 2. The other subfamilies of Hamamelidaceae, however, are subject to extensive rearrangement with the ad- dition of random error. Thus, the phylogenetic reconstruction of the basal Hamamelidae presented in Figure 1 is questionable if we assume that there is at least 5% homoplasy in the origina! data set. A possible source of error in the hypothetical phylogeny presented in Figure | is the use of the two families of Trochodendrales as outgroups in the cladistic analysis. Fossil evidence of Platana- ceae in the Albian, upper Lower Cretaceous (Ku- tuzkina, 1974; Hickey & Doyle, 1977; Dilcher & Тће first records of TABLE 2. Character state reversals for the tests of cladistic relationships in the basal Hamamelidae. Character state reversals are identified, by taxon, for each of the four tests presented in the body of the paper. Numbers given after each taxon correspond to the characters listed in Table 1. Random Error Test One (Fig. 2A) Disanthoideae: 29, 33, 34, 54, 70, 89 Exbucklandioideae: 30, 38, 60, 80 Hamamelidoideae: 4, 14, 23, 27, 38, 57 Altingioideae: 68, 75 Rhodoleioideae: 2, 4, 23, 27, 31, 48, 52, 69, 82, 86 Platanaceae: 5, 76, 85 Eupteleaceae: 23, 56, 59, 64, 68, 71 Tetracentraceae: 59, 61 Trochodendraceae: 65 Random Error Test Two (Fig. 2B) Disanthoideae: 30, 49, 71, 88 Exbucklandioideae: 25, 34 Hamamelidoideae: 23, 45, 52, 91 Altingioideae: 5, 15, 18, 20, 45, 51, 54, 86 Rhodoleioideae: 20 Platanaceae: 5, 59, 66, 78 Eupteleaceae: 16, 28, 43, 44, 46, 84, 89 Tetracentraceae: 18, 32, 42, 67, 73 Trochodendraceae: 1, 27, 40, 55, 64, 73 Random Error Test Three (Fig. 2C) Disanthoideae: 3, 31, 38, 59 Exbucklandioideae: 4, 19, 53 Hamamelidoideae: 18, 22, 34, 61, 77, 79 Altingioideae: 48, 52, 70, 80, 86, 91 Rhodoleioideae: 16, 18, 30, 43, 48, 52, 59, 76 Platanaceae: 5, 33, Eupteleaceae: 32, 43, 66, 68 Tetracentraceae: 81, 91 A 19, 34, 56, 63, 69 Random Error Test Four (Fig. 2D) Disanthoideae: 7, 75, 77, 84, 87, 92 Exbucklandioideae: 4, 6, ae 50 Hamamelidoideae: 3, a Altingioideae: 10, 73, Rhodoleioideae: 19, 5 po 65, 68, 71, 81 Platanaceae: 26, 28, Eupteleaceae: 23, a 54, Tetracentraceae: 4, 12, M A 48, 66, 67, Trochodendraceae: 40, 55 Eriksen, 1983; Upchurch, 1984; Crane et al., 1986; Schwarzwalder, 1986), predates the occur- rence of any other member of the Hamamelidae. Hickey & Wolfe (1975) noted the early occurrence of the platanoids and suggested them as “possible early members of the trend toward the hamamelid line." If we assume that their early occurrence and morphological similarity to other members of Volume 78, Number 4 Schwarzwalder & Dilcher Systematic Placement of 967 Platanaceae А) ER 3 DISANTHOIDEAE 6 | Го EXBUCKLANDIOIDEAE 5 ALTINGIOIDEAE HAMAMELIDOIDEAE 15 Le ALTINGIOIDEAE 15 = А E 3 17 HAMAMELIDOIDEAE RHODOLEIOIDEAE 58 be 39 53 Мел" 50 56 EXBUCKLANDIOIDEAE TETRACENTRACEAE | pe 57 TROCHODENDRACEAE ыы 60 35 64 39 DISANTHOIDEAE a 69 72 20 74 79 91 83 87 RHODOLEIOIDEAE B) PLATANACEAE FIGURE 3. Adams consensus tree of two equally par- IRR simonious trees where the Platanaceae serve as ut- EXBUCKLANDIOIDEAE group. Synapomorphies are designated by rectangles. баке Numbers in rectangles refer to character numbers i ALTNGIODEAE Tab PLATANACEA RHODOLEIOIDEAE EUPTELEACEAE . the group demonstrate an early evolution from a TETRACENTRACEAE : :3 |; м Бар узин hamamelid ог prehamamelid lineage, and use the Platanaceae as an outgroup for analyzing the Ham- с amelidaceae, we obtain an alternative hypothesis ) A Я "m i of evolution in these two families (Fig. 3). РЕ Figure 3 suggests that Altingioideae were the EXBUCKLANDIOIDEAE _ earliest-derived of the hamamelidaceous subfami- HAMAMELIDOIDEAE | таайы : | lies. Later evolution in the family, according to the RHODOLEIOIDEAE ALTINGIOIDEAE EUPTELEACEAE TETRACENTRACEAE TROCHODENDRACEAE DISANTHOIDEAE EXBUCKLANDIOIDEAE HAMAMELIDOIDEAE ALTINGIOIDEAE PLATANACEAE RHODOLEIOIDEAE EUPTELEACEAE TETRACENTRACEAE TROCHODENDRACEAE FIGURE 2. Adams consensus trees of the basal Hama- melidae with 5% randomly induced error; Tetracentra- steps). — C. trees (174 steps steps). equally parsimonious s). — D. Consensus tree of four trees (174 analysis, resulted in the evolution of the other four subfamilies with Disanthoideae and Rhodoleioideae being the latest-derived. While the consensus tree (Fig. 3) demonstrates the difficulty in resolving the node connecting Hamamelidoideae, Exbucklan- dioideae, and the Disanthoideae-Rhodoleioideae group, there is support for some of the relationships proposed by this cladogram. The position of Altin- gioideae as the earliest-derived subfamily of Hama- melidaceae is supported by 12 synapomorphies, which include features of phytochemistry, epider- mal anatomy, venation, petiole anatomy, ovule and seed morphology, and floral structure (Fig. 3). here is corroborative evidence from the fossil record for this phylogenetic scheme (Fig. 3). As mentioned previously, Platanaceae can be traced to the Lower Cretaceous, Altingioideae to the Up- per Cretaceous, and Hamamelidoideae to the Low- er Paleocene. Foliar megafossil records of Ex- bucklandioideae date from the Upper Oligocene (Lakhanpal, 1958) and Lower Miocene of Oregon (Brown, 1946) and the Middle Miocene of Wash- ington State (Brown, 1946). We could find no records of fossil leaves, flowers, or fruits of the Disanthoideae or Rhodoleioideae. However, Tiffney 1986) reviewed records of dispersed seeds of Rho- doleia Champ. from the Upper Еосепе (Mai & Walther, 1985) and Disanthus Maxim. from the 968 Annals of the Missouri Botanical Garden Lower Oligocene (Mai & Walther, 1978). These data suggest that Platanaceae may have been the earliest offshoot of a lineage that later gave rise to Altingioideae and, still later, Hamamelidoideae and other subfamilies of the Hamamelidaceae. Hickey € Wolfe's (1975) suggestion that “Сег- cidiphyllales" and Trochodendrales form a clade separate from the remainder of the subclass that was derived from a prehamamelid or hamamelid lineage is in need of elaboration. Although our data suggest that the Eupteleaceae might be better placed in the Trochodendrales than the Hamamelidales, they lend some support for the derivation of the Hamamelidales from the "*platanoids" proposed by Hickey & Wolfe (1975). The early occurrence of Platanaceae along with their cladistic proximity to Altingioideae and distance from Trochodendrales does not support inclusion of Trochodendrales in the lineage that led directly to the evolution of Platanaceae and Hamamelidaceae. he cladistic analysis of the Hamamelidaceae using the Platanaceae as an outgroup results in a cladogram that differs drastically from current con- ceptions of evolution in the family (Cronquist 1968, 1981; Takhtajan, 1969) but agrees overall with the fossil record. We are not suggesting that this reconstruction represents the one true phylogeny of Platanaceae and Hamamelidaceae. Yet, we pro- pose that the current practice of rooting the Hama- melidae in the Magnoliidae is in need of reevalu- ation. Paleobotanical (Dilcher, 1979; Retallack & Dilcher, 1981) and phytochemical data (Giannasi, 1986) suggest that Hamamelidae are at least as primitive as Magnoliidae. Examination of the Ha- mamelidae without assumptions as to their pre- sumed ancestors may facilitate our understanding of the subclass and of angiosperm evolution. LITERATURE CITED ADAMS, Е. N. Consensus techniques and the comparison of taxonomic trees. Syst. Zool. 21: 390- 397. BAILEY, І. W. € C. Nast. 1945. Morphology and relationships of А and Tetracentron, 1. stem, root, and leaf. J. Arnold Arbor. 26: 143-153. BLUME, К. 1928. Florae dee nec non Insularum Adjacentium, J. Frank, Brussels. ‚ А. 1970. Flor al morphology and vascular of the Hamamelidaceae: the apetalous gen- ra of Hamamelidoideae. J. Arnold Arbor. 51: 310- & C. T. Рнивкіск. 1980. A generic atlas of hamamelidaceous pollens. Contr. Gray Herb. 210: 99- iori р, L. E. 1930. The uii ud and anatomy the inflorescence and flower of the Platanaceae. er. J. Bot. 17: 678-6 Brown, К. . А cretaceous sweet gum. Bot. Gaz. pa 94: 611-615. Alterations in к fossil and living floras. 1 Wash. Acad. Sci. 36: 344-355. CHANG, С. Т. 1959. The pollen morphology of Liquid- ambar L. and Altingia Nor. Bot. Zurn. (Moscow & Leningrad) 44: 1375-1380. 1964. The pollen morphology of the i se Hamamelidaceae and Altingiaceae. Trudy Akad. Nauk SSSR, Ser. 4, Eksper. Bot. 13: 173- 32 CHESTERS, К. I., Е. В. Смлоск & М. К. HUGHES. 1967. Angiospermae. /n W. B. Harland et al. e The Fossil Record. Burlington House, Lon CRANE, P. R., E. M. Fris & К. В. iade 1986. Lower Cretaceous angiosperm flowers: fossil evidence Science 232: 85 on early CRONQUIS IST, А. 1968. The Evolution and Classification lowering Plants. Houghton Mifflin, Boston 1981. An Integrated System of Classification of Flowering Plants. Columbia Univ. Press, New Yor DAHLGREN, R. 1977. A commentary on a diagrammatic presentation of the angiosperms in relation to the distribution states. Pl. Syst. Evol., Suppl. 1: ее ‚ S. RosENDAL-JENSEN & B. J. NIELSEN. 1981. A revised classification of the angiosperms with com- Phytochemistry and Angiosperm Phylogeny. Praeger Publ., New York DiLcHER, D. L. Early angiosperm reproduction: an introductory report. Rev. Paleobot. Palynol. 27: 291-328. & L. ERIKSEN. 1983. Sycamores are ancient trees. Mus. W. Colorado Uu Spring 1983: 8- l ENDRESS, P. K. 1977. Evolutionary € in the ira melidales-Fagales-group. Pl. Sys . Evol, Suppl. 1 321-347. FELSENSTEIN, J. 1982. о methods for inferring evolutionary trees. Quart. Rev. Biol. 57: 379-404 GiaNNAsI, D. Е. 1986. оа aspects s of phy- m in Hamamelidae. Ann. Missouri Bot. Gard 73: UNS E E J. B. A. BouN & R. DaHLGREN. 1979. T ге n of flavonoids in the angiosperms. Hip Еј В. 967. "Comparative Biochemistry of the Flavonoids. Academic Press, New York. 0. HARMS, Н. Hamamehdavbae, ngler & K. Prantl, е ое А oni, nd edi- tion. Band 1 HAYNE, К. С. Ted wee von Storaxbaumen. Flora (Stuttgart) 13: 171-172. Hickey, L. J. & J. A. DOYLE. Ln Early Cretaceous fossil evidence for angiosperm evolution. Bot. Rev (Lancaster) 43: 3-104. . WOLFE. 1975. The bases of angiosperm phylogeny: vegetative morphology. Ann. Missouri Bot. 538-589. Gard. 62: HOLLICK, A. T The Tertiary floras of Alaska. U.S. Geol. Sur. Prof. Pap. 182: 1-185. Jav, M. 1968. Distribution des flavonoides chez les Hamamelidacées et familles affines. Taxon 12: 136- 147 Volume 78, Number 4 Schwarzwalder & Dilcher 969 1991 Systematic Placement of Platanaceae ЈНА, V RETALLACK, С. & D. Г. DILCHER. 1981. Arguments for 1977. e PT of the Hamamelida- ceae. J. Indian Bot. Soc. -48. KocH, B. E. 1963. Fossil s from the lower Paleo- cene of the Agatdalen (Angmartussut) area, central Nugssuaq peninsula, northwest Greenland. Meddel. renland 172: 1-120. KUTUZKINA, E. 1974. Platanaceae. Jn: А. Takhtajan (e zs uu срна Plants of the U.S.S.R., Vol- um Magnoliaceae- Eucommiaceae. Nauka, Len- ingra ^ LAKHANPAL, R. N. 8. The Rujada flora of west central Oregon. Dae Calif. Publ. Geol. Sci. 35: 1- LINDLEY, J. 1836. А jai sim System of Botany, 2nd edition. Longman, Lon Mar, D. H. & H. WALTHER. ` 1978. Die Floren der Haselbacher Serie im Weisselster-Becken (Bezirk Leipzig, DDR). vow Staatl. Mus. Mineral. Geol. Dresden 28: 1-20 & —. а Die зен lid des Weisselster-Beckens und seiner Randge Abh. Staatl. Mus. Mineral. Geol. Dresden 33: 1-260. MATHIESEN, F. J Notes on some fossil plants from east Greenland. Meddel. Grønland 85: 1-62 MELIKIAN, A. P Anatomical structure of the spermodermis in the members of the genera Liquid- ambar L. and Altingia Nor. in cm to their taxonomy. Izv. Biol. Nauki 24: 50- 732. Seed coat types of З ИИА d allied families in relation to their systematics. Bot. Zurn. (Moscow & Leningrad) 58: 350-359. 1973b. The anatomy of seed coat and sys- tematics of the family Hamamelidaceae. Izv. Biol. Nauki 26: 104-105. METCALFE, С. В. & L. CHALK. 1965. Anatomy of the Dicotyledons, 2nd edition. Oxford Univ. Press, Ox- ord. NAKAI, T. 1943. Ordines, Familiae, Tribi, Genera, Sec- tiones, Species, ара ates et Combinationes Novae. Imperial Univ., Мазт, C. С. & I. W. d! 1945. Morphology and relationships of Trochodendron and Tetracentron. II. Inflorescence, flower, and fruit. J. Arnold Arbor. 26: 267-275. Morphology of Euptelea and comparison with Trochodendron. J. Arnold Ar- bor. 27: 186-191. шуу J.S. 1898. The later extinct в 2 и епса. U.S. Сео]. Survey Mon. 35: 5. [А р ен work edited by А. Ho llic 1 PRAGLOWSKI, J. he pollen morphology of the Trochodendraceae, Tetracentraceae, Cercidiphylla- ceae and оне with a ence to taxonomy. Pollen & m 9-4 Rao, P. В. M. 1974. Seed a y in some Hama- Се сагта and а. Е 24: Као, B A. & 0. P. € sclereids in various Indian Acad. Sci., 1974. Typology of foliar a of Hamamelidaceae. Proc Sect. B, 79: 127-138 a Glossopterid ancestory of angiosperms. Paleobiol- ogy 7: 54- Конте, Е. J. 1982. Consensus indices for SU classifications. Math. Bioscience 59: 131-14 iiie “Tables, 2nd sper diy c Publ., New Yor SCHWARZWALDER, R. N. Systematics and Early Evolution of the S PHA Ph.D. a Indiana Univ., Bisexuality in the pistillate Ногой of Platanus occidentalis L. Proc. Indiana Acad. Sci. 90: 89-98. SEIGLER, D. 5. 1981 Terpenes and plant phylogeny. In D. А. Youn . S. Seigler (editors), Phyto- chemistry and Angiosperm Phylogeny. Praeger Publ., ew York. SkvonTsOvA, М. T. 1960. The anatomical structure of the conducting system of leaf petioles е са batt d Alt aceae Dokl. Akad. Nauk S.S.S.R. 133: 1231- 1234. Sworrorp, D. 1984. PAUP: Phylogenetic Analysis Us- ne Parsimony. Version 2.2. Illinois Natural History rvey, Champaign- Urbana, Illinois. s e ] хина, А. 1969. Flowering Plants, Origin Dispersal. Oliver & Boyd, Edinburgh. [Translated n . Jeffre M T. 943. Systematic anatomy of the wood of e Hamamelidaceae. Bull. Fan. Mem. Inst. Biol. 14: iria. B. H. 1986. Fruit and seed dispersal and the evolution of the Hamamelidae. Ann. Missouri Bot. ard. 73: 394-416 TIPPO, 0. 1938. Comparative anatomy of the Moraceae and their presumed allies. Bot. Gaz. (London) 100: -99. Орсноксн, С. К. 1984. Cuticle evolution in early Cre- taceous angiosperms from otomac Group of Virginia and Maryland. Ann. Missouri Bot. Gard. 71: -550. WILLIS, J. С. 66. A Dictionary of the Flowering Plants and Ferns, 7th edition. Cambridge Univ. Press, ambridge. WILSON P. 1905. Altingiaceae. In: М. Г. Britton & Г. Underwood (editors), Ln American Flora. New York Botanical Garden, Bronx, New York. WOLFE, J. A. 6. Tertiary slants from the Cook Inlet Region, Alaska. Prof. Pap. U.S. Geol. Sur. 398-B: -p32. 1973. Fossil forms of the Amentiferae. Brit- tonia 25: 334-355. ТАМАРА, М. 5. & D. Г. DILCHER. 1986. Comparative pollen morphology and its relationship to phylogeny of pollen in the ا‎ lidae. Ann. Missouri Bot. Gard. 73: 348-38 ANALISIS MULTIVARIADO Clara Inés Orozco* DEL COMPLEJO DE TRES ESPECIES DE BRUNELLIA (BRUNELLIACEAE)! RESUMEN Analisis de agrupamiento de e principales (АСР) y el árbol del valor minimo (PRIM) fueron usados para esclarecer la sistemática de tres taxa de Brunellia: B. antioquensis (Cuatrec.) Cuatrec., B. comocladifolia Humb. & Bonpl., y B. sibundoya Gee ue aleni estudiada la validez de las subespecies para los dos últimos taxa. Diez y siete Unidades Taxonómicas Operacionales (UTO) re PAD TU los taxa, y se muestrearon 32 caracteres i dA se en es posiblemente el factor importante en la ын зү а especifica. А nivel nn ЊЕ pie cl diagnosticos se semilla. para el reconocimiento a nivel паа а tienen que ver соп el nümero de lobulos del cáliz у el nümero de carpelos; la humedad es posiblemente el fac e mayor importancia en la variación floral. Los resultados del análisis fenético muestran que В. antioquensis с. a una subespecie de B. sibundoya, y los taxa infraespecificos B. comocladifolia subsp. boyacencis y B. sibundoya subsp. sebastopola son sinónimos de B. comoc cladifolia subsp. cundinamarcensi is y B. sibundoya subsp. о respectivamente. El estudio de la estructura de la exina bajo geográficas. eL con cuatro diferentes conibinscionel de caracteres y que corresponden a la distribución geográfica para Colombia de los cuatro taxa infraespecificos reconocidos bajo este estudio. El número cromosómico, — 28, es reportado por primera vez para B. comocladifolia subsp. cundinamarcensis y B. sibundoya subsp. sibundoya. ~ БЫ ABSTRACT Clustering analysis, Principal Component Analysis (PCA), and Minimum Пе Tree (PRIM) were used to define three taxa of Brunellia: B. antioquensis (Cuatrec.) Cuatrec., B. comocladifolia Humb. & Bonp . sibundoya Cuatrec. The subspecies were studied also. Seventeen OTUs represented the de taxa, ER 32 m orphological of six were recognized. The two spec les are localized at different кш кү characters at the specific се comocladifolia Humb. & Bonpl. subsp. cundinamarcensis Cuatrec. and B. sibundoya Cuatrec. subsp. sibundoya. ' Agradezco al Fondo de Investigaciones Cientificas y кү Especiales Francisco José de Pond COLCIENCIAS; a Jorge Victor Crisci, Maria Fernanda López Armengol de la Universidad Nacional de La Plata, a Vicki Funk Claudia Sobrevilla del Smithsonian Institution por las За ни и y criticas del manuscrito; a José Gina ses con quien discuti la sistemática de los taxa y la distribución de ellas; a Amy Jean Gilmartin mis especiales oum рог la prontitud en la lectura del manuscrito y sus invaluables criticas; a los profesores del Instituto de Cienc ; у Amsterdam рог las fotografias del grano de polen; а los _ | de los herbarios COL, MEDEL, MO, US, у VALLE; a los revisores del trabajo, José Cuatrecasas y Pau y por sus críticas y sugerencias. Parte de este trabajo fue presentado como requisito parcial para ced al titulo de Magister r en Sistemática Botánica, dirigido por Enrique Forero y Jorge V. Crisci a quienes doy tambien mi agradecimiento, ? Universidad Nacional de ы decim de Ciencias Naturales, Herbario Nacional Colombiano, Apartado Aereo 7495, Bogotá, Colom ANN. MISSOURI Вот. GARD. 78: 970-994. 1991. Volume 78, Number 4 1991 Orozco 971 Analisis Multivariado de Brunellia Brunellia Humb. & Bonpl. es el unico género de la familia Brunelliaceae, completamente neo- tropical y con una amplia distribución y especiación en los bosques Andinos de Colombia. En el estado actual del conocimiento de la familia, el género comprende dos secciones y subsecciones (Cuatrecasas, 197 85). De las 60 especies registradas ei para el neotrópico, 39 han sido reportadas para Colombia (Cuatrecasas, 1970, 1985; Orozco 1981, 1985, 1986) En la revisión que se adelanta para Brunelliaceae se encontró un complejo de tres especies muy semejantes morfológicamente con superposición e inconsistencia de caracteres. Los tres taxa se en- cuentran ubicados en la sección Brunellia Cuatrec. subsección Comocladifoliae Cuatrec. La subsec- ción esta definica por la presencia de hojas opuestas e imparipinadas, foliolos con 10-17 pares de ner- vios secundarios, margen doble o simple, dentada, con indumento patulo. Las flores son pequeñas, y el caliz de 2.5-7.0 mm en diámetro con 4-6 piezas. El endocarpo es cartilaginoso en formo de U, y el exocarpo de pelos derechos y rigidos. El actual estudio tuvo como objetivos: establecer los límites de variación y la validez de los tres taxa conocidos como Brunellia antioquensis (Cuatrec.) Cuatrec., Brunellia comocladifolia Humb. Вопр!., y Brunellia sibundoya Cuatrec.; recer la existencia de subespecies cuya limitación y escla- geográfica para algunas de ellas discrepa del con- cepto clásico de subespecie Mayr (1963). Otro de los objetivos consistió en el estudio de caracteres no revisados anteriormente, tales como polen cromosomas, con el propósito de conocer su utilidad y aplicación en la sistemática de la familia. Por otra parte, con este estudio se pretende aclarar los problemas encontrados por Cuatrecasas (1985) en cuanto a la determinación de las poblaciones de В. comocladifolia у B. sibundoya, y provenientes de Cundinamarca, señalando entre ellas una posible hibridación. Problemas similares a los planteados en este trabajo han sido resueltos por medio del análisis multivariado, el cual permite un conocimiento mas exacto de los caracteres. MATERIALES Y MÉTODOS Muestras de Brunellia antioquensis, Brunellia comocladifolic, y Brunellia sibundoya fueron co- leccionadas en la región Andina colombiana, tra- tando de cubrir la mayor parte de ella para registrar el rango total geográfico de distribución y la va- riacion completa de las especies. El material fue determinado con base en las claves, en la distri- bución geográfica, y las colecciones preexistentes en los herbarios de С ALLE, y MEDEL de- terminadas por Cuatrecasas. n la elección de las 17 Unidades Taxonómicas Operacionales (Tabla 1) se tuvo en cuenta que las mismas representaran la variación total de los taxa y cada UTO exhibiera el estado adulto de desarrollo (en fruto). Para algunos casos las UTO están con- formadas por varias muestras de un individuo y para otros casos muestras de dos o mas colecciones de varios individuos presentes en un mismo sitio. Esta última consideración fue hecha por la varia- ción sexual exhibida en estos taxa Un total de 32 caracteres РЕТ florales y vegetativos fueron registrados para cada UTO (Tabla 2) de los cuales 24 son cuantitativos con- tinuos, 3 cuantitativos discontinuos (caracteres 5, 6, y 8), 2 cualitativos doble estado (caracteres 16 y 23) y 3 cualitativos multiestado (caracteres 30, 31, y 32). Varios de estos caracteres habian sido usados como diagnósticos, otros fueron elegidos teniendo en cuenta si el estado del caracter diferia por lo menos en una UTO. Los caracteres de con- sistencia del foliolo, forma y base del foliolo, y nümero de nervios secundarios utilizados en previos trabajos para diferenciar los taxa, no fueron tenidos en cuenta en este estudio por la dificultad en definir con exactitud, para el caso del primer caracter, los estados coriáceo-subcoriáceo o cartáceo de mo- deradamente coriáceo; para el segundo y tercer caracter debido a la variabilidad intra-UTO; y para el cuarto por la ausencia de variación entre las . Los caracteres cualitativos multiestados fue- ron codificados con base al razonamiento de una secuencia lógica de variación del caracter. Para los caracteres cuantitativos continuos el valor pro- medio fue anotado directamente. La matriz básica de datos (Apéndice 1) fue so- metida a distintos análisis multivariados utilizando el programa NT-SYS de Rohlf et al. (1971). Dos análisis de agrupamiento, a partir de una ma- una matriz de correlación; un análisis de ordenamiento (ACP); el analisis del árbol de la minima distancia (PRIM); triz de similitud de distancia y de u y los arboles de consenso fueron los análisis prac- ticados en este trabajo (Fig. 1). En los análisis de agrupamiento, para el cálculo de la matriz de dis- tancia, se usó el coeficiente taxonómico (TD) de Sokal (1961) y para el de la matriz de correlación el momento producto de Pearson. Las técnicas de agrupamiento del método de grupo par (Sneath & Sokal, 1973), ligamiento promedio ponderado (WPMGA), ligamiento completo (LC), y ligamiento promedio no ponderado (UPMGA) fueron utilizadas 972 Annals of the Missouri Botanical Garden TABLA 1. Relación de las ОТО con sus respectivas determinaciones y referencias. Material de origen para los análisis de polen B y de cromosomas C. Cada UTO fue coleccionada en Colombia. Cauca, Municipio, Jamundi, Villa Colombia, cá, Muncipio La Victoria hacia Orozco, Lozano & Vidal ОТО 1 Brunellia comoc = subsp. comocladifolia. Valle del Cauca, Municipio de Dagua, Pichendé, 1,650 m, Orozco, Lozano & Vidal 1156 (COL) (6 especimenes). UTO 2 Brunellia comoc eda a subsp. comocladifolia. Cauca, Municipio de Popayán, Piendamó, 1,800 m, »zco, Vidal & Lozano 1160 (COL), B (5 especímenes). UTO 3 Brunellia comocladifolia subsp. comocladifolia. Valle del 1,700 m, Orozco, Vidal & Lozano 1228 (COL) (3 especimenes) ОТО 4 Brunellia comocladifolia subsp. cundinamarcensis Cundinamarca, Municipio de Yacopi, 1,730 m, Orozco, Franco & Lozano 1105 (COL), B, C (7 especimenes). ОТО 5 Brunellia comocladifolia subsp. cundinamarcensis. Cundinamarca, Municipio de La Palma, 1,680 m, Orozco, Franco & Lozano 1102 (COL), B (8 especimenes). UTO 6 Brunellia comocladifolia subsp. boyacensis. Boyacá, Municipio de La Victoria, Alto del Chapón, 1,800 m, Orozco, Sanabria & Sanchez 1258 (COL), C; departamento de Boya Quípama, 1,350 m, Orozco, Sanabria & Sanchez 1267 (COL), B (11 especimenes). UTO 7? Brunellia sibundoya subsp. sibundoya. Putumayo Valle de Sibundoy, 1,650 m, 1200, 1203 (COL); 1,650 m, Bristol 1247 (COL); 2,000 m, Schultes 3202 (COL) (7 especimenes). ОТО 8 Brunellia sibundoya subsp. sibundoya. Cauca, Municipio del Puracé, Moscopán, 2,230 m, Orozco, Lozano & Vidal 1162 (COL) (3 especimenes). ОТО 9 Brunellia sibundoya subsp. sibundoya. Cauca, Municipio del Tambo, 2,300 m, Orozco, Vidal & Lozano 1173 (COL), B (6 especimenes). UTO 10 Brunellia sibundoya subsp. sibundoya. Valle del Cauca, Municipio de Tulua, Barragán, 2,250 m, Orozco, Vidal & Lozano 1120 (COL) (6 especimenes). UTO 11 Brunellia sibundoya subsp. sebastopola. Cundinamarca, Municipio de Bojacá, 2,550 m, Orozco, Lozano & Franco 1091, 1092, 1093 (COL), B (17 especimenes). UTO 12 Brunellia е subsp. sebastopola. Cundinamarca, Municipio de San Francisco Sabaneta, 2,500 т, Orozco, Franco, Lozano & Barrera 1098 (COL), C (6 especimenes) UTO 13 Brunellia sibundoya subsp. sebastopola. Cundinamarca, Municipio de Zipacón, Estación de Sebastopol, 2,480 m, Orozco & Carbonó 1271, 1272, 1273 (COL) (13 especímenes). UTO 14 Brunellia antioquensis. Antioquia, Municipio de Medellín, Cerro del Boquerón, 2,420 m, Orozco, Tobón & Henao 1320 (COL), B (6 especimenes). UTO 15 Brunellia antioquensis. Antioquia, Municipio La Ceja, Las Lomitas, 2,170 m, Orozco, Tobon & Henao 1290 (COL) (4 especimenes). UTO 16 Brunellia antioquensis. Antioquia, Municipio de Medellín, Piedras Blancas, 2,250 m, Orozco, Tobón & Henao 1280 (COL) (5 especimenes). UTO 17 Brunellia antioquensis. Antioquia, Municipio de Río Negro, Sajonia, 2,290 m, Orozco, Tobón & Henao 1286 (COL), C (6 especimenes). para agrupar los datos de similitud de cada una de las matrices (Métodos 1, 2). rrelación cofenética fue calculado para cada uno El coeficiente de co- de los fenogramas y es el resultado de comparar la matriz de similitud, derivada del fenograma, con la matriz de similitud de la que se originaron. El uso de este coeficiente es un estimado de la forma como cada técnica de agrupamiento traduce los valores de similitud. El análisis de los componentes principales, Mé- todo 3, parte de una matriz de correlación entre caracteres, sin embargo, los resultados finales son las unidades taxonómicas operacionales represen- tadas en el espacio y cuyas dimensiones corres- ponderian al resumen de los caracteres en vectores o componentes principales. El análisis permite tam- bien conocer el valor sistemático de cada uno de los caracteres. El árbol de la minima distancia, Método 4, parte de una matriz de similitud de distancia entre UTO, y los resultados finales son las mínimas distancias fenéticas encontradas entre ellas; ver Sneath & Sokal (1973) y Crisci & Lopez Armengol (1983) para una mejor explicación de los procedimientos. Los árboles de consenso (Adams, 1972) se usa- ron para comparar los agrupamientos obtenidos en los fenogramas; en otras palabras vienen a ser un resumen de la manera como operan las técnicas de agrupamiento para cada una de las matrices de similitud. Volume 78, Number 4 Orozco 1991 Analisis Multivariado de Brunellia 973 Caracteres utilizados en los análisis taxonómicos numéricos, registrados para cada una de las 17 ОТО. TABLA 2. Codificacion de los caracteres cuantitativos discontinuos y cualitativos. Codifi Caracteres Estados cación 1. Longitud de la hoja (cm) = = 2. Longitud del pecíolo (cm) — — 3.* Longitud de los folíolos superiores (cm) = — 4. Ancho de los foliolos superiores = == 5.* Nümero de estípulas 2-3 estipulas 1 4-7 estipulas 2 6. Número de lóbulos del cáliz (3)4 1 (4)5(6) 2 5y6 3 7.* Diámetro cel cáliz (mm) — — 8. Numero de carpelos (3)4(5) | (4)5(6) 2 5y 6 3 9. Separación de los nervios (mm) — — 10.* Longitud del pedicelo (mm) — — 11. Grosor del pedicelo (mm) = 12. Longitud del pedúnculo (cm) — = 13. Longitud de la inflorescencia (cm) = = 14. Longitud де. foliculo (mm) — 15. Ancho del foliculo (mm) -— — 16. Indumento del foliculo Laxamente hirsuto 1 Densamente hirsuto 2 17. Ancho de la semilla (mm) — 18. Longitud de la semilla (mm) — — 19. Alto del endocarpo (mm) = = 20. Grosor де les ramas de primer grado de la inflorescencia (mm) = — 21. Grosor de les ramas de segundo grado de la inflorescencia (mm) — 22. Grosor de las ramas de tercer grado de la inflorescencia (mm) = = 23. Forma de la semilla Elíptica 1 Conica 2 24. Longitud de los foliolos inferiores (ст) — — 25. Ancho де lcs rie ea (cm) — — 26. Longitud de los filamentos (mm) = == 27. Longitud de las ramas dd primer grado de la inflorescencia (cm) == = 28. Longitud de las ramas de segundo grado de la inflorescencia (cm) — — 29. Longitud de las ramas de tercer grado de la inflorescencia (mm) — — 30.* Ресто ов Presente 1 Presente- Ausente 2 Ausente 3 31. Superficie del foliolo en estado adulto iso 1 Liso o rugoso 2 Rugoso 3 32.* Margen del foliolo Frecuentemente serrada, 1 recta-diente delgado Frecuentemente serrada, 2 ncava-diente delgado Frecuentemente serrada, 3 óncava-diente grueso Frecuentemente crenado 4 o serrado * Caracteres diagnósticos en previos trabajos. ( ) escaso Annals of the 974 Missouri Botanical Garden 'sooureurnu sismeue soj әр еотдојоројош eíiZojens, ү VANDA 2 OGOLIW | 0QOL3W n | t OGOLAW | ооцеџејоо | OSN3SNOO за $злоану | | OSN3SNO2 30 53708 УУ | upI2D[94402 эр әүиә!2|}әо2 | | ueÁnqi4jUOO spw V2113N34 VOIL3N33 | | уо!1змз4 V2113N34 V2113N34 NOI2V13uH могоулани | |могомланн NOI2V13uH 9J9/9DJD НЕО -02 30 зї -09 за зі | |-09 за ai -0930 зі T N31914309 -N31913309 | |-N31913309 -N31214302 -N31214302 IAI sajuauodwoy WA] sajuauodwosy ALO 21 MLO 21 n10 21 ALO Li 'lvNOIS 577 30 сут за АЕТ sv130 sv130 VWVH9ON34 VAVH SONIA КЕБ» ГЕР] ЕЕЕ VAWVH90N34 “-NSWIGIG = O1DVd 30 NOIDD3A08d VOWdN 010 - 3NOUd OL 7N3INVv9I1 VONdN ош -3WMONd OL 7N3INV51!?1 V9dW^A 010 ¬ атома OL 7N3INV911 213149409 OL 7N3INV9!I?1 D1 алатоо O | 7N3INV911 (а2у) sajodidurd se(ueuoduio2 soj ay V2113N34 V9MdM Old ¬ зона OL -N3IN V917 O.LN3INVN3Q [ | -Y0 30 SISIIVNV OLN3IWVdnuov OLN3IN VdfiH9v | за SVOINO3L за SV2INO3L | OWIN “IW NO IVA 330 08V 2920109 әдиә чогооја105 NLO eques up!2D|8110) | | NLO eque 012004510 GNLITIWIS 30 ZIHLVW ONLITIWIS 30 ZIYVLVA ONLIMINIS 30 71949 \ L | | | URS | за SISITUNV 39420202 X J949DJD) = М D9IU23J* Sepppiun | NLO X NLO =O рәшә; 'sepppiun | ошоо $ә1зӘу20102 soj DJedo әс ошо? NLO SD Dado as ЕК чоя | NOIOV ZIYVONVLSS | зајајоојој 2€ х NLO LI SOLVO за VOISV8 ZIM.IVN Volume 78, Number 4 1991 Orozco 975 Analisis Multivariado de Brunellia Muestras de granos de polen de las colecciones Orozco et al. 1087, 1092, 1105, 1158, 1160, 1173, 1227, 1257, 1320, 1375, y Bernal 626 corresponden a diferentes muestras de las pobla- ciones (Tabla 1). Las muestras fueron enviadas al laboratorio Hugo de Vries, Amsterdam. Se utilizó la técnica Erdtman (1969) y el microscopio de barrido para obtener los micropreparados y foto- grafias. Directamente de las fotografias se tomaron datos del ancho del muro, ancho del lumen, longitud y ancho del grano, longitud del colpo, intercolpo y longitud y ancho del poro; estos datos fueron tambien complementados con los de las prepara- ciones que estaban en buen estado. Los datos del grosor de la exina se tomaron directamente de las fotografias. Semillas de las poblaciones de las especies es- tudiadas, con excepción de las de Sibundoy, fueron colocadas en el germinador que consistia en una caja de petri con doble papel de filtro y agua destilada. El segundo papel de filtro fue previa- mente perforado en ocho huecos y en cada uno de ellos se colocaron las semillas. En estado de plántula fueron transplantadas en tierra y posteriormente llevadas al invernadero. Cuando las plantas estaban fuertes (6-7 meses de edad) se sometian a la poda de las puntas de la raíz. Este proceso se repetia cada 15 a 20 días para lograr raices gruesas y figuras mitóticas claras en donde se pudieran contar los cromosomas. Las puntas se trataron con HCI 0.1 N y Carnoy 3:1. RESULTADOS Y DISCUSIÓN CONSIDERACIONES GENERALES Los resultados de los análisis de agrupamiento con los dos matrices de similitud muestran dos grandes grupos de UTO: Brunellia comocladifolia (UTO 1-6) y B. sibundoya (UTO 7-17). Cada grupo está conformado por las mismas UTO a excepción de los grupos obtenidos en el fenograma derivado de la matriz de correlación y la técnica del ligamiento completo (LC). En todos los fenogramas se observa una mayor constancia de asociación (igual topología) para los subgrupos formados dentro del grupo de las UTO 1, 2, 3, 4, 5, y 5 (ver consensos Figs. 3-6) mientras que para el otro grupo, integrado por las restantes UTO la constancia de asociación entre las UTO es solamente observada en los fenogramas derivados de la matriz de correlación (Figs. 5, 6). La amplia variación dentro de este último conjunto sumado al efecto propio del coeficiente de distancia son las razones por las cuales cada uno de los subgrupos resultantes en los fenogramas de distancia varian en su totalidad en el contenido de las UTO (Figs. 3, 4). Aunque los coeficientes de correlación co- fenética son ligeramente superiores a los obtenidos en los fenogramas derivados de la matriz de co- rrelación, las verdaderas similitudes son enmas- caradas. Sin embargo, se aprecia en los fenogramas de distancia la asociación constante entre las UTO Ту 12, 8, 9, y Ту 14, 16. Е fenograma МРМСА derivado de la matriz de correlación (Fig. 2) fue escogido como el que refleja con mas exactitud las similitudes morfoló- gicas. Este agrupamiento de las UTO da un coe- ficiente de correlación cofenética (C.C.C.) de 0.7017. Los dos componentes principales del aná- lisis de los componentes principales (ACP) (Tabla 3) representan el 58.13% de la variación entre las UTO (Fig. 7). El árbol de la minima distancia (PRIM) está superpuesto sobre la figura del análisis de componentes principales; las minimas distancias estan representados por los vinculos de unión entre las UTO y por los cuales se explica el porqué de los agrupamientos obtenidos en los fenogramas de distancia. GRUPOS TAXONÓMICOS Los dos mayores grupos del análisis de agru- pamiento corresponden a las especies ВгипеШа comocladifolia (ОТО 1-6) y B. sibundoya (ОТО 7—17); ver la Figura 2. También se observan los dos grupos por el componente I del análisis de los componentes principales (Fig. 7). Existe una estrecha afinidad fenética entre las UTO 1, 2, 3 y 4, 5, 6, formando claramente dos subgrupos que corresponden, respectivamente, a las subespecies Brunellia comocladifolia subsp. comocladifolia y B. comocladifolia subsp. cun- dinamarcensis. En el análisis de los componentes principales por el componente II, tambien se re- conocen las dos subespecies (Fig. 7). La posibilidad de diferenciar a la UTO 6 como representante de una tercera subespecie boyacencis se ve bastante disminuida por cuanto existe una alta similitud fe- nética entre las UTO 4, 5, 6, y de ahi la constante asociación entre ellas (Figs. 3—6). Por otra parte, la similitud también fue observada en la estructura de la exina entre las subespecies cundinamarcensis y boyacensis (Fig. 9). La cercania geográfica y la ausencia de barreras geográficas (Fig. 14) explican la variación continua de las dos poblaciones y por ende el reconocimiento de un sólo taxon B. co- mocladifolia subsp. cundinamarcensis. Una marcada superposición fenética se observa entre las UTO que representan los taxa B. sibun- doya subsp. sibundoya (UTO 7, 8, 9) y B. si- 976 Annals of the Missouri Botanical Garden S20- 0 ъъ AYU O FIGURA 2. WPMGA C.C.C. 0.7017 bundoya subsp. sebastopola (UTO 11, 12, 13). La superposición fenética es medida por la fre- cuente asociación en nücleos entre UTO que pro- vienen de los diferentes grupos taxonómicos (Figs. 2-6). En el análisis de los componentes principales por el componente II, las UTO se muestran tambien relacionadas (Fig. 7). La estructura de la exina soporta estos resultados; no se encontró diferencia 0 ~ alguna еп la estructura de los dos taxa (Fig. 1 Los resultados muestran una se cie, B. sibundoya subsp. antioquensis, UTO 15, 16, 17, y la UTO 10 (Figs. 2, 5-7). La UTO 13, en los análisis de correlación está relacionada gunda subespe- || con В. sibundoya subsp. antioquensis pero рог el análisis de los componentes principales se encuen- cl'o- 00 ОО 090 TOON YOU BUN _ o N сл Fenograma де las 17 UTO basado en el coeficiente de correlación у la técnica de ligamiento ponderado tra mas asociada con B. sibundoya subsp. sibun- doya. En la Figura 7 (cuadrantes derecho, inferior y superior), las interconexiones debidas a las minimas distancias fenéticas entre las UTO de los dos taxa infraespecificos B. sibundoya son el resultado de mayor relación fenética entre las subespecies que la encontrada en B. comocladifolia; en la misma figura cuadrante izquierdo inferior y supe- rior se observa sólo una interconexión, entre las U y 6, por tratarse de una unidad especifica. Bajo diferente concepto taxonómico que el mos- de este trabajo, Cuatrecasas (1985) anota una estrecha afinidad entre B. si- bundoya y B. antioquensis. trado por los resultad Volume 78, Мигпбег 4 Orozco 977 1991 Analisis Multivariado de Brunellia TABLA 3. Eigen-valores. Porcentaje de traza. Acumulación de porcentaje. Análisis de los componentes principales (ACP). Valores propios Componentes principales Eigen-Valores % de Traza Acumulación de % Primer componente 14.185 44.33 44.33 Segundo componente 4.417 13.80 58.13 Tercer componente 3.527 10.44 68.57 Cuarto componente 2.267 6.71 75.28 Quinto componente 1.963 5.81 81.09 Sexto componente 1.518 4,49 85.58 Séptimo componente 1.334 3.95 89.53 Octavo componente 0.987 2.92 92.45 Noveno componente 0.696 2.06 94.51 Décimo componente 0.485 1.43 95.94 Décimo primer componente 0.398 0.84 96.78 Décimo segdo. componente 0.308 0.91 97.53 Décimo tercer componente 0.255 0.75 98.28 Décimo cuarto componente 0.180 0.53 98.81 Décimo quinto componente 0.153 0.45 99.26 Décimo sexto componente 0.118 0.34 99.60 Décimo séptimo componente 0.037 0.10 99.70 Décimo octavo componente 0.026 0.07 99.77 Décimo noveno componente 0.016 0.047 100.24 12 4563 7128 9 11 1416 IO I5 13 17 123456 7128 9 11 IO I3 17 14 16 15 124562 71213178 9114161005 | 23456 pé FIGURAS 3 y 4. Consenso entre las técnicas de agru- FIGURAS 5 y 6. Consenso entre las técnicas de agru- C pamiento basadas en el coeficiente de se cia.— 3. Con- pamiento а en el coeficiente de correlación. — 5. senso Estricto. — 4. Consenso de Mayoria. Consenso Estricto. — 6. Consenso de Mayoria. 978 Annals of the Missouri Botanical Garden ~ == | س کے – س ت ee‏ O ~ сл 58.13% = ч | у 2 ‚57 | a 6 |. 79 6 =„ 5 |. 35 5 « 4 1.28 | « 3 1.95 | „|4 1.97 14 =» 16 1.61 14 « 8 1,85 8 „ 9 1.97 16 , 7 1.92 Figura 7. Anális las UTO representan las distancias minimas entre e Los agrupamientos de distancia entre las UTO que integran las dos subespecies de Brunellia si- bundoya (Figs. 3, 4) asi como las interconexiones fenéticas son debidas a una amplia variación de los caracteres cuantitativos continuos. Asi, los extre- grupos taxonómicos infraespecificos, por ejemplo las mayores longitudes promedio de los foliolos su- periores (15.16-19.04 cm) son exhibidas por las UTO 10, 15, 17 (B. sibundoya subsp. antio- quensis) y las UTO 13, 12 (B. sibundoya subsp. sibundoya); las menores longitudes (9.69- 12.30 UTO 14, 18 (B. sibun- doya subsp. antioquensis) y las UTO 8, 7, 9, 11 (B. sibundoya subsp. sibundoya). En la Figura 7 cm) se encuentran en las ~ = ~ ————ÓÓ — — — ~ ‚55 1.83 1.85 1.92 13 . 10 2.02 17 « 15 2.31 isis de los componentes prisa ad (ACP) y el árbol del valor minimo (PRIM). Las líneas entre las estas UTO se encuentran unidas por las distancias fenéticas minimas. CARACTERES DISCRIMINANTES PARA EL RECONOCIMIENTO DE LOS GRUPOS TAXONÓMICOS En el análisis de los componentes principales Fig. 7), el componente I separa a los taxa Bru- nellia comocladifolia y Brunellia sibundoya. De los 32 caracteres morfológicos 3 d ~ e ellos (carac- teres 23, 10, 5) en el componente I son importantes para diferenciar las especies (Tabla 4). La forma de la semilla en B. comocladifolia es elíptica, el indumento sobre los foliculos es laxo y el número de estipulas no es superior a 3 (Figs. 16, 17). En Volume 78, Number 4 1991 Orozco 979 Analisis Multivariado de Brunellia contraposición, la semilla en B. sibundoya es có- nica, el indumento sobre los foliculos es denso, y el nümero de estipulas es mayor de 3 (Figs. 18, 19). L ос carac terec titat T | grosor de las ramificaciones de la inflorescencia ей teres 20, 21, 22) y el ancho de los foliolos supe- riores e inferiores (caracteres 4, 25), cuyos aportes al componente 1 son superiores a 0.8, no son muy ütiles como diagnósticos por cuanto hay superpo- sición de los limites superiores e inferiores. Sin embargo, en promedio el grosor de las ramifica- ciones y el ancho de los foliolos son menores en B. comocladifclia y pueden considerarse como tendencias. El ancho de los foliolos está relacionado con la forma lanceolada o angostamente oblonga observada con relativa frecuencia en B. comocla- difolia. De manera similar en B. sibundoya los caracteres de grosor de las ramificaciones y ancho de los foliolos pueden considerarse como tendencia hacia mas gruesos y anchos, respectivamente. El ancho de los foliolos tiene que ver con las formas ovada, largamente ovada, ampliamente eliptica o ampliamente oblonga de los foliolos; estados ob- servados con frecuencia en cada unidad taxonó- mica (UTO). Dos caracteres (6, 8) en el componente II son utiles para la diferenciación infraespecifica (Tabla ): B. comocladifolia subsp. comocladifolia ex- hibe 4 lóbulos del cáliz y 4 carpelos, en contra- posicion con la subespecie cundinamarcensis con 5 lobulos del cáliz у 5 carpelos. En B. sibundoya subsp. sibundoya se encuentra con mayor fre- cuencia 5 lobulos del cáliz y 5 carpelos, pero tam- bién se observó en algunas muestras hasta 6 lobulos del cáliz y 6 carpelos. Brunellia sibundoya subsp. ioquensis se reconoce por la frecuencia de 4 lóbulos del cáliz y 4 carpelos y con menos fre- cuencia 5 lóbulos del cáliz y 5 carpelos. Tres aspectos de los caracteres diagnósticos son importantes de mencionar en el momento: (1) los caracteres se encuentran correlacionados con las áreas o microáreas geográficas ocupadas por los taxa infraespecificos (Fig. 15); (2) existe una alta correlación (análisis de agrupamiento de caracte- res, técnica R) entre los caracteres diagnósticos tanto para la diferenciación especifica como in- fraespecifica (Orozco, 1989); y (3) el escaso nú- mero de caracteres diagnósticos está relacionado con la estrecha afinidad morfológica entre las dos especies. ESTRUCTURA DE LA EXINA No se encontraron diferencias significantes en el tamaño del grano de polen, longitud del colpo, TABLA 4. Analisis de los componentes principales. Aporte de cada uno de los caracteres a los componentes principales. Carac- Componentes teres 1 2 3 1 0.541 0.491 0.461 2 0.735 0.349 0.493 3 0.595 0.373 0.596 4 0.855 0.270 0.271 5 0.908 0.089 0.208 6 0.107 0.880 0.323 1 0.658 0.170 0.430 8 0.187 0.791 0.344 9 0.644 0.634 0.154 10 0.699 0.065 0.020 11 0.626 0.348 0.399 12 0.630 0.430 0.377 13 0.033 0.399 0.066 14 0.731 0.230 0.358 15 0.663 0.072 0.251 16 0.906 0.049 0.243 17 0.532 0.343 0.318 18 0.255 0.207 0.277 19 0.681 0.025 0.072 20 0.852 0.139 0.182 21 0.868 0.246 0.131 22 0.930 0.139 0.066 23 1.008 0.028 0.170 24 0.754 0.130 0.519 25 0.884 0.049 0.311 26 0.507 0.427 0.114 27 0.397 0.418 0.023 28 0.377 0.264 0.452 29 0.730 0.316 0.168 30 0.467 0.720 0.210 31 0.268 0.251 0.827 32 0.741 0.186 0.072 intercolpo, y longitud y ancho del poro. Las dife- rencias se observan a nivel de la estructura de la exina. Brunellia comocladifolia exhibe una estructura reticulada con tendencia en la subespecie como- cladifolia a lúmenes más amplios y muros conti- nuos y aplanados (Fig. 8). La subespecie cundi- namarcensis exhibe una exina reticulada irregular con lümenes menos amplios y muros parcialmente fracturados y realzados (Fig. 9). Se observa en B. sibundoya dos patrones di- ferentes de la estructura de la exina. Uno de ellos corresponde a la subespecie sibundoya con una estructura perforada rugulosa (Fig. 10); el otro patron a la subespecie antioquensis con una es- tructura reticulada irregular (Fig. 1 1) y muy similar a la observada en B. comocladifolia subsp. cun- dinamarcensis. 980 Annals of the Missouri Botanical Garden FIGURA 8. Forma y estructural del grano del polen de Breall comocladifolia Humb. & Вопр!. subsp. со- mocladifolia. a y b, Orozco et al. 1227; c, Orozco et al. 1158. Cuatrecasas (1985) se refiere también a las va- riaciones infraespecificas de la estructura de la exina encontradas por Marticorena (1970), bajo el microscopio de luz, en otras subespecies neotro- picales de B. comocladifolia. Muller (1979) y No- wicke & Skvarla (1979), quienes han observado las variaciones estructurales de la exina a nivel infraespecifico, atribuyen dic plio componente de adaptación del grano de polen, ha variación a un am- especialmente el de la superficie del tectum. Po- siblemente también sea el caso de la variación encontrada en Brunellia. CROMOSOMAS El número cromosómico 2n = 28 es reportado para Brunellia comocladifolia subsp. cundina- marcensis y B. sibundoya subsp. sibundoya. Para las dos otras subespecies no se reporta el nümero cromosomico debido a que las plantulas no tuvieron éxito. Los resultados presentados fueron obtenidos sin pretratamiento alguno (Figs. 12, 13). Las di- ferencias en el tamano де los cromosomas se deben a diferentes estadios metafásicos. Los datos del патего de cromosomas coinciden con los reportados por Ehrendorfer et al. (1984 para Brunellia mexicana Standl. y B. comocla- difolia subsp. funckiana (Tul.). Ehrendorfer et al. (1984) senalan la dificultad en obtener un cariotipo claro, dificultad tambien observada en este estudio por la escasa precisión en la morfologia de los cromosomas. DISTRIBUCIÓN GEOGRÁFICA Las dos subespecies de Brunellia comocladi- folia se encuentran en las selvas subandinas entre 1,000-1,8 registra a alturas mayores entre 2,000- 2,900 m, 00 m, mientras que В. sibundoya se ocupando las selvas andinas y subandinas. Brunellia comocladifolia subsp. comocladifo- lia no sólo se distribuye en el sur de Colombia y Ecuador (Cuatrecasas, 1985) sino que se extiende hacia el norte por la cordillera occidental hasta Antioquia. Brunellia comocladifolia subsp. cun- dinamarcensis se distribuye sobre la cordillera oriental y se reporta para el norte en el Huila, Cundinamarca y en Boyacá; posiblemente se ex- tiende más hacia el sur y más hacia el norte en los departamentos de Santander. Los datos dispo- nibles para el estudio permiten visualizar la sepa- ración geográfica entre las dos subespecies por la cordillera central (Fig. 14). Se infiere que la di- ferenciación infraespecifica pudo haber ocurrido en el sur de Colombia en los sitios geográficos que nd a la ramificación de los Andes co- lombiano Brunellia sibundoya subsp. sibundoya tiene una más amplia distribución (Fig. 14). Original- mente la subespecie fue reportada en el Nudo de los Pastos (Putumayo), en la cordillera central flan- co oriental, e interconexiones con la cordillera oc- cidental, Cauca (Cuatrecasas, 1970), sitios donde se registraron las menores alturas (2,000-2,300 m). Nuevas colecciones en el Tolima senalan el ascenso de la subespecie a lo largo de la vertiente Volume 78, Number 4 1991 Orozco 981 Analisis Multivariado de Brunellia FIGURA 9. Ferma y estructura del grano del polen de Brunellia comocladifolia Humb. & Bonpl. dinamarcensis. Cuatrec. a y b, Orozco et al. 1105; c y comocladifolia subsp. boyacensis. oriental de la cordillera central. Los resultados del actual estudio indican también la presencia del taxon en la cordillera oriental (Cundinamarca) y por los datos disponibles se infiere que la presencia de la subespecie en la cordillera oriental se deba más bien al ascenso por la misma desde el Nudo de los Pastos. Esta inferencia se apoya en el registro del taxon para el sur del Huila en los limites con el departamento del Cauca, cercano al lugar del nacimiento del ramal de la cordillera oriental y la separación entre la cordillera central y oriental (Rio Magdalena, Fig. 14) contraposición con los restantes subespecies los pis disponibles permiten reconocer el taxon B. sibundoya subsp. antioquensis en una área geográfica mas estrecha y con una amplia distri- bución hacia el norte de las cordilleras central y occidental. Е] taxon en previos trabajos fue regis- trado hacia el norte en la cordillera central (Antio- quia), sitios en donde la cordillera no tiene sus más altas elevaciones; nuevas colecciones indican su distribución en la cordillera occidental (Caldas y Antioquia) y más hacia el sur en la cordillera cen- tral-vertiente occidental (Valle del Cauca y Caldas). d, Orozco et al. subsp. cun- 67, colección muestreada como ВгипеШа La presencia del taxon en la cordillera occidental y en la cordillera central puede estar relacionada con la diferenciación infraespecifica en el sur de Colombia o, por la subespeciación mas hacia el norte, en la cordillera central. Esta ültima pudo haber sido favorecida por la disimetria climática en los dos flancos de la cordillera a causa de los efectos de la glaciación (Salomons, 1986). Bajo esta consideración, debe de haber existido una am- plia dispersión por su ocurrencia en la cordillera occidental. Las subespecies antioquensis y sibun- doya se encuentran en los sitios más elevados de la cordillera central y donde las vertientes son más acentuadas (Fig. Al parecer la diferenciación especifica está re- lacionada con el gradiente altitudinal, mientras que la infraespecifica lo está con la cantidad de hu- medad. Estos dos factores climáticos (Sarmiento, 1986) vienen a ser los factores de mayor impor- tancia ecológica variando a través y a lo largo de las cordilleras andinas. La humedad, que carac- terza a la cordillera occidental y vertiente occi- dental de la cordillera central, aunque debida a diferentes causas tiene que ver con la reducción 982 Annals of the Missouri Botanical Garden FicuRA 10. del nümero de partes florales: B. comocladifolia subsp. comocladifolia y B. sibundoya subsp. anti- oquensis tienen solamente cuatro lóbulos del cáliz y cuatro carpelos. Los datos de distribución permiten establecer cuatro áreas o microáreas geográficas libres de superposición y correlacionadas con cuatro dife- rentes combinaciones de caracteres morfológicos y de estructura de la exina (Fig. 15). TRATAMIENTO SISTEMÁTICO CLAVE PARA LAS ESPECIES Y SUBESPECIES DE BRUNELLIA ESTUDIADAS la. Estipulas 2 lel foliculo | hirsuto; didis a 1,000-1,800 m. óbul el cáliz y carpelos 4; cordillera occidental e ا‎ шава su a кашне central y occidental. ....................... comoc ladifolia рер, сотос о 2b. Lábulos del cáliz ^ ue iid 5; er tal mocladifl bsp. сипа пататс ensis lb. Estipulas 4-7; indumento del folículo densa- mente hirsuto; semilla cónica; 2,000-2, en m. За. Lóbulos del cáliz y carpelos 5-6; en el sur del Macizo colombiano e interconexiones Forma y estructura del grano de polen de Brunellia sibundoya Cuatrec. subsp. sibundoya. a Orozco et al. 1173; c y d, Orozco et al. 1087, colección muestreada como Brunellia sibundoya subsp. seba stopola. de cordilleras ат y central y vertiente oriental de la cordillera central . sibu ion € car nda : Lábulos del cáliz y carpelos 4(5); h el orte en la cordillera occidental y аке сасна) de la central 2b. B. sibundoya subsp. antioquensis о = 1. Вгипе а comocladifolia Humb. & Bonpl. P > 1- "ri pl. 5 . 1825; RERUM Fl. Neotropica, 2: 63- = D Supplement 43-46. 1985 Arbol 5-12 m. Estipulas 2-3. Hojas compuestas opuestas, 12.4-73.2 cm de largo, imparipinadas 1-7 yugos, 3.15(-18) foliolos; pecíolo 3.3- cm de largo. Foliolos peciolulados о sésiles, los superiores 3.6–16.0 cm de largo, (1.4-)2.0- 7.4 cm de ancho, oblongo-elipticos, elipticos, lanceo- lados; base del foliolo obtuso o cuneado. Foliolos inferiores ovados (1.7-)2.5-10.5 cm de largo (1.0-)1.6-5.0 cm de ancho; márgen de los foliolos crenado-serrada, serrada o márgen cóncavo-serra- > c „а E. = o 5 > da; nervios secundarios prominentes por el envés; (6-)9-18(-20) pares, distantes entre si (1.5-)3.0- , Моште 78, Митбег 4 1991 18 mm; indumento sobre los nervios de pelos lar- gos, flexibles. Superficie de los foliolos lisa y menos frecuente rugosa. Inflorescencia tirsoide, 6.3-30.0 cm de largo; ramificaciones delgadas, en cimas dicasiales, isotónicas, ramas de primer grado 1.2- de grueso, de segundo grado 1.0-2.0(- 3.0) mm de grueso, de tercer grado de 0.8-1.5(2.0) mm de grueso. Pedünculo 1.5-12.5 cm de largo. Flores hermafroditas o unisexuales, pediceladas en fruto, en flores femeninas y con menos frecuencia n flores masculinas; pedicelos 2.0-3.5 mm de lareo, 0.2-0.5(--1.0) mm de grueso. Cáliz pentá- mero o tetrámero, diámetro del cáliz en fruto (3. 5. M..0-5.5(-6.0) mm. Estambres 10, filamentos (1.5-)2.0-3.0 ram de largo. Carpelos 4 o 5. Fo- liculos 4 o 5, con indumento laxamente hirsuto, ovados, 2.2-3.0 mm de largo, 2.0-3.0 mm de ancho. Епдосагро seco en forma де U, 2.0-2.7(- 2.9) mm de alto. Semilla 1 desarrollada por folículo, elíptica, 1.8-2.2 mm de largo, 1.5-2.0 mm de ancho la. Brunellia comocladifolia Humb. & Bonpl. subsp. comocladifolia Fl. Neotropica 2: 66– 67. 1970; Supplement 43-45. 1985. TIPO: Colombia. Cauca: Popayan, 1906, Humboldt & Bonpland s.n. (ћојопро, P; isótipo, P n.v.). Figura 16. Arbol 5-12 m. Foliolos peciolulados 1.0-4.0 mm de largo; foliolos inferiores de una misma rama de menor tamano comparados con los superiores; margen de los foliolos en su máximo desarrollo tendiente a ser cóncava y serrada. Cáliz tetrámero. Carpelos 4. Foliculos 4 con muy escasa frecuencia en menor número. Granos de polen con exina re- gularmente reticulada. Distribución. Brunellia сотос!аа оба subsp. comocladifolia tiene una amplia distribución desde el sur (Ecuador), y a lo largo de de la cordillera occidental en Colombia, hasta el departamento de Antioquia. También al sur en la intersección entre la cordillera central y la occidental. Ejemplares representativos. COLOMBIA. ANTIOQUIA: Municipio La Concordia, via Urrao, 1,750 m, 3 dic. 1984 (fr), Orozco, Tobón & Henao 1341 ein RISARALDA: Municipio de Pueblo rico, carretera al cerro Montezu 1,350 m, 5 1986 (1 8), Bernal, Саган по, Tobón & Henao 1101 (COL). VALLE DEL CAUCA: cordillera oc- cidental, hoya del rio Calima, El Cairo entre Darién y Mediacanoa, 1,650-1,750 m, 7 ene. 1943 (fr), Cuatre- casas 13867 (COL); Municipio de Dagua, Pichindé, 1,650 m, 22 sep. 1983 (fr), Orozco, Vidal & Lozano 1156 (COL). Orozco 983 Analisis Multivariado de Brunellia a b C FIGURA 11. de Brunellia sibundoya subsp. aan Cuatre Forma y estructura del grano de қ Orozco et al. 1320; b y c, Orozco et al. 2 muestreadas como Brunellia о Й li 1b. Brunellia comocladifolia Humb. 4 Bonpl. muco, 1,000 m, dic. 1855 (fl), Triana 3720 (holótipo, W; isótipos, BM, C, US, W). Figura ET. oe ig Mgr Humb. & Bonpl. subsp. bo- sis Cuatrec. Fl. Neotropica 2: 67-68. 1970. TIPO: E Colombia. Boyacá: Monte del Chapón, 1,230 Annals of the Missouri Botanical Garden Ficura 12. m, 21 jul. 1932 pr жо 347 (ћојопро, МУ; isótipos, BM, Е, С ‚ МО, 5, U, US). Arbol 6-10 m. Foliolos sésiles o muy cortamente peciolulados, 0-1.5 mm de largo, base obtusa o cuneada, margen serrada o crenado-serrada. Ló- bulos del cáliz 5. Carpelos 5. Granos de polen con exina irregularmente reticulada. Distribución. Cordillera oriental en los de- partamentos de Cundinamarca, Boyacá y en Huila. Ficura 13. Cromosomas metafásicos de Brunellia sibundoya subsp. sibundoya. Existe en COL un especimen de la colección Little 7253 procedente de Huila que parece más bien corresponder por los datos de altura, 2,000-2,030 m, a B. sibundoya subsp. sibundoya. Ejemplares representativos. COLOMBIA. BOYA Municipio La Victoria, Alto del e 29 jul. 1984 (fo), rozco, Sanabria & S 6 (COL); hacia la que- brada de la Colichona, vereda San Martin, 1,500 m, 29 jul. 1984 (fr), Orozco, Sanabria & Sanchez 1265 (COL); hacia Quipama, 1,350 m, 30 jul. 1984 (fl, 8), Orozco, F Cromosomas premetafásicos de Brunellia comocladifolia subsp. cundinamarcensis. Volume 78, Number 4 Orozco 985 1991 Analisis Multivariado de Brunellia 6°00' - 4°00' Va Пе ceras | A e. ыу 2 Pm ا‎ | - I7UTOS) =. m ни "x Ф B. comocladifolia subsp. comocladifolla 1000 - 1800 Ó B. comocladifolia subsp. cundinamarcensis 2000-2750 O B. Sibundoya subsp. antioquensis ? 000 - 2.800 ө B. sibundoya subsp. sibundoya Es Y "X Nude: dels. MM | astás У ext: S. de И Region andina desde los 1000 m. 2 = 0 Ficura 14. Distribución ка де Brunellia comocladifolia Humb. & Bonpl. Brunellia sibundoya Cuatrec. Los nümeros se refieren a las Sanabria & Sanchez 1267 (COL). CUNDINAMARCA: Ми- 2. Brunellia sibundoya Cuatrec. Revista Acad. nicipio de la Palma, 1,730 m, 30 mayo 1983, den 0, Colomb. Ci. Exact. 5: 35. 1942: Cuatrec. Fl Lozano & Franco 1102 (COL); Sasaima, 1,600 - | - 70 _ . к И 5 епе. 1973 (8, 8), Сагста Ваггіва 20379 (COL). m HUILA: Neotropica 2: 78-79. 1970; Supplement 46 Rio Venadito, 25 km SE de la Bodega, 1,780 m, 2 dic 48. 1985 1944 (fr), Little 9024 (COL). Nombre vulgar. Tambor, en Cundinamarca у Arbol 6.0-20 m. Estipulas 4-7. Hojas com- Boyacá. puestas opuestas, (9.0-)20-74.3(-83.8) cm de 986 Annals of the Missouri Botanical Garden 78°00' -6*00' NV WEA OCEANO PACIFICO уы ын - 4900' ON \ \\ \ МА» : NS \ => : metros [——4À TAa (C 1000 - 1950 B.comocladifolia subsp. comocladifolia х V end | E nO 7 1000 - 1.800 SS 8.comocladifolia subsp. cundinamar – + + к EH censis ECUADOR ш \ 7 2000-2150 È B.sibundoya subsp. antioquensis X X | | way! 2 000 - 2.800 E B.sibundoya subsp. sibundoya L Rx 1 1 FIGURA 15. Distribución geográfica y caracteres diagnósticos de las especies de Brunellia estudiadas. largo, imparipinadas, 1-7 yugos, 3-15 folíolos. cia en tirso o en dicasio, 6.5-31.0 cm de largo; Peciolo 3.0-19.5 ст de largo. Foliolos sésiles o ramificaciones gruesas, en dicasios de una casi per- peciolulados, los superiores 6.0-28.0 ст de largo, fecta isotonia, ramas de primer grado, (1.5-)(2.0— (2.5-)3.3-10.6 cm de ancho, ovados, ovado- 4.2) (-5.5) mm de grueso, de segundo grado oblongos, elipticos u oblongo-elípticos; base obtusa (1.0-)2.0-3.5 mm de grueso, de tercer grado 1.2- o cuneada; margen de los foliolos crenado-serrada 3.0(-4.0) mm de grueso. Pedünculo 1.5-18.0 cm o cóncavo-serrada; nervios secundarios prominen- long. Flores hermafroditas o unisexuales, pedice- tes por el envés, (6-)9-22 pares, se apartan entre ladas en fruto о en flores femeninas, sésiles en sí (2.0)3.0-21.0 mm. Foliolos inferiores ovados; flores masculinas, pedicelos 1.0-2.0(- 2.5) mm de 4.3-17.0 ст de largo, 1.8-9.9 cm de ancho. largo, (0.8-(1.0-1.5)(- 2.0) mm de grueso. Caliz Superficie de los folíolos lisa o rugosa. Inflorescen- pentamero о tetrámero (hexamero), diametro en ~ Volume 78, Number 4 Orozco 987 1991 Analisis Multivariado de Brunellia brio Ies Y) FIGURA 16. Brunellia comocladifolia Humb. & Bonpl. subsp. comocladifolia. A. Variación en forma y tamario de los foliolos.— Aa-Ad. Folíolos superiores.— Ae, Af. Foliolos inferiores. — B. Detalle de las estipulas y yema.— Margen del foliolo. — Da. Inflorescencia completa. — Db-Dd. Ramificación de la inflorescencia. — E. Partes de la flor y fruto. — Ea. Folículos y estambres. — Eb. Nümero de lóbulos del cáliz. — Ec. Semilla. — Ed. Flor masculina. A (a, d, f), B, € (b), D (a-d), y E (a-c) Orozco et al. 1156; A y Orozco et al. 1160; A (c), Duque- Jaramillo 3893; A (e) y С (a), Cuatrecasas 13857; E (d), Orozco et al. 11 fruto (5.0—)5.5- 7.0(- 7.5) mm. Estambres 10, fi- ancho. Endocarpo seco en forma de U, (2.0-)2.5- lamentos (2.0—)2.8-5.0 mm de largo. Foliculos 4 — 3.0(-4.0) mm de alto. Semilla 1 desarrollada por o 5(6), con indumento denso e hirsuto, ovados, folículo, cónica, (1.2-)2.0-2.5(-3.0) mm de largo, 2.8—4.0(-4.2) mm de long, (2.2-)3.0-4.0 mm de ` (1.2-)1.8-2.0(- 2.2) mm de ancho. 988 Annals of the Missouri Botanical Garden | Ca Ficura 17. Brunellia comocladifolia Humb. & Bonpl. subsp. cundinamarcensis Cuatrec.— A. Forma, tamano de los folíolos y estado del peci de la inflorescencia. D. Partes de la flor y fruto. — Da. Foliculos iolulo. — B. Margen del foliolo. — Са. Inflorescencia completa. — Cb, Cc. Ramificación . Fo m 2 C (a-c), y D (a, b), = et al. 1102; А (c, d), Garcia: Barriga 20379; A (e, f), y D (c), Orozco et ul ; B, Orozco et al. 12 2a. Brunellia sibundoya Cuatrec. subsp. si- bundoya Fl. Neotropica 2: 80. 1970; Sup- plement 46. 1985. про: Colombia. Comisaría del Putumayo: Cuenca alta del río Putumayo en el Valle de Sibundoy, 2,200 m, | ene. 1941 (fr), Cuatrecasas 11570 (ћојопро, US; isotipos, F, US). Figura 18. Brunellia sibundoya Cuatrec. subsp. sebastopola. Fl. entre Sebastopol y Alto de las Escaleras, 2,300- 2,400 m, 21 таг. 1942 (fl, 6), Cuatrecasas 13586 (holotipo, US; isotipos, COL, F). Arbol 8-20 m. Foliolos sésiles, generalmente Volume 78, Number 4 Orozco 1991 Analisis Multivariado de Brunellia / > а | o Е РЕГ ООН ОБН f AL, 3 | 3mm / FIGURA 18. Brunellia sibundoya Cuatrec. subsp. sibundoya. — А. Forma y tamaño de los folíolos. — B. Detalle de las estipulas.—— C. Margen del foliolo.— Da, Db. E. Partes de la for y fruto. —Ea. Койсшоз. — Eb. А А : : masculina. A (a), D (a, b), y E (a-c), Orozco et al. 1273; A (b), Cuatrecasas 11570; А (c), Schultes 3202; А (d), Orozco et al. 239; C (a), Vidal 113; C (b), Idrobo et al. 10325; C (c) y E (d), Cuatrecasas 13586; D (c), Bristol 990 Annals of the Missouri Botanical Garden Da ү АУ n." Ficura 19. Brunellia sibundoya ab subsp. antioquensis Cuatrec.— А. Forma, їатайо de los foliolos y estado del pecidlulo (a, foliolo terminal). —B. Detalle de las estipulas.—C. Detalle de la margen del fo liolo. — Da. Inflorescencia completa. — Db. Ramificación ea la inflorescencia. — E. Partes de la flor y fruto. — Ea. Folículos. — Eb. Volume 78, Number 4 1991 Orozco 991 Analisis Multivariado de Brunellia crenado-serrados, con menos frecuencia dentados. Caliz 5-6 lóbulos. Carpelos 5-6. Foliculos todos desarrollados. Granos de polen con exina perforada rugulosa. Distribución. Se registra desde el Macizo co- lombiana hasta Cundinamarca en la cordillera oriental y para el Tolima en la cordillera central. Ejemplares representativos. COLOMBIA. CUNDINA- MARCA: Municipio de San Francisco, Vereda las Minas, 2,600 m, may. 1981 (fr), Orozco & Barrera 239 (COL); Municipio de Zipacón, Sebastopol, 2,480 m, 20 nov 1984 (fr), Orozco & Carbonó 1273 (COL); cerros de los alrededores de Bogota 19 jul. 1963 (fr), Bristol 1247 (COL); 2,250 m, 12 feb. 1942 (fr), Schultes 3202 (COL El isótipo estudiado de Brunellia sibundoya subsp. sebastopola es un especimen de flores mas- culinas y foliolos delgados. La condición sésil ob- servada en previos trabajos y utilizada como ca- racter diagnóstico para diferenciar las subspecies sebastopola y sibundoya es constante en las in- florescencias masculinas. La condición delgada de los foliolos utilizada también como caracter diag- nóstico representa solo un estado de la variación. 2b. Brunellia sibundoya subsp. antioquen- sis Cuatrec. Fl. Neotropica 2: 85. 1970. Bru- nellia antioquensis (Cuatrec.) бай. FI. 1985. TIPO: Colombia. Antioquia: Cerca de Santa Elena, entre Medellin y Rio Negro, 2,300-2,500 m, (8, fr), Gutierrez & Deslisle 1126 (holótipo, F; isotipo, F). Figura 19. Neotropica 2; Supplement 48. Arbol 6-15 m. Foliolos sésiles о cortamente peciolulados, crenado-serrados o fuertemente den- tados. Cáliz 4 lóbulos, menos frecuente 5. Carpelos 4, raro 5. Foliculos todos desarrollados. Granos de polen con exina irregularmente reticulada. Distribución. Se extiende hacia el norte de Colombia en las cordilleras occidental y central (vertiente occidental). Se registra en los departa- mentos del Valle del Cauca, Caldas, y Antioquia. Ejemplares representativos. COLOMBIA. ANTIOQUIA: Municipio de Giraldo, carretera a Urabá, Boquerón de Toyo, 2,100 m, 19 mayo 1983 (fr), Bernal 626 (COL); Municipio de Medellín, cerro del Boquerón, 2,420 m, 29 nov. 1984 (fr), Orozco, Tobón & Henao 1320 (COL); Municipio de Río Negro, Sajonia, 2,290 m, 27 nov. 1984 (fr), Orozco, Tobón & Henao 1286 (COL); Municipio de Santa Fé de Antioquia, Canasgordas, Boquerón de Toyo, 2,000 m, 18 feb. 1985 (fl, 8), Henderson & Bernal 145 (COL); Municipio de Urrao, Alto de Caicedo, 2,270 m 4 dic. 1984 (fl, 8, fr), Orozco, Tobón & Henao 1355 (COL). CALDAS: San Antonio de Padua, 2,430 m, 6 oct. 1981 (fl), Orozco, Rivera, Torres, Pinto & Lozano 259 OL). © CONCLUSIONES De los tres taxa considerados al iniciar el estudio sólo dos especies, Brunellia comocladifolia y Bru- nellia sibundoya, se reconocen por los análisis de agrupamiento, componentes principales (ACP) y el árbol de la minima distancia (PRIM). Los dos taxa son el resultado de los estudios de 17 Unidades Taxonómicas Operacionales y 32 caracteres mor- fologicos. Los caracteres diagnosticos para reconocer las dos especies son en escaso numero. Sin embargo, los taxa son fenéticamente distintos con base en la similitud total de los 32 caracteres. Los caracteres cuantitativos continuos como el ancho de los folio- los, directamente relacionado con la forma (Orozco, 1989), y el grosor de las ramificaciones de las inflorescencias pueden ser considerados como ten- dencias en el reconocimiento de las especies, mas no como caracteres diagnósticos por presentar su- perposición de los limites de variación. Las diferentes interconexiones fenéticas (PRIM) observadas entre las unidades taxonómicas que in- tegran В. sibundoya, у los agrupamientos de dis- tancia entre diferentes miembros de las dos subes- pecies son atribuidas a la variación de los caracteres cuantitativos continuos (24 en total). Posiblemente, la amplia distribución del taxon está relacionada con la plasticidad fenótipica de estos caracteres. Para B. comocladifolia se registra una mayor ho- mogeneidad medida por la constante asociación entre las UTO que integran cada una de las su especies; en otras palabras siempre se encuentran las mismas UTO conformando los grupos infraes- pecificos, por otra parte en el PRIM se encuentra sólo una interconexión entre miembros de los dos grupos infraespecificos. Mientras que la temperatura en relación con el gradiente altitudinal es un factor determinante para ma de la semilla. Pis Flor con estambres y Ee. poco йен е E (c), Henderson et al. 14 or : 25 y E (а, b), О o et al. 1286; A (b) y — Ed. Foliculos y estambres. A (а), С А (d), Orozco la al. 1289; B, Orozco 186; C (a), Benmi et al. 626; D (b), dan et al. 1290; E (d), ipia et al. 132 992 Annals of the Missouri Botanical Garden la diferenciación específica, la humedad es posi- blemente para el caso de la variación infraespecifica el factor predominante. Los taxa infraespecificos con la tendencia hacia el menor nümero de partes florales están distribuidos en las zonas de mayor humedad, occidental de la cordillera central. r los datos de distribución se infiere que la diferenciación infraespecifica esta muy estrecha- la cordillera occidental y la vertiente mente relacionada con la bifurcación de los ramales de la cordillera de los Andes en el Sur de Colombia. Sin embargo, la subespeciación para el caso de В. sibundoya se registró hacia el norte de la cordillera central. LITERATURA CITADA ADAMS, Е. М. 1972. Consensus techniques and the comparison of taxonomic trees. Syst. Zool. 21: 390- 397. Crisci, J. У. & M. Е. Lopez ARMENGOL. 1983. Intro- ducción a la Teoria y Práctica de la taxonomia nu- mérica. Serie de biología: Monogr. 26. Organización de los Estados Americanos, о D.C. CUATRECASAS, J. 1970. Brunelliaceae. Fl. Neotrop. onogr. p 1985. Brunelliaceae. Fl. Neotrop. Monogr. 2. [Supplement. ] EHRENDORFER, , W. Morawetz & J. 84. he neotropical angiosperm и аи ls and Caryocaraceae: first ошеа data af- finities. Pl. Syst. Evol. 145: 183-191. ERDTMAN, С. 1969. Handbook of ibo Hafner, New MARTICORENA, C. 1970. Pollen morphology. Pp. 27- 32 in J. Cuatrecasas, Brunelliaceae. Fl. Neotrop. Monogr. 2. Mayr, E. 1963. Especies Animales y Evolucion. Edi- ciones Univ. Chile, Ariel, Chile. MULLER, Ј. 1979. Form and function E есе аш ollen. Ann. Missouri Bot. Gard. 66: 32. NOWICKE, J. & J. SkvaRLa. 1979. Pollen Ка the potential influence in higher order systematics. Ann. Missouri Bot. Gard. 66: 633-700. Onozco, C. I. . Contribución al ir de las Bruneliaceas de Colombia — I. Mutisia 50: | Dos especies nuevas de B deed de (Colonia. Pp. 20-24 m Cuatrecasas, Fl. Neotrop. Е т [Supplem 1986. ши ап al estudio de Brunellia- ceae е o. III. Caldasia 15: 176-185. 989. Correlación de caracteres para algunas espe cies de Brunellia. Revista Acad. Colomb. Ci. Exact. 17(65): 357-365. PEARSON, К. 1926. On the coefficient of racial likeness. Biometrika 18: 105-117. Вонте, F., J. KisuPAUGH & D. Kirk. 1971. NT-SYS. Numerical taxonomy system of multivariate statis- tical programs. Techn. Rep. State Univ. New York, Stony Brook, New York. SALOMONS, J. B. 1986. Pp. 157-171 in Paleoecology of Volcanic Soils in the Colombian Central Cordillera (Parque Nacional Natural de los Nevados). J. Cramer, Berlin & Stuttgart. SARMIENTO, C. 1986. Ecological features of climate in high tropical mountains in high altitude tropical bio- geography. Pp. 11-45 in Vuillemier & Monasterio (editors), High Altitude е Biogeography. Ох- ford Univ. Press, New Y SNEATH, P. & R. SOKAL. miu Nu merical Тахопоту: The Principles and Practice of Numerical Classifi- cation. Freeman, San Francisco. SoKaL, К. Distance as a measure of taxonomic similarity. Syst. Zool. 10: 70-79. Volume 78, Number 4 Orozco 993 Analisis Multivariado de Brunellia APÉNDICE I. Matriz básica de datos. Información obtenida para cada Unidad Taxonómica Operacional (ОТО). | г 3 4 5 8 9 10 1 12 16 17 26 83| 29.70 46.30| 34.15 "um 2 00 = 31.93) 51.22 | | 4.70 7.06 6.17 7.69) 5. 84| 7.16 | 9.06) 9.59|14.97, 909 11.11 12. зо) 7.29 A 8.61 13.00 22. з 2292 31.52 24.50/35. a ав B4 T | | |!® 68| 9.98| 9.09|12.01|13.60| 12.30| 12.30) 11.67 |198.04 | I1. 51 | 15.16 "m 11.53 | 16. 20 9.90 15.24 | | 4 4.80 2.41) 3.84| 3.16| 420| 425 6.44| 4.77) 4.6! 6.90 4.44| 5.23, 6.71 3.79 ме 4616.22 5 | | | | | i ? 2 2 2 2 2 2 2 2 | 2 | о 6 | 1 | 2 2 2 2 2 2 | 3 2 2 | 1 1 Ф | Ф © 7 5.8 4| 4.83 5.22 4.27 4.03) 3.93 6.14 542| 645 612 540| 5.75| 5.87| 6.59 520 5.93 6.6! 9 | о | о 9 8.58| 6.35 8.18 - 7.01 | 760 8.50 712| 7,62|10.78| 7.21| 798| 9.83, 8.41) 9.43 В.84 9.05 10 2.55 2.48| 2.67 12.00 2.30 1.81 2.24| 1.28) 131 1.26| 1.43| 1.76) 1,65, 1.42 0.98| 1.38 1.30 11 .56| .36| .86| .80. .50| .56| 1.18 1.14 1.24| .88| ,B8| 1.07 N.C ‚95 .65| .95 .83 12 3.23| 2.80| 456| 2.91| 4,98 | 4.94 4.68 425, 5.08 B.15| 423 453| 6.37, 3.25| 5.95| 405, 9.87 13 16.2:'5| 10.53 [28,62 | 10,36 16.47 18,30 13.84 | 11. 87 |28.65 19.00 13.20 13.73, 1913|12,06/24 50/22.66| 9.05 | 14 2.85) 2.90| 300| 270 2.75, 2.78 3.56) 3.03} 3.61; 3.02) 3.19) 344) 361 3.11) 2.89) 3.61 3.33 N.C = No comparable UTO | 2 3 4 5 6 7 8 9 10 " 12 13 14 15 16 17 | 15 2.58| 2.66 287| 2.68! 267 2.55 336 2.50 308, 2.70 293 3.27 иа 278) 208 ЖАТ 3.82 | | | | | le | | 16 | | | | | | | 2 г г г г N.C] 2 |2 2 2 | 2 о 5 | Ф | = 2! 1.63| 1.20| 1.20| 1.69) 1.54] 1.72, 277! 2.05 2.04| 172, 2.03| 2.97, 2.40, 1.93| 2.25, 2.24| 270 о | 5 | о 22 87 92, 1.01| | . .00| 1.25| 1.35| 2.07, 2.00, 1.80; 1.62) 1.55) 2.30| 1.99) 1.34] 1.91| 1.63) 2.25 24 4.97| 3.60| 2.63, 6.16! 8.90, 8.67 | 11.21 8.50, 8.61) 1412) 7.27) 11.74) 11.99 7.27, 745) 5.92! 944 25 2.57, 1.65| 1.96| 2.63, 372, 3.27) 5.67| 3.93 3.90, 6.12| 3.79| 4.84| 570, 335| 3.62, 378| 3.26 26 342, 2.90 249 247 уг | 2.90 3.77) N.C | 3.16] 3.08| 509| 2.81| 430, 302| N.C | N.C |N.C 2T 2.311 1.15 2.91/ 2.58 1.41) 2.58| 268| 3.0, 243, 3.12| 1.15| 3.14 | 2.30, 178| 3.68| 2.70| 2.56 28 1.70} 1.30| 1.18| 1.84| 1.20) 1.66) 2.04| 1.55, 1.12| 1.16| 1.08) 145| 1.87] 199) 2.27) 2.621! 2.15 Annals of the 994 Missouri Botanical Garden UTO | 2 3 4 5 6 7 8 9 10 T 12 13 14 15 16 17 29 1.05; .85| 1.36| 1.09 n 1.29! 178| 1.81 1 35| 148] 1.10] 1.81! 1.70] 1.58] 2,71 146| 1.86 Y || Ф Ф 30 | | 3 2 2 3 E 3 2 3 3 3 2 2 г 2 о о 5 3| 2 2 г \ | | 3 |? 3 | | | ! 3 3 3 3 о 32 2 2 2 | | 4 4 4 4 4 4 4 4 3 5 3 3 HYBRIDITY AND PARENTAGE Ching-! Peng? and Yung-Kuan Chen? OF BEGONIA BUIMONTANA YAMAMOTO (BEGONIACEAE) FROM TAIWAN! ABSTRACT Begonia buimontana Yamamoto has been considered an endemic species with a restricted distribution at elevations between 1,000 and 1,600 m in southern Taiwan. Studies of morphology, flowering habit, pollen stainability, and meiotic chromosome behavior suggest a hybrid origin. Ws nn das comparisons, distribution patterns, chromosome cytology, and experimental hybridization show that B. buimontana consists of F, hybrids between B. palmata D. Don and B. taiwaniana Hayata. The hybrid nature and meiotic abnormalities in B. buimontana may account for its sterility and explain in part its rarity in nature. А description and line drawing of this hybrid are given. The genus Begonia is represented by ten species distribution at elevations between 1,000 and 1,600 in Taiwan (Liu & Lai, 1977; Lai, 1979, 1990; т in southern Taiwan (Fig. 1). We were initially Liu & Ou, 1982; Peng et al., 1988; Peng & Chen, puzzled by not finding any opened staminate flowers 1990). In our systematic revision of Begonia of іп herbarium specimens of this species, although Taiwan, our attention was drawn to a peculiar habit opened pistillate flowers are common. In our field of B. buimontana, an endemic with a restricted trips to the type locality of B. buimontana, we FIGURES 1 & 2. Distribution maps for Taiwan. — 1. Begonia buimontana. — 2. Begonia taiwaniana (stars), and B. ds oen (dots). о ir part by a research E from the Academia Sinica, Taipei. We thank Luc Brouillet for helpful comments on the manuscript and Wen-Pen Leu for technical assistance Тае. of Botany, Academia Sinica, а Taipei, Там п 11529, Republic of China. ANN. Mes Bor. GARD. 78: 995-1001. 1991. 996 Annals of the Missouri Botanical Garden Volume 78, Number 4 1991 Peng & Chen 997 Hybridity and Parentage of Begonia buimontana FIGURE 5. Meotic chromosome spreads of Begonia buimontana. — disoriented bivalents, univalents, and multivalent associations. All from Chen 671 (HAST). Bar equals 10 џ observed many live plants. Like all other species of Begonia in Taiwan, B. buimontana is androg- ynous and produces staminate flowers prior to pis- tillate flowers. All plants of B. buimontana at an- thesis produced abundant flowers of both sexes. АП staminate flowers observed developed normally up to their late bud stage and dropped off before they were open. By contrast, pistillate flowers were al- ways fully open at anthesis, and bud drop did not occur. Plants brought back from the wild and cul- tivated in the experimental greenhouse behaved similar Although staminate flowers of B. buimontana were shed prematurely, the plants appeared to set abundant fruits. We suspected initially that, in the field, on some plants staminate flowers may open normally to make fruit set possible. An examination . Metaphase I, showing vA often of the pollen fertility of B. buimontana estimated by determining the percentage of stainable pollen using the malachite green-acid fuchsin-orange stain of Alexander (1969), however, revealed near- ly complete pollen abortion (Fig. amination of mature capsules of B. buimontana revealed that, although the fruits were well formed, they contained very few plump seeds (Fig. 4). Cytological studies showed that B. buimontana has a somatic chromosome number of 2n — 30. Chromosome configurations at meiotic metaphase I typically consisted of some sticky, often disori- ented bivalents and up to at least 11 univalents; . Careful ex- multivalents are often present (Fig. e aberrant phenomena led us to suspect a hybrid origin in Begonia buimontana. In our search for putative parents, B. palmata D. Don and B. E E 3. Aberrant and aborted pollen of Begonia buimontana (A) vs. mostly normal and stainable pollen of IGUR other species of Begonia, e.g., taiwaniana (B). Bar eq FIGURE uals 20 um. A portion of the seeds from a capsule of B. buimontana: all but one aborted. Bar equals 0.5 mm. 998 Annals of the Missouri Botanical Garden palmata ÉS. B. buimontana B. taiwaniana FIGURE 6. Comparison of Begonia buimontana with putative peso» — А. Leaf. — B. Leaf, cross section. — C. Petiole and hair t taiwaniana Hayata, the only species that com- monly associate with B. buimontana, came to mind. Begonia palmata and B. taiwaniana are wide- spread and sympatric in the mountainous areas of central and southern Taiwan (Fig. 2). An analysis of vegetative and floral characteristics showed that B. buimontana is indeed intermediate between B. Table 1). Furthermore, cytological data showed that B. pal- palmata and В. taiwaniana (Fig. 6; mata has n — and В. taiwaniana has n = 19; a somatic chromosome number of 2n = 30 would be expected in F, hybrids of the two species (Fig. 7) In order to substantiate our hypothesis that B. buimontana is derived from natural hybridization between В. palmata and B. taiwaniana, we crossed the putative parents and obtained artificial F, hy- brid plants. These plants closely resembled В. bui- montana in aspect and morphology. Furthermore, as in wild plants of B. buimontana, they she type. — D. Fruit, cross section, showing placentation staminate flowers precociously. They have a so- matic chromosome number o — 30 and neg- ligible stainable pollen, too. In conclusion, we consider that B. buimontana from Taiwan represents a hybrid, which, as a result of discrepancy in chromosome number and/or dis- harmony in genomes of the parental species, is greenhouse but failed to set fruit. B mature fruits with a few seeds were often observed in plants of B. buimontana in the field. These were probably derived from back crosses with the pu- tative parental species. Some of these seeds ger- minated in the experimental greenhouse but died at the cotyledonous stage. Although both putative parents are widespread and sympatric in the mountainous areas of central and southern Taiwan, (Fig. 2), B. buimontana is Volume 78, Number 4 Peng & Chen 999 1991 Hybridity and Parentage of Begonia buimontana TABLE l. Comparison of Begonia buimontana with putative parents. Characters B. palmata B. buimontana B. taiwaniana Creeping rhizome present lacking lacking Plant height (cm) to 100 to 200 Leaf width : length ratio 1:1-1.5 1:2-2.5 E Pubescence Densely scabrous Scabrous Sparingly scabrous to nearly glabrous Number of styles 2 3 Number of ovary cells 2 (2-)3 3 Length of abaxial wing of capsule (cm) ca. 2 ca. 1 ca. | Chromosome number п = 11 2n = 30 n = 19 known only from the type locality (Yamamoto, 1933) and three other stations (Fig. 1). Recently, Ying (1988) described Begonia fenchihuensis. Àn examination of his original description and type specimens reveals that it is synonymous with В. buimontana. The type locality of B. fenchihuensis represents the northernmost station known for B. uimontana. Їп this study we collected B. bui- montana from two additional localities (Fig. 1 and see below). Although B. buimontana is relatively rare on Taiwan, plants of this hybrid were fairly abundant where found, frequently intermixing with both pu- tative parents, at least in the three localities we visited. The physiological characteristics and eco- logical tolerances of the putative parents may have recombined in the hybrids in such a way as to limit its distribution range. egonia buimontana is most likely represented only by F, hybrids. Populations are very uniform morphologically, and experimental F, hybrids are comparable to wild plants. Although such hybrids are sterile, once established, they are able to persist in a given location due to their perennial habit. The expansion of their populations on the island, however, can опу be achieved by recurrent natural hybridizations. The taxonomic treatment and an illustration of this hybrid species are given below. Begonia x buimontana Yamamoto (pro sp.), J. Soc. Trop. Agric. 5: 353. 1933. Liu & Lai in Li et al., Fl. Taiwan 3: 792. 1977. TYPE: Taiwan. Pingtung County (“Hsien”): Wuwei- shan. Matuda 322 (holotype, TAI). Figure 8. Tune i gr e ы syn. nov. Color Ill. Fl. Tai- 8. TYPE: Taiwan. Chiayi Count UA js n E Shihcho (as Shi-Chao), elev. 1,400 m, 16 Jan. 1988, S. S. Ying s.n. (holotype, Herb., Dept. Forest., Eo Taiwan Univ.; isotype, Herb., Dept. Forest., Natl. Taiwan Univ.) Erect, perennial, succulent herb; horizontal rhi- zomes lacking. Stems glabrous or sometimes re- motely pilose, 40-125 cm high, to Stipules glabrous, caducous, ovate to broadly ovate, mm diam. to 23 mm long, 18 mm wide, the apex acuminate, the margins entire. Leaves densely scabrous or hirsute on both surfaces, obliquely lanceolate to narrowly ovate, 8-22 cm long, 4-8 cm wide, the acuminate to cuspidate, the base obliquely cordate, the margins irregularly denticulate or ser- rulate; venation palmate, the veins 5-8, sometimes B. palmata B. taiwaniana 2n = 22 2n = 38 B. buimontana B. palmata 2n = 30 | | ка B. taiwaniana 2n= 30 FIGURE 7. Camera lucida drawings of somatic chro- mosome im. of B. palmata, B. taiwaniana, B. bui- montana, and an naga hybrid, B. palmata x B. taiwaniana. Bar equals 10 um. 1000 Annals of the Missouri Botanical Garden FIGURE 8. Begonia x buimontana. — А. Habit. — B. Bract. — C. Stipule. — D. Stamen. — E. Pistillate flower. — F. Staminate flower, from a bud opened and spread with : forcep.— 6G, С’. Style. — H. Fruit, external morphology. — . Fruit, cross section, showing placentation. — J. Seed. Volume 78, Number 4 Peng & Chen 1001 Hybridity and Parentage of Begonia buimontana villous; petioles rusty villose, 7-16 cm long, to 5 Additional specimens examined. TAIWAN. КАО- mm diam. Bracts in pairs, thickly papery, glabrous, j HR narrowly ovate to ovate, to 24 mm long, 17 mm wide, the apex mucronate to acuminate, the margins entire. Inflorescence 6-11 cm long; peduncles erect to pendulous, glabrous or some- times pilose, to 7 cm long, 2 mm diam. Flowers pink, abaxial surfaces of the outer tepals densely scabrous. Staminate flower: tepals 4, never open, the outer two obovate to orbicular, to 14 mm long, 13 mm wide, the inner two oblanceolate to narrowly obovate, to 13 mm long, 6 mm wide; stamens numerous, to ca. 140, some of which may be reduced, yellow, golf-club- the anthers nar- rowly obovoid, m long, 0.9-1.2 mm across, the ee 1.5- as mm long. Pistillate flower: tepals 5(-6), unequal to subequal, oblan- ceolate to orbicular, the largest 11—19 mm long, 9-18 mm wide, the smallest 9-17 mm long, 4-9 mm wide; styles 2-3, yellow, 4.7-5.3 mm long, at their base fused ca. 1.4 mm long, each bifid; ovary inferior, ellipsoid, (2—)3-locular, longitudi- nally shallowly grooved between the locules 3-winged; placentae axile, bilamellate. Infructes- сепсе to 17 mm long, the fruit-bearing stipes 18- 34 mm long; the abaxial wing triangular, 12-19 mm long, 11-16 mm wide, the lateral wings much narrower, 4—6 mm long, 10-14 mm wide. Well. formed seeds few, never exceeding 35 per mature fruit, broadly ellipsoid, 0.35-0.38 mm long, 0.19- mm thick, the apex rotund, the seed lip nipple- shaped, the micropylar end constricted. Distribution. Southern part of the Central Mountain Range of Taiwan, at elevations between 1,000 and 1,600 m. Known from Pingtung, Kao- hsiung, and Chiayi counties. HSIUNG CO.: Tengchi, Chen & "t 363 (HAST), Chen 726 (HAST). PINGTUNG CO.: mountain-hiking entrance of Peitawushan, Peng 10617, 13240 (HAST), Chen 644, 662 (HAST), Lai 9819 (TAI); en route from Peitawushan to Taiwu, Chen 671, 672 (HAST Begonia X buimontana consistently has 5(-6) tepals in pistillate flowers. The report of Liu & Lai (1977) of two pistillate tepals in this hybrid species was incorrect. In Taiwan, the only species of Be- gonia with two tepals in pistillate flowers is В. ravenii Peng & Chen (Peng et. al., 1988 LITERATURE CITED ALEXANDER, M. P. 1969. Differential PPS eain and nonaborted pollen. Stain Technol. 4 -122. Lar, M. J. 1979. Critical studies on some Ps from Taiwan. Taiwania 24: 35-37. Lar, M. J Begonia tarokoensis Lai, species from eastern Taiwan. Landscape perrea 1977. Begoniaceae. Pp. 791- 798 in H. L. Li, T. S. Liu, T. C. Huang, T. Koyama 8 C. E. DeVol (editors), Flora of Taiwan, Volume 3. Epoch Publ. Co., Taipei. & 1982. Contributions to the " dicotyledonous plants of Taiwan (VII). Bull. Exp. Forest. Natl. Chung Hsing Univ. 4: 1-1 Penc, C. I. € Y. К. CHEN. 1990. Begonia austro- taiwanensis (Begoniaceae), a new species from south- rn Taiwan. J. Arnold Arbor. 71: 567-574. H EN. 1988. Begonia ravenii (Begoniaceae), a new species from Taiwan. Bot. Bull. Acad. Sin. – 222. YAMAMOTO, Y. E Observation ad Floram cuam sanam VIII. J. Soc. Trop. Agric. 5: 346-3 YING, S. S. 1988. Coloured Illustrated Flora of MR with the Introduced Plants. [Published by the author, ipei.] CARAIPA (GUTTIFERAE) DEL PERU! Rodolfo Vásquez M? RESUMEN € ciones recientes de Сагагра de la Amazonia peruana elevan el nümero de especies conocidas para el pais de p jaramilloi y C. utilis son especies nuevas, y ésta última es la madera de construcción mas es à Ваен еп Iquito ABSTRACT Recent collections of Caraipa from Amazonian Peru raise the number of s pecies known from that country from three to eight. Caraipa jaramilloi and C. utilis are new to science, the latter ide the most important construction timber in Iquitos. Caraipa es un genero casi desconocido para el Рега debido en mayor parte a la falta de colec- ciones en la región amazónica. El presente trabajo reporta ocho especies de Caraipa de las cuales tres están en el estudio de Kubitzki (1978), y cinco son nuevas para el Perü, incluyendo dos hasta ahora inéditas. Las ocho especies ahora conocidas para el pais están aqui distinguidas por una clave, y solo se describen las especies nuevas; he obser- vado todas las especies menos C. m yrcioides en el campo, y las descripciones incluyen carac teristicas de la planta viva además de las de herbario, algunas notas de ecologia, nombres vernaculares, y usos de las especies también estan incluidos. Las especies peruanas de Caraipa se caracteri- zan por falta de aletas; fuste cilindrico; corteza externa generalmente blanca parduzca con escazo ritidoma en placas (en C. utilis marrón oscuro o negro); corteza interna rosada o rojiza de textura ligeramente fibrosa, con látex amarillento viscoso o amarillo fluido, u oleoso transparente; y xilema diferenciado semiduro, semipesado con albura ro- sada, duramen rojizo, textura media, y grano recto. Caraipa habita casi todos los tipos de bosques primarios de la Amazonia peruana; pero cada es- pecie está ecológicamente restringida a un solo tipo de suelo, un buen ejemplo de la 8-diversidad que caracteriza a muchas plantas lenosas de la Ama- zonia (Gentry, 1981, 1986). Dos especies pueden vivir juntas como C. utilis y C. tereticaulis que habitan suelos infértiles de arena blanca, o como tambien C. jaramilloi y C. myrcioides que habitan suelos arcillosos-arenosos o arcillosos y relativa- mente fértiles. Por lo general Caraipa tiene madera resistente al ataque de insectos y hongos xilófagos, razón por la que los fustes de casi todas las especies se usan como madera redonda para la construcción de vi- viendas, tanto en las áreas rurales como en las ciudades, llegándose а comerc ializar en Iquitos, bajo ésta modalidad a: C. utilis, C. tereticaulis, y C. densifolia subsp. densifolia а 1987). Cerca del 70% de la madera redonda que se usa para construcción de viviendas en la zona de Iquitos viene de la primera especie (Реги, 1986). También es importante destacar el gran valor ecológico de C. utilis y C. tereticaulis como pro- tectores de los suelos de arena blanca. Las caracteristicas del género, respectivas si- nonimias de las especies, distribución, y otros usos no se presentan, porque están bien detallados en Kubitzki (1978). > DE ESPECIES Y SUBESPECIES DE CAR AIPA RE PARA EL PERU la. сх и las hojas pubescente, con células epi- T s protuberantes, buliformes; cápsulas Оле о Е trigonas-piramiformes, si- metric as o ligeramente curvadas. Pelos estrellados sésiles y pedunculados; -28(-40) cm de иш cápsulas пика (2.5 ~> on s о con dimi- nates Fels estrellados muy dispersos, has- n de longitud. 3a. ној eliptico-ovadas u oblongas, 5- 12.5 cm de longitud, ápices redon- deados u fees: y frecuentemente ' Realizado, dentro del и de actividades del ae to Flora del Perú con la supervision del A. Gentry. ? Proyecto Flora del Perú, Apartado 280, Iquitos, Pert ANN. Missouni Вот. GARD. 78: 1002-1008. 1991. Volume 78, Number 4 1991 Vásquez M. 1003 Caraipa del Perü iid ur oi cápsulas 1.1-1.4 c ongitud C. is dh 3b. Hojas е lanceoladas о lanceo das, (10-)14-21 cm de longitud à ápi- ces А y frecuentemente acumi- nados; cápsulas 1.6-2.4 cm longitu C. punctulata lb. Envéz de las hojas glabro o glabre scente, sin células epidérmicas protuberantes; cápsulas to- mentosas, trigonas-oblicuas, trigonas- subglo- bosas, o trigonas-ovoides, asimétricas. 4a. Filotaxia en espiral, hojas coriáceas, я ticas ovadas а ovadas, 5-8(-10) ст de longitud; cápsulas trigonas-oblicuas o tri- gonas-subglobosas 1.5-2.0 cm de longitud C. utilis 4b. Filotaxia distica, hojas cartáceas o subcori- gonas-oblicuas, trigonas subglobosas, o tri- gonas-ovoides. Hojas 13-18(-20) cm de longitud, venación terciaria subparalela c gonas-subglobosas 3.0 cm de longitud C. valioi 5b. Hojas 5.5-14 cm de longi tud, vena- 6a. у ие vistosa а 1.7-2.0 ст de trigonas-ovoides, longitu сы eii ini pens UM беоне poco vis ›осаз flores, venación terciaria reticulada; с cápsulas trigonas-obli- 8 [ej > Ed de las hojas obtusas, pe- iolos 3-4 mm de longitud C. jaramilloi Paniculas con las ramas la- ~] = las hojas agudos, pecíolos 5- 10 mm de longitud . C. myrcioides EsPECIES NUEVAS DE CARAIPA REPORTADAS DEL PERU Caraipa jaramilloi Vásquez, sp. nov. TIPO: Perú. Loreto: Provincia de Maynas, Distrito de Iqui- tos, Puerto Almendras, bosque primario no undable, Vásquez 6528 (holotipo, AMAZ; isotipos, MO, USM, Е, С). Figuras 1, 3. Arbor, 20 m alta. Foliis lanceolatis vel ovato lanceolatis, be pilis parvis stellatis mayoribusque simplibus praeditis, ex- ocarpio ex endocarpio secedente Arbol mediano 20 m de alto, fuste con látex amarillento muy viscoso de lenta segregación, cer- ca del cambium ligeramente rosado. Ramitas pu- bescentes o glabrescentes. Hojas disticas, subcoriá- ceas, haz glabro, envéz glabro y liso, sin células epidérmicas protuberantes; lanceoladas а оуадо- 2-4 cm de ancho, ápice agudo, generalmente atenuado o ter- lanceoladas, 6-12 cm de longitud, minando en acumen de 5-10 mm de longitud, base obtusa; nervadura media con escaza pubescencia en el envéz, impresa en la haz, y ка еп el envéz, nervios secundarios 10-12(-14) pares, terciarios reticulados inconspicuos; márgen plano o ligeramente revoluto; puntos translucidos no vi- sibles o dificilmente visibles; peciolo glabro, raro pubescente, 3-4 mm de longitud. Inflorescencia terminal o lateral, paniculas de 6-9 cm de longitud, raquis, ramitas, pedicelos, y brácteas tomentosas ferrugineas con pelos simples; brácteas y bractéolas caducas, lanceoladas; bractéolas 1.5-2.0 mm de longitud; pedicelos 4-6 mm de longitud. Yemas orales antes de la antesis globosas 4-6 mm de longitud. Sépalos unidos en la base, con los lóbulos triangulares, 0.8-1.0 mm de longitud por 1.0-1.3 mm de ancho, con los bordes ciliados, tomentosos por fuera, y pubescentes por dentro. Pétalos oblon- gos con un borde ciliado, tomentosos por fuera y glabros por dentro, 5-8 mm de longitud, por 3.5 mm de ancho. Ovario densamente pubescente hir- suto, 1 mm de longitud, por 1.7 mm diàm., estilo 1 mm de longitud. Cápsula trigona-oblicua, curva, contraida en el ápice, 1.5-1.8 ст de longitud, hasta 1.2 cm de ancho en la base, superficie ligera- mente rugosa, puberulenta, con pelos = hirsutos caducos de 0.3 mm de longitud; exocarpo separable del endocarpo; 1-2 semillas por fruto. Paratipo. PERU. LORETO: Provincia de Maynas, Dis- trito de TN Mosis e bosque primario no ‚ае ble, Grández 682 (AMAZ). Ecología. Bosque primario no inundable, con suelos arcillosos-arenosos o arcillosos, relativa- mente ricos en nutrientes; florece en noviembre y frutifica en mayo. Localmente se conoce Nombre vernacular. con el nombre brea caspi de altura. 1004 Annals of the Missouri Botanical Garden F 1cm | 5mm G E à SS 1cm | NUR 5mm FIGUR Caraipa jaramilloi ae ar —A. Habito. [Vasquez 6528 (AMAZ, MO, USM, F, G).]—B. Inflores- cencia. ms 682 (AM Ж . Fruto visto рог un costado y por la base respectivamente. — E. Yema floral. —F. Pétalo. —G. Ova Usos. Los fustes se usan como madera redon- шр utilis Vásquez, sp. nov. TIPO: Peru. Lore- da, para la construcc ion de viviendas rusticas. Caraipa jaramilloi difiere de las otras especies por hojas con venación reticulada inconspicua, ba- ses obtusas, y peciolos 3-4 mm de longitud y por inflorescencias poco vistosas. Las cápsulas trigonas oblicuas, asimétricas, y tomentosas son similares a las de C. densifolia subsp. densifolia pero menos robustas, más contraidas, y con pelos simples dis- látex amarillento también se persos; el VISC OSO, asemeja al de ésta subespecie. : Provincia de Maynas, Distrito de Iquitos, Р, uerto Almendras, bosque primario no тип- dable, con suelos de arena blanca, Vásquez 5513 (holotipo, AMAZ; isotipos, MO, USM, F, С). Figuras 2, З Arbor 25 m alta. Foliis spiralibus, ovatis vel rd ө ovatis, coriaceis, glabris, 5-8(-10) ст longis, 2.3-4 cm latis; pe stolis: glabris 6-10 mm longis. dsl на terminales vel laterales, paniculatae 6 ст longis, bracteis bracteolisque caducis; pedicellis 5- 10 mm longis, glabris. Volume 78, Number 4 1991 Vásquez M. Caraipa del Perü 1005 FicURA ks Caraipa utilis Vásquez. — A. Hà C. Pétalo (x 2). — D. Estambre (x12). — E. Ovario (x6). — tivamente. е 6144 (MO, AMAZ, USM).]— prs trigonis obliquos, ferrugineo мо: 1.5-2.0 п longis, exocarpio ex endocarpio seceden Arbol mediano a grande, 25 m de alto, fuste negruzco con ritidoma en placas pequenas, latex transparente oleoso con puntos blancos. Ramitas glabras, en jóvenes las ramitas puberulentas. Hojas en espiral, coriáceas, haz glabro, envéz glabro y liso, sin células epidérmicas protuberantes; ovadas a elipticas-ovadas, 5-8(-10) ст de longitud, 2.3— 4.2 cm de ancho, ápice redondeado o con pequeno acümen obtuso de 3 mm long, rara vez emarginado, base aguda o a veces obtusa; nervadura media Q Ay f 7mm у | 2mm bito. [Vásquez 5543 (AMAZ, MO, USM, Е, G).]— B. Flor (x 2).— tos vistos por un costado y por la base respec- К Н. Нојаз de diable juveniles. algunas veces escazamente pubérulas en el envéz, impresa en la haz y prominente en el envéz, se- cundarias 5-7(-9) pares, terciarias subparalelas, visibles en ambos lados; márgen ligeramente re- voluto; puntos translúcidos visibles solo en las hojas de los individuos jóvenes; en éstos las hojas hasta 17 cm de longitud, 5.5 cm de ancho, lanceoladas, abruptamente acuminadas 1.5 cm long, base agu- da, con 11 pares de venas secundarias; peciolo glabro, 6-10 mm de longitud. Inflorescencia ter- minal y lateral, paniculas o racimos de 6 cm de longitud, raquis, ramitas, pedicelos, y brácteas gla- bras; brácteas y bractéolas caducas; pedicelos ro- 1006 Annals of the Missouri Botanical Garden C. densifolia subsp. densifolia O C. grandifolia subsp. grandifolia © С. jaramilloi O C. myrcioides @ C. punctulata Ж C. tereticaulis © C. utilis % C. valioi ж FIGURA 3. Distribución de las especies y subespecies bustos y engrosados hacia el ápice, de 5-10 mm de longitud. Yemas florales antes de la antesis ovoides, 6-7 mm de longitud. Sépalos obtusos, glabros por fuera y pilosos por dentro, ciliados en el borde, lóbulos libres 2.5 mm de longitud, 3 mm de ancho en la base. Pétalos oblongos puberulentos por fuera y uno de los bordes ciliado, hasta | cm de longitud, 4-6 mm de ancho. Ovario tomentoso, surcado longitudinalmente, 1 mm de longitud, 0.8 mm diám., estilo 1 mm long. Cápsula trigona- oblicua o trigona-subglobosa, contraída en el ápice en un mucrón de 1 mm de longitud, la cápsula 1.5-2.0 em de longitud, 1.2-1.5 ст de ancho еп la base, superficie, ligeramente rugosa, cubierta de Caraipa en el Perú. con tomento marrón, exocarpo separable del en- docarpo, 1(-2) semillas por fruto, ovoides apla- nadas 1.7 cm de longitu Muestras adicionales ее PERU. LORETO: Provincia de Maynas, Distrito de Iquitos, Mishana, bosque d no inundable, : sion € con arena blanca, Vásquez 6144 AZ, ue e , Puerto Almendras, Tad 5518, 5544 (MO, SM); Requena, Jenaro Herrera, vols 9986 (MO, AMAZ, USM). Ecología. | Bosque primario no inundable, con suelos de arena blanca, algunas veces de extrema infertilidad, c PL como gregaria, con in- dividuos « formand un bosque llamado loc 'almente a que es un poco más Volume 78, Number 4 1991 Vásquez M 1007 Caraipa del Peru alto que el formado рог С. tereticaulis. Florece en septiembre, y frutifica en enero. Nombres vernaculares. Localmente se con- осе con los nombres siguientes: aceite caspi, aceite caspi colorado y aceite caspi negro. Usos. Los fustes se usan como madera redon- da para la construcción de viviendas, tanto en las áreas rurales como en las ciudades, llegándose a comercializar en gran cantidad, bajo ésta modalida en Iquitos, también se asierra. Caraipa utilis difiere de las otras especies por las hojas ovadas a elipticas-ovadas, en espiral, muy coriáceas, aunque las hojas de los individuos јоуе- nes son dificiles de separar de las de C. densifolia subsp. densifolia. A diferencia de las otras especies peruanas tiene las inflorescencias glabras; las cáp- sulas trigonas-oblicuas o trigonas- -subglobosas, asi- métricas y tornentosas (no son tan pequeñas у contraidas como las de С. densifolia subsp. den- sifolia, ni tan grandes como las де С. valioi); el látex transparente oleoso con puntos blancos con- trasta con el las otras especies; y su preferencia por habitar en bosques con suelos de arena blanca. OTRAS ESPECIES DE CARAIPA REPORTADAS DEL PERU Caraipa densifolia subsp. densifolia Mart. Látex amarillo viscoso. Hojas disticas glabras subcoriaceas, envéz liso, sin células epidérmicas protuberantes. Capsula tomentosa, trigona-oblicua, trigona-subglobosa, o trigona-ovoide. Habita suelos inundables estacionales por aguas negras, llamados localmente tahuampas. Florece en mayo, у fru- tifica en septiembre. Á esta especies se conoce localmente con los nombres vernaculares sig- uientes: brea caspi, casharana, y sacha parinari. Es usado como madera redonda en la construcción de viviendas, llegándose a comercializar bajo esta modalidad LORETO: Provincia de Muestras examinadas. PERU. L (MO, AMAZ, 3161 (A и 7652, А AMAZ, USM); Mishana, Vasquez 5. 382, , USM); Moronillo, dior 55: ; Вю Mazán, Gentry 1665 AMAZ, USM); din Arévalo 62 (AMAZ) Caraipa grandifolia subsp. grandifolia Mart. Látex transparente oleoso en el fuste, y blan- cuzco en las ramitas. Hojas disticas, subcoriaceas, envéz con pelos estrellados sésiles, las hojas jovenes con pelos estrellados pedunculados y con buli- formes células epidérmicas protuberantes. Capsula tomentosa, trigona- -piramiforme, geramente curvada en el ápice. Habita suelos hú- simétrica о |- medos con mal drenaje llamados localmente: agua- jales o sacha aguajales, muchas veces sometidos a los extremos de las inundaciones estacionales. Frutifica de diciembre hasta abril. A ésta especie se conoce localmente con el nombre vernacular aceite caspi. Se usa como madera redonda en la construcción de viviendas rurales. Muestras examinadas. PERU. LORETO: Provincia de Maynas, Mishana, Gentry 28947 (MO, AMAZ, USM); Puerto Almendras, Vásquez 6211 (MO, AMAZ, USM); Provincia de Ramón Castilla, Quebrada Ponaico, Vásquez 7332 (MO, AMAZ) Caraipa myrcioides Ducke Hojas disticas, subcoriáceas, glabras, envéz liso, sin células epidérmicas protuberantes. Habita sue- los aluviales inundables. . MADRE DE DIOS: Pro- Maestru 5 esaminalas. PERU Сова де топо, gom 41648, 41727 (MO, AMAZ, USM) Caraipa punctulata Ducke Latex amarillo, + fluido. Hojas disticas, carta- ceas o subcoriáceas, envéz con pelos estrellados sésiles y con buliformes células epidérmicas pro- tuberantes. Cápsula glabra o escazamente cubierta con diminutos pelos estrellados dispersos, trigona- piramiforme, simétrica o ligeramente curvada en el ápice. Habita suelos con mal drenaje, localmente llamados aguajales o sacha aguajales, preferente- mente en la transición entre éstos y el bosque no inundable por embalse, como también las pen- dientes de los riachuelos. Florece en octubre, y frutifica en enero. А esta especie se conoce con el nombre vernacular azufre caspi. Muestras спаљени PERU. LORETO: Provincia de ан саг 1 construcción Requena а Јепаго rrera, коз 4816 (МО, AMAZ, USM); Provincia e Maynas, Allpahuayo, Vasquez 5751, 6184 (MO, WTCC USM). 1008 Annals of the Missouri Botanical Garden Caraipa tereticaulis Tul. Látex amarillento viscoso, oxidando a blanco y solidificandose al contacto con el aire. Hojas dis- ticas, subcoriáceas a coriáceas, envéz con dimi- nutos pelos estrellados sésiles dispersos, y buli- formes células epidérmicas protuberantes. Cápsula glabra o con diminutos pelos estrellados dispersos, trigona-piramiforme simétrica o ligeramente cur- vada. Habita preferentemente suelos de arena blan- ca, muchas veces de extrema infertilidad, com- portandose como gregaria con individuos de reducido diámetro, formando bosque enano de 4- 5 m de alto, llamado localmente chamizal. Florece en agosto, y frutifica en enero. A ésta especie se conoce localmente con los nombres vernaculares siguientes: aceite caspi blanco, brea caspi del varillal, y boa caspi. Se usa como madera redonda para la construcción de viviendas llegándose a co- mercializar bajo esta modalidad. Muestras examinadas. PERU. LORETO: Provincia de Maynas: Puerto Almendras, | d@sques 5519, 5520 (MO, AMAZ, USM); Mishana, Гахдис 0143, 79 AMAZ, е, пизта localidad. а USM); Provincia de Requena, Vásquez 7661 (MO, AMAZ, USM); Freitas ОЗ, р? LVII (AMAZ). > aro H , misma localidad, Caraipa valioi Paula Es la especie peruana de Caraipa con árboles más grandes siendo su altura regular cerca de los 30 m y diámetro de 70 cm. Tiene látex oleoso, viscoso, y UN poco más transparente que las otras especies. Hojas son disticas, coriaceas a subcori- aceas, con envéz lizo, sin células epidérmicas pro- tuberantes. Cápsulas tomentosas, trigonas-oblicuas o trigonas-subglobosas, asimétricas. Habita suelos arcillosos arenosos, frecuentemente en zonas pan- tanosas o en los nacimientos de las quebradas. Florece en septiembre. Á esta especie se conoce con el nombre vernacular aceite caspi. Se usa como madera redonda en la construcción de vi- viendas, y también en aserrio. Muestras e ен PERU. LORETO: Provincia de Maynas, Puerto Almendras, Vásquez 6309 (MO, AMAZ, USM); Río Tahuayo, Vásquez 6754 (MO, AMAZ. USM); Provincia de Requena, Jenaro Herrera, Vásquez 9618 (MO, AMAZ, USM) LITERATURA CITADA Gentry, А. Н. 1981. Distributional patterns and an additional species ME the Passiflora vitifolia complex; Amazonian species diversity due to mdr T ou ferentiated communites. Pl. Syst. Evo : 105 ә. со с Endemism in tropical vs. temperate plant communities. Pp. 153-181 in M. Soulé nd tor), Conservation abd Sinauer Press, Sunder bond: Mns KUBITZKI, K. 1978. aceae). Highland Carta and Mahurea (Bonneti- 3. Maguire, The Botany of the ( n Part X. Mem. New York Bot. Gard. 2 T. о DE AGRICULTURA. 1986. Administra- ción Técnica del Distrito Forestal y Fauna de Iquitos, e 83), control para iuc y otros extrac- V "n EZ, 1987. Identificación de las especies usadas como madera redonda, para la construcción de viviendas, en la zona de Iquitos. Tesis. UNAP- FIF, Iquitos. PHYLOGENETIC ASSESSMENT Mark A. Hershkovitz" AND REVISED CIRCUMSCRIPTION OF CISTANTHE SPACH (PORTULACACEAE)! ABSTRACT examination of C ои s cladistic analysis of Portulacaceae confirms that Cistanthe Spach should be segregated rey & A. Gray Re from Calandrinia Kunth a equivocal, and the relationships among the s ^ Amarantoideae (Reiche) Carolin ex Hershkovitz, C. Calyptridium (Nu ovitz, C. pu Philippiamra (Kuntze) Hershkovitz, and C. sect. Strophiolum (B. Mathew) Hershkovitz The phy Boe relationship of prota to phere Portulacaceae indicated in Carolin’s cladogram ctions remain po should include Philippiamra Kuntze. Calyptridium Nutt. in Torr A. : i therefore recognized stern American species in five sections, ПГ Cistanthe m is, upon reanalysis, orly understood. Biogeographic understanding of the taxa here included i in Cistanthe has A аи by earlier, unnatural, and sometimes erroneous classifications of these plants Cistanthe Spach is a putatively monophyletic genus of Portulacaceae comprising five sections and perhaps 45-50 western American species for- merly classified in Calyptridium Nutt. in Torrey Gray, Philippiamra Kuntze (= Silvaea Phi- lippi), Spraguea Torrey, five sections of Calan- drinia Kunth, and one subgenus of Lewisia Pursh (see Table 1 for sectional key, citations, and syn- onymy). In the present paper, 1 describe the tax- onomic history of Cistanthe, evidence for its nat- uralness as a genus, its relationship to other Portulacaceae, evidence on phylogenetic relation- ships among the species, and its biogeography. TAXONOMIC HISTORY Spach (1836) segregated the Chilean species Calandrinia grandiflora Lindley, C. glauca Schrader, C. discolor Schrader, and C. Lehm. tanthe. Virtually all subsequent workers (e.g., Franz, 1908; Kelley, 1973; McNeill, 1974; Nyanyano, spec iosa into a new genus, which he named Cis- 1986, 1990; Pax & Hoffmann, 1934; Reiche, 1897, 1898) continued to recognize these species 1987). Calandrinia in the broadest sense includes more than 100 species of western North America, west- ern South America, and Australia (Carolin, 1987, in press; Kelley, 1973) Reiche (1897, 1898), in a revision of Chilean Portulacaceae, recognized 12 sections of Calan- in Calandrinia sensu lato (see Carolin, drinia s.l., including Calandrinia sect. Cistanthe, which corresponded to Spach’s (1836) generic cir- cumscription of Cistanthe. Reiche’s (1897, 1898) keys and descriptions make evident a close inter- relationship among Calandrinia sects. Andinae, Arenariae, Cistanthe, and Rosulatae (collective- ly, Cistanthe sect. Cistanthe) based on the com- mon presence of distinctive, resinous-appearing, black bract and sepal markings (Carolin, 1987) and often pubescent seeds (Kelley, 1973). Reiche did not propose a common category comprising these four sections, which differ primarily with respect to plant size and branching habit (Carolin, This manuscr ipt. was adapted from the first of five that constitute a doctoral dissertation in botany completed at California, — Varied astro-Shields Scholarships from d, a Smithsonian Pre-doctoral Fellowship, and two travel awards from the Department of Botany, University of forms of assistance from the following individuals are gratefully acknowledged: J. A. Doyle, ‚ D. 1. Ford, P. Hershkovitz, W. A. Kelley, В. J. McCaskill, J. W. Nowicke, С. Е. Russell, К. Schmid, J. L roig G. L. Webster, and S. Yankowski. ANN. Missouni Вот. GARD. 78: 1009-1021. 1991. 1010 Annals of the Missouri Botanical Garden TABLE 1. Taxonomy of Cistanthe Spach. Provided below is a key to the sections of Cistanthe, along with full citations, synonymy, and distributions (see Hershko- vitz, 1990a). la. pe and capsule чч 3(-4); sepals her- 2 aceous, marked marked; petals usually r то xe: seeds «rophilate or estrophiolate, dala or glabrou 2a. Sepals be а marked or unmarked; pedicels mostly = 2 mm; seeds strophiol- ate, pubescent or glabrous; sinuous and ribbonlike veins present or absent; western m America and western South Ameri- 3a. Sepals marked with black etchings; bonlike veins rarely absent; western South America and mae North ic tanthe sect. Cistanthe о c "о E т = = 3: > = añ Ф = "о o = 25 B SE xL. Fé 12); © ee кы with British Columbia. KU EIN Cistanthe sect. Strophiolum У se membranous, unmarked; pedicels ostly = 2 mm; Seeds сторам gla- N c ns present; western South Amer Ci тона А sect. Amarantoideae lb. Stigmas 2 and capsule valves 2 or none; sepals usually membranous at least along margin, un marked; petals 4 or fewer; seeds ни. glabrous. 4a. Fruit dehiscent by two valves; sepals rarely nearly completely herbaceous; ovules most- ly more than 1; P3 ja America. ........................ anthe sect. Calyptridium . Fruit indehiscent or ا‎ dehiscent; sepals membranous; ovule 1; South Amer- Cistanthe sect. Philippiamra > = Cistanthe Spach, Hist. Nat. Vég. 5: 229. 1836. Cistanthe sect. Cistanthe. LECTOTYPE: Calandrinia grandiflora Lindley. Calandrinia sect. Cistanthe Reiche, Ber. Deutsch. 1897 Calandrinia sect. Arenariae Reiche, Ber. Deutsch. ot. 5: 502. 1897. Calandrinia sect. Andinae Reiche, Ber. Deutsch. Bot. Ges. 15: 502. 1897. Calandrinia sect. Rosulatae Reiche, Ber. Deutsch. t 02. 1897. Ca. 30 spp. from Chile, ои апа Реги; 2 ѕрр. from СаШогша/ Мех 2. Cistanthe sect. Strophi iet _ Mathew) Hersh- kovitz, Phytologia 68: 268. 1990. Lewisia subg. Strophiolum B. Mathew, The rie Lewisia 139. 989. LECTOTYPE: Calandrinia tweedyi A. Gray TABLE 1. Continued. 1 sp. from Washington and British Columbia. . Cistanthe sect. Amarantoideae (Reiche) Carolin ex Hershkovitz, Phytologia 68: 269. 1990. Calandri- nia sect. Amarantoideae Reiche, Ber. Deutsch. Bot. Ges. 15: 501. 1897. LECTOTYPE: Calandrinia sal- soloides Barnéoud in Gay. Ca. 4 spp. from Chile, Argentina, and mn 1 sp. from 0 lifornia, Arizona, and Cistanthe sect. Calyptridium (Nutt. in Torrey & A. Gray) Hershkovitz, Phytologia 68: 267. 1990. eo ng in Torrey & A. Gray, Fl. N. r. 1(2): 198. 1838. TYPE: pun mon- dun ue in т & A. Spraguea Torrey, Pl. Frémont. Smithsonian Contr. Knowl. 6(2) [егг., 5(1)]: 4. 1853. Ca. 8 spp. from western North ju ; . Cistanthe sect. Philippiamra (Kuntze) Hershko- vitz, Phytologia 68: 269. 1990. Philippiamra Kuntze, Revis. Gen. Pl. 1: 58. 1891. LECTOTYPE: Silvaea celosioides Philippi. Silvaea Philippi, Fl. Atacam. 22. 1860 (non Silvaea J. D. Hook. . ex Baillon, 1858). 2-4 spp. from Chile ы Peru. о + сл 1987; Reiche, 1897, 1898). Reiche (1897) noted the similarity between Cistanthe sects. Amaran- toideae and Philippiamra, but he classified the former in Calandrinia sect. Amarantoideae and the latter in Si/vaea (see Table 1). ax Hoffmann (1934), in their taxonomic revision of Portulacaceae, classified Cistanthe sects. Amarantoideae and Cistanthe (as part of Cal- s..) ш Calandrinieae-Calandriniinae; Cistanthe sect. Calyptridium (as Calyptridium and Spraguea), in Calandrinieae-Calyptridiinae; and Cistanthe Philippiamra (as Philip- piamra) in an informal group described as inter- mediate between Portulacaceae and Basellaceae. andrinia sect. rhe sections of Cistanthe, in their former generic assignments, were each taxonomically associated with other relatively distantly related members of Portulacaceae (Carolin, 1987). Thus, Caladriniinae also included Talinum Adans., Anacampseros L., and other genera; Calyptridiinae included Mono- cosmia Fenzl [= Calandrinia sect. Monocosmia (Fenzl) Hershkovitz (Hershkovitz, 1990b, 1991b, in press а) |; and Philippiamra was associated with Ceraria Pearson & Stephens and Portulacaria Jacq. (Carolin, 1987; Pax & Hoffmann, 1934). As noted above, Cistanthe sect. Strophiolum was thought to belong in Lewisia, which Pax & Hoff- mann placed in tribe Portulaceae along with Por- tulaca Volume 78, Number 4 1991 Hershkovitz Assessment and Circumscription of Cistanthe 1011 Lewisia Calandrinia § Acaules * Clade Calandrinia $ al a Calandrinia § Hirsuta Rumicastrum * Calandrinia § Calandrinia * Calandrinia 8 Monocosmia Portulacaria Сегапа Calyptrotheca Schreiteria * Portulaca © FIGURE 1. Simplified reproduction of Carolin's (1987: 4 major clades and among members of Cistanthe. Talinopsis Grahamia Talinaria Anacampseros Claytonia Montia Lenzia Cistanthe § Calyptridium Cistanthe § Cistanthe * y Clade Cistanthe $ Amarantoideaé’ “Е” Cistanthe $ Philippiamra ж 2, fig. 7) most parsimonious cladogram of Portulacaceae, e The numbered characters and “X” are Carolin's (see Table 2). All of the character state changes shown are regarded here as problematic (see text). Kelley (1973), in his survey of seed and tri- chome morphology in the genus, maintained that Calandrinia s.l. was a natural assemblage. The traits Kelley cited as common to all the species (two sepals, three carpels, free-central placenta- tion) are, LIA Is symplesiomorphic with many to all elley recognized the close interrelationship between Calandrinia sects. Ап- ortulacaceae. dinae, Arenariae, Cistanthe, and Rosulatae (1.е., among members of Cistanthe sect. Cistanthe). Like Reiche (1897), Kelley remarked on the mor- phological similarities between Cistanthe sects. Amarantoideae and Philippiamra, but he did not recommend including these in a common genus. McNeill's (1974) tribal classification of Portu- lacaceae is similar to that of Pax & Hoffmann (1934) in that the components of Cistanthe are placed in four different tribes. Nyanyano (1986), in a revision of Portulaca- ceae, divided Calandrinia s.l. into three sections, are anomalous (cf. Carolin, 1987, in press; Kelley. 1 1973; Reiche, 1897, 1898), nor with the characterizations in his sectional key. Furthermore, many species of Cal- including all of the Australian ele- ment, are not accounted for in his treatment. In апапта s.l., Nyanyano's three-section classification of Calan- 1012 Annals of the Missouri Botanical Garden drinia s.l., members of Cistanthe sect. Cistanthe are included in two of the sections, while Cistanthe sect. 4marantoideae is included in the third. Each of Nyanyano's three sections includes taxa that, according to Carolin Hershkovitz (1990b, in press a), belong in different genera. Although Nyanyano (1986, 1990) recognized the relationship between Cistanthe sects. Calyptri- dium and Philippiamra (as distinct genera) and included both in tribe Calyptridieae, his circum- scription of Calyptridieae is otherwise unnatural, and he went so far as to include Monocosmia in Calyptridium (cf. Hershkovitz, in press a). Carolin (1987), in his phylogenetic analysis of Portulacaceae, divided the existing sections of Cal- andrinia s.l. (see Cullen, 1953; von Poellnitz, 1934; Heiche, 1897, 1898) into 11 OTUs (excluding Calandrinia sect. Partitae von Poellnitz, which belongs in 4nacampseros; see Carolin, 1987; Ke ley, 1973). Carolin's OTU Cistanthe comprised Calandrinia Cis- tanthe, and Rosulatae. Perhaps the most signifi- cant finding of Carolin's analysis was that the OTUs sects. Andinae, Arenariae, of Calandrinia s.l. showed cladistic relationships with various genera throughout Portulacaceae, in- dicating that its erstwhile circumscriptions were e OTU Cistanthe, along with Amarantoideae (= Cistanthe sect. unnatural (Fig. 1). ' the OTUs Amarantoideae) and Silvaea (= Cistanthe sect. Philippiamra), formed clade “К” in Carolin's (1987, fig. 7) cladogram. Although lacking an un- equivocal synapomorphy (see below), clade “К” corresponded with Carolin's (1987, in press) pro- posed generic circumscription of Cistanthe. Car- olin (in press) proposed combining the OTUs Ата- rantoideae and Silvaea in a single section named for the former. In Carolin's (1987) analysis, clade “F” formed the sister group to the OTU Calyp- tridium (— Cistanthe sect. Calyptridium). Clade F” + Calyptridium |] was diagnosed by one unique and nearly universally shared feature: the presence of unequal bracts associated with the in- florescence nodes (Fig. 2). Carolin (1987, in press) maintained Calyptridium as a distinct genus, how- ever. The present paper basically follows Carolin's (1987, in press) recommendation that Cistanthe be segregated from Calandrinia s.l., but the cir- cumscription is modified to include Calyptridium and Lewisia subg. Strophiolum as distinct sec- tions, and Carolin’s OTUs Amarantoideae and Sil- vaea are each assigned to their own section (see Table 1). PHYLOGENETIC ASSESSMENT OF CISTANTHE Figures | and 3 depict the relationships among Portulacaceae as determined in Carolin’s (1987, fig. 7) cladistic analysis and subjectively reanalyzed here. As discussed below, these figures emphasize evidence pertaining to the monophyly of Cistanthe, the relationship of the genus to other Portulaca- ceae, and relationships within the genus. The tax- onomy in Figures | and 3 follows Carolin (1987, in press), except for Cistanthe and Calandrinia, which follows Hershkovitz (1990a, b, 1991b, in press a). The numbered characters are defined in Table 2. Figure 1 depicts Carolin's consensus tree and shows the character state changes occurring between the root of the tree and the terminal OTUs of Cistanthe, as well as selected character state changes that are relevant to the present discussion but involve other branches of the tree. All of the indicated. synapomorphies аге regarded here as equivocal or unwarranted, as explained below. Fig- ure 3 emphasizes possible phylogenetic relation- ships among the sections of Cistanthe as speculated here on the basis of the data in Table 3 and shows the relationships of Calandrinia sects. Calandri- nia and Monocosmia I have proposed elsewhere (Hershkovitz, 1990b, in press a). Cistanthe sect. Strophiolum is not shown in Figure 1 because Carolin was unaware of the relationship of this taxon to Cistanthe rather than Lewisia (Hersh- kovitz, 1990b, in press b). 1. THE MONOPHYLY OF CISTANTHE Although the monophyly of Cistanthe in the circumscription proposed here is evidenced by only a single synapomorphy (Figs. 1, 3), additional traits reinforce the close interrelationships between the sections. For example, leaves of Cistanthe gen- erally have winged (if present) petioles, clasping leaf bases, festooned brochidodromous venation (if sufficiently broad to show this trait), a primary vein that becomes obsolete at the leaf apex, ribbonlike veins, sinuous veins, and predominantly brachy- paracytic stomata (Hershkovitz, 1990b, in press b, 1991c). absent among other Portulacaceae, although their The combination of these leaf traits is value in cladistically diagnosing the genus remains uncertain. 2. PHYLOGENETIC RELATIONSHIP OF CISTANTHE TO OTHER PORTULACACEAE Subjective reanalysis of Carolin's (1987) data in light of additional evidence from leaf morpho- Volume 78, Number 4 1991 Hershkovitz 1013 Assessment and Circumscription of Cistanthe logical studies (Hershkovitz, 1990b, in press b, in press c, in prep.) indicates that the western Amer- ican members of Portulacaceae (Fig. 3; Carolin, 1987, fig. 8; Hershkovitz, 1990b, 1991a, c, in prep.), including Cistanthe, form a cohesive and possibly monophyletic group. The western Amer- ican taxa share similar habit, leaf base, petiole, and stomatal morphology (Hershkovitz, 1 a, in prep.). The distinction between the western Amer- ican group and the eastern American/ African group of Portulacaceae with respect to these characters suggests that one or more of the characters may help cladistically diagnose either group. The Aus- tralian endemic Rumicastrum Ulbrich (cf. Fig. 3), however, is morphologically intermediate between these two groups (Hershkovitz, 1990b, in press a, in prep.). The precise relationships of Cistanthe to other western American taxa remain uncertain pending additional analysis. The phylogenetic relationships of Cistanthe in- ig. 1) are more equivocal than they appear to be. The synapo- morphies that diagnose Portulacaceae (including dicated in Carolin's cladogram “a »* Cistanthe) minus Carolin's “secondary” clade (not a clade in Carolin's consensus tree, Fig. 1) are problematic for several reasons. The synapomor- phies are the change from more than five to five or fewer petals (character 22), and the two-step change from five or more to four to three or fewer stigmas (characters 28 and 29). Five petals and/ or three stigmas characterize several species in each of the excluded OTUs, however (see Mathew, 1989, for Leuisia; and Reiche, 1898, for the HA sections of Calandrinia), and these char- acters show considerable reversal and rederivation in the b diagnosed by low petal and stigma numbers (Carolin, 1987). The variability of these characters within and among the highly critical the accuracy of the character po- larity assessment. Carolin based his polarity as- s renders sessments for these characters in Portulacaceae (and, presumably, his assumptions on primitive states within OTUs) on the outgroup relationships proposed in Rodman et al.'s (1984) cladistic anal- ysis of Centrospermae. In fact, based on that study, the polarities of these characters would be equiv- оса! because one outgroup (Aizoaceae + Cacta- ceae) was scored for having high petal and stigma numbers, whereas the other (Basellaceae + Didieri- aceae) has low petal and stigma numbers exclu- sively. This point is moot, however, because Rod- man et al.’s (1984) character definitions and/or scores for floral characters and, for that matter, most aspects of their phylogenetic analysis have been heavily criticized (Levin, 1985; Bittrich, 1990; Bittrich & Hartmann, 1988; Hershkovitz, 1989; Rodman, 1990). Also, I have elsewhere (Hersh- kovitz, 199 la, in prep.) maintained that Cactaceae, Didieriaceae, and Basellaceae are each phyloge- netically nested among Portulacaceae. Carolin's (1987) treatment of the petal and stig- ma characters is problematic for other reasons as well. For example, it is not clear whether petals in Portulacaceae are homologous to the sepals or sta- mens of apetalous Centrospermae, to similar struc- tures in other ““petaloid”” Centrospermae, or even to petals in the currently unknown outgroups of Centrospermae (see Cronquist, 1981, 1988; Levin, 1985; Rodman, 1985). Also, the change from five to three stigmas should not require two steps, as it does in Carolin's (1987) analysis. n the absence of unequivocal outgroup rela- FIGURE 2. Inflorescence bract morphology in Cistanthe. - MO). Terminal portion of inflorescence showing unequal bracts at the patines. nodes and US). Portion of inflore maller bract is displaced apically on the ec axis. (Modified from Hershkovitz, in press ие calycina (Johnston 5313, US). ищу. bracts — B. C. (Strophiolum) iiid (Thompson 6463, flowering node. The b.)—C al but dh he E flowers M duda a short t distance below. EF. 5. The бү with the node; Ь?, wers ak rit) luin eic with the flow — -А. C. (Cistanthe) ко و‎ & — 497, and sepal m 5 сепсе showing а окна! аеро at the cag portion of inflorescence showing (inset, with smaller of the pair of bracts associated with the node; m, black bract/sepal markings; p, pedicel; s, sepal; x, scar from removed or abscised flower. (Drawings by Lesley Randall.) 1014 Annals of the Missouri Botanical Garden Volume 78, Number 4 Hershkovitz 1015 1991 Assessment and Circumscription of Cistanthe 1016 Annals of the Missouri Botanical Garden Australia D 5 : 5 e. America/ Africa Calyptrotheca Talinum Schreiteria Portulaca Talinopsis Grahamia Talinana Anacampseros w. America 9 Lewisia Calandrinia $ Dianthoideae Calandrinia $ Hirsutae Calandrinia $ Acaules Calandrinia $ Calandrinia Calandrinia $ Monocosmia еш Monta Lenzia Cistanthe § Strophiolum Cistanthe § Cistanthe Cistanthe § Amarantoideae Cistanthe § Calyptndium Cistanthe § Philippiamra FIGURE 3. Revised cladogram of Portulacaceae derived by eliminating the questionable synapomorphies indicated The ho in dim 1 (see text). The numbered and lettered characters are defined in Table Calandrinia sects. Calandrinia and Monocosmia are after Hershkovitz (гооо, in press a), and the shown among the members of Cistanthe are discussed in the present text. the regions of endemism or greatest endemism of the inc Juded taxa (Carolin, 1987; Hershkovitz, in circum press a, In prep .). iscribe tionships for Portulacaceae, the evolutionary po- larity of petal and stigma numbers for this family cannot be determined a priori, nor can primitive states be arbitrarily designated in those genera variable for these characters. Carolin's cladogram, therefore, the synapomor- phies diagnosing the basal node must be regarded as unwarranted Because the basis for the exclusion of Carolin's In the context of 2. The relationships shown for relationships e boxes enclosing portions of the diagram (1987) clade “a” from the remainder of Portula- caceae (Fig. 1) is questionable, the sole synapo- morphy that diagnoses the clade (Lenzia Philippi + Cistanthe (in the present sense) + tribe Mon- tieae Dumort.), the derivation of a monochasial inflorescence from “dichasia passing into mono- chasia" (character 14), also becomes problematic. Monochasia and/or solitary flowers occur in sev- eral members of clade “a.” Dichasial inflorescence Volume 78, Number 4 Hershkovitz 1017 1991 Assessment and Circumscription of Cistanthe TABLE 2. Characters referred to in Figures 1 and 3. The numbered characters and character “X” correspond - to Carolin's (1987: 392, 412, appendix 3). The character 9 M е з states given first are those supposed by Carolin to be = E 5 5 Lu ne E = primitive. The polarities for characters 1, 2, and 4 аге S 317% E g Se | eg de A a a erived from Carolin’s text because those indicated in his 5 = SE BE = = ~ Кре _ 2 == = appendix appear erroneous. S rolin's polarity " E EN Y E т dà E g g 2:5 a Sy a Lovo determinations are discussed in the present text. Char- = А УЕ ANSE >» >» Y acters A a efined here; the state given first is -2 considered aa £ = v Е " Char- $ e Б = Eb acter E E = е ы = по. Character: primitive state vs. derived state 5 Е = Е 8 5 M ص‎ & N E № Ф 1 Growth form: not phanerophyte vs. phanero- = | зу 2 2 E o 8 РД 0 Ф ~ — ћуте E ора» A 9 = оо © 2 С crm | i ДР 33533 =БЕБЕБЕ 2 Growth form: not suffruticose and/or passive = TSF ¥ SIGE == chamaephyt ff d/ o y saan” 1 TELA оо o phyte уз. suffruticose and/or passive о} оч саз бб ~ ~ ~ chamaephyte = 5 4 Growth form: not therophyte vs. therophyte = 5 a 14 анаа "dichasial passing into топос s E E 5 Е ~ СУ sia" vs. monochasial or solitary Е ° Ws ed 5 v 5 15 ны die not solitary vs. solitary os | 2 | У 2 и = © 16 Inflorescence bracts: equal vs. unequal ^ v Е 5 ~ 4 Е Dada я = 3 = = ao & С 21 Sepal markings: absent vs. present mB E| ЗУЕВ 289398 Р" зо ~ on д @ 22909 0 ا‎ 22 Petals: more than 5 vs r less 2873 ‘Osa E nms Eww GRR v o. 25 Pollen: trizonocolpate vs. panporate = $ 28 — Stigmas: less than 5 E 2 tigmas: 5 or more vs. less than a 29 Stigmas: 4 or more vs. less than 4 FE = 31 Fruit: dehiscent vs. indehiscent = в 13| 2 E i А S =| o Inflorescence: open with long pedicels vs. con- LE ES 2 = densed with short pedicels ~ | ЗВ $ e бык oS NEO |= ЕБ $ AI RF E CE 8 A Carpel number: 3 or more vs. с Sou 2 он LT. я. го ыле Шы ы Фф фб © B Fruit: indurate vs. membranous . eee ж A ЕУ ОЁ 23 Jod ЕЕ a branching in Lewisia is restricted to species of L. 5 2 11% Е 8 x " > © о = = sect. Cotyledon J. Hohn ex B. Mathew, and FE p ES. ES dr - E 5 А | 4 са $ mo - there is no strong evidence that section is otherwise 5 = = ла ~ „ белее E neas | cT © s Es Я "а primitive within the genus (Hershkovitz, 1990b,in Û g # . . ~ = press b; but see also Hohn, 1975); i.e., the dichasia o g E . . á Y might be secondarily derived. Reiche (1897) and S 5 ^ ЖА . . . uuo a Y Hershkovitz (1990b) reported that dichasial “У = ~ 2 Е . : um $ = = Е o Ф g branching rarely occurs in Cistanthe, but this is on la Es $ 85 = В : | 8 > 5 | = a m questionable (Hershkovitz, in press b). Closer ex- ас пее аи озш AT : : IE LE е ^ a.c ~ Al > 5 в ara с. с. amination of herbarium specimens reveals that su- 2H . . . . . o perficially dichasial branching results when the in- = B я 5. wn ternodes are not elongated. Б 2 E us The scoring of inflorescence types in Carolins 5 & о ~ a en ; © (1987) analysis is problematic for other reasons as Но = Ed Ф Б 5 5 5 = = = E Ф 2 Ё well. Because the primitive state could not be as- У. |5 - 55 „5925 BD < . Е А Ф с Ф Qs > > се sessed by outgroup comparison, Carolin based his 3 || 5 8 So q 5 E ZE 2 543 8 А 4 o и =] Я Ф O = o Ф inflorescence character transformation series on 4 5 ЕЛ Б کے‎ 9 x ЕВ pa = 5 = = = = Hm a Е 4 $ E Hom x чи = l| = Ф = general principles," presuming that metaxytriads & 3 5554 88227 2033 5 nut EL язь Sfaas о 2 are primitive among Portulacaceae. І suspect that 5 Боке боо о bS =: р 5 P e SAY RO O A Carolin is correct in this regard; nevertheless, in his cladogram, the taxa scored for having metaxy- 1018 Annals of the Missouri Botanical Garden triads (Portulacaria, Ceraria, Calyptrotheca Gilg, Portulaca) are nested among taxa having the sup- posedly more ee inflorescence types (see Car- olin, 1987, . Also, ( series requires bo solitary flowers be derived from Carolin's transformation dichasia via monochasia; i.e., in two steps rather than one. This requirement appears to have influ- enced both the topology and number of steps їп the most parsimonious cladogram. For example, except for the assumption that solitary flowers are derived from monochasia, Lenzia would not be included in the clade with Cistanthe and Montieae (see Fig. 1). Other problems in Carolin's (1987) cladogram have an indirect effect on the assessment of the position of Cistanthe among Portulacaceae. For example, both synapomorphies (Fig. 1, characters 2, 25) supporting the clade comprised of Rumi- castrum, Calandrinia sect. Calandrinia (= Cal- andrinia sect. Compressae Reiche), Calandrinia Monocosmia, Ceraria, and sect. Portulaca riq, Calyptrotheca are problematic. The shared loss of the chamaephytic growth form (character 2) is essentially redundant with the subsequent deriva- tions of the phanerophytic growth form (character 1) in Calyptrotheca, Ceraria, and Portulacaria, and the therophytic growth form (character 4) in Rumicastrum and Calandrinia. Carolin treated different growth forms as different characters rath- er than different states of a single character. The derivation of pantoporate from tricolpate pollen . 1, character 25), which subsequently reverses in lbs Portulacaria + Ceraria clade, is equivocal: two independent derivations of pantoporate pollen would be equally parsimonious. In this case, two independent derivations seem more plausible, how- ever, because there is no evidence elsewhere among angiosperms that tricolpate pollen can be rederived from pantoporate. Also, additional reanalyses of Carolin’s (1987) cladogram and corroboration with leaf morphological evidence indicate that Calan- Calandrinia and Monocosmia ac- Acaules drinia sects. tually belong near Calandrinia sect. Reiche, rather than next to the Rumicastrum clade (Hershkovitz, 1990b, in press a, in press с) Pending a thorough reanalysis of phylogenetic relationships among members of Portulacaceae and affiliated families, the position of Cistanthe will remain uncertain. If the problematic synapomor- phies in Carolin's (1987) cladogram described above are disregarded, however, the basal polytomy shown in Figure 3 results. Figure 3 presumes that, in a formal reanalysis of Portulacaceae, Cistanthe would remain monophyletic. Preliminary reanalyses (D. Ford, unpublished; Hershkovitz, unpublished) in- dicate that this will be the case. Figure 3 is not only highly unresolved, it is inherently inaccurate because of the probable paraphyly of Portulaca- ceae with respect to Basellaceae, Didieriaceae, Hectorellaceae, (Hershkovitz, 1989, two other members of Portulacaceae, Baillon and Атрћре ит Bacigalupo in Spichi- ger, were not included in Carolin's (1987) analysis and possibly also Cactaceae 1991а, unpublished). Also, Talinella and are not shown in Figures 1 and : 3. PHYLOGENETIC RELATIONSHIPS WITHIN CISTANTHE The relationships among the sections of Cis- tanthe indicated in Figures 1 and 3 presume that all of the sections of Cistanthe are monophyletic. The monophyly of Cistanthe sect. Cistanthe is evidenced only by the universally shared presence of bract and sepal markings (Fig. 2A), although the trait is apparently lacking in aberrant speci- mens (Hershkovitz, 1991b). The possibility that this section is paraphyletic, therefore, should not be excluded, especially because this group other- wise shows variation for several of the characters listed in Table 3. Cistanthe sect. Strophiolum is monotypic, hence monophyletic. Cistanthe sect. Philippiamra appears to be monophyletic based on its fruit morphology. According to Table 3 an Figures | and 3, sects. Amarantoideae and Ca- lyptridium lack an autapomorphy. Á potential aut- apomorphy for sect. Calyptridium is the tetraploid condition found in all members of this section (see Hinton, 197 Based on their character states for carpel num- ber, capsule dehiscence, and capsule texture, sects. Amarantoideae, Cistanthe, and Strophiolum ap- М pear to represent the primitive element of Cis- tanthe (see Fig. 3). Because the precise outgroup relationships of Cistanthe are not known, however, the data in. Table 3 cannot corroborate an un- equivocal hypothesis of relationship among these three sections. If the outgroups lack arils and have cauline leaves, sinuous and ribbonlike veins, and few petals (e.g., some Rumicastrum spp.), then Amarantoideae likely a the basal- t. Strophiolum sect. most clade in the genus, and s would arise collaterally with sect. оте If the outgroups possess the alternative states of these characters (e.g., some Lewisia spp.), then sect. Strophiolum would likely represent the basalmost ershkovitz, in press b)—a result that would have interesting biogeographic implications taxon (see (see below). Most of the potential outgroups of Cistanthe (see Fig. 3), however, do not possess the Volume 78, Number 4 1991 Hershkovitz 1019 Assessment and Circumscription of Cistanthe entire suite of traits that could unequivocally show either sect. Amarantoideae or sect. Strophiolum as the basal taxon in Cistan The difference between Carolin’ s (1987) and the present interpretation of the position of sect. Ca- lyptridium can be accounted for, in part, by a problematic ie ena! indicated in Carolin’s cladogram (Fig. 1). Carolin’s cladogram indicates that the absence black bract and sepal markings (character 21) excludes sect. Calyptridium from the remainder of the genus. According to Carolin (1987), this trait is weakly evident in members of sect. Amarantoideae, but I have found no such evidence in the specimens | examined (see Hersh- kovitz, 1991b). (Carolin (in press), in his generic key to Portulacaceae, also reported markings in Rumicastrum and Schreiteria Carolin—I cannot confirm whether markings exist in these taxa.) Even if the markings occur in sect. Amarantoideae, however, their absence must be regarded as the primitive state unless this section is presumed to have arisen from within sect. Cistanthe. In any case, Carolin misscored sect. Philippiamra for having the markings. With this in mind, it shou be clear from Figure 1 that sect. Calyptridium cannot be excluded from Cistanthe on the basis — of this character. The putatively derived position of sect. Calyp- tridium indicated in Figure 3 is based on its pos- session of a two-carpellate rather than three-car- pellate gynoecium (shared with sect. Philippiamra), and a condensed rather than more open inflores- сепсе (character “X” in Figs. 1, 3; not unique to Amarantoideae and Philippiamra, contra DI 1987). Other possibly significant charac- ters (not indicated in Fig. : sepal texture, and pollen morphology. The fruit texture in sect. € 'alyptridium is intermediate be- Amarantoideae, 3) include fruit texture, tween the more indurate of sects. Cistanthe, and Strophiolum, and the more mem- branous of sect. Philippiamra. Like sects. Ama- rantoideae and Philippiamra, the sepals in sect. Calyptridium are usually at least partially mem- branous (Munz & Keck, 1973) rather than her- baceous (i.e., ch orophyllous). The last character is difficult to evaluate from herbarium specimens— even herbaceous bracts probably become more membranous with age and more so with herbarium processing. Finally, Nilsson (1967) noted a simi- larity in pollen morphology between sect. Calyp- tridium and sect. Philippiamra, although he did not extensively survey other members of Cis- tanthe. In order to resolve phylogenetic relationships further among the sections of Cistanthe, it will be necessary to resolve interspecific relationships with- in the sections. Only limited inferences can be made а! present. example, thr Calyptridium (C. umbellata (Torrey) Hershkovitz, С. monosperma (Е. Greene) Hershkovitz, and pulchella (Eastwood) Hershkovitz) that collecte constituted Spraguea (Hinton, 1975) lack the smaller of the two bracts that characterize all other Cistanthe (Fig. 1 E). The close relationship of these three species to the remainder of sect. Calyptri- dium seems beyond doubt (Hinton, 1975), so it seems reasonable to conclude that the former con- stitute a clade within the latter (and that the con- troversial genus Spraguea does, in fact, represent a natural grouping; cf. Hinton, 1975). Cistanthe calycina (Philippi) Carolin ex Hershkovitz and C. ambigua (S. Watson) Carolin ex Hershkovitz, both ree species of sect. of sect. Amarantoideae, have a somewhat more open inflorescence (plesiomorphic?, see Table 3) than other species of the section. Cistanthe am- bigua, the only North American species of sect. Amarantoideae, has especially long pedicels (ple- siomorphic?, see Table 3) compared to the other species of this section (Kelley, 1973), but its leaf venation pattern is clearly derived (Hershkovitz, 1990b, 1991c). Thus, the classification of C. am- bigua in sect. Amarantoideae deserves additional scrutiny. Cistanthe sect. Cistanthe is the most polymorphic and widely distributed section of Cis- tanthe, and while some apparently closely inter- related species groups can be identified (Hersh- kovitz, 1991b; Kelley, 1973; Reiche, 1898), the overall phylogenetic interrelationships remain ob- scure. The South American species of sect. Cis- tanthe are especially in need of revision (Hersh- kovitz, 1991b). Until this is accomplished, attempts to circumscribe taxonomic units for purposes of phylogenetic analysis may be futile. Phylogenetic resolution in sects. Philippiamra and Strophiolum is trivial. The former is monotypic, and while the number of species in the latter is not well estab- lished, the degree of polymorphism is clearly limited 91] b) (Hershkovitz, 19 BIOGEOGRAPHY OF CISTANTHE Understanding of the biogeography of Cistanthe was obscured by pre-Carolin (1987) phylogenetic and biogeographic concepts of Portulacaceae, in which much of Cistanthe was included in Calan- drinia s.l. The latter genus was regarded as ba- sically western South American and Australian (Kelley, 1973; Raven & Axelrod, 1978). The only two North American species of Cistanthe formerly included in Calandrinia s.l., C. ambigua (Mojave 1020 Annals of the Missouri Botanical Garden and Sonoran deserts) and C. maritima (Nutt. in Torrey & A. Gray) Carolin ex Hershkovitz (coastal California and Baja California), were presumed to have arrived from temperate South America via relatively recent (Quaternary ?) long-distance and/ or stepwise dispersal (Kelley, 1973; Raven & Ax- elrod, 1978). Cistanthe sect. Philippiamra was, presumably, conceived of as disjunct with the South African genera Ceraria and Portulacaria (see Car- olin, 1987). Cistanthe sects. Calyptridium and Strophiolum, as Calyptridium and a species of Lewisia, respectively, are presumably among those North American Portulacaceae believed to have ultimately had a (temporally and phyletically non- specific) South American origin (see Raven х- elrod, 1978) The present synthesis of Cistanthe as illumi- nated by Carolin (1987) and emended here pro- vides a previously unappreciated example of an amphitropical temperate disjunct taxon having considerable endemism in North and South Amer- ica. Particularly critical to this biogeographic rev- elation is the determination that sect. Calyptri- dium, with eight species in North America, is not only related to other members of Cistanthe (Car- olin, 1987; also implicit їп Nilsson, 1967, and Nyanyano, 1986, 1990) but is also cladistically inextricable from the remainder of the genus (see above). Also significant is the recognition of the relationship of the Cascade-endemic C. (Strophio- lum) tweedyi (A. Gray) Hershkovitz to Cistanthe (Hershkovitz, 1990b, in press b). Finally, the re- alization that Cistanthe guadalupensis (Dudley in D. Jordan) Carolin ex Hershkovitz (Guadalupe Is- land) belongs in C. sect. Cistanthe (Carolin, 1987; Hershkovitz, 1990a) rather than 7alinum (Dudley, 1899) contributes an additional species to the North American element of Cistanthe. The misassign- ment of-this species to Talinum, perpetuated in all pre-Carolin (1987) taxonomic treatments (e.g., Rydberg, 1932; Pax & Hoffmann, 1934; Wiggins, 1980), yielded a spurious interpretation of the dis- tribution of Talinum (cf. Raven & Axelrod, 1978: Although the bulk (ca. 35 spp.) of Cistanthe remains South American, the degrees of morpho- logical variation, ecological specialization, and geo- graphical extent among the North American spe- are The diversity distinctiveness of North American Cistanthe would cies comparable. and not corroborate hypotheses that these species are either all recent immigrants from South America or descendants from a single common ancestral immigrant. Cistanthe maritima and C. guadalu- pensis may, in fact, represent relatively recent dai arrivals—these species are morphologically very similar to South American members of sect. Cis- tanthe, and both have very limited distributions in open, hence invasible, coastal habitats. Because of its morphological distinctiveness, C. ambigua fails to fit the profile of a recent immigrant (see above). Cistanthe sects. Calyptridium and, especially, Strophiolum are also morphologically distinct from their North and South American counterparts. The Strophiolum represents the sister group to the remainder of the genus is par- possibility that sect. ticularly significant, because it implies that Cis- tanthe was present at least as early in North Amer- ica as, if not earlier than, in South America. Overall, no simple biogeographic scenario emerges to ac- count for the diversity and distribution of North American Cistanthe. LITERATURE CITED BiTTRICH, У. 1990. Systematic studies T оаа Mitt. Inst. Allg. Bot. Hamburg 23b: 4 —— ——— QU Н. E. К. HARTMANN. сы Аш» aceae — a new approach. J. Linn. Soc., Bot. 97: 239- 25 CAROLIN, R. C. 1987. A review of the amily Portu- lacaceae. Austral. J. Bot. 35: 384-41 . Portulacaceae. /n K. Kubitzki cir, Families and Genera of Vascular Plants, Volume 2. Springer Verlag, Heidelberg. As press.) CRONQUIST, A. 1981. An Integrated System of Classi- fication of Flowering Plants. Columbia Univ. Press, 1988. The Evolution and Classification of Flowering Plants, 2nd edition. New York Bot. Gard., Bronx, New ы CULLEN, D. С. A. S. DE. 1953. Las especies Argentinas del enero Calandrinia. Bol. Soc. Argent. Bot. 5: 29. DUDLEY, W.R. 1899. Report of the plants of Guadalupe Island. Pp. 280-283 in D. S. Jordan (editor), The Fur Seals and Fur-seal Islands of the Northern Pacific Ocean, Part 3. Gov. Printing Off., Washington, D.C. FRANZ, E. 1908. Beitrage zur Kenntis Portulaceen und Basellaceen. we Jahrb. a (Beibl. 97) 42: 1-46. HERSHKOVITZ, hylogenetic studies in yT a brief lo Taxon 38: 602- 08. ÉL Nomenclatural changes in Portula- Ма ы. 68: 267-270 990b. Phylogenetic and Morphological Stud- ies in ie Ph.D. Dissertation. University of California, Davis. 1991a. More Centrospermae I: the portula- caceous ae Amer. J. Bot. 78(6, pt. 2): 191. 91b. Taxonomic notes on Cistanthe, Cal- NEA and Talinum (Portulacaceae). Phytologia 09-225 1991c. Leaf ОА ~ ^ istanthe Spach. jin Bot. Gard. 78: 60. . In press a. | Пе А апа sectional taxonomy of Calandrinia Kunth (Portu- lacaceae). Ann. Missouri Bot. Gard Ann. Š. Volume 78, Number 4 1991 Hershkovitz Assessment and Circumscription of Cistanthe 1021 s b. Leaf morphology and taxonomic appl vais of P жаа tweedyi (nee Lewisia tweedyi; Portulacaceae). Syst. Bot n press с. Leaf morphology of Calandrinia Kunth (Portulacaceae). Ann. Missouri Bot. Gard. HiNTON, W. F. Systematics of the E A umbe latum ‘complex (Portulacaceae). Brittonia 27: 19 ; Нонм, " 2 1975. Biosystematic Studies of the Genus Lewisia, вес по n Cotyledon (Portulacaceae). Ph.D. е sity of Washington, Seattle, . A. 1973. Pollen Morphology and и а in the Genus Calandrinia H.B.K. (Por lacaceae). M.S. Thesis. California State University, Northridge, California LEVIN, G. A. Character кз and cladistics: a response to Rodman et al. Syst. Bot. 10: 4 500. MATHEW, B. 1989. The Genus Lewisia. Botanic Gardens, Kew, England McNEiLL, J. Synopsis of a revised classification of Portulacaceae. Taxon 23: 725-72 Munz, P. A. & D. D. Keck. 1973. A Е Нога and Supplement, combined edition. Univ. California Press, Berkeley, California. NILSSON, О. 1967. Studies in Montia L. and Claytonia L. and allied m^ III. Pollen morphology. Grana Palynol. 7: NYANYANO, B. E j 454 “Tribal and Generic Relationship and Classification of the Portulacaceae (Centrosper- mae). Ph.D. Dissertation. University of Reading, En- gland. The Royal 1990. Tribal and generic relationships in the = Portulacaceae ње Feddes Repert. Већ. 101: -24 F. & K. EM 1934. Portulacaceae. Pp. 234-262 in A. Engler & K. Prantl (editors), Die Natürlichen Pflanzenfamilien, 2. Auflage, Band 16c. Engelmann, Berlin. POELLNITZ, К. VON. 1934. Die о Arten Aus- traliens. Feddes Repert. Beih. Raven, P. & D. A. AXELROD. Eu pisa and rela- tionships of ph California flora. Univ. Calif. Publ. ot. 72: 1- Весне, K. i Zur Systematik der Chilenischen Arten der actui Calandrinia. Ber. Deutsch. Bot. Ges. P Pax, 898. Fin de Chile, Volume 2. Univ. Chile, Santiago. RODMAN, И E. 1985. Reply to Levin: "Honest effort was better than none at all." Syst. Bot. 10: 500- 03. 1990. Centrospermae revisited, Part I. Taxon 39: ce 393. ‚ M. К. OLIVER, В. К. Nakamura, J. U. Mc CLAMMER & A. H. BLEDSOE. 1984. A taxonomic analysis and revised ран of Centrospermae. Syst. Bot. 9: 2 RYDBERG, P. А. 1932. Lora Pp. 2 36 in R. Britton et al. (editors), North — Flora, Volume 21. New York Bot. Gard., Bronx, New Yor SPACH, Е. Histoire Naturelle des Végétaux: Pha- Pdl ape Volume 5. Librairie encyclopédique de Roret, Paris. W С INS, I. L Flora of EU California. Stanford Jniv. Press, Stanford, California LEAF MORPHOLOGY OF Mark A. Hershkovitz? CISTANTHE SPACH (PORTULACACEA E)' ABSTRACT Leaves of Cistanthe are typic ally characterized by broad, more or less clasping leaf bases, winged Зен n broc hudodrameus venation, apically Áo primary veins, sinuous veins, "'ribbonlike" veins, and pre nantly 'ytic and similar stomata. This icular combination of о does not occur in any о iter taxa most of the traits occur in individual specie Rumicastrum. Leaf morphology is similar in м sections of Cistanthe except кш the monotypic sect. Е which lacks sinuous veins and ribbonlike Individual species may, however, possess distinctive foliar traits or combinations of traits. Overall, the leaf ыйы cal evidence а кү hypothesis of monophyly of Cistanthe, but does not substantially clarify the phylogenetic position of this genus, nor relationships therein. Leaf characters may have utility for additional phylogenetic study. Although the systematic foliar morphology of the dicotyledons remains poorly surveyed com- pared to reproductive morphology, its potential for providing phylogenetic information is becoming in- creasingly appreciated (see, e.g., Gifford & Foster, 1989). Instrumental in the apparent resurgence of foliar morphological study was the refinement of the classification of foliar features, especially for venation pattern (Hickey, 1973) and stomatal fea- tures (for review, see Baranova, 1987). When stud- ied in light of the more sophisticated terminology, leaves revealed long-overlooked features that have contributed considerably to the understanding of angiosperm origins and phylogenetic trends (Hick- ey & Wolfe, 1975). The classification of venation features, in particular, has occasioned analyses of foliar morphology in familial and subfamilial tax- onomic studies, for example, by Gillespie (1988) for Omphalea (Euphorbiaceae); Keating & Ran- drianasolo (1988) for Rhizophoraceae; Levin (1986a, b, c) for Phyllanthoideae (Euphorbiaceae); Dickison (1975) Todzia (1988 for Hedyosmum (Chloranthaceae); and others (see ~ for Cunoniaceae; Levin, a) Some studies (especially Levin, 1986c) used leaf characters in cladistic analyses, which permitted evaluation of the significance of foliar evolution during the course of phylogeny. Despite the relative technical simplicity of leaf morphological investigation, systematic studies of Portulacaceae have heretofore made limited use of The first detailed апа extensive survey of foliar features in the family was by Becker (оћаг features. (1895), who studied leaf histology, stomatal mor- phology, and trichomes in more than 30 species. The systematic significance of trichomes in Рог- tulacaceae was subsequently considered by Chorin- sky (1931), Franz (1908), Pax & Hoffmann (1934), and Reiche (1897, 1898). Later, Kelley (1973; see 1987, in Hershkovitz, 1990a, 991с) surveyed essentially all species of Calan- Carolin, press; drinia s.l. for trichome morphology. McNeill (1975), in his numerical phenetic analysis of the tribe Montieae, incorporated gross leaf and epi- dermal features in the database. Hershkovitz (1986) presented a preliminary assessment of leaf venation patterns and their potential taxonomic significance in Portulacaceae. Nyanyano (1986a, b, 1988) de- scribed stomata, trichomes, and leaf bundle sheath anatomy in ca. 100 species of Portulacaceae. Car- olin (1987), in his phylogenetic analysis of Por- tulacaceae, included three trichome characters in the database. Otherwise, foliar features have been used primarily in monographs, revisions, and flo- ristic treatments of various Portulacaceae, e.g., by Award, a Smithsonian Pre-doctoral Fellowship, and t California, Davis. I than DAV, F, JEPS, K, L, M, MO, NA, P, RSA, U, " UC. У: gratefully ac knowledged: J. А. Doyle, R. H. E Now icke, G. F Russell, R. Schmid, J. L. ны С. L. Wel s manuscript was adapted from the second of five that constitute a doctoral dissertation in botany complete avel awards from the Department of Botany, University of ВМ, CAS the curators of the following ae for loans and/or access to collections: B, AS, aried forms of assistance Ps Ms urge: individuals are , B. J. McCaskill, J. W. * Department of Botany, University of California, Davis, California 95616, U.S.A ANN. Missouni Bor. GARD. 78: 1022-1060. 1991. Volume 78, Number 4 1991 Hershkovitz 1023 Leaf Morphology of Cistanthe Nilsson (1966, 1967, 1970, 1971a, b) for Mon- tieae, Reiche (1898) for Chilean Calandrinia s.l., and Rydberg (1932) for several genera. For various reasons, the use of foliar characters in these studies did not significantly advance the understanding of taxonomic relationships among Portulacaceae. For example, the most distinctive of the observed traits, multicellular trichomes and "Kkranz" vascular bundles, have a highly restricted incidence in the family and occur in species that were already regarded as closely interrelated (Kel- ley, 1973; Nyanyano, 1986a, 1988; Reiche, 1897, 1898). Carolin (1987), however, misscored some OTUs for trichome characters, which may have contributed, at least partially, to some spurious results (Hershkovitz, 1990a, in press a, in press b). The stomatal data (Nyanyano, 1986a, b) was interpreted according to Metcalfe & Chalk's (1950) simplistic scherne that recognizes few basic sto- matal types. All but Carolin's (1987) study were executed using “pre-cladistic”” taxonomic con- cepts. As a result, the monophyly of existing su- praspecific taxon circumscriptions was not seri- ously questioned, and interpretations of variation did not incorporate the concepts of character po- larity and parsimony. Thus, my earlier interpre- tations of leaf venation pattern evolution in Por- tulacaceae (Hershkovitz, 1986) reflected not only inadequate sampling, but also unnatural generic and tribal circumscriptions. The monographic and floristic considerations of foliar characters empha- sized primarily species delimitations rather than interrelationships. The present work, which stems from my prelim- inary survey of leaf venation patterns in Portula- caceae and represents a revision of a chapter of my dissertation (Hershkovitz, 1990a), provides a systematic characterization of gross leaf morphol- ogy, leaf venation pattern, and leaf epidermal mor- phology of Cistanthe Spach. Cistanthe consoli- dates ca. 47 species of temperate western North America and South America that had formerly been classified in as many as five genera and four tribes (Hershkovitz, 1990a, b, 1991c). Five sections of Cistanthe are recognized here (see Hershkovitz, 1990a, b, 1991c; cf. McNeill, 1974), including C. sect. Cistanthe, C. sect. Amarantoideae (Reiche Carolin ex Hershkovitz, C. sect. Philippiamra (Kuntze) dva C. sect. Calyptridium ids in Torrey & ay) Hershkovitz, and C. Strophiolum (B. Mathew) Hershkovitz. га (1987), in his cladistic analysis of Portulacaceae, determined that the first four of these formed a monophyletic grouping (Fig. 1) evidenced by the nearly universally shared presence of unequal in- florescence bracts. Cistanthe sect. was later cladistically associated with this group rather than Lewisia, based in part on its possession of unequal inflorescence bracts (Fig. 2; Hershko- vitz, 1990a, in press c). Elsewhere, I have described the leaf morphology of sect. Strophiolum (Hersh- kovitz, 1990a, in press c) and discussed relation- Strophiolum ship among the sections of Cistanthe and of this genus to other Portulacaceae (Hershkovitz, 1990а, 19910). Тһе purpose of such a detailed consideration of leaf morphology in Cistanthe is severalfold. First, this study seeks to establish whether the circum- scription of the genus, itself representing a radical departure from pre-Carolin (1987) taxonomies of Portulacaceae, dence. Second, by defining leaf characters and evaluating character states in Cistanthe, this study contributes to the database that can be used for resolving phylogenetic relationships within the ge- nus and of this genus to other Portulacaceae. Third, the data presented in this study contribute to the existing body of information on foliar morphological receives support from leaf evi- phenomena and is thus potentially useful for studies of foliar morphological evolution, of the relationship between leaf form and function, and of leaf mor- phogenesis. MATERIALS AND METHODS 1. GROSS LEAF MORPHOLOGY Gross leaf morphological variation of Cistanthe was surveyed primarily from herbarium pea and described according to the terminology p posed by Dilcher (1974) and Hickey (1971, 1973, 979). Figures 3-21 illustrate the range of vari- ation in gross leaf morphology in Cistanthe; Fig- ures 22-38 illustrate the gross morphology of the leaf apices. These illustrations are arranged ac- cording to sectional taxonomy. Table 1 lists vouch- ers, putative identifications, and full taxonomic ci- tations of all specimens examined for anatomical features. Table 1 also lists additional representative specimens of North American Cistanthe examined for gross leaf morphology. 1 have provided else- where a listing of ca. 175 pres pla collections — and putative determinations of ~ merican Cistanthe (Hershkovitz, 1991а) aud of specimens of Cistanthe (Strophiolum) tweed yi examined for leaf morphology (Hershkovitz, 1990a, in press c). 2. LEAF VENATION PATTERN Leaf venation was examined in ca. 250 cleared leaves representing essentially all species of Cis- tanthe and all leaf sizes and shapes in the genus. 1024 Annals of the Missouri Botanical Garden BLE l. Specimens examined for leaf morphology. Listed below by section are specimens of Cistanthe that were examined for leaf morphology, including vouchers of specimens illustrated and/or cited in the present paper. Additional specimens of South American species examined only for gross leaf morphology are listed elsewhere (see Hershkovitz, 199 1a), as are specimens of Cistanthe sect. Strophiolum (see Hershkovitz, in press c). Representative specimens of North American species examined only for gross leaf morphology are denoted below with an asterisk (*). Specimens examined for epidermal bic шш but not venation pattern are denoted with a dagger (T). The remaining specimens were sampled for leaf venation stud- ies and some also for epidermal morphology (see Table 2). Identifications are revised from Hershkovitz (1 990a) and remain tentative (see Hershkovitz, 1991). Cistanthe sect. Amarantoideae (Reiche) Carolin ex Hershkovitz, Phytologia 68: 269. 1990 (“Атагап- toides ая ambisu 1a (5. Mes Carolin ex Sid shko- z, Phytologia 68: 1990. U.S.A. CALIFORNIA: M “E Jones s.n. (l N du 9982 (UC); а & Nelson 3287 (US). Cistanthe calycina T Carolin ex Hershkovitz, Phytologia 70: 220. 1991. CHILE. ANTOFAGASTA: ия 3590 (US); Johnson 5318 (US). ATACAMA: Werdermann 418 (US). Cistanthe densiflora (Barnéoud in Gay) Hershkovitz, Phytologia 70: 220. 1991. ARGENTINA. SAN JUAN Cabrera 29553 (1 15): Castellanos 15520 (US). Cistanthe salsoloides Seige in Gay) Carolin ex Hershkovitz, Phytologia 70: 221. 1991. CHILE. ANTOFAGASTA: TM d 1048 (leg. Francke; F, US). Cistanthe sect. nC ed ie (Nutt. in Torrey & A. Gray) Hershkovitz, Phytologia 68: 267. 1990. — ар (Nun. a MARS & A. Gr sin itz, Phytologia 67. 1990. Моа к: Wiggins 8155 (Г 5). U. S.A. ARIZONA: Har- ка 1477 7 (05); Тоитеу s.n. (US). CALIFORNIA: Heller Cistanthe monosperma (E. v Hershkovitz, Phy- tologia 68: 267. 1990. U.S.A. CALIFORNIA: Elmer 3733 (US); Heller 10804 (US C i parryi (A. Gray) Hershkovitz, Phytologia 68: 268. 1990. Ме ү parryi var. arizonica M T. Howell) d Gandhi, Phytolog ia 71: 62. 1991. U.S.A. А NA: Griffiths 35 US). Cistanthe parryi var. hessae (J. H. E ч & Gandhi, Phytologia T 62. 1991. IFORNIA: Thomas & Ernst 6001 (1 ш Cistanthe parryi var. нај СА Т. с Gandhi, Phytologia 71: 62. 1991. U.S.A DA: Beatley 5 о к C чо parry var. i. U.S.A. CALIFORNIA: Munz 5726 (UC)*; P 3081 (UC)*; Parrish 3725 05); Peirson 3124 (UC)* ее pulchella (Eastw.) о Phytologia 8: 268. 19 S.A. CALIFORNIA: Congdon s.n. (US); Hamon 80- ү (UC)*; "n 80-194 (UC)*; Hoover 3442 (US). Cistanthe pygmaea (Parish ex Rydberg) Hershkovitz, ~ TABLE 1. Continued. Phytologia 68: 268. U.S.A. CALIFORNIA: Howell 17427 (CAS)*; Parish 1803 (holotype, CAS)*; Twis- selmann 16891 (CAS)*. C ео quadripetala (S. Seid | iue tologia 68: 268. 1990. U.S.A. CALIF¢ : Bak 3075 n уў; Hoover 3571 (US); Parish 3082 (0 S); Sharsmith 4345 (U Cistanthe rosea (S. Watson) и Phytologia 68: 268. 1990. U.S S.A. CALIFORNIA: Alexander « Kellogg 4352 (US ode 2805 (US); Reveal & Reveal 424 (UC)*. oREGON: Cusick 2585 (US Cistanthe umbellata (Torrey г) Hershkovitz, DE ( 990. U.S.A. CALIFORNIA: Heller 12062 (US) Jones 2460 (US). мене ми 11351 (US) Coville & Applegate 422 (US). 2 pa Cistanthe sect. Crane Phytologia 70: 211. 1991. d aa hing! 'agen- knecht 18444 (Е, UC). RUBLE: Joseph 3990 (US). Cistanthe cephalophora (1. M. Johnston) Carolin ex ershkovitz, Phytologia 70: d 1991. CHILE ANTOFAGASTA: Werdermann 85. Cistanthe coquimbensis (Barnéoud i in Gay) Carolin ex Hershkovitz, Phytologia 70: 212. 1991. CHILE. COQUIMBO: Werdermann 881 (Е). Cistanthe cymosa Mini Li Hershkovitz, Phytologia 70: 213. 1991. CHILE. ANTOFAGASTA: Werdermann 853 (05); Worth & p iae 15816 (NA Cistanthe fenzlii (Barnéoud in Gay) Carolin ex Hersh- kovitz, Phytologia 70: 213. 1991. CHILE. вю-вю: Neger s.n. (M); Philippi s.n. (B). Cistanthe grandiflora (Lindley) Carolin ex Herebkovitz, Phytologia 68: 269. 1990. CHILE. ACONCAGUA: Mor- rison et al. 16872 (NA); West 3959 (F, UC, US). ANTOFAGASTA: Worth & Morrison 16133 (МА). ATA- САМА: Werdermann 405 (К, UC). вю-вю: Hutch- inson 234 (UC, US). eae ee Zollner 10284 (NA). U.S.A. cuLTIVATED: Peele ) Cistanthe guadalupensis (Dudley i in D. Jordan) Carolin ex Hershkovitz, Phytologia 68: 269. 1990. MExico. GUADALUPE IS.: уе 2635 (UC)*; Moran 5991 (US)*; Wiggins & Ernst 174 (UC). Cistanthe eh Lo Par Lopez & Pavón) Neid nita hytologia 70: 214. 1991. PERU. ANCASH: Ferreyra = 532 (US) LA LIBERTAD: López Miranda. 374 (US). : Ferreyra 10486 (US). И longisca apa (Barnéoud in Gay) Carolin ex nis shkovitz, Phytologia 70: 215. 1991. CHILE. АТА- : Johnston 5034 (US ); Werdermann 445 (Е). Cite maritima (Nutt. in Torrey and A. Gra Carolin ex Hershkovitz, Phytologia 68: 269. 1990. MEXICO. ESTADO DE BAJA CALIFORNIA: Bacigalupi 3045 (ОС); Webster 21615 (DAV)*; Wiggins & Ernst 207 (UC Cistanthe paniculata (Ruiz Lopez & Pavón) Carolin ex Hershkovitz, Phytologia 70: 216. 1991. Peru. AREQUIPA: Ferreyra 12022 (US). Cistanthe picta (Gillies ex Arn. in Cheek) Carolin ex He e var. picta, Phytologia 70: 217. 1991. CHILE. ACONCAGUA: Hutchinson 98 (US). ATACAMA: prone 6218 (US). METROPOLITANA: Kuntze s.n. US); Morrison et al. 16786 (МА). O'HIGGINS: o. ne ll 12279 ). Cistanthe picta var. frigida (Barnéoud in Gay) Hersh- Volume 78, Number 4 1991 Hershkovitz Leaf Morphology of Cistanthe 1025 TABLE l. Continued. kovitz, Phytologia i Ap 1991. CHILE. COQUIMBO: Morrison et al. 16992 (NA). Cistanthe w b eee [^ Carolin ex Hershkovitz, Phytologia 70: reyra 12006 (US). 1 LIMA: Cistant e sp. cf. C. arenaria. CHILE. COQUIMBO: Zoller 10636 (NA: the smaller-leav specimens of this ا‎ ч tion—see Hershkovitz, 1991a). МЕТ- ANA: Joseph 2785 (US— see Hershkovitz, la " Cistanthe sp. cf. C. longiscapa. CHILE. ATACAMa: Worth & Morrison 16184 (NA; see Hershkovitz, 19913). Cistanthe sp. CHILE. COQUIMBO: Zollner 9807 (NA; see Hershkovitz, 1991a 9] O % ЕН CaF = ES 3 Мл МАЈ N — — — Cistanthe sect. Philippiamra (Kuntze) Hershkovitz, Phy- tologia 68: 269. 90. C Seni i celostoides о Carolin ex Hershkovitz, Phytologia 68: 269. ¿HILE. ANTOFAGASTA: ve 613 (UC); RD 862 (UC, US); m est 3859 (UC); Worth & Morrison 15820 (UC). САМА: Werdermann 477 (US). Leaves of herbarium (rarely ethanol-preserved) specimens were cleared by successive treatments with Fisher Aerosol OT, ca. 10% aqueous sodium hydroxide, and 200% aqueous chloral hydrate, and stained with 1% The leaves were destained as necessary in ca. 70% basic fuchsin in. 10045 ethanol. ethanol, dehydrated, and mounted in standard me- dia (e.g., Fisher Permount) using appropriately sized glass slides and cover slips. Prior to mounting, selected specimens were counterstained in ca. 0.1% fast green in 1:1 absolute ethanol: xylene. Gross venation patterns illustrated in Figures 3-54 were photographed by inserting the cleared leaf speci- men into a photographic enlarger, sometimes with a green acetate filter, and projecting the image on photographic paper. Highly magnified venation de- tails were photographed using a compound micro- scope and brightfield optics sometimes enhanced with variable degrees of Nomarski interference The optimal preparatory regimen and photographic technique varies with the taxon, and additional information mav be obtained upon request. Venation is described according to the Hickey system" (Dilcher, 1974; Hickey, 1971, 1973, 1979; Hickey & Wolfe, 1975). I follow Levin (1986a), however, in consistently referring to brochidodromous and higher-order “loops,” rather than "arches" and loops (cf. Hickey & Wolfe, 1975). The Hickey system was undoubtedly in- spired by, and best adapted to, leaves of taxa having large numbers of “iterations” of architectural fea- tures, especially those of woody perennials that bear numerous (hundreds to many thousands of) leaves having more or less uniform morphology, and in which particular venation features are re- peated many times in each leaf. Members of Cis- tanthe, however, like most Portulacaceae, are her- baceous perennials and annuals, bearing few (sometimes only 10-20) leaves per season or life- time, and the major leaf venation features iterate relatively few times. The paucity of leaves per plant can increase the degree of morphological variance in the sample because of heteroblasty. Likewise, the paucity of regularly iterated venation features within a leaf results in greater morphological vari- ance among the iterations. As a result, attempting to discern a leaf architectural “mode” for a species of Cistanthe is difficult. Nevertheless, the Hickey system, because of its classification of a large num- ber of venation details, provides a useful framework for the present discussion. 3. EPIDERMAL MORPHOLOGY Epidermal morphology was examined in more than 37 specimens representing ca. 20 species or species complexes (i.e., including segregate species elsewhere recognized; see Hershkovitz, 199 а) of Cistanthe. Emphasis in this study is on the mor- phology of the stomatal complex on the abaxial intercostal region. Figures 80-94 illustrate the epi- Table 2 provides brief descriptions for all examined spec- dermal morphology of selected species. imens, which are vouchered in Table 1. Epidermal peels were obtained from herbarium specimens by treating leaf fragments successively with ca. 10% sodium hydroxide, distilled water, 30% hydrogen peroxide, КАА, and 50% ethanol. This harsh pro- cedure readily separates the abaxial epidermal lay- er from the mesophyll, although often part or all of the adaxial epidermis is separated as well. Epi- dermal tissue from the abaxial intercostal region was studied on wet-mounts and subsequently pre- served in 50% ethanol. Epidermal peels were рћо- tographed using brightfield optics with variable de- grees o omarski interference. The optimal preparatory regimen and photographic technique varies with the taxon, and additional information may be obtained upon request. RESULTS 1. GROSS LEAF MORPHOLOGY Leaves of Cistanthe are simple, entire, and somewhat to quite succulent, although the precise degree of succulence could not be determined from herbarium material The typical leaf area in different species varies 1026 Annals of the аи Botanical Garden from less than 10 mm? (C. picta var. frigida, Fig. 55; and C. pulchella), to over 3,000 mm? (С. grandiflora; C. paniculata, Vig. 3; C. tweedyi, see Hershkovitz, 1990а, in press с). According to the leaf area classes defined by Dilcher (1974), the leaf area in Cistanthe ranges from microphyllous to mesophyllous. The leaf shape in species of Cistanthe ranges from approximately wide obovate (Figs. 3-6), 14, 19), t essentially linear and/or terete (Figs. 13, 18). The through narrow oblanceolate (Figs. 12, lamina may be essentially straight and symmetrical (Figs. 8, 9, slightly curved to S-shaped (Fig. 10). Leaf sym- 20) to more often asymmetrical, and metry usually varies in a species or individual. The shape of the leaf apex may be acute (Figs. 8, 9, 27), acuminate (Fig. 37), to rounded (Figs. 25, 26, 34), and weakly emarginate (Fig. 17) to weakly mucronate (Fig. 24). Characterization of leaf apex shape is sometimes relative to the point of reference; for example, the leaf shown for С. weberbaueri (Figs. 10, 28) is more or less acu- minate toward the apex, but rounded at the very ip. Petiolar development varies in the genus. Typ- ically an individual will have distinctly petiolate and essentially sessile leaves, along with leaves of in- termediate morphology. Leaf bases in petiolate leaves constrict more abruptly (Fig. 6) than grad- ually. Such leaves appear to be differentiated into an elliptical, ovate, orbicular, rhombic, or lanceo- late blade portion, and a narrower, winged, petiole portion. In petiolate and sessile leaves, the leaf base broadens at the point of attachment to the stem so that it is nearly or quite clasping. Am- plexicaul leaves occur in Peruvian plants that have been referred to С. 199 Та). While the leaves illustrated here more ог less paniculata (Hershkovitz, typify the species or species complexes shown, the variability is such that putatively closely related species exhibit no obvious consistent distinctions in gross leaf morphology. Likewise, no distinctions among the sections of Cistanthe are evident, ex- cept to the degree that the more polymorphic sec- tions (1.e., those with more species) vary more than the less polymorphic. Not all species intergrade, however; leaves of Cistanthe grandiflora, C. pic- ta, C. ambigua, and C. quadripetala, for example, are readily distinguishable from one another. 2. LEAF VENATION PATTERN The venation features of Cistanthe (except for sect. Strophiolum, described and illustrated in Hershkovitz, 1990a, in press c) are selectively illustrated in Figures 3-79. These illustrations are arranged according to subject, from gross venation pattern to finer anatomical details. Within each grouping, illustrations are arranged according to sectional taxonomy. Because of the apparent variation in gross leaf morphology in species or species complexes, the the paucity of iterated venation features in individual concomitant variation in venation features, leaves, and the difficulties encountered in identi- fying material (see Hershkovitz, 1991a), useful and reliable descriptions of venation in individual spe- cies was largely unobtainable. The descriptions pre- sented here emphasize the venation features char- acteristic of Cistanthe as a whole and the distribution and range of variation of these features. Particular features encountered in only a few taxa are also noted. Venation type. Venation in Cistanthe is usu- a irregularly festooned brochidodromous (Figs. , 15-17, 20, 21, 39-46, 50, 51, 53, 54; ef. Hickey & Wolfe, 1975: 547, fig. 5), sometimes nly weakly festooned (Figs. 12, 19, 47, 52), less often irregularly brochidodromous with only a hint of festooning (Figs. 13, 14, 48, 49). (See also discussion of secondary veins, below. Primary vein. Тһе primary vein is prominent basally (but more so in some taxa than in others; e.g., compare Figs. 13, 17) and much diminished to obsolete apically (Figs. 22-38). vein may be evident as a protruding veinlet at the leaf tip (Figs. 24, 27, 29) or not at all evident at 25, 26). Hickey’s (1973) parameter for primary vein size— The primary the apex (Figs. 2 relative width midway be- tween the blade base and apex— cannot be satis- factorily evaluated in Cistanthe, because the blade is not readily definable (see above). The primary vein course varies, but usually shows a slight zig- zag or angular shift in course away from the de- parting secondary veins (Figs. 6, 15, 55). (See also comments on general vein course, below.) Secondary veins. Prominent secondary veins are generally few— three to four pairs per leaf— and their number is apparently not strongly cor- | 55). Гће basalmost secondary veins are often less prom- related with leaf size (e.g., compare Figs. 6, inent and more decurrent or more acutely angled than the apical secondaries. The secondary veins generally change course abruptly and form loops with suprajacent secondary veins. These secondary loops are often situated one-half to two-thirds the distance from the primary vein to the leaf margin in broader leaves (Figs. 3-6) and proportionally closer to the leaf margin in narrower leaves (Fig. 10). Sometimes secondary loops fail to form or form only weakly in leaves otherwise possessing Volume 78, Number 4 1991 Hershkovitz Leaf Morphology of Cistanthe 1027 such loops (Figs. 15-17). Tertiary loops are barely evident in the narrowest leaves (Figs. 48, 49) but are present, sometimes along with 4? and 5? loops, in broader leaves (Figs. 39-46, 50-54) Intersecondary veins. Simple intersecondary veins are sometimes present in the intercostal regions, especially in broader leaves (Figs. 4-7, 10, 11, 20, : Tertiary veins. Tertiary veins are irregularly developed, variable in size, and, because of inter- gradation in prominence between lower and higher vein orders, sometimes difficult to identify through- out the leaf. Those arising from the apical side of the secondary veins are generally more prominent than those arising from the basal side. The branch angle varies but is generally more acute on the apical side of the secondary veins than the basal. The tertiary vein pattern is ramified to reticulate but sometimes varies in different parts of the same leaf. Rarely, the tertiary veins are nearly percur- rent (Fig. 16 Higher-order veins (excluding freely ending the leaves, the highest vein order is essentially equiv- alent to the tertiary veins (Figs. 47-49, 55). In the largest leaves, a fourth order of veins is present. Fourth-order veins are not always readily distin- guishable as such throughout the lamina because ramifications of the tertiary veins can be highly variable in prominence. Fourth-order veins, ten- tatively identified as such in Figures 39-43, 45, 46, 50, 51, 53, 54, are usually ramified or rarely more orthogonally reticulate, as in C. longiscapa (Fig. 40). Fifth-order veins are rarely distinguish- able as such, e.g., in С. paniculata (Fig. 3 longiscapa (Fig. 40), and С. monosperma (Fig. 51). veinlets). In smallest and/or narrowest Marginal venation. The marginal venation ranges from incomplete (Figs. 42, 44) to somewhat looped (Fig. 39) but is usually intermediate between these forms. Freely ending veinlets. The freely ending veinlets are highly variable in length and degree of branching (Figs. 56-67) and may be promi- nently dilated with respect to the penultimate veins (Figs. 64-66) cr hardly dilated at all (Fig. 60). Тће veinlets are sometimes distally coalesced with adjacent veinlets and/or veinlet branches (Figs. 56, 61-63). The terminal tracheary elements of the dilated vein'ets are usually numerous, short, and densely clustered. Similarly short tracheary elements (““tracheoids””) also occur rarely along the highest-order veins (Fig. 56). The terminal ele- ments of the less dilated veinlets are more elongate and less densely clustered (Figs. 60, 67). Areoles. — Areoles are difficult to define in Cis- tanthe because the vein orders are not well dif- ferentiated, but they are probably best character- ized as imperfect to incomplete, and least well developed in the smaller or narrower leaves (Figs. 48, 49, 55). ge y illustration of “incomplete” areolation (1979: : ences in venation T Such differences are 1 7) shows local differ- generally not present in larger leaves of Cistanthe, їп which the venation is more or less uniformly dense, although the shape and size of ultimate enclosures in a leaf are highly variable (Figs. 56- 61). The number of veinlets per ultimate enclosure is typically 0-2 but may be more in the elongate enclosures that adjoin major veins. Other venation features a. Vein course. The veins are sinuous in most or all vein orders in most species, although indi- viduals of some species may have less sinuous or straight veins, especially in C. umbellata (Fig. 16), С. monosperma (Figs. 17, 51), C. tweed yi (Hersh- kovitz, 1990a, in press c), and in some leaves and/ or higher vein orders of C. picta (Fig. 58). Sinuous venation is possibly an artifact of specimen shrink- age during drying or preparation, but it neverthe- less represents a diagnostic trait because similarly prepared leaves of other succulent-leaved members of Portulacaceae do not have sinuous veins (e.g., Lewisia spp.; Hershkovitz, in press c). b. Three-dimensional venation. A few species of Cistanthe have veins that interconnect in more than one plane, i.e., have three-dimensional ve- nation. Three-dimensional venation occurs in leaves that are especially succulent but not necessarily more succulent than leaves with two-dimensional (planar) venation. In the C. grandiflora complex 1991a), the finer veins in the central and basal portion of the leaf blade form a (see Hershkovitz, complex three-dimensional network (Fig. 68). In C. paniculata, the fine veins in the central and basal portion of the leaf blade may form a planar reticulum adaxial to the plane of the major veins (Fig. 69). In the cylindrical leaves of C. ambigua, the primary vein occurs toward the adaxial surface, and the departing veins follow the contour of the leaf toward the abaxial surface (Figs. 18, 35, 70). Toward the leaf apex, the marginal venation in- terconnects abaxially, forming а basketlike retic- ulum. Gibson (1982) estimated the density of venation by dividing the number of vas- c. Vein density. cular bundles present in a leaf cross section by the length of the section, and by measuring the distance between bundles. He observed that vein density 15 much lower in succulent leaves of desert perennials 1028 Annals of the Missouri Botanical Garden than in nonsucculents. Vein density in members of Cistanthe, which are succulent, was here estimated by counting the number of veins crossing arbitrarily delimited transects of the photographs in Figures 39-54 and dividing by the length of the transect, yielding data equivalent to that obtained by Gib- son’s method. Cistanthe grandiflora (Fig. 44) was not measured because of its extensive three-di- mensional venation. The density values averaged slightly less than 5 veins/mm and were mostly lower than the 7 veins/mm minimum reported by Gibson for nonsucculent desert perennials. The highest value observed in Cistanthe was ca. 9.5 veins/mm in the small, linear leaves of C. fenzlii (Fig. 49). The lowest values, са. 2.5 veins/ mm, were found in C. paniculata (Fig. 39) and C. quadripetala (Fig. 50). Mostly helical but also annular, scalariform, and reticulate d. Tracheary element wall pitting. wall thickenings occur (see also below). Scalariform to reticulate thickenings predominate in vessel el- ements in the primary (and sometimes higher-or- der) veins of some species (e.g., C. lingulata, Fig. 7l, and C. Fig. 72), whereas helical thickenings predominate in the prominent veins of ambigua, others. e. Fein anatomy. Тһе finer veins in most spe- cies of Cistanthe are flat, or “ribbonlike” (Figs. 73-79). These veins are often sinuous and usually one cell-layer thick and up to eleven (perhaps more ~ tracheary elements wide (Fig. 76). The elements on one side have annular wall thickenings, and those toward the other have progressively less steep helical and, in some cases, scalariform (Fig. 75) to nearly reticulate (Fig. 74) thickenings. Some of the more prominent veins may also be ribbonlike but are more than one cell-layer thick and approach the "normal" condition in vascular plants, in which the tracheary elements are arranged in clusters rather than ribbons and have more or less the same type of secondary wall thickening throughout the vein. Ribbonlike veins appear to be common and characteristic. of essentially all the fine veins in some taxa (e.g., C. quadripetala, Fig. 60) or rel- atively rare or absent among the fine veins of others (e.g., C. picta, Fig. 58, and С. Fig. 61). They are absent in sect. Strophiolum (Hersh- kovitz, 1990а, in press c). Ribbonlike leaf veins apparently have not been described in any other umbellata, vascular plants. 3. VARIABILITY IN VENATION FEATURES A thorough analysis of the infraspecific variation in venation features of putative species of Cis- tanthe was not undertaken in the present inves- tigation because of poorly defined species limits and the large amount of herbarium material neede for such a study. The limited sampling, however, suggested constancy of venation features in some taxa. This is demonstrated by a comparison of large and small leaves of C. picta (Figs. 6, 55), C. grandiflora (Figs. 8, 9), and the C. umbellata species complex (Figs. 16, 17; see Hinton, 1975). The illustrations show that the larger and smaller leaves of each pair resemble each other more than either does similar-sized leaves of other species (although no other leaves as small as that shown in Fig. 55 are illustrated). The leaves of C. grandiflora appear distinguish- able by the co-occurrence of the following traits: (1) a small protrusion of the weakened primary vein at the apex (Fig. 27); (2) relatively fine sec- ondary veins and vein loops, the latter forming proximal to the primary vein; (3) a relatively nar- row — < 30°) secondary vein angle; (4) prominently dilated terminal veinlets, often juxtaposed with ex- tremely fine penultimate veins; and (5) three-di- mensionally reticulating veins in the central and basal portion of the leaf. Leaves of members of the C. umbellata complex are distinguishable on the basis of: (1) primary veins that are particularly prominent toward the leaf base and nearly to quite obsolete at the apex; (2) prom- inent secondary veins, some of which are promi- nently looped while others are ramified and not looped; (3) numerous festooning loops; (4) rela- tively well developed third and fourth vein orders; and (5) a paucity of sinuous veins and ribbonlike veins. The venation of the leaf of C. picta shown in Figure 55 does not closely resemble that of the larger counterpart (Fig. 6). Nevertheless, the for- mer appears to resemble the latter more than it does any other leaf shown in Figures 3-21 because of: (1) the massiveness of the primary vein toward the base and its obsolescence at the apex; (2) the zig-zagging character of the primary vein; (3) the relatively broad angle of the secondary veins; and 4) the paucity of sinuous and ribbonlike veins. Note that venation in C. picta is similar to that in the C. umbellata complex but has less regularly festooned loops and apparently broader secondary vein angles. Preliminary measurements also indi- -~ cate a greater vein density in C. picta (6-7 veins/ mm) than in c ЕВ leaves in the С. umbellata complex (3.5-5.5 veins/mm). In all of the cases described above, the smaller leaves differ from the larger leaves primarily in having lower “rank” (Hickey, 1971); i.e., the high- er vein orders and, concomitantly, higher orders Volume 78, Number 4 1991 Hershkovitz 1029 Leaf Morphology of Cistanthe of festooning loops are less or not distinguishable in the smaller leaves. 4. EPIDERMAL MORPHOLOGY Nyanyano (1986a, b) referred to stomata in Cistanthe sects. Calyp- a. Stomatal morphology. tridium and Philippiamra as paracytic. Nyanyano (1986a) illustrated paracytic stomata їп C. (Cis- tanthe) grandiflora; otherwise his results for Cis- tanthe sects. Cistanthe and Amarantoideae can- not be determined because he did not report results for individual species, and he included these sec- tions of Cistanthe in Calandrinia s.l., in which both paracytic and tetracytic stomata were re- ported. Contrary to what Nyanyano (1986a, b) claimed, his characterizations of stomatal types correspond to tae simplistic classification proposed by Metcalfe & Chalk (1950), rather than the more precise classification proposed by Wilkinson (1979), in which different types of paracytic (e.g., brachy- paracytic) stomata are distinguished. Becker (1895) also reported the presence of two lateral subsidiary cells (= paracytic s.l.) in sect. Calyptridium. The principal morphological features of stomata in Cistanthe are illustrated for several species in Figures 80-94 and diagrammed in Figure 95. Ta- ble 2 summarizes stomatal morphology for several specimens vouchered in Table 1. The stomata are described below following Wilkinson's (1979) ter- minology except as noted. e most common stomatal type found in Cis- tanthe is brachyparacytic (Fig. 958), in which the guard cells are flanked by a pair of subsidiary cells but not enclosed by them. The morphology of the lateral subsidiary cells varies: the shape ranges from reniform (i.e., contoured to the shape of the guard cell; Fig. 951, J) to more rectangular (Fig. 95L, M) to polygonal (Fig. 950, P). The width of the subsidiary cells ranges from relatively narrow (Fig. 951, L, О) to broad (Fig. 95Ј, M, P). The broader cells are intermediate between distinct sub- sidiary cells and ordinary epidermal cells. A similar range of variation exists when two subsidiary cells are adjacent to a guard cell (Fig. 95K, М, 0), although the outer subsidiary cell 15 generally broader than the inner. The opposing subsidiary cells in a brachyparacytic stomatal complex may differ in morphology with respect to the variants noted in Figure 95. To some degree, the variation described above might reflect the optical plane of section. More likely, however, a three-dimensional shape analysis will reveal an even greater degree of variation in subsidiary cell morphology. Infraspecific variation in stomatal type, not re- ported by Nyanyano (1986a, b), was found in all specimens, although some specimens were more variable than others. Aside from brachyparacytic, the variants in stomatal type include (cf. Wilkinson, 1979): hemiparacytic (Fig. 95A; only one of the guard cells with a distinct subsidiary cell); amphi- brachyparacytic (Fig. 95D; two pairs of flanking subsidiary cells); hemi-amphibrachyparacytic (type proposed herein; Fig. 95C; a total of three lateral subsidiary cells); tetracytic (Fig. 95E; two pairs of subsidiary cells, one pair parallel and one pair per- pendicular to the long axis of the guard cells); hexacytic (Fig. 95F; similar to the preceding but with two pairs of lateral subsidiary cells); stauro- cytic or anomotetracytic (Fig. 956; three to five subsidiary cells surrounding the guard cells); and anomocytic (Fig. 95H; no distinct subsidiary cells present). In addition, variants more or less intermediate between pairs of the above listed types and having no formal names occur. In one relatively common variant, one lateral subsidiary cell (or two adjacent cells) is divided transversely (Fig. 95R, S). In rare instances, both flanking cells are divided. Although Wilkinson (1979: 100) noted that this type should not be confused with laterocytic stomata, no sharp structural distinction can be made. Such stomata in Cistanthe are here arbitrarily designated brachy- paracytic with one or both lateral cells split. Cis- tanthe stomata are thus distinguished from those typifying certain Lewisia species, in which the guard cells are flanked by four to eight subsidiary cells (Hershkovitz, 1990a, in press c). In another variation, an elongate subsidiary cell perpendicular to the guard cells occurs at one pole but not the other (Fig. 95T). When such cells occur at both poles, tetracytic and hexacytic stomata may be formed. A character of the stomatal complex that cor- relates with morphological type is the number of epidermal cells, whether defined as subsidiary cells or not, that directly border the guard-cell pair (“contact cells," cf. Esau, 1977: 94). This char- acter is useful independently of stomatal type be- cause the distinction between subsidiary cells and ordinary epidermal cells is often arbitrary. Because most examined species of Cistanthe have predom- inantly brachyparacytic stomata (or structurally similar types shown in Fig. 95A-F), the number of contact cells is usually four: two lateral subsid- iary cells plus one subsidiary or ordinary epidermal cell at each pole. Occasionally, two ordinary epi- dermal cells border the guard-cell pair at one ог, exceptionally, both poles, so that the number of contact cells is five or, less often, six. It seems potentially possible for more than six contact cells to occur in Cistanthe (e.g., if both flanking cells 1030 Annals of the Missouri Botanical Garden TABLE 2. а к, in some species of Cistanthe. Listed below аге stomatal types and other epidermal features found on th xial intercostal area of selected species and specimens of Cistanthe. The species are listed according to section, en the vouc cher are cited in Ta ble 1. о е of sect. ње is described elsewhere (Hershkovitz, in press c e stomatal types, described in the text, are listed, with the suffix “-cytic” дин The frequency of the к types in the specimens аге e denoted as follows: ++, common to dicii type; +, present and recurring throughout the specimen; (+), observed but rare and not recurring thro e specimen. The evaluations are subjective and intended only to кие the general pattern of variation in ds genus. Hemi- Hemi- amphi- Amphi- Section brachy- Brachy- — brachy- — brachy- Species (voucher) Anomo- para- para- para- para- Tetra- sect. Amarantoideae C. ambigua (Nelson & Nelson 3287) + + + + + z calycina (Werdermann 4 16) + + ` densiflora (Cabrera 29553) + + Hd : densiflora (Castellanos 15520) + ++ ++ C. salsoloides (Werdermann 1048) + + + + + sect. Calyptridium C. monandra (Harris 1477) 4 4 4 C. monosperma (Elmer 3733) + + + + + + C. monosperma (Heller 10804) + + + SESS $ C. parryi var. parryi (Beatley 5732) (+) ++ + + (+) C. parryi var. parryt (Parrish 3725) ++ ++ C. pulchella (Congdon s.n.) + ++ ++ C. quadripetala (Hoover 357 4 C. quadripetala (Sharsmith 4345) + + ++ + С. rosea (Duran 2805) ++ ++ С. umbellata (Abrams 11351) + + + + Тарт + С. umbellata (Heller 12062) + ++ ++ С. umbellata (Jones 2460) (+) + + ++ (+) sect. Cistanthe C. cymosa (Werdermann 853) + + + + С. cymosa (Worth & Morrison 15816) + + + + + + + С. fenzlii (Neger s.n.) (+) + + + + + C. grandiflora (Hutchinson 234) + cae + + С. grandiflora (Morrison et al. 16872) + + + + + C. grandiflora (Peele 154) (+?) (+) ++ d+ А grandiflora (Zollner 10284) + + + + Б | D (Ferreyra 10480) (+) + + d а с longiscapa (Johnston 5034) + + + + + С. al (Ferreyra 12022) ++ Е ++ C. picta var. picta (Kuntze s.n.) + + + + + j C. picta var. picta (Morrison et al. 16786) + + + + + d Cistanthe sp. cf. longiscapa (Worth & Morrison 16184) + + + + Cistanthe sp. (Zollner 9807) ++ T (*) "E sect. Philippiamra C. celosioides (Werdermann 477) + + + + + + C. celosioides (Werdermann 862) + + + + ++ Моште 78, Митбег 4 1991 Hershkovitz Leaf Morphology of Cistanthe 1031 TABLE 2. Continued. Split lateral З соп- subsi tact Epidermal Hexa- Stauro- iary cell cells cell outline Comments + orthogonal + sinuous sinuous poor specimen orthogonal orthogonal single polar subsidiary cell sometimes present; papillar cells along leaf margin + orthogonal small sample + orthogonal single polar subsidiary cell sometimes present; stomata sunken + + + orthogonal stomata highly variable; some papillar cells; stomata sunken (+) orthogonal papillar cells present; stomata sunken + * orthogonal to arcuate few papillar cells present; stomata sunken orthogonal to sinuous few papillar cells present; stomata и sunken T small sample; stomata sunker (+) orthogonal stomata sunken orthogonal papillar cells present; stomata sunken + + T orthogonal stomata highly variable; stomata sunken (*) (+) + sinuous to orthogonal o p subsidiary cell sometimes present; papillar near midri esent; stomata sunken orthogonal papillar cell present; stomata sunken orthogonal papillar cells present; some stomata sunken orthogonal + + + orthogonal to sinuous stomata highly variable; papillar cells present along leaf margin orthogonal + (+) (+) orthogonal to arcuate some polar subsidiary cells present (+) orthogonal to sinuous (+) orthogonal stomata somewhat sunken + (+) sinuous other odd stomatal variants | (+) (+) sinuous other stomatal variants pre orthogonal stomata somewhat sunken; pen sample + + (+) sinuous stomata highly variable T + (+) sinuous stomata highly variable (*) orthogonal large imd cells present + + (+) somewhat sinuous other stomatal variants present; small papillar ceils orthogonal orthogonal fice leaf margin 1032 Annals of the Missouri Botanical Garden TaBLE 3. are cited in Table 1. For all measurements, (N) — the number of stomata sampled. Stomatal index, stomatal density, and mean stomatal length in selected members of Cistanthe. Vouchers All measurements are based on photographs. Stomat index Stomata/ Stoma length (um) Section Species (voucher) (N) mm? (N) [x = SD (N)] sect. 4marantoideae С. ambigua (Nelson & Nelson 3287) 3.6(22) 100(37) 21.9 + 2.2 (20) C. salsoloides (Werdermann 1048) Um (22 69(24) 24.8 + 2.4 (20) sect. Calyptridium C. parryi var. parryi (Beatley 5732) 13.7(44) 144(101) 27.3 € 4.2 (20) C. quadripetala пне 4345) 26.7 (34) 157(55) 26.4 + 2.5 (20) С. umbellata (Abrams 21) 9.5 (23) 240(48) 23.8 + 2.0 (20) sect. Cistanthe C. cymosa (Werdermann i 13.5 (36) 60(53) 17.8 + 2.3 (17) C. cymosa (Worth & Morrison 15816) 15.6 (45) 233 (63) 15.8 + 1.1 (20) C. fenzlu (Neger s.n.) 10.7 (52) 203(78) 25.1 + 1.9 (20) C. grandiflora (Werdermann 405) 11.1 (22) 30(24) 35.1 + 3.5 (20) C. grandiflora и th & Morrison 16133) 8.8(29) 23(36) 35.6 + 3.3 (20) C. picta var. picta (Kuntze s.n.) 12.3(33) 250(45) 19.0 + 1.0 (20) C. lingulata (Ferreyra 10486) 13.3 (42) 203(78) 30.1 + 1.8 (20) C. sp. cf. longiscapa (Worth & Morrison 13.3(20) 115(33) 15.6 + 1.9 (20) 16184 C. sect. Philippiamra C. celosioides (И erdermann 477) 15.2(12) 139(12) 22.7 + 1.69 (11) C. celosioides (Werdermann 862) 12.4 (11) 111 (15) 17.9 + 1.48 (9) are divided and more than one ordinary epidermal — radically surveyed in Cistanthe (Table 3). The cell borders the guard-cell pair at one or both poles), but no more than six cells bordering the guard-cell pair were observed, and this number is rare. While all examined material of Cistanthe pos- sessed some variation in stomatal morphology, the pattern and degree of variation differ among the taxa (see Table 2). The least variation occurred in (Fig. 80), in which amphi- and hemi-amphibrachyparacytic members of the C. grandiflora complex stomata predominate, while other types are rare. Relatively little variation also characterizes C. lin- gulata (Fig. 85), С. salsoloides (Fig. 90), and С. celosioides (Fig. 94). Greater variation occurs in C. picta (Figs. 82, 83), C. fenzlii (Fig. 84), and C. umbellata (Fig. 92). High proportions of hemi- — brachyparacytic stomata (e.g., C. cymosa, Fig. 88 or anomocytic stomata (e.g., С. quadripetala, Fig. are uncommon in the genus. "Double" sto- mata, in which two guard-cell pairs adjoin, were observed in С. cymosa and С. quadripetala (Figs. 88, 91). Sunken stomata occur in several taxa (Table 2) and are often variably present in an epidermal specimen. Sunken stomata are most prominent and frequent in sect. Calyptridium, e.g., С. rosea (Fig. 93) and C. umbellata (Fig. 92). Sunken stomata are illustrated in C. grandiflora (Fig. 80), C. lon- giscapa (Fig. 81), and C. cymosa (Fig. 88). Biometric stomatal parameters (stomatal index, stomatal density, and stomatal size) were only spo- preliminary data suggest only limited taxonomic The stomatal index values are similar to those noted for other angio- sperms (cf. Wilkinson, 1979). The lower stomatal densities observed are typical of those found in utility of these parameters. other succulent-leaved species (cf. Gibson, 1982). Likewise, the stomatal lengths are in the “normal” range found in other plants (cf. Wilkinson, 1979). The stomatal index in C. quadripetala is greater than in other species of Cistanthe, but the stomatal density and length in C. remarkable. The high stomatal index in this case apparently reflects the high incidence of anomo- quadripetala are not cytic stomata іп C. quadripetala; i.e., the total number of subsidiary cells will correlate with the ratio of guard cells to subsidiary and ordinary epi- dermal cells. The correlation between low stomatal density and high stomatal length in the C. gran- diflora complex has been reported for other xero- phytic herbs (Wilkinson, 1979). ther biometric parameters, including the vari- ation in stomatal index or density on the abaxial versus adaxial surface of the leaf and angle of orientation of the guard cell axis relative to the leaf axis, were not measured. Leaves of Cistanthe, like all Portulacaceae, are amphistomatic (cf. Nyanyano, 1986a, b), and the stomatal index and / or density might be expected to be lower on the adaxial surface, although superficial observations of a few adaxial epidermal layers did not indicate Volume 78, Number 4 Hershkovitz 1033 Leaf Morphology of Cistanthe that the difference was marked. In the intercostal region, the guard cell axis appears to be oriented randomly (see Figs. 80-94). No average or modal angle of divergence of the guard-cell axis was cal- culated here. Toward the leaf midrib, however, the guard-cell axis appeared to be more commonly oriented in the direction of the leaf axis. b. Stomatal ontogeny. Ontogeny of the sto- matal complex was not studied, nor was the precise sequence of cell division generally inferable from the mature stomatal morphology, because it is known that particular structural types of stomata are derivable from different ontogenetic pathways (Farooqui, 1961a, b; Rasmussen, 1981). Super- TOS however, ontogeny of the stomatal c omplex appears to vary in Cistanthe. For example, in C. | aede (Fig. 90), the structure of the stomatal complex suggests mesogenous development, whereas the subsidiary cells of C. 9) appear to be mesoperigenous or perigenous. (Payne (1979) doubted whether stomata can be truly perigenous (i.e., ambigua (Fig. the guard cells developing directly from an otherwise undifferentiated proto- dermal cell), but this hypothesis does not seem to require that any or all structurally distinct subsid- iary cells be derived mesogenously.) There appears to be no reason why stomatal ontogeny could not vary in a species. Indeed, variation in mature sto- matal morphology in individual leaves of Cistanthe suggests that this may be the case. c. Other epidermal cell features. outline of ordinary epidermal cells in the leaf in- Тће surface tercostal region in Cistanthe ranges from orthog- onal and nearly isodiametric in surface view (Figs. 80, 86, 90, 92-94 sinuous radial walls (Figs. 81, 84, 85). The entire ) to irregular in shape with range of variation occurs in individuals of some taxa, e.g., in C. picta (Figs. 82, 83). As might be expected, ordinary epidermal cells toward or over- lying the midrib tend to be more elongate and quadrangular, presumably reflecting the pattern established by the young, rapidly elongating leaf primordium. Ordinary epidermal cells of some Cistanthe spe- cies are more or less dimorphic, with large cells interposed among the smaller cells associated (or intergrading) with subsidiary cells (Figs. 89, 90). n extreme cases, the large cells form epidermal alveolae or papillae, such as in a specimen referable to Cistanthe longiscapa (Worth & Morrison 16184, see Table 1, Figs. 86, 87) and in species of sect. Calyptridium (Table 2), including C. rosea (Fig. 93). The epidermis of C. also appears somewhat papillate. Alveoli and/or “Cal- = cymosa (Fig. papillae have been reported previously in andrinia lamprosperma 1. M. Johnston” (aff. Cis- tanthe grandiflora; Hershkovitz, 1991a; John- ston, ; but see also Kelley, 1973), in sect. Calyptridium (Hinton, 1975), and in C. lophora (1. M. Johnston) Carolin ex Hershkovitz (Kelley, 1973). The presence of papillate /alveolate cepha- cells in species of sect. Calyptridium varies among individuals (Hinton, 1975; pers. obs.). DISCUSSION |. LEAF MORPHOLOGY AND THE CIRCUMSCRIPTION OF CISTANTHE Table 4 outlines a syndrome of nine leaf traits that typify Cistanthe and summarizes the distri- bution of these features in this genus and elsewhere in Portulacaceae. Leaf morphological data and il- lustrations for species of Lewisia and Calandrinia 1990a, in press b, in press с), and the data for the remaining are presented elsewhere (Hershkovitz, Portulacaceae will be presented in future publi- cations. The taxonomic circumscriptions used in Table 4 are based on those proposed by Carolin (1987, in press), except for Cistanthe, which fol- lows the present paper, and for Calandrinia, which follows Hershkovitz (1990a, in press a) As indicated in Table 4, typifying Cistanthe are not universally present in most of the leaf traits the genus, and all occur elsewhere in the family. Hence, none can be designated a priori as a syn- apomorphy diagnostic of the genus, nor can any of their evolutionary polarities in the genus be determined a priori. Table 4 also indicates, how- ever, that no other genus of Portulacaceae pos- sesses the entire Cistanthe leaf-trait syndrome. Moreover, many of the traits of Cistanthe are rare in the other genera in which they occur. Therefore, the leaf morphology of Cistanthe appears to cor- roborate other morphological evidence for the nat- uralness of the genus as circumscribed here — kovitz, 1990a, 1991c). That no distincti sect. Calyptridium and the remainder of Еви etween are discernable from Table 4 should not be over- looked. Leaf morphology provides no evidence in support of Carolin's (1987, in press) retention of sect. Calyptridium as a distinct genus. 2. LEAF MORPHOLOGY AND THE RELATIONSHIP OF CISTANTHE ТО OTHER PORTULACACEAE Leaves of Cistanthe share features with j taxa throughout Portulacaceae (compare Table : and Fig. 2). The greatest degree of similarity ap- pears to exist between Cistanthe and species of the Australian genus Rumicastrum. The absence 1034 Annals of the Missouri Botanical Garden TABLE 4. Leaf traits of Cistanthe. Listed below are several leaf traits that occur commonly to universally among species of Cistanthe, their distribution in the genus, and their distribution elsewhere among Portulacaceae (Hershkovitz, 1991c, in press a, in press b, in press c, unpublished E Trait Distribution in Cistanthe Distribution elsewhere among Portulacaceae Winged petiole differentiated om blade Leaf base broad, nearly or quite clasping Festooned brochidodromous ve- c nation pattern Primary vein weak to obsolete at leaf apex 4? veins distinct Sinuous veins Ribbonlike veins Predominance of brachypara- a (incl. hemi-, ‘hemi’’-amphi- 95 cytic stoma КА and * riants, see Fig. Contact cells mostly 4, other- wise mostly 5, rarely 3 or 6, probably never 2 or more than 6 Rarely absent in broader-leaved spe- cies All species All species except those in which the leaves are extremely small and/or narrow Weak to obsolete in all species Distinct in several larger-leaved spe- cies throughout the genus, absent in smaller- and narrower-leaved species In nearly all examined species except tweedyi, but hardly evident in the С. umbellata complex ar and in- dividuals of th picta complex Distribution same as sinuous veins Characteristic of nearly all species In all species Common in Montieae and Calandrinia Calandrinia and Monocos- mia; rare in Lewisia spp.; in one umicastrum sp. Montieae, Calandrinia, Lewisia, Lenz- ia; otherwise apparently absent Weakly evident in broader-leaved 7ali- num spp., Ceraria spp., and Portu- laca spp.; rarely and weakly evident in Lewisia spp.; Talinella Weak but usually evident in Rumicas- trum and Talinopsis; weak to obso- lete in Portulacaria, ee Ceraria spp., Portulaca spp., Calandrinia sect. а "PP. Evident in broader leaves of Cerari spp.; rarely distinct in Болса. аз in P. lutea; weakly evident in broad leaves of Claytonia megarhiza, broader-leaved Lewisia spp., ralan- drinia sects. Acaules and тта ia spp., and Talinum sect. Talinum spp- Evident in some Ceraria spp., Tali- num spp., and Rumicastrum spp. Sica in (all?) species of Talinum . Phemeranthus but not co mon nor well-developed in Talinum sect. Talinum; in many, but not all Rumicastrum spp.; and in Talinop- sis and Grahamia Characteristic of many Montieae spp.; Calandrinia sects. e Hirsutae, and a few Acaules spp.; Lenzia; stems of Talinum sect. Ta- linum spp. but в not leaves; rare in Lewisia spp. Rumica trum spp. (see also CES 1986b) In all examined Montieae, un and Lenzia spp.; in at lea Rumicastrum and one Lana isia spp.; otherwise probably rare these genera; in Talinum sect. “Tali num stems, but probably not leaves of both festooned brochidodromous venation and strongly differentiated higher vein orders in Rumi- castrum species is not surprising because their leaves are usually small and/or relatively narrow. Most species of Rumicastrum have a type of three- dimensional venation not evident in Cistanthe (Hershkovitz, unpublished), but the planar venation patterns in other species are hardly distinct from Volume 78, Number 4 1991 Hershkovitz Leaf Morphology of Cistanthe 1035 those in the very narrow-leaved species of Cis- tanthe. The phylogenetic e of this sim- ilarity remains to be deter The large number of тине leaf traits for Cis- tanthe and Rumicastrum indicated in Table 4 may be misleading because some traits typical of the former occur only rarely in the latter and not in combination with other traits. For example, the petiolar morphology characteristic of many species of Cistanthe occurs only in Rumicastrum pumila (F. Muell.) ined. (= Calandrinia pumila F. Muell.). This species is also similar to typical Cistanthe in its stomatal morphology, but it lacks ribbonlike veins and has а primary vein that is prominently dilated at the leaf apex. Moreover, this species has pantoporate-operculate pollen, which is regarded in terms of aperture number and morphology as derived in Rumicastrum (Kelley, 1973; Carolin, 1987; Hershkovitz, 1990a, in press a). Otherwise, I have found Cistanthe-like stomatal morphology Black) ined. = Calandrinia remota 1. M. Black), a species only in Rumicestrum remota (1. M. with 12-15-pantocolpate pollen, which is the pu- tatively primitive condition in Rumicastrum (Kel- ley, 1973; Hershkovitz, 1990a, in press a; but see also Carolin, 1987). Rumicastrum remota has a three-dimensional venation pattern typical of many species of Rumicastrum and unlike three-dimen- sional venation in any species of Cistanthe. Other species of Rumicastrum possess various stomatal types having three, rather than four or five, contact cells; e.g., anisocytic (Hershkovitz, unpublished). Leaves of Cistanthe are mostly less similar to those of other genera of Portulacaceae than to those of Rumicastrum. With respect to their leaf morphological similarity to Cistanthe, most Por- tulacaceae fall into one of two categories: (1) those taxa having in common with Cistanthe characters of gross and epidermal morphology, e.g., Montieae, Calandrinia, and Lewisia; and (2) those taxa hav- ing in common with Cistanthe venation features, e.g., Talinum, Ceraria, and others listed in Table 4. These two groups are otherwise potentially dis- tinct cladistically (see Fig. 2), as well as biogeo- graphically: the first group, along with Cistanthe, comprises the genera of Portulacaceae that occur predominantly westward from the American cor- illera, and the second group comprises most of the taxa that occur predominantly eastward from the American cordillera and in southern Africa (Hershkovitz, 1990a-c, in prep.). Thus, while the leaf data presented here do not resolve the phy- logenetic position of Cistanthe, they seem to sug- gest a pivotal position of this genus in Portulaca- ceae, potentially linking the eastern. American/ African, western American, and Australian mem- bers of the family. 3. LEAF MORPHOLOGY AND RELATIONSHIPS WITHIN CISTANTHE Leaf morphology appears to have limited utility for diagnosing or assessing relationships among the sections of Cistanthe. Sections of Cistanthe are not distinguishable on the basis of leaf features, although, as noted, the monotypic sect. Strophio- lum lacks sinuous veins and ribbonlike veins. Dis- tinctions in venation pattern and epidermal mor- phology are not evident among Reiche’s (1897 1898) subdivisions of Cistanthe sect. лаик (1.e., Calandrinia sects. Cistanthe, Rosulatae, Аг- enariae, and Andinae), except to the degree that Reiche's circumscriptions of Calandrinia sects. Cistanthe and Andinae consist of the Cistanthe grandiflora and Cistanthe picta species complex- es, respectively (Hershkovitz, a). Kelley’s (1973) assignments of the extra-Chilean species of Cistanthe to Reiche’s (1897, 1898) sections of Calandrinia cannot be independently corroborated by venation and epidermal evidence. The mono- phyly of Cistanthe sect. Calyptridium might be evidenced by the pervasiveness of sunken stomata and leaf surface papillae (see Table 3), but both traits occur elsewhere in the genus. Likewise, the absence of sinuous and ribbonlike veins in Cis- tanthe sect. Strophiolum is not unique to this taxon. Nevertheless, the sectional status of sect. Strophiolum is recognized partially on the basis of these traits (Hershkovitz, 1990а, in press с). The utility of leaf morphology in elucidating species-level phylogeny in Cistanthe is limited to the degree that the variable characters show con- stancy in otherwise diagnosable taxa and are po- larizable. The utility 15, therefore, inherently re- stricted given the poorly defined species limits in Cistanthe and uncertainty regarding outgroups of 1990a, 19912). (1897) emphasized gross leaf morphology in his the genus (Hershkovitz, Reiche keys to the Chilean species of Cistanthe, but the apparent high degree of plasticity and continuous range of variation in gross leaf morphology as ob- served here contribute to the inadequacy of Reiche's keys for identifying the Chilean material. The re- sults presented above, however, provide at least some evidence that species or species complexes (e.g., Cistanthe grandiflora, C. picta, C. umbel- lata) can be distinguished by their combined leaf traits. It is possible, therefore, that multivariate morphometric analyses of leaf morphological ра- 1036 Annals of the Missouri Botanical Garden E 5 Leaf traits having a restricted occurrence in Cistanthe. Listed below are leaf traits rare in Cistanthe, their distribution in the genus (see text), and their distribution elsewhere among Portulacaceae (Hershkovitz, in press c, in press a, in press b, in press c, unpublished). Trait Distribution in Cistanthe Distribution elsewhere among Portulacaceae Leaves over 3,000 mm? б; lata complex, and C. tweedyi Fifth-order veins distinct grandiflora complex, C. panicu- C. umbellata complex; weakly evi- Talinum sect. Talinum, Lewisia cong- donii, L. cotyledon, and Claytonia spp. Absent dent in C. longiscapa and C. panic- ilata Three-dimensional venation lata complex, and С. ambigua Large proportion of C. quadripetala anomocytic stomata Papillae /alveoli on leaf surface tanthe (see text) Amplexicaul leaves C. grandiflora complex, C. panicu- Variably present in sect. Calyptri- dium and specimens of sect. Cis- Specimens of C. paniculata complex ee text) Talinum sect. Phemeranthus, nearly all Portulaca sect. Portulaca, Talinop- sis, Grahamia, Sc a Lewisia ds some, ot all Anacamp- eros spp. and е. зрр. Lewisia triphylla, L. kelloggii, and Montia spp. (see also Nyanyano, 19 ) Anacampseros spp., one species of Cal- andrinia sect. Acaules, one species of Rumicastrum Claytonia spp., but probably not homol- ogous (see text) rameters may help refine ideas on species limits in Cistanthe. A few leaf traits in certain species of Cistanthe are possibly or likely derived in the genus (Table 5): (1) leaves larger than 3,000 mm"; (2) fifth- order leaf venation; (3) three-dimensional leaf ve- nation; (4) a large proportion of anomocytic sto- mata; (5) papillate or alveolate cells throughout the leaf surface; and (6) amplexicaul leaves. There is no compelling structural or phylogenetic evidence at these traits are symplesiomorphic with similar traits in other Portulacaceae. The leaf-size char- acter is most troublesome, because leaf size varies in all taxa and, presumably, is influenced by en- vironmental factors. The size of the largest leaves exceeds that found in all The rela- tively strong differentiation of higher vein orders in Cistanthe, however, but a few scattered taxa in the family. found in a few species of Cistanthe is almost cer- tainly derived because this trait is absent elsewhere in the family. Three-dimensional venation is a com- plex characteristic because more than one mani- festation is present in Cistanthe and in other Por- tulacaceae. [n at least some cases, other members of Portulacaceae with three-dimensional venation seem to be most closely related to species with planar venation. For example, in Portulaca, three- dimensional venation is restricted to P. sect. Por- tulaca (Hershkovitz, unpublished). This section is thought to be derived in the genus (Geesink, 1969, 1987). Therefore, three-dimensional venation has probably evolved several times in the family. In Cistanthe, anomocytic stomata are common only in a species of sect. Calyptridium. This section is probably relatively derived in the genus (Hersh- kovitz, 1990a, 1991c). The leaf surface рарШае/ alveoli that occur in Cistanthe appear to be very large cells that protrude beyond the smaller epi- dermal cells; those in Anacampseros and Calan- drinia sect. Acaules are more hairlike, and those in the latter are also ribbed (Hershkovitz, 1990a, in press b). Unribbed papillae have been reported in Rumicastrum quadrivalvis (F. Muell.) ined. (— Calandrinia quadrivalvis F. Muell.; cf. Kelley, 1973) but the specimen was not vouchered and other specimens of К. quadrivalvis lack papillae (Hershkovitz, 1990a, in press b). The amplexicaul leaves that occur in C. paniculata are not ho- mologous with superficially similar “perfoliate” leaves in species of Claytonia, which are formed by the congenital fusion of two leaves at a node. Table 5 are mostly restricted to recognized species or species complexes, they provide little insight into the phy- logeny of Cistanthe. For example, epidermal pa- pillae/alveoli may represent a synapomorphy of sect. Calyptridium, but these species are presum- ably closely related in any case (Hershkovitz, 1990a, Because the characters listed in Volume 78, Number 4 1991 Hershkovitz 1037 Leaf Morphology of Cistanthe 1991c; Hinton, 1975). An especially close rela- tionship between papillate-leaved species in sect. Calyptridium and sect. Cistanthe is not supported by additional characters. Large leaves and similar three-dimensional venation may be indicative of relationship between the C. grandiflora and C. paniculata complexes. These species complexes occupy adjacent desert habitats in northern Chile and southern Peru, respectively. Both possess long- haired seeds, which are otherwise of restricted oc- currence in Cistanthe (Kelley, 1973), although glabrous seeds also occur in both complexes (Hersh- 1990a, 1991a; Johnston, 1929). Other- wise, leaves of C. paniculata (Figs. 3, 22, 39), kovitz, which have very wide, obovate laminae and rela- tively well developed vein orders, more closely re- semble leaves of the northern Chilean C. longisca- pa complex (Figs. 4, 23, 40; cf. Figs. 8, 9, 27, 44), in which long-haired seeds also occur (Hersh- 1990a, 1991a). Kelley (1973) assigned Cistanthe paniculata to Calandrinia sect. Cis- kovitz, tanthe Reiche, which includes the Cistanthe gran- diflora complex (Reiche, 1898), whereas Reiche 1898) assigned Cistanthe longiscapa to the high- ly variable Calandrinia sect. Rosulatae Reiche. 4. LEAF FORM. FUNCTION, AND EVOLUTION IN CISTANTHE While the aim of this survey of leaf morpho- logical diversity in Cistanthe was to help resolve phylogenetic and taxonomic questions, the data pose questions on the relationship of leaf form to function and evolution. Such questions cannot be answered from a survey of herbarium specimens alone, but preliminary data such as that generated here might provide a focus for future ecophysio- logical and evolutionary investigations. Only a few aspects of leaf form that have been studied with respect to function will be considered in the data presented here for Cistanthe, including gross leaf morphology, leaf venation pattern, sinuous and rib- bonlike veins, and subsidiary cell morphology. Givnish (1979, 1982, 198 eral models that relate aspects of leaf form (e.g., 4) has developed sev- leaf shape, arrangement, venation pattern) to the optimization of photosynthetic ability and the min- imization of the metabolic cost of leaf support and supply. Givnish (1979) suggested rosette-forming herbs might be expected to have obovate leaves that, collectively, form a circular photosynthetic a that mimics an optimally efficient light-cap- E. structure. Cistanthe species have predom- inantly basal and/or suprabasal leaves and usually have relatively few cauline leaves except in species that form secondary rosettes along prostrate branches. The leaves are, in accordance with Giv- nish's model, typically obovate to oblanceolate (see above). In the Cistanthe picta complex, which has more diffusely arranged cauline leaves (Hershko- vitz, 1991a; Reiche, 1898), the basal leaves are sessile and oblanceolate to obovate while the cauline leaves tend to be more petiolate with orbicular to ovate blades. The morphology of the cauline leaves might be explained with reference to Givnish’s (1979) stem-leaf packing model, which purports that stem-leaf bases are evolutionarily honed to eliminate the portion that would be shaded by other leaves. Веји ври from the predicted form of Cis- tanthe leaves (e.g., linear leaves or petiolate leaves with a small, Pian rhombic blade portion, as in the Cistanthe arenaria complex; Hershkovitz, 199 la: Reiche, 1898) might best be explained as adaptations to xeric conditions, in which the leaf area is reduced to resist overheating and reduce water loss Civnish (1979) also noted that obovate leaves of rosette-forming plants, which are supported by the ground, or especially thick cauline leaves, which are supported independently of the vascular tissue, may tend to evolve parallel venation in order to maximize the efficiency of water and nutrient sup- ply. He supposed that the pinnate venation pattern typical of dicotyledons provides an optimal support system (but a suboptimal supply system) for the cantileverlike form of cauline leaves. Superficially, it would appear that leaves of Cistanthe are sup- ported primarily by their thickness and broad, clasping leaf bases, or, in the case of rosette leaves, by other leaves and/or the ground. The primary vein in Cistanthe does not seem to be well-suited for support because of its irregular, zig-zag, or sinuous course and its relatively small size in the broader portion of the leaf blade. Yet venation in Cistanthe is basically pinnate, although the sec- ondary veins may arise at a particularly narrow angle (e.g., in C. guadalupensis, Fig. 5, a species with all т leaves; Hershkovitz, in press с; Ryd- berg, 1932). Thus, the pinnate venation patterns of Cistanthe appear to represent an exception to Givnish's (1979) model. The functional significance of the peculiar rib- bonlike veins of Cistanthe is difficult to evaluate without more detailed anatomical and develop- mental data. In the vascular bundles of primary stem tissue, the centrifugally developing sequence from protoxylem elements with annular thickenings to metaxylem elements with scalariform thicken- ings 15 thought to facilitate vascular supply during the course of stem elongation and maturation. A Annals of the 1038 Missouri Botanical Garden and guard cells. This hypothesis seems plausible in instances where the perimeter of a stomatal com- plex (i.e., the guard and apparent subsidiary cells) conforms to the dimensions of the adjacent ordi- nary epidermal cells, e.g., in Figures 80 and 81. similar function might be suspected for ribbonlike veins if the succulent leaves continue to grow for a prolonged period after becoming functional. Whether the sinuousness of the veins serves any function cannot be evaluated without prior deter- mination that this is not a preparatory artifact. The sinuousness might reflect the ability of the leaf to expand and contract under oscillating water re- glmes. The functional significance of subsidiary cell shapes and arrangements does not appear to have been extensively studied. Reviews of stomatal func- tion and cell biology (e.g., Sack, 1987; Salisbury & Ross, 1985) tend to emphasize the structure and physiology of the guard cells, while reviews of stomatal supracellular morphology (e.g., Wilkin- son, 1979) emphasize gross form and development. Payne's (1979) brief review of stomatal morphol- ogy is unusual in considering both. Reviews of stomatal function emphasize that guard cells are КЕ ed ally isolated from other leaf cells (Sac ). and that the "surrounding cells" (= contact cells) serve as а source of po- tassium ions that are transported into guard cells and affect stomatal opening (Salisbury & Ross, 1985). The morphology and spatial distribution of the contact cells undoubtedly influence this phys- iological process (Payne, 1979; Salisbury & Ross, 1985: 62, figs. 3-9), but whether extreme differ- ences in stomatal complex n hexacytic vs. anomocytic stomata) have any net effect on this process is not clear. Payne (1979) observed a correlation between mesogenous sto- matal development (presumably resulting in guard cells surrounded by distinct subsidiary cells) and the ability of the plant to withstand water stress, e.g., in succulent leaves. Leaves of Cistanthe are unquestionably succulent and presumably dry- adapted, yet mesogenous stomatal development does morphology (e.g.. not appear to be the rule, or perhaps even es- pecially common, in the genus (see above). In any case, Cistanthe (and/or other Portulacaceae in which marked differences in stomatal morphology occur in closely related taxa; Hershkovitz, in press b, in press c), might provide a natural system for investigating subsidiary cell/guard cell physiolog- ical interactions. The development of the stomatal complex has been studied extensively among angiosperms, but as with physiology, the functional significance о the various ontogenetic pathways is apparently poorly understood. Payne (1979) proposed a sort of "t function for subsidiary cells; 1.e., their development compensates for the difference in cell enlargement between ordinary epidermal The hypothesis does not, however, explain the an- omocytic condition (Fig. 91), in which the guard cells are directly adjacent to the much larger epi- dermal cells, or instances where the perimeter of the stomatal complex appears to have greater di- mensions than those of the adjacent epidermal се! (e.g., Fig. 84). Тће goal of understanding the functional or ad- aptational value of leaf morphological features ren- ders an understanding of the taxonomy and phy- logeny of the plants in question all the more important. The fact that the leaf traits of Cistanthe conform to simple functional models in some ways but not others indicates that additional factors, be they genetic or epigenetic, sometimes influence leaf morphology in the genus. The genetic factors, col- lectively, reflect evolutionary history. Correlations are evident between the leaf traits of Cistanthe and environmental or developmental factors, but one should not overlook the fact that the collection of leaf traits shared M most or des members of the the co-occurrence genus seem tí of functionally unrelated traits, e.g., unequal inflo- rescence bracts. In other words, the leaf mor- phology of Cistanthe correlates best with other evidence that members of the genus share a com- mon ancestor. Collectively, leaf traits permit mem- bers of Cistanthe to be distinguished from other members of Portulacaceae, including those that eral geographic region, occupy at least t and probably more so from other angiosperms, including those that might superficially resemble members of Cistanthe and/or occupy the same habitat. Resolution of phylogenetic relationships within Cistanthe (and Portulacaceae in general) will help clarify which traits were inherited from a more remote ancestor and which evolved more recently as members of the genus came to occupy their present habitats. LITERATURE CITED Historical iier nien of the es of sto- BARANOVA, M. 198 present TT of ар са| 1 mates. Bot. Rev. (Lancaster 3 > 1895. Beitrag zur vergle 'ichenden Anatomie der бе а е Dissertation. Universitàt нш d pir CAROLIN, R. 987. А ion a л иш Portu- lacaceae. pre ral. J. Bot. 41 Portulacaceae. [л К. E or [edito Families and Genera of Vascular Plants, Volume 2. Springer Verlag, Heidelberg. (in press da 53-7 Volume 78, Number 4 Hershkovitz 1039 1991 Leaf Morphology of Cistanthe CHORINSKY, К. 1931. Vergleichend-anatomische Un- Kunth (Portulacaceae). Ann. Missouri Bot. Gard. tersuchung der Haar- "go bei deor ua und Cactaceen. Osterr. Bot. Z. 80: 308- DicKisoN, W. С. 1975 MORE noniaceae. J. Linn. Soc., Bot. DILCHER, D. L. 1974. i to o the ¡dembificalión of angiosperm leaf remains. Bot. Rev. (Lancaster) E 1-157 Leaf 5 of Си- 275-294 Esau, К. 1977. Anatomy of the Seed Plants, 2nd P tion. John Wiley & Sons, New York. FAROOQUI, P. 198la. Ontogeny of the tricytic stoma ~ variations and modifications. Proc. Indian Acad. Sci эмр Sci.) 90: 85-89. 1981b. Ontogeny of the anomocytic stoma — variations and modifications. Proc. Indian Acad. Sci. ci.) 90: 245-252. Beitráge zur Kenntnis m js TER н Bot. Jahrb. Syst. (Beibl. 97) 42: os R. 1969. An account of the genus Portulaca 1 Indo-Australia and the Pacific (Portulacaceae). Blumea 17: 275-307. 987. Piante of classification of organisms. Pp. 91- 126 in E. F. de Vogel (editor), Manual of ory and Practice. UNESCO/MAB, Jakarta. To A. C. 1982. The anatomy of succulence. Pp. 7 in I. P. Ting & M. Gibbs (editors), Crassulacean ۴ Metabolism. Waverly Press, Baltim GIFFORD, E. M., JR. & A. S. FosTER. 1989. and Evolution of Vascular Plants, H. Freeman & Co., New York GILLESPIE, L. J. 1988. A Revision and Phylogenetic Analysis of Отрћа!еа a (Euphorbiaceae). Ph.D. Dis- sertation. University of California, Davis, California. GIVNISH, T. J. 1979. On the adaptive significance of leaf form. Pp. 375-407 in O. T. Solbrig et al. (ed- itors), Topics in Plant Population Biology. Columbia Univ. Press, New York. 982 Мог наа 3rd edition. W. n the adaptive significance of leaf height in forest herbs. Amer. Naturalist 120: 353- 381. 1984. Leaf and пе кдын in tropical forests. Pp. 51-84 i Medina editors), Physiological Ecology of Plants of Du Wet Tropics. Dr. Junk, The Hague HERSHKOVITZ, M. А. 1986. Notes on leaf architecture and phylogeny in the Portulacaceae. Amer. J. Bot 73: 766. [Abstract Оа. Phylogenetic and Morphological Stud- ies in Portulacaceae. Ph.D. Dissertation. University of California, Davis, California. | rá 2m in Portula- P e o 68: 26 — la. с Е notes on Cistanthe, Cal- ema and Talinum (Portulacaceae). Phytologia 25. lb. More Centrospermae I: the portula- caceous alliance. Amer. J. Bot. 78(6, pt. 2): 191. [Abstract 199]c. Phylogenetic assessment and revised circumscription of Cistanthe Spach (Portulacaceae). Ann. Missouri Bot. Ga n press a sed circur mscription and sectional taxonomy of — R Kunth (Portu- lacaceae). Алп. Missouri Bot In press b. Leaf каше udin of Calandrinia 1 press c. Leaf morphology and taxonomic analysis of Cistanthe tweedyi (nee Lewisia tweedyt; oe Syst. Bot. Hickey, L. J. 71. Evolutionary significance of leaf sicot rie ur in the woody dicots. Amer. Bot. 58: 469. [Abstract.] Classification of the architecture of Две айнан leaves. Amer. Ј. Bot. 60: 17-33. 979. A revised Vea cia of the archi- tecture of dicotyledonous leaves. Pp. 25-39 in C. R. Metcalfe & L. Chalk (editors). Anatomy of the Dicotyledons, 2nd edition, Volume 1. C larendon Press, Oxford. —— & . WOLFE. 1975. The bases of angio- sperm phylogeny: ш morphology. Ann. Mis- souri Bot. Gard. 538-589. HINTON, W.F Systematics of the y dbi umbe pr complex (Portulacaceae). Brittonia 27: тиннин. T M. 929. МЕ on the flora of northern Chile. Contr. Gray Herb. n.s. 85: 1-164. C. & V. RANDRIANASOLO. 1988. The con- tribution of leaf architecture and wood anatomy to the classification of the Rhizophoraceae and Aniso- phyllaceae. Ann. Missouri Bot. Gard. 75: 1368. KELLEY, W. 1973. Pollen Morphology and Rela- NUM. in the Genus Calandrinia H.B.K. (Portu- lacaceae). M.S. | California State University, Northridge, _ aliforn Levin, С. A. a. Sumus foliar morphology of Pina apre 'ae). 1. Conspectus. Ann. -85. Missouri Bot. 986b. pn ated foliar morphology of Phyl- А by due II. Pheneti Ann. Missouri Bot 73: 86- c analysis. Systematic foliar morphology of Phyl- тод (Exphorbiaces e). III. Cla Syst. Bot. 11: 515-530. McNrnL, J. 1974. Synopsis of a revised classification of Portulacaceae. Taxon 23: 28. 1975. А generic revision of Portulacaceae tribe Montieae using techniques of numerical tax- onomy. Canad. J. Bot. 53: 789-809. METCALFE, С. R. & L. CHALK. 1950. Anatomy of the Dicotyledons. Clarendon Press, Oxford. NILSSON, O. 66. Studies in Montia L. ien Cla yto- nera, Mona distic analysis 725- 72 Shades in Молда L. and Claytonia Lis sud allied Borer III. Pollen morphology. Grana Palynol. 7: 279-363. => 1970. Studies in Montia L. and Claytonia L. and allied genera. e genus Crunocallis Rydb. Bot. Not. 123: 119-148. : e in Montia L. and Claytonia L. and allied genera. The genus rey) коза Вог. Мог. e 87-121. Studies in Montia L. and Claytonia і, ы allied genera. VI. The genera Limnalsine eg and Maxia О. Nilss. Bot. Not. 124: 187 Montiastrum AE B. L. 1986a. Tribal and Generic Relation- ship and Classification of the Portulacaceae (Cen- trospermae). Ph.D. Dissertation. University of Read- ing, England 1040 Annals of the Missouri Botanical Garden = Taxonomic significance of the sto- — complex in the Portulacaceae. Feddes Repert. Be h. 97: 763-766 1988. Leaf anatomical studies in the Portu- laca aceae (Centrospermae) with regards to o thetic pathway. Folia Geobot. Phytotax. 23: 99-101 15. 1 P S. Pax, К. Pg 1934. Portulacaceae. Pp. 234- 262 in А. er & K. Prantl (editors), Die Natürlichen PRanzenfamilien, 2. Auflage, Band 16c Engelmann, Berlin. PAYNE, W. W. 1979, Stomatal patterns in embryo- phytes: i Puro ontogeny, and interpretation. Tax 117 Как H : кее and o of stomata and stomatal development — a critical sur vey. J. Linn. Soc., Bot. 83: 212 REICHE, К. 1897. Zur Systematik de. Chilenischen Arten der E. Calandrinia. Ber. Deutsch. Bot Ges. 15: 493-503 1898. Flora de Chile, Volume 2. Univ. Chile, Santiago. RYDBERG, P. А. 1932. Portulacaceae. Pp. 279-336 in R. Britton et al. (editors), North American Flora, V olume 21. New York Bot. Gard., Bronx, New York Sack, F. D. 1987. The puru es and structure of stomata. Zie eger et al. (editors), Р. е ш EC Stanford Univ. Press, Stanford, . W. Ross. 1985. Plant Physi- E 3rd idis Wadsworth Publ., Belmont, Cal- or nia io C. А. 988. Chloranthaceae: Hedyosmum. Fl. Ne e eA 48. New York Bot. Gard., Bronx, ics. "н: 1979. Тће | surface (mainly leaf). Pp. 97- 165 in С. R. . Chalk (editors), Anatomy of the Dic cotyledons, : Volume 1. Clarendon Press, Oxford. 2nd edition; Volume 78, Number 4 Hershkov 1041 1991 Leaf e BN of Cistanthe Lewisia Calandrinia $ Acaules Calandrinia $ r wakna Calandrinia $ Hi Rumicastrum Calandrinia $ Calandnnia —Ü Calandrinia $ Monocosmia Portulacaria Сегапа Calyptrotheca Talinum Schreiteria Portulaca о Talinopsis alinaria Anacampseros Claytonia Montia Lenzia Cistanthe $ Calyptridium Cistanthe $ Cistanthe Cistanthe $ Amarantoideae 0 Cistanthe $ Philippiamra REl. Simplified reproduction of Carolin's (1987: 402, fig. 7) most pu cladogram of Portulacaceae, Ec the relationships among major clades and among members of Cistanthe. The taxonomy follows Carolin press), except for Cistanthe and Calandrinia, which follows Palito (1990a, 1991c, in press a). The | deco supporting the branching structure shown are problematic (see Hershkovitz, 1990a, 1991c) 1042 Annals of the Missouri Botanical Garden Australia AE Rumicastrum [ex AAPP ي‎ НЕВА) Calyptrotheca Talinum Schreiteria Portulaca Talinopsis e. America/ amara Africa Anacampseros Lewisia Calandrinia $ Dianthoideae Calandrinia $ Hirsutae Calandrinia $ Acaules Calandrinia $ Сајапапта Сајапапта $ Monocosmia { Claytonia Monta A A A ы A A A ae | AS Cistanthe $ Strophiolum Cistanthe $ Cistanthe Cistanthe $ Amarantoideae Cistanthe $ Calyptridium Cistanthe $ Philippiamra Lenzia (2) w. America IGURE Revised cladogram of Portulacaceae emphasizing relationships among the sections of Cistanthe vis Неон. 1990а, 1991c) and showing relationships of Calandrinia sects. Calandrinia and Monocosmia I hav proposed elsewhere (see Hershkovitz, 1990a, in press а). The taxonomy ne Carolin (1987, in press), except for Cistanthe and Calandrinia, which follows Hershkovitz (1990a, 1991c, in press a). The boxes enclosing portions of the diagram circumscribe the regions of endemism or greatest endemism of the included taxa (Carolin, 1987; Hershkovitz, 1990a, 1991c, in press a, in prep.). — FIGURES 3-55. Gross venation features in Cistanthe. All vouchers are listed in Table 1. x specimens in Figs. 22-54 are the same as those in Figs. 3-21 and are not reidentified except where necessary. l, primary vein, 2, secondary vein, etc.; 12, secondary loop, L*, tertiary loop, etc.; IS, ens vein; X, the е position of the constriction beiwean the blade iud winged petiole. Volume 78, Number 4 Hershkovitz Leaf Morphology of Cistanthe FIGURES 3, 4. Cleared whole leaves of Cistanthe sect. Cistanthe. —3. C. paniculata (Ferreyra 12022).— 4. С. pe ue nston 50. 34). The репоје, ca. half as long as the blade, is not shown. Scale bars are (mm) ca. 6.2 nd 3.0, respe ~ FIGURES 5-7. Cleared whole leaves of Cistanthe sect. Cistanthe. —5. C. guadalupensis (Wiggins & Ernst 174). — 6. C. picta var. picta (Kuntze s.n.).— 1. C. coquimbensis (Werdermann 881). Scale bars are (mm) ca. 4.6 2.0, and 2.5, respectively. Annals of the Missouri Botanical Garden FIGURES 8-12. Cleared whole leaves of Cistanthe sect. Cistanthe. —8. C. grandiflora (Werdermann 405). C. grandiflora (Morrison et al. 16872).—10. C. weberbaueri (Weberbauer ders 11. C. sp. cf. arenaria (Zollnar 10636). —12. C. lingulata (Lopez 374). Scale bars are (mm) ca. 4.5, 2.8, 3.3, 3.0, and 2.1, respectively. Fig URES 13-17. Cle и whole leaves of Cistanthe un Cistanthe (13, 14) and Calyptridium (15-17 sp. (Zollner 9807). . C. fenzlii (Philippi s.n.). — 15. C. о e (Hoov er 3571). — 16. C. monosperma пе ller 10804). The pe Я ca. as long as the blade, is not show ` umbellata (Jones 2460). Scale bars are (mm) ca. 2.8, 1.6, 3.4, 3.8, and 2.2, respectively. Volume 78, Number 4 Hershkov 1045 1991 Leaf Ait dad of Cistanthe 21)- 18. С. ambigua (Nelson & Nelso 87). —19. C. calycina (Johnston 5318).— 20. C. salsoloid s (We aer 1048).— 21. C. celosioides (W ен Fg pia 15820). Scale bars are (mm) ca. 3.0, 1.3, 2.4, and L5 , respectively. FIGURES 18-21. Cleared n l1. of Cistanthe sects. Amarantoideae (18-20) and dot curia ~ LIA 294 Uem F N ~ 4 DICEN SESS ME $e TN ОҢ S RN У LI У ES > YER Spe. * FIGURES 22-27. С Леагеа le af apic es of Cistanthe sect. Cistanthe. C. guadalupens 25. C. picta var. d 26. C 16872). Scale bars are j firn oa a. 2.7, 1.2 . C. panic ulata. — 23. C. longisc apa. 24. : Wan ву nsis. — 27. C. grandiflora а et al. , 0.8, 0.8, and 0.8, respectively. NA € мі 3-38. Cleared leaf apices of Cistanthe sects. oe (33, 34), Amarantoideae (35-37), and Philippiamra aa 33. C. quadripetala. —34. C. monosperma. . C. ambigua. — 36. C. calycina. — 37. С salsoloides. — 38. C. celosioides. Scale bars are (mm) са. 0.9, 1.0, . А T 7, 0.8, and 0.6, respec tively. FIGURES DO Cleared leaf central- marginal portions of Cistanthe sect. Cistanthe. —39. C. paniculata. In Figures 39 and 40, the €— and fifth- gii veins are difficult to discern, but their existence can be inferred by е эры cm the tertiary veins. The fourth-order veins form the reticulum adjacent to the tertiary Mm bs: he fifth-order veins traverse pd areoles oe by the fourth-order veins and give rise to the free- vi veir 0. C. longiscapa. —41. C. guadalupensis. CAM . picta var. picta. —43. C. coquimbensis. — 44. Maec b erdermann 405). Scale bars are (mm) ca. , 1.4, 1.7, 0.9, 0.8, add 1.6, respectively Volume 78, Number 4 Hershkovitz 1991 Leaf Morphology of Cistanthe FIGURES 50-54. Cleared leaf central-marginal portions of Cistanthe sects. Calyptridium (50, 51), Amarantoideae (52, 53), and Philippiamra (54).—50. C. quadripetala. —51. C. monosperma. —52. C. calycina. —53. C. sal. soloides. — 54. C. celosioides. Scale bars are (mm) ca. 0.8, 1.5, 0.6, 0.8, and 0.6, respectively. — FIGURE 55. Cleared whole leaf of Cistanthe (Cistanthe) picta var. frigida (Morrison et al. 16992). Scale bar = ca. 1 mm. FIGURES 56-59. Fine venation features in Cistanthe. – . Ultimate venation in C. Seo mida каги чейн (West 3959). Note the “tracheoids” (T) along the fine veins | the coalescence of veinlet branches (large a 57. Ultimate venation and coalesced veinlets (arrows) in C. (Cistanthe) coquimbensis (Werdermann "88 . — 58. Ultimate venation in C. (Cistanthe) picta var. picta (Hutchinson 08) Note the lack of sinuous veins. — 59. Ultimate venation in C. (Philippiamra) celosioides (Worth & Morrison 15820). Scale bars = ca. 200 um. FIGURES 60-65. Fine venation features in Cistanthe. —60. Ultimate venation in C. pans Lir) MT (Baker 3075). Note ribbonlike venation (arrows) and undilated veinlets. — 61. Ultimate venation in alyptridium) umbellata (Jones 2460). Note the lack of sinuous veins.— 62. Terminal, coalescent veinlets in C. (Cistanthe) guadalupensis (Wiggins & Ernst 174). —63. Terminal, coalescent енин іп C. (Cistanthe) grandiflora (Werder- mann 405). Note the numerous densely clustered short tracheary elements. — 64. Branched veinlet in С. (Cistanthe) grandiflora (Werdermann 405). Note the fineness of the vein supplying En dilated veinlet (arrow). — 65 е dilated veinlet of C. (Cistanthe) picta var. picta (Pennell 12279). Scale bars are (um) 200, 200, 200, 100, 10 ly. and 50, respectively 1054 Annals of th Missouri ESI Garden FIGURES 66-70. Fine venation features in Cistanthe. —66. Dilated veinlets in C. (Cistanthe) fenzlii (Philippi s.n.). Note the fineness of the proximal end of the veinlet and the penultimate vein (arrows). — 67. Branched veinle of С. (. ни ae) ambigua (Nelson & Nelson 3287). Note the lesser degree of veinlet dilation and more e elongat terminal tracheary elements relative to other species. — 68. Three- Lie ridus venation in C. (Cistanthe) grandiflora (Werdermann 4 35), The veins reticulate freely in more than one plan 69. Three-dimensional venation in C. (Cistanthe) paniculata (Ferreyra 12022). The plane of focus is on the more prominent, lower-order veins, and the finer, higher-order veins form a reticulum in an adaxial plane. — 70. Three-dimensional venation in C. (Amarantoideae) "ылы (Nelson & Nelson 3287). The finer, higher- order veins form a reticulum in a plane abaxial to the primary 1) and secondary (2) veins. Scale bars are (um) 100, 100, 200, 200, and 200, respectively. ПРИ У УДАТА "ts. FIGURES 71-72. Primary veins of Cistanthe. — 71. C. (Cistanthe) lingulata (Lopez 374) showing helical thick- E of the vessel element walls toward the adaxial side and scalariform to reticulate thickenings toward the abaxial. 2. (Ат ат ка (Nelson & Nelson 3287) showing alae thickenings of the vessel осика walls, Scale bars — Volume 78, Number 4 Hershkovitz 1055 1991 Leaf Morphology of Cistanthe Ме?) 5-в[ jg | | 1 | | j ў | ! 0 FIGURES 73-79. Ribbonlike veins in Cistanthe. — 73. C. (Cistanthe) ко (Weberbauer 5321), showing vein branching, sinuousness of the veins, and a free- sap ain dien „Мара дела) ambigua (Nelson & Nelson 3287) showing sinuousness of ки vein. — 75. C.( ingi ipd (Lopez 374). — 76. C. (Сайера шы ker 3075) showing ribbonlike vein 11 e nin elements wide.— 77. C. (Amarantoideae) ambigua (Nelson & Nelson 3287) showing scalariform to reticulate vessel element ie thickenings on the abaxial side of the | aug — 78. C. el T celosioides (Wort h & Morrison 15820).— 79. С. (Cistanthe) На bus rder- ma 05). A == annular wall thickenings; Н = helical wall thickenings; s — scalariform wall thicken | pide to не wall thickenings. Scale bars are (um) 1,000, 50, 50, 50, 50, 50, апа 50, mai E 1056 Annals of the Missouri Botanical Garden => Figures 80-94. Epidermal morphology in Cistanthe. мена id where noted, refer to specific features diagrammed in Figure 95 evident in particular stomatal complexe RES 80-85. Epidermal morphology in Cistanthe sect. ee All scale bars = са. 100 um.— 80. C. grandiflora (Werdermann 405). The plane of focus is subsur —the stomata are so dm. sunken: a, more or less hexacytic stoma; 2, hemi-amphibrachyparacytic stoma; 3, ое Were i stoma.— 81. C. longiscapa (Johnston 5034). The stomata are slightly sunken. 1, brachypa ПАК stoma with fou aa cells, 4 Figure 95M and P; 2, amphibrachyparacytic stoma with six contact cells, cf. Fig 95N; 3, ا‎ to 1 above but with five contact cells; 4, more or less brachyparacytic stoma with very broad ا‎ subsidiary cells; 5, ih а нын stoma with three contact cells; 6, stoma with a split lateral subsidiary cell, cf. Figure 95R; 7, amphibrachyparacytic stoma. 82, 83. C. picta var. picta (Kuntze 5.п.).— 82. 1, brachyparacytic stoma with polygonal picos cells, cf. Figure 95P; 2, aeg amphibrachyparacytic stoma, cf. Figure 95P and 0; 3, brachyparacytic stoma, cf. Figure 95M and О; 4, brachytetracytic stoma; 5, brachyparacytic stoma, cf. Figure 951; 6, anomocytic stoma with three contact cells. —83. 1, brachyparacytic ies cf. Figure 951 and P; 2, stoma о five contact cells, cf. Figure 955; 3, stoma with id s cells, cf. Figure 958; 4, amphibrachyparacytic stoma; 5, unclassifiable stoma with five contact cells. —84. C. fenzlii (Neger s.n. 1) 1, staurocytic stoma with four contact cells, cf. Figure 95G; 2, hemi-amphibra- ce stoma; 3, brachytetracytic stoma; 4, amphibrachyparacytic stoma; 5, stoma with six contact cells but showing essentially the characteristics of Figure 95R. —85. С. lingulata (Ferreyra 10486). 1, hemi-amphibrachy- paracytic stoma; 2, stoma with five contact cells, cf. dubi 95N iE R; 3, amphibrachyparacytic stoma, cf. Figure 95K; 4, stoma with five contact cells, cf. Figure 95J and S FIGURES 86-90. Epidermal morphology in Cistanthe sects. Cistanthe and Amarantoideae. All scale bars = a i um. 86, 87. C. (Cistanthe) x cf. longisc ара (Worth & Morrison 16184).— 86. Epidermal papillae (P). — ‚ brachyparacytic stoma, cf. Figure 951 and L; 2, hemi-amphibrachyparacytic stoma, cf. Figure 95] and N. B e (Cistanthe) cymosa (Werdermann 853). The largest epidermal cells are more or less papillar. The stomata are somewhat sunken. 1, double stoma; 2, alle stoma, cf. Figure 95M; 3, anomocytic stoma with four contact cells; 4, brachyparacytic stoma, cf. Figure 951.—89. C. Si Maa did ambigua (Nelson & Apa 3287). i, свел stoma, cf. e 95L and О; 2, stoma with five contact cells, cf. Figure 95L and R; : ibrachyparacytic stoma, cf. Figure 95Q; 4, br e stoma, cf. Fig e 95L.— 90. C. о i salsoloides Werder rmann 1048). 1, amphibrachyparacytic stoma; 2, hen E O stoma with a polar subsidiary cell, cf. Figure 95T; 3, brachyparacytic stoma with a polar subsidiary cell, cf. Figure 95T. Ficures 91-94. Epidermal morphology in Cistanthe sects. Calyptridium and Philippiamra. АП scale bars = са. 100 um.— 91. C. Ar A е (Sharsmith 4345). 1, brachyparacytic stoma with five contact cells; 2, anomocytic stoma with fou ntact cells; 3, anomocytic stoma with three contact cells; 4, double stoma; hemibrachypara cytic stoma. — 92. C (C odia rar и ан 11351). The stomata are slightly to markedly sunken. 1, stoma with five contact cells, cf. Figure 95N and 5; 2, brachytetracytic (but nearly hexacytic) stoma, с Figure 95E and F; 3. hemi-amphibrachyparacytic stoma with one especially narrow polar subsidiary cell; 4, vong ы stoma with four contact iind 5, dudes nds id stoma.— 93. C. му а rosea (Duran 2805). T stomata F sunken. l, anomocytic stoma; 2, hemibrachyparacytic stoma, cf. Figure 95N; P, н а cell. — 94. C. (Philippiamra) celosioides e 862). 1, hemi- a stoma (the epidermis is torn on the left side between the subsidiary cell and the guard cell pair); 2, amphibrachyparacytic stoma with five contact cells; 3, amphibrac al stoma with four contact cells. сл 1057 Hershkovitz Volume 78, Number 4 1991 AS Leaf Morphology of Cistanthe 10 Annals of the Mun Botanical Garden — = r^ LI. ? a pee irs я 50: Volume 78, Number 4 Hershkovitz Leaf Morphology of Cistanthe |" bi ` ч ч n 1060 Annals of the Missouri Botanical Garden 0 0 00 00 р о се се о го се Р о r FIGURE 95. Stomatal complex configurations in Cis- tanthe. A-H, stomatal types, cf. Wilkinson (1979). — A. Hemibrachyparacytic. — B. Brachyparacytic. — C. Hem amphibrachyparacytic (see text). — D. Amphibrachypara- супе. Е. Tetracytic. —F. Нехасуйс. —С. Staurocyt- :. . Anomocytic. 1-0. Examples of morphological variation of lateral subsidiary cells. — I. Contoured, nar- гом. — J. Contoured, broad. — К. Contoured, double. — L Rectangular, double. —O. Polygonal, narrow. — P. Polyg- onal, broad. — Q. Double, the outer polygonal. R- T. Other subsidiary cell types. — R. Lateral cell divided transverse- ly. — 5. Two lateral cells, both divided transversely. — T. Polar, elongate cell perpendicular to the guard-cell axis. PARSIMONY ANALYSIS AND CLADISTIC RECLASSIFICATION OF THE RELHANIA GENERIC GROUP (ASTERACEAE- GNAPHALIEAE) Arne A. Anderberg' and Kare Bremer? ABSTRACT Parsimony analysis of the Relhania group of genera соон миа yields a revised hypothesis of iiri phylogeny. The genera Leysera L. and Oedera L. are demonstrated to have their closest relatives within Relhan , which is shown to be paraphyletic as Visum circumscribed. A oci generic classification is proposed. Relhania aud Leysera are redefined more narrowly. Rhynchopsidium DC. and the monotypic Nestlera Sprengel are reestablished. Oedera is amended to include some former Relhania species. The new genus Comb Anderb. & Bremer, with 2 species, is described, and 14 new combinations in Oedera and Comborhiza are made. Because parsimony analysis produces better phylogenetic buses than limited selections of alleged synapomorphies, and because the latter approach hitherto has been more common in generic classification, it is argued that, with improved phylogenetic hypotheses derived from parsimony analysis, generic reclassifications will become necessary also in other groups. In this paper we present a parsimony analysis and propose a cladistic reclassification of a group of South African Asteraceae genera belonging to the tribe Gnaphalieae. The genera were revised during the 1970s and were then the subjects of some of the first cladistic approaches in botany (Bremer, 1976a, b, 1978a, b). At that time, uniquely derived, qualitative characters that could be interpreted as synapomorphies, unequivocally indicating moncphyletic groups were sought. Thus, selected features, mainly of the pappus structure crowning the fruits, were used as putative synapo- morphies defining the genera. Today, cladistics is methodologically much more sophisticated and is also considered a standard technique for analyzing phylogenies (Hull, 1989). Application of the par- simony criterion (Farris, 1983) in computerized programs has made it possible to use even large and homoplastic data sets including all kinds of information, minimizing a priori assumptions. Par- simony analysis of the Relhania group of genera yields a revised hypothesis of their phylogeny, and hence a basis for a cladistic reclassification. It is not only the theoretical and methodological background to the study of the phylogeny that has changed during the past 15 years, but also the knowledge of the phylogeny of the tribe Gnaphalie- ae as a whole (Anderberg, 1991). Furthermore, another genus, not considered in Bremer's 1976- 1978 revisions, has recently been added to the group (Anderberg & Kallersjo, 1988). Bremer (1976a, b, 1978a, b) revised the tax- onomy and nomenclature of the genera Relhania, Rosenia, Leysera, Oreoleysera, and Antithrixia. In trying to circumscribe correctly these genera as monophyletic groups, Bremer relied on partic- ular characters as synapomorphies defining them. Thus, Relhania, with 29 species, was distinguished by its pappus of more or less connate scales and no bristles. Bremer also pointed out the hypothet- ical nature of the generic delimitation: Admittedly the loss of pappus bristles might have occurred several times. However, since there is no evidence that Relhania should be polyphyletic in its present circumscription, I believe we must for the time being rely on this character as uniquely derived and keeping the genus together (Bremer 1976a: 9) Relhania was amended by Bremer to include most of the species of the genus Nestlera, which prior to Bremer's work was used to house all the species with epaleate receptacles. Bremer showed ' Department cf Phanerogamic Botany, Swedish Museum of Natural History, P.O. Box 50007, S-104 05 Stockholm, d Sweden. ? Department of Systematic Botany, University of Uppsala, P.O. Box 541, 5-751 21 Uppsala, Sweden. ANN. Missouni Bor. GARD. 78: 1061-1072. 1991. 1062 Annals of the Missouri Botanical Garden that epaleate receptacles were not uniquely derived and could not be used to separate Nestlera from Relhania. Thus, Bremer (1976b) placed much of Nestlera in Relhania and transferred some species to Rosenia. The latter genus comprises four species and has a pappus of scales as well as 1—4 scabrid bristles. Later, Bremer (19782) revised the related genus Leysera. It has а pappus of scales and five mostly plumose bristles and was assumed to be monophy- letic due to a synapomorphy in the very long pe- duncles. Again, Bremer commented on the pappus characters used to distinguish the three genera: In Leysera the disc-floret pappus consists of scales and five (or sometimes fewer) bristles, in Rosenia it consists of scales and generally 1-2 bristles, and in ne should be polyphyletic or о ин а (Сезе 1978а: 371). Bremer (1978) also removed one former Ley- sera species with several pappus bristles and de- scribed it as the monotypic genus Oreoleysera. Another monotypic genus, Antithrixia, was noted as a close relative of Relhania, Rosenia, Leysera, and Oreoleysera. Recently, Anderberg & Kallersjo (1988) dis- cussed the similarity between Oedera and some species of Relhania. А transfer of Oedera, which comprised six species, from the Anthemideae to the Relhania generic group was proposed. In a review of the tribe Gnaphalieae, Anderberg (1991) recognized the subtribe Relhaniinae comprising two major clades, viz. the Metalasia group, with 14 genera, and the Кећата group, with the genera Relhania, Rosenia, Leysera, Oreoleysera, Ап- tithrixia, and Oedera. Anderberg considered Oe- dera to be the sister group of Кећата. However, no synapomorphy could be found to define Rel- hania if Oedera was kept separate. Infraspecific relationships between the two genera were not in- vestigated, but the possibility that A'elhania could be paraphyletic was noted. Furthermore, Anderberg showed that the mono- typic genus Antithrixia is the most plesiomorphic representative of the Кећата group. The other ve genera share a synapomorphy in their pappus scales, absent in Antithrixia and the Metalasia group. (Bremer, (1978b) erroneously stated that Oreoleysera lacks pappus scales, which, albeit very narrow, do exist in Oreoleysera). The presence of a monotypic genus as the sister group of the rest of the Relhania group made ап analysis of the entire Relhania group at the species level conceivable. Based on the characters pre- sented in Tables 1 and 2, we have performed a parsimony analysis of the recognized species of the group, using Antithrixia as outgroup. The analysis and its taxonomic implications are presented below. METHODS DATA The analysis was based mainly on the characters used by Bremer (1976a, b, 1978a, b) in his re- visions. The character list has been amended with further characters by investigating the specimens housed in the Swedish Museum of Natural History (S). The characters are listed in Table 1. Only two species have not been available for study, viz. Oede- ra muirii and Oedera laevis. Most of the char- acters of Oedera muirii were possible to extract from the protologue of this species (Smith, 1927), but Oedera laevis has been omitted. However, judging from De Candolle's (1838) and Harvey's (1865) descriptions, it seems to be a typical rep- resentative of Oedera ANALYSIS The data matrix in Table 2 was computed with а parsimony program (Hennig86, version 1.5) writ- ten by and obtained from J. S. Farris (1988), and using a standard PC-AT. The following options were used: multiple hennig (mhennig*, construct- ing several initial cladograms by adding the taxa of the character matrix in several different se- quences, retaining the shortest cladogram of each) with a subsequent branch-breaker command (bb*, generating all the multiple, equally parsimonious cladograms that can be found). The cladograms were rooted with a hypothetical ancestor corre- sponding to the outgroup, Antithrixia. The mul- tistate characters 13, 16, 29, and 45 were coded as additive (cc +) because the character states were hypothesized to represent a gradual о series. The characters 8, 21, 23, 43, ‚ оп the other hand, were coded as amie (сс—), since no such hypothesis could be formulated in these cases. RESULTS The analysis resulted in 200 equally parsimo- nious cladograms, each 123 steps long, with a consistency index (ci) of 0.45. One of these clado- grams is shown in Figure 1. The strict consensus Моште 78, Митбег 4 1991 Anderberg & Bremer 1063 Analysis and Reclassification of Relhania TABLE 1. morphic states 1, 2, and 3 See Methods for coding of multistate characters. Characters used in the analysis of the Rel- hania-group. Plesiomorphic states are coded 0 and apo- with Antithrixia as outgroup. 1. (0) Perennial half-shrubs or shrubs, (1) annual herbs. 2. (0) Subterranean woody o absent, (1) subter- ranean woody rhizome prese . (0) Stem not e em branched, (1) stem subdichotomously branched. (0) branches unarmed, (1) branches subspinescent or ~ = spinescent. . (0) Brachyblasts present, (1) brachyblasts absent. (0) Leaves decussate, (1) leaves alternate. . (0) Leaves spreading, (1) leaves squarrose-recurved. (0) Leaves linear—oblong, (1) leaves obovate-spathu- late, (2) leaves widely cordate to orbicular 9. (0) Leaf margin entire, smooth, (1) leaf margin ap- parently denticulate to sparsely serrate. 10. (0) Leaves blunt, (1) leaves pungent. 11. (0) Leaves mid-nerved, (1) leaves with at least three main nerves. . (0) Leaves adaxially more = pessum than abaxially, (1) leaves adaxially gla . (0) Leaves glandular-hairy with pean! glands, (1) leaves glandular-punctate, (2) leaves glandular-punc- tate, with glands sunken in pits. 14. (0) Leaf margin not involute, (1) leaf margin involute. 15. (0) Capitula sessile or on short peduncles (generally 5 mm), (1) capitula on long peduncles (generally од сол — N — о m). 16. (0) Capitula solitary, (1) capitula paired to cymose- corymbose or congested, (2) capitula many in dense secondary heads 17. (0) Synflorescense not surrounded by a leafy invo- lucre, (1) synflorescense surrounded by a leafy in- volucre. 18. (0) Involucral bracts with spreading limb, (1) invo- lucral bracts with straight lim 19. (0) At least some involucral bracts spathulate, (1) involucral bracts not spathulate. 20. (0) Receptacle without long squamae, (1) receptacle with long squamae 21. (0) Receptacle naked, (1) receptacle paleate, with paleae abaxially to the florets, (2) receptacle deeply 22. (0) Paleae deciduous, (1) paleae persistent. 23. (0) Paleae entire, (1) paleae apically serrate to la- ciniate, (2) paleae with two lateral teeth . (0) Ray-florets monomorphic, (1) ray-florets dimor- ~ к hic. . (0) Ray-floret lamina usually 4-veined, (1) ray-floret lamina frequently with up to 10 veins. . (0) Ray-floret tube сене (1) ray-floret tube ~ сл N с somewhat triquet . (0) Disc-florets bet (1) disc-florets functionally male, style undivided . (0) Disc-floret corolla somewhat funnel-shaped, in- distinctly divided in tube and limb, (1) disc-floret corolla distirictly divided in tube and limb. ~ = ~ со TABLE 1. Continued. 29. (0) Floret tubes eglandular or with few scattered glands, (1) floret tubes regularly glandular, (2) floret tubes with subulate-triangular hairs 30. (0) Anther vide nd acute to iesus: (1) anther appendage tr 31. (0) Anther tails гоа branched. 32. (0) Style branches in disc-florets apically penicillate, truncate, (1) style branches in disc-florets dorsally and apically penicillate, rounded-obtu 33. (0) Cypselas oblong-elliptic, (1) ue linear. 34. (0) Cypselas of ray-florets terete or angular, (1) cyp- selas of ray-florets sharply triquetrous. 35. (0) Cypselas of disc-florets terete or angular, (1) cyp- selas of disc-florets flattened. 36. (0) Cypselas qs (1) cypselas at least in ray- florets pubesc 37. (0) Cypselas i? pm ray-florets pilose, (1) cyp- selas of pubescent ray-florets villose 38. (0) оре trichomes straight, (1) gpi trichomes apically co 39. (0) Cypselas generally eglandular, (1) cypselas reg- ularly very glandular. 40. (0) Cypsela epidermis smooth, (1) cypsela epidermis with acute papi 41. (0) Cypselas eed 5 vascular bundles ашшы 10), (1) cypselas with 2-3 vascular bun 42. (0) Pappus bristles in ray-florets present, a pappus bristles in ray-florets absent . (0) Pappus bristles in disc-florets numerous, (1) pap- pus bristles in disc-florets five, (2) pappus bristles in disc-florets one to four, (3) pappus. bristles in disc- florets absent. . (0) Pappus bristles apically scabrid to barbellate, (1) pappus bristles apically plumose . (0) Pappus scales free, (1) pappus scales connate, (2) pappus tubular. . (0) Basic chromosome number x = 7 (2n = 14, 28, 56), (1) basic chromosome number x — 5 (2n — 10), (2) basic chromosome number x = 4 (2n = 8, 16). (1) anther tails un- ~ о ~ > + сл > On tree in Figure 2 was computed with the nelsen option and shows the clades consistently present in all the equally parsimonious cladograms. In Fig- ures 1 and 2 the terminal taxa are named as in Table 2, following the earlier generic classification, whereas the generic reclassification proposed below is indicated with names in capitals. In the following discussion, Relhania s.l., Leysera s.l., and Oedera s.s. refer to the earlier circumscriptions, whercas Relhania s.s., Leysera s.s., and Oedera s.l. refer to the proposed new delimitations. Variation among the 200 equally parsimoniously cladograms is restricted to 100 alternative topol- ogies within the screened right part of Figures 1 and 2 (Oedera s.l.) as well as to two alternative 1064 Annals of the Missouri Botanical Garden Data matrix. The characters are listed in Table 1. The taxon names follow the earlier classification. TABLE 2 Inapplicable, unknown, or variable character states are coded - Character number 000000000111111111 144444 89012345678901 789012345 Antithrixia flavicoma reoleysera montana Leysera longipes Leysera gnaphalodes Leysera tenella Leysera leyseroides Rosenia glandulosa osenia spinescens Казына oppositifolia Rosenia Relhania dieterlenii Relhania fruticosa Relhania rotundifolia Relhania spathulifolia Relhania corymbosa Relhania tricephala Relhania relhanioides Relhania genistifolia > % ~ = ~] 3 5 5 о a Relhania muli рана tata Relhania resinifera Relhania sedifolia Relhania foveolata Relhania nordenstamii Relhania uniflora Relhania silicicola Relhania pumila Relhania sessiliflora Relhania biennis Relhania tuberosa Relhania acerosa Relhania decussata Relhania pungens Relhania calycina Relhania speciosa Oedera capensis Oedera hirta de icata Oedera nus Oedera тишти 000011 010011000001001000000--00-0110101000--0001100- 0000-1000000001000-10--00001211010010001111102 100011000000001000110--00001211010010001111102 100011000000001001110--00001211010010001111102 001 001 00-00000000-00000000 0000100000001 10000000 00013-1 0000110000001 10000001000000100100000--00013-2- 0000110100001 00013-2- 0000110100001 10000000000001 100100001 1000013-10 0000110100001 10100000000000100100001100001301- 0000110100001 10100000000001 100100001 1000013-10 0000110100001 00100001 1000013-1 - 000011- 00001 1001000010000001 1010000100- 0013-10 D 13-10 00001110 00001 10000:10010000100000011010000-—10112 1- 00000--00013-10 ODD о о, 000011000001200101001000000110111000---0013-1- 11010000001 3-10 0000110000012001011010000001101-10010000013010 011000001 00000000000 1 2001011010000001101000010000013-1- 2 10 000010000001100100000--0000110100000—-10013- l- d -11 10001 3-11 100011000000000000100--0-0012110100100000 3-21 ] 13-10 001011000101100000000--00001101110010000113-1- 0000100001001000000011000001001100010000113-10 000011000110000000001100-100200101110000113-10 000011-00111000000001100-100200101110000113-10 000011-00111- 00000001 1001100200- 01110000113-10 0000111011 00120100000--00013-1- I E. 1- 00001112111 00013-1- ОООО ош 10 000010101111000210001011100100100000--00-13-1- sister groups to Relhania tricephala, either К. corymbosa or the species pair R. relhanioides + R. spathulifolia, within the screened left part of Figures 1 and 2 (Relhania s.s.). The majority of the clades in the cladogram in Figure 1 are present in all the other equally parsimonious cladograms, as shown by the strict consensus tree in Figure 2. Several analyses were also performed with slight modifications of the data matrix, omitting some characters or including others not accepted in the final matrix (ultimately rejected for various rea- sons, for example, because of unreliable informa- tion in some taxa or because of strong correlation with other included characters such as presence of 1065 ‘sedeo ш поцеошевејо оџопов mau розодола “UONBIYISSPIO ошопов лэцлеэ оцу Sutwo[oj ‘Z e[qe] Ul Se seureu uoxej јешшлој *] ојдеј, ш злојовлецо ayi оу риодволлоо siequinu au 'sumo[e1ed = seu [o[[e1ed 'арезлолол = sosso1o “шелдорејо од uo ѕүеѕләләг цим Aydsourodeuks = ѕәлепЬѕ uado ‘sarydiourodeuds = so1enbs мовја ‘с qL J, ut хеш вер aU} Jo sisA[eue шолу роитејдо sure1dope[o snoruoursied Аепђо 005 991 Jo 900. сү ANNOY Моште 78, Митбег 4 1991 Anderberg & Bremer Analysis and Reclassification of Relhania | унзазо VZIHHO8NOO VH3SAdlO3uHO ISSoS|oH pundi9y VH3lLS3N weqeu WNIGISdOHONAHY VIXIBH.LLLNV 1066 Annals of the Missouri Botanical Garden џаширао леџлрео umuipeo eunupeo uedeopeg = џдезвјен || эштән | ajuoojey « ерзошен a шәлен o урезјен о јупшјен E —3À ¿enbsiay озвијен stuebtey оәлодән —— mm ичйән —— 1биојәј VZIHYO8NOO aqyay E snoopy озаовјен Strict consensus tree based on all the 200 equally parsimonious cladograms obtained from analysis of the data matrix in Table 2. Terminal taxon names as in Table 2, following the earlier generic classification. Proposed new generic classification in capitals. Shaded areas show the circumscription of Relhania and Oedera, respectively. VIXIGHLILNY — т ul эзАэАэ7 за ¡auajAa7] _] u ydeubAa7] _| VH31lLS3N eeu Issasjay WNIGQISdCOHONAHY eee vH3SA3103H0 —— < ривјбзон = ri euidssoH О пшпцѕон " soddosoy а = 3 e= Volume 78, Number 4 1991 Anderberg & Bremer 1067 Analysis and Reclassification of Relhania glands on various organs). These analyses showed, as might be expected, that the clades supported by single parallelisms are weak and easily broken up even with slight changes in the data matrix. However, in all analyses, Rosenia and Oreoleysera successively came out at the base of the cladogram with the remaining genera as one monophyletic roup. Furthermore, Leysera and Oedera were always сш. nested within a paraphyletic Rel- hania s DISCUSSION The consistency index 0.45 and the frequent occurrence of parallelisms and reversals on the cladogram in Figure 1 show that there is consid- erable homoplasy in the data, as is usually found in similarly sized analyses of other groups. The impact of homoplastic characters on the phylo- genetic hypothesis is important, and the most par- simonious way in which these characters can be distributed on a cladogram is virtually impossible to detect without computerized programs. scientific theory, the phylogeny reconstructed by parsimony analysis is superior to any phylogeny sa obtained from a limited selection of characters only, since the former takes into account all available information. Parsimony analysis is increasingly em- ployed in taxonomic revisions, and it offers possi- bilities for improved, phylogenetically more infor- mative classifications, where the named taxa reflect evolutionary units rather than artificially composed groups defined only by a few characters. In this case, the analysis clearly demonstrates that Relhania s.l. is paraphyletic with Leysera and Oedera excluded. Delimitation among the three genera has to be altered in order to meet the demand for monophyletic taxa. To lump all genera into a single genus, Leysera sensu latissimo (Ley- sera being the oldest name), would result in a very heterogeneous taxon, defined only by characters such as absence of brachyblasts, alternate leaves, glandular floret tube, unbranched anther tails, ray- florets without pappus, and corolla divided in tube and limb. The characters are weak and can, in a wider context, also be shown to be parallelisms (e.g., by investigation of a larger number of taxa of the Gnaphalieae). A better alternative is to split Relhania and Leysera and recognize smaller monophyletic genera that are more morphologi- cally homogeneous. This would mean the separa- tion of the apomorphic Leysera s.s., Rhynchop- sidium, Comborhiza, and Oedera from each other, and from Relhania. The definition of Oedera is hereby weakened due to the transfer to this genus of 10 Relhania species. This is unfortunate, but no other alternative is at hand if we want to main- tain monophyletic genera. Relhania itself is poorly defined but still constitutes a monophyletic group. By accepting this second classification, we gain a detailed, predictive phylogenetic hypothesis with more explanatory power, because more taxonomic statements are made. To reduce all this information of phylogenetic interrelationships and conceal it in a lumping such as outlined above would hardly improve scientific progress. Consequently, we have resurrected Nestlera in its original monotypic sense, reestablished the old ditypic genus Rhynchopsidium, described another ditypic genus Comborhiza, and transferred a num- ber of the former Relhania species to Oedera (see Figs. 1, 2). Relhania is restricted to 13 species, mostly with hairy and glandular-punctate (not glandular-hairy) leaves and solitary capitula. The genus comprises two distinct groups, recognized also by Bremer ~ — 76a), as well as a more isolated species, К. garnotii. The two groups are К. acerosa to R. speciosa, with five species (including the distinct subgroup comprising К. pungens, К. calycina, and К. speciosa), and К. dieterlenii to К. spa- thulifolia, with seven species (see Figs. 1, 2). Leysera is restricted to the three typical species and is thus the most well-defined of all the genera, with a whole set of synapomorphies, e.g., plumose pappus bristles (see Fig. 1). The former Leysera longipes and Relhania tuberosa are grouped as sister species in the new genus Comborhiza. They have a unique synapomorphy in their thick sub- terranean rhizomes. Although different in several characters, the two species are habitually similar, Comborhiza virgata (= Relhania tuberosa) hav- ing pedunculoid stems with reduced leaves beneath the capitula, approaching the condition in C. lon- gipes. The genus has adaxially glabrous, dorsally glandular-hairy leaves, not glandular-punctate as in Relhania s.s. and Oedera s.l. The OBO genus Nestlera and the ditypic genus to Leysera s.s. ee Figs. 1, 2). бе for the (secondarily) perennial L. gnaphalodes, they are ter groups annuals or biennials with reduced chromosome base numbers, x = 5 in Rhynchopsidium and Nestlera, and x = 4 in Leysera. (Humphries et al. (1978) reported 2n — 14 for L. leyseroides, a count that must be doubted in the light of Blanca's (1983) detailed investigation of the karyology of the spe- cies; Blanca showed that L. leyseroides is a tet- raploid with 2n — 16. Oedera, finally, is amended to include 12 former 1068 Annals of the Missouri Botanical Garden Relhania species, and as such Ое4ега s.l. ıs cnar- acterized by the generally glabrous, glandular- punctate leaves, and the cymose-corymbose to con- gested capitula. Infraspecific relationships within Oedera are still uncertain. The analysis produced 100 equally parsimonious topologies within Oedera s.l. Nevertheless, it must be concluded that Oedera s.s. is firmly nested within the larger monophyletic group that we have circumscribed as Oedera s.l. Hence, Oedera s.s. cannot be maintained at the generic level, since it would leave an unresolved paraphyletic residue of former Relhania species. CONCLUSION The former generic classification of the Rel- hania group was based on selected, alleged syn- apomorphies, the usual cladistic approach during the middle-1970s. Thus, Bremer (1976a, 1978a) defined genera on some pappus features, at the same time pointing out the possibility that the sup- posed synapomorphies could be false. Parsimony analysis of all the characters has indeed produced a revised picture of the phylogeny. We have re- circumscribed genera according to the best phy- logenetic hypothesis available, as expressed in the cladograms produced by the analysis. Reclassifi- cation of genera like Relhania s.l., formerly de- fined by unique, alleged synapomorphies, which in parsimony analysis of larger groups are shown to be false, is likely to become necessary in many other generic groups as well The present classification of the Relhania group is summarized in the following key: la. Plants compact, cushion-forming Oreoleysera lb. Plants not cushion-formin 2a. Pappus of capillary bristles, absent Antithrixia 2b. Pappus at least in ray-florets with conspicuous scales, bristles present or absent. 3a. Pappus bristles en р! Leysera 3b. res bristles scabrid- iw or absent. ual or biennial herbs. 5a. you. densely Sud wiih long, apically coiled hairs 5b. Cypselas almost glabro Rhynchopsidium Nestlera 4 = ба. Disc-florets with 1- 6b. Disc-florets without pappus bristles 7a. Plan generally cymose glabrous, + distin Plants + ha ctly one "di E Perennial half-shrubs, LEE or shrubs. 4 pappus bristles ... Rosenia nts glabrous (if hairy then with capitula congested in secondary heads), capitula -corymbose to cluste и іп secondary heads (if solitary then це» ved lea airy or cades г (if ике then with distinctly 3-9-nerved leaves), "sias generally solitary (if corymbose then with hairy leaves). aves with long gla Antithrixia DC., Prodr. 6: 277. 1838. ТУРЕ: А. flavicoma DC. Shrublet. Leaves opposite or crowded on brach- yblasts, sessile, linear, adaxially concave and to- mentose, glandular-hairy, apically mucronate; margin entire. Capitula solitary, terminal, het- erogamous, radiate. Involucral bracts with spathu- late laminas. Receptacle flat, epaleate. Female florets yellow with purple bands dorsally, pistillate, radiate, in one row, fewer than the disc-florets. Cypselas and pappus as in disc-florets. Disc-florets perfect. Corolla yellow. Anthers ecalcarate with short tails; endothecial tissue polarized; apical ap- pendage acute. Style branches truncate with obtuse sweeping-hairs apically. Cypselas rodlike, with ten (2 x 5) vascular bundles, glabrous or with a few scattered, elongated twin hairs. Pappus of slightly ndular hairs Bb. Leaves without long glandular hairs .. Comborhiza Relhania connate, barbellate, capillary bristles in one to two rows. Chromosome number unknown. Monotypic: А. flavicoma DC. Reference: Bre- mer, 1978b. Rosenia Thunb., Nov. Сеп. 161. 1800. ТУРЕ: А. glandulosa Thunb. Synonym: Polychaetia Less. Shrublets. Leaves decussate or rarely alternate, straight, adaxially concave, tomentose, glandular- hairy; margin entire. Capitula heterogamous, ra- diate, solitary or few together. Involucral bracts with spathulate laminas. Receptacle flat, often pa- leate. Female florets yellow with purple bands dor- sally, radiate, pistillate, in one row, fewer than the disc-florets. Cypselas and pappus as in disc-florets. Volume 78, Number 4 1991 Anderberg & Bremer 1069 Analysis and Reclassification of Relhania Disc-florets perfect. Corolla yellow. Anthers ecal- carate with short tails; endothecial tissue polarized; apical appendage acute. Style branches + truncate with obtuse sweeping-hairs apically. Cypselas rod- like with five vascular bundles, sparsely hairy with elongated twin hairs. Pappus a crown of + free scales, and 1--4, barbellate, capillary bristles with flattened apical portion. Chromosome number 2n = 14, 28,5 Four species: R. glandulosa Thunb., R. humilis (Less.) Bremer, R. oppositifolia (DC.) Bremer, R. spinescens DC. Reference: Bremer, 1976b. Oreoleysera Bremer, Bot. Not. 131: 450. 1978. TYPE: O. montana (Bolus) Bremer. Compact woody perennial, forming dense tufts. Leaves alternate, sessile, narrowly oblong, flat- tened, grayish tomentose on both surfaces, eglan- dular; margin entire. Capitula solitary, terminal on long leafless peduncles, heterogamous, radiate. Involucral bracts with lanceolate laminas. Recep- tacle flat, epaleate. Female florets yellow with pur- ple bands dorsally, pistillate, radiate, in one row, fewer than the disc-florets. Cypselas and pappus as in disc-florets. Disc-florets perfect. Corolla yel- low, glandular-hairy. Anthers ecalcarate with short tails; endothecial tissue polarized; apical appendage acute. Style branches truncate with obtuse sweep- ing hairs apically. Cypselas rodlike with five vas- cular bundles, moderately hairy with elongated twin hairs. Pappus of free, barbellate, capillary bristles in one row with an outer row of very narrow scales. Chromosome number unknown. Monotypic: O. montana (Bolus) Bremer. Ref- erence: Bremer, 1978b. Rhynchopsidium DC., Мет. Soc. Phys. Сепеуе 7: 283. 1836. ТУРЕ: R. sessiliflorum (L. f.) DC. Synonym: Rhynchocarpus Less., nom. Шер. Annual herbs. Leaves alternate or sometimes opposite, straight, adaxially concave, tomentose, glandular-hairy; margin entire. Capitula heterog- amous, radiate, solitary, sessile or pedunculate. Involucral bracts with lanceolate laminas. Кесер- tacle flat-convex, paleate; paleae lanceolate with two lateral teeth. Female florets yellow with purple bands dorsally, radiate, pistillate, in one row, fewer than the disc-florets. Cypselas and pappus as in disc-florets. Disc-florets perfect. Corolla yellow, of- ten with robust multicellular trichomes. Anthers ecalcarate, with short tails; endothecial tissue po- larized; apical appendage acute. Style branches truncate with obtuse sweeping-hairs apically. Сур- selas narrowly elliptic with three vascular bundles, densely hairy with elongated, apically coiled twin hairs. Pappus a crown of = free scales. Chro- mosome number 2n = 10. Two species: К. pumilum (L. f.) DC. (= Rel- hania pumila (L. f.) Thunb.), К. sessiliflorum (L. DC. (= Relhania sessiflora (L. f.) Thunb.). Reference: Bremer 1976a (spp. 21-22). "^ ~ Nestlera Sprengel, Anleit. Kennt. Gewachse 2(2): 68. 1818. TYPE: N. biennis (Jacq.) Sprengel. Synonyms: Columellea Jacq. nom. illeg., Stephanopap- pus Less. Biennial herb. Leaves alternate or occasionally opposite, straight, adaxially concave, tomentose, glandular-hairy; glands with multicellular heads; margin entire. Capitula heterogamous, radiate, solitary or few together. Involucral bracts with lanceolate laminas. Receptacle flat, often paleate. Female florets yellow with purple bands dorsally, radiate, pistillate, in one row, fewer than the disc- florets. Cypselas and pappus as in disc-florets. Disc- florets perfect. Corolla yellow, often with robust multicellular trichomes. Anthers ecalcarate, with short tails; endothecial tissue polarized; apical ap- pendage truncate. Style-branches + truncate with obtuse sweeping-hairs apically. Cypselas rodlike with five vascular bundles, sparsely hairy with elon- gated twin hairs. Pappus a tubular crown of connate scales. Chromosome number 2n = 10. Monotypic: N. biennis (Jacq.) Sprengel (= Rel- hania biennis (Jacq.) Bremer). Reference: Bremer, 1976a (sp. 23). Leysera L. Sp. Pl. ed. 2, 2: 1249, 1763; Amoen. Acad. 6: 104, 1763; Gen. Pl. ed. 6: 431, 1764. ТҮРЕ: L. gnaphalodes (L.) L. Synonyms: ie quce (Vaill.) Adans., Callicornia Burm. C ‚ Leptophytus Cass., Longchampia Willd., Page Кш Leyssera auct., ortogr. , Leyseria Necker, Annual herbs or a perennial half-shrub or shrub- let. Leaves alternate or occasionally opposite, ses- sile, linear, straight, mucronate, adaxially concave, tomentose, glandular-hairy; margin entire. Capit- ula heterogamous, radiate, solitary, terminal on long leafless peduncles. Involucral bracts with lan- ceolate laminas. Receptacle flat, marginally with prominent squamae. Female florets yellow with 1070 Annals of th Missouri ы Garden purple bands dorsally, radiate or miniradiate, pis- tillate, in one row, fewer than the disc-florets; tube often with robust, conical trichomes. Cypselas as in disc-florets. Pappus of scales only. Disc-florets perfect. Corolla yellow, often with robust multi- cellular trichomes. Anthers ecalcarate, with short tails; endothecial tissue polarized; apical appendage Style branches truncate with obtuse sweeping-hairs apically. Cypselas rodlike, with three truncate. vascular bundles, glabrous or with scattered, elon- gated twin hairs. Pappus of free, apically plumose, an outer row of obtuse scales. Chromosome number 2n = 8, 16 Three species: L. gnaphalodes (L.) L., L. ley- seroides (Desf.) Maire, L. tenella DC. Reference: Bremer, 1978а. capillary bristles in one row wit P y Relhania L'Hér., nom. cons.. Sert. Angl. 1: 22, 1789. ТУРЕ: К. fruticosa (L.) Bremer. Synonyms: Osmites L., Lapeirousia Thunb., nom. illeg. Shrubs, shrublets, or half-shrubs. Leaves alter- nate or decussate, sessile, straight, adaxially concave, tomentose, margin entire. Capitula het- erogamous, radiate, solitary or sometimes cymose- corymbose. Involucral bracts with spathulate lam- inas. Receptacle flat to convex or rarely deeply alveolate, often paleate. Female florets yellow with purple bands dorsally, radiate, pistillate, in one row, fewer than the disc-florets. Cypselas and pappus as in disc-florets. Disc-florets perfect or function- ally male. Corolla yellow, sometimes with robust multicellular trichomes. Anthers ecalcarate, with short tails; endothecial tissue polarized; apical ap- pendage acute. Style branches truncate with obtuse sweeping-hairs apically (rarely obtuse with sweep- ing-hairs dorsally). Cypselas rodlike with 5 or 3 vascular bundles, glabrous or sparsely hairy with elongated twin hairs. Pappus a crown of free to connate scales without bristles, occasionally with 1 or 2 barbellate bristles in odd florets. Chromosome number 2n = Thirteen species: R. acerosa (DC.) Bremer, R. calycina (L. f.) L'Hér., К. corymbosa (Bolus) Bre- mer, R. decussata L'Hér., К. dieterlenii (E. Phil- lips) Bremer, К. fruticosa (L.) Bremer, R. garnotii (Less.) Bremer, К. pungens L'Hér., R. relhan- ioides (Schltr.) Bremer, R. rotundifolia Less., R. spathulifolia Bremer, К. speciosa (DC.) Harv., R. tricephala (DC.) Bremer. Reference: Bremer, 1976a (spp. 1-7, 13, 25-29) Comborhiza Anderb. & Bremer, gen. nov. TYPE: C. virgata (N. E. Br.) Anderb. & Bremer. Eponymy: The name is deduced from the Greek words combos, meaning knot, tuber, and rhi- za, meaning root. Fruticulus vel suffrutex. Caules rhizomate vel tubere pogaeo, crasso, lign brachyblastis congesta, sessilia, concava, inferne integra. Capit tula heterogama, caulium paucifoliatorum vel p li radii flavi sed dorso: interdum purpureo-fasciati, pistillati, uniseriati, fl ores. Cypselae e eaedem in flosculis disci. ciae ex squam- glabr rae. · Pappus ex setis Fiber capillaribus, ee is scariosis plus minusve connatis solum aan, Shrublet or suffrutex. Stems arising from sub- terranean, thick woody tubers or rhizomes. Leaves alternate or crowded on brachyblasts, straight, adaxially concave, glabrous above, glandular-hairy below, sessile, argin entire. Capitula heterogamous, radiate, solitary, on pedunculoid few- leaved stems or on long leafless peduncles. Invo- lucral bracts with spathulate laminas. Receptacle flat or flat to convex, epaleate. Female florets yellow with purple bands dorsally, radiate, pistillate, in one row, fewer than the disc-florets. Cypselas as in disc-florets. Pappus of narrow scales. Disc- florets perfect. Corolla yellow. Anthers ecalcarate, with short tails; endothecial tissue polarized; apical appendage acute. Style branches truncate with ob- tuse sweeping-hairs apically. Cypselas narrowly elliptic with five or three vascular bundles, glabrous or sparsely hairy with elongated twin hairs. Pappus of 5 barbellate capillary bristles and narrow scales, or of + connate scales only. Chromosome number 2n = 14 Two species: C. longipes (Bremer) Anderb. & Bremer, comb. nov. Basionym: Leysera longipes Bremer, Bot. Not. 131: 381. 1978. C. virgata (N. E. Br.) Anderb. & Bremer, comb. nov. Basionym: Nestlera virgata М. r., Kew Bull. 1895: 25 (= Bremer). References: Bremer, 1976a (sp. 24), 1978a (sp. 4 Relhania ав Volume 78, Number 4 1991 Anderberg & Bremer 1071 Analysis and Reclassification of Relhania Oedera L., nom. cons. emend. Anderb. & Bremer, emend nov. Mant. Pl. 159, 1771. TYPE: О. capensis (L.) Druce. ] C + 1 Synonyms: Eroeda Levyns, nom. illeg., E Shrubs or shrublets. Leaves alternate or decus- sate, ovate to linear, mucronate, adaxially concave, glabrous or sometimes sparsely pilose, glandular; margin entire but sometimes provided with prom- inent, robust teethlike hairs. Capitula generally densely cymose-corymbose or even forming sec- ondary heads surrounded by a common involucre of leaves, heterogamous, radiate, few-flowered. In- volucral bracts with spathulate laminas. Recep- tacle flat to conical-convex, paleate or occasionally epaleate. Female florets yellow with purple bands dorsally, radiate in some species miniradiate or almost tubular in different positions in the head, in one row, fewer than the disc-florets. Cypselas and pappus as in disc-florets. Disc-florets perfect. Co- rola yellow, sometimes with robust multicellular trichomes. Anthers ecalcarate, with short tails; en- dothecial tissue polarized; apical appendage acute. Style branches generally truncate with obtuse sweeping-hairs apically. Cypselas rodlike, with 5 vascular bundles, glabrous or with a few elongated twin hairs. Pappus a crown of = connate scales, Chromosome number 2n — 14. ighteen species: without bristles. O. capensis (L.) Druce, O. hirta Thunb., O. im- bricata Lam., О. intermedia DC., О. laevis DC., O. muirii C. A. Smith. O. conferta (Hutch.) Anderb. & Bremer comb. nov. Basionym: Кећата conferta Hutch., Ann. S. African Mus. 9: 381. 1917. O. foveolata (Bremer) Anderb. & Bremer, comb. nov. Basionym: Relhania foveolata Bremer, Opera Bot. 40: 52. 1976. O. genistifolia (L.) Anderb. & Bremer, comb. nov. Basionym: Athanasia genistifolia L., Syst. Nat. 2 Ed. 12: 540. 1767 (= Relhania genis- tifolia (L.) L'Hér.). O. multipunctata (DC.) Anderb. & Bremer, comb. nov. Basionym: Relhania multipunctata DC., Prodr. 6: 286, 1838. О. nordenstamii (Bremer) Anderb. & Bremer, O. resinifera (Bremer) Anderb. & Bremer, comb. nov. Basionym: Кећата resinifera Bremer, a Bot. 40: 48, 1276. O. sedifolia (DC.) Anderb. & Bremer, comb. nov. Basionym: Eclopes sedifolia DC., Prodr. 6: 288, 1838 (7 Relhania sedifolia (DC.) Harv.). . silicicola (Bremer) Anderb. & Bremer, comb. nov. Basionym: Relhania silicicola Bremer, Opera Bot. 40: 56, 1976. . squarrosa (L.) Anderb. & Bremer, comb. nov. Basionym: Santolina squarrosa L., Cent. 2. Plant. 30, 1756 (7 Relhania squarrosa (L.) L'Her.). O. steyniae (L. Bolus) Anderb. & Bremer, comb. nov. Basionym: Relhania steyniae L. Bolus, Ann. Bolus Herb. 1: 191, 1915. O. uniflora (L. f.) Anderb. & d x nov. Е Athanasia uniflora L. f., Suppl. Pl. 362, 1781 (= Relhania ere (L. f.) O O ce). O. viscosa (L.Hér.) Anderb. & Bremer, comb. nov. Basionym: Relhania viscosa L'Hér., Sert. Angl. 1: 23, 1789. References: Harvey, 1865; Smith, 1927; Mans- field, 1935; Bremer, 1976a (spp. 8-12, 14-20), Anderberg & Kallersjo, 1988. LITERATURE CITED pc А. . Taxonomy and phylogeny of the VELA MAN Rydb. (Asteraceae). Opera Bot & KALLERSJO, M. . The tribal position of Oedera L. ав, J. Linn. Soc., Bot. 96: -332. BLANCA, С. leyseroides (Dest | Maire. Trab. Dept. Вог., Consideraciones sobre la Leysera Univ. BREMER, К. 1976a. Opera Bot. 40. 1976b. с. е Rosenia (Compositae). Bot. 129: 97- . 1978а. The genus Leysera (Compositae). Bot. 131: 369-383. —— ——. 19785. Oreoleysera and Antithrixia, new and old South African genera of the Compositae. Bot. Not. 131: 449-453 DE CANDOLLE, А. P 38. Prodromus Systematis Na- turalis Regni Vegetabilis, Volume 6. Treuttel et Würtz, Pari "The genus Relhania (Compositae). Not. Not. aris. Farris, J. S. 1983. The logical basis of phylogenetic analysis. Pp. 7-36 in N. I. Platnick & V. A. Funk (editors), Advances in Cladistics 2. New York Bo- tanical Garden, Bronx, New York. 1988. Hapus 86, Version 1.5, Program and Univ. of New York Stony Ur s Reference. State Brook, New York. Harvey, W. H. 1865. Compositae. Pp. 44-530 in W. H. Harvey & O. W. Sonder (editors), Flora Capensis, Volume 3. Hodges, Smith & Co., Dublin. Hut, D. L. The evolution of phylogenetic sys- 1072 Annals o Missouri Botanical Garden tematics. Pp. 3-15 in B. Fernholm, K. Bremer & H. Jórnvall dud The INNO of Life. Elsevier Science don Amster HUMPHRIES, С. J., B. С. Murray, С. BocouET & К. N. VASUDEVAN. 1978 . Chromosome numbers of pha- nerograms from Morocco and Algeria. Bot. Not. 131: 04. adr d In: А. Reh- atherby, R. Mansfeld & M. L. Green. I Conservation of ез generic homonyms. Bull. Misc. Inform. 6-9: 422-455. SMITH, С. А. 1927. Four interesting species of Com- positae. Bothalia 2: 360-365. MANSFELD, R. ег, vt POPULATION STRUCTURE OF Richard D. Laven,’ Robert B. Shaw, THE RECENTLY Patricia P. Douglas, and ictor E. Diersing* REDISCOVERED HAWAIIAN SHRUB TETRAMOLOPIUM ARENARIUM (ASTERACEAE)! ABSTRACT A demographic monitoring study was initiated in the sole known population of Tetramolopium arenarium, a taxon that was considered extinct until recently rediscovered on the Pohakuloa Training Area, Island of Hawaii, € Spatial distribution, size structures, life history stage structure and reproductive output were determined for population. This sole population is restricted to a 100 m by 300 m area along a mesic ridge system and is compri Po of 134 individuals. Size class frequency distributions are bell-shaped with the population pa ani ie by large individuals. In spite of large flower and seed production, interpretation of these РА reveals that this = is either on the verge of extinction or that episodic establishment is necessary to ensure the long-term persistenc of this ra taxon. Їп order to safeguard this species, we recommend that military пса Е. hunting activity {= а and that the area. be fenced to exclude feral animals. As part of the U.S. Army's Land Condition- Trend Analysis Program, a comprehensive vege- tation inventory was conducted of the Pohakuloa Training Area (PTA) on the Island of Hawaii, Ha- waii (Fig. 1). During the inventory, Te arenarium? (A. Gray) Hillebr. was collected (Doug- las et al., 1989). This taxon was considered extinct because no collections had been made since the late 1800s (Ayensu & DeFilipps, 1978; Lowrey, 1986) tramolopium In order to minimize the risk of extinction of this rare species, specific management strategies must be developed for the U.S. Army. A first step in this process is to assess the viability of this population. Because basic population biology may e of more imraediate importance than population genetics in determining minimum viable sizes of wild populations (Lande, 1988), we initiated a de- mographic monitoring study to determine if the population is expanding, stable, or declining. Our specific objectives were: (i) to determine the geo- graphic (site-specific) distribution of the population; (ii) to describe the population size structure and life history stage structure; and (iii) to assess the flowering and reproductive output of the popula- поп. NATURAL HISTORY AND MORPHOLOGY BRANCHING PATTERN AND PLANT ARCHITECTURE Tetramolopium arenarium is an upright shrub less than 2 m tall. It has three-dimensional, axillary branching, i.e., sympodial branching related to the development of a terminal, compound capitules- cense (Lowrey, 1986). Continuation of longitudinal shoot growth from usually two or three axillary ! This research was funded by the U.S. Army Corps of Engineers, Engineering & Housing Support Center (CEHSC- FN), Donald M. Bandel, Chief of Natural and Cultural Resources. We men and women under his command. We thank Greg Aplet for field assistance Commanding Officer, PTA, and the appreciate the assistance of Richard S. Beahm, and are especially grateful to Christine Bern for data synthesis and presentation. We also appreciate the critical review comments and suggestions of Peggy Fiedler, Greg Aplet, and David Tazik ? Department of Forest Sciences, Colorado State University, Fort Collins, Colorado 80523, U.S.A. 3 ее of Range Science, Colorado State University, Fort Collins, Colorado 80523, U.S.A Army Corps of Engineers, Engineering and Housing Support Center (CEHSC-FN), Fort Belvoir, VA 22060, USA The name of this t However, because of a typification. problem, the plants referred to as subsp. laxum from E arium, leaving this subspecies from the Island of Hawaii without a name. Therefore, we will not called subsp. aren tly Tetramolopium arenarium (A. Gray) Hillebr. subsp. arenarium var. arenarium. ast Maui must be use intraspecific designations and will refer to the plant discussed in this paper as simply Tetramolopium arenarium. ANN. Missour! Bor. GARD. 78: 1073-1080. 1991. 1074 Annals of the Missouri Botanical Garden Pohakuloa Training Area Critical Habitat Area Palila Kipuka Kalawamauna Endangered Plants Га і | | [ Lal Impact Area N | di ===> Scale oS 1000 о 1000 3000 5000 Meters 1 о 1 2 3 Mies Hawaii Ест Map of Pohakuloa о Area, Island of Hawaii, Hawaii, showing the location of the Kipuka ees. Endangered Plants Habita buds leads to development of a terminal capitules- cense. Capitulescense development leads to sym- podial branching, which occurs repeatedly on the plant (Lowrey, 1986). This distinctive branching pattern is represented in Figure 2 and is consistent with Leeuwenberg's architectural model of tropical trees (Hallé et al., 1978). Because branching phe- nology is not known, we included the number of tiers for each plant (i.e., the number of sympodial branch nodes) in а model of plant architecture. Our observations indicate that the epicotyledonary tier is the largest, and each successive tier de- creases in both length and diameter. We also ob- served that plants growing in relatively open site conditions were more erect, retained more live and dead foliage, and had longer internodes than plants Volume 78, Number 4 Laven et al. 1075 Population Structure of Tetramolopium arenarium 30 ст 0 ст Seedling Juvenile FIGURE 2. tier and flowering architecture. growing in relatively dense site conditions. Open- grown plants developed a free-standing, upright ""сћапдећег"' crowded situations were sprawling, and commonly morphology. Plants growing in their crowns were interlaced with crowns of other species. We are trying to determine the relationship between the number of tiers, terminal shoots, or capitulescences, and age. POPULATION ВЕБАМОК Mortality, growth, and reproduction data can be used in a transition matrix to project population behavior over time (Lefkovitch, 1965). This matrix technique can project population dynamics based on size structures or life history stages alone and, consequently, can be used with populations that cannot be aged (e.g., the lack of annuli) or where destructive sampling is not desirable (e.g., extreme- ly small population size). The projection starts with the size/stage struc- ture at one time and follows the change in structure over one time step as individuals either remain at the initial stage, grow to another stage, or die. Size/stage-specific survivorship, fecundity, and transfer rates project the future of the population (Menges, 1986) For this species, we have divided the life history into four stages: seedlings, juveniles, vegetative and reproductive (after Menges et al., 1985). Seedlings were arbitrarily defined as those individuals less than 10 cm in height; juveniles were defined as individuals taller than 10 cm and incapable of flowering (i.e., unbranched); vegetative plants have reached the minimum flowering size (i.e., at least one tier) but are nonreproductive (i.e., currently not flowering, but have flowered in the past); and reproductive plants (i.e., This paper describes the initial population struc- currently flowering). Reproductive Graphical representation of sympodial branching pattern for Tetramolopium arenarium, indicating ture. Population projections will be made after changes in the structure are determined, and life history stages may be redefined using Vandemeer’s (1978) optimization criterion. STUDY AREA The Tetramolopium population was rediscov- ered in the Kipuka Kalawamauna Endangered Plants Habitat (KKEPH) on the northwestern side of PTA (Fig. 1). KKEPH was designated cooper- atively by the U.S. Army, the U.S. Fish and Wild- life Service, and the Division of Forestry and Wild- life, Department of Land and Natural Resources, State of Hawaii. Two other endangered plants have previously been located within the Kipuka (Herbst & Fay, 1979): Haplostachys haplostachya (A. Gray) St. John and Stenogyne angustifolia A. ray. These two species were once more widely distributed; the extirpation of historic populations apparently has been due to effects of human dis- turbance, feral animals, and introduced species, rather than volcanic activity (Herbst & Fay, 1979). Kipuka Kalawamauna is a vegetated, older lava flow surrounded by younger, sparsely vegetated lava flows (Fig. 3). The elevation of the Kipuka is approximately 1,545 m, and it receives about 380 mm of precipitation annually (Douglas et al., 1989). The Tetramolopium population is situated on a mesic ridge system that runs west-northwest (Doug- as et al., 1989) with a rockland (rRO) soil type (USDA, 1973). Tetramolopium is part of a shrub- land community dominated by Dodonaea viscosa Jacq. and Dubautia linearis (Gaud.) D. Keck. — METHODS An initial field reconnaissance was conducted to determine the spatial extent of the population. Spe- 1076 Annals of the Missouri Botanical Garden Kipuka | " | L ; | = о AS 222. | RS Kalawamauna Legend = 4000-1500 yr. B.P. (pahoehoe lava) E3 4000-1500 yr. B.P. (aa lava) Ра 1500-750 yr. В.Р. (pahoehoe lava) c gza 1500-750 уг. В.Р. ===3 (aa lava) 34 750-107 yr. ВР. (aa lava) < 4000 y P. (pahoehoe lava) Historical- ША 1859 lava flow ш auna Кеа volcanic rock inliers FiGURE 3. et al., 1988). cial care and attention were taken to locate and identify all individuals so that few, if any, were overlooked. All individuals (both alive and dead) were located, affixed with numbered aluminum tags, and mapped. A center line, with known azimuth, was established along the long axis of the popu- lation, and individuals were mapped in coordinate space by measuring the distance along the center line and the offset distance (i.e., either left or right of, and perpendicular to, the center line). The beginning and ending points of the center line were permanently marked. During July and August of 1989, the following data were collected for each individual: total height, basal caliper, life history stage, number of branch- ing tiers, number of capitulescences, and the num- ber of flowering heads per capitulescence. Destruc- tive sampling, in order to assess the possibility of a relationship between size and age, was not con- ducted because of the small population size. RESULTS AND DISCUSSION The sample population is comprised of 134 live and 25 dead plants. As far as we know, this is the sole population of this species; our data are there- Ages of lava flows adjacent to the Kipuka Kalawamauna Endangered Plants Habitat (from Lockwood fore species data and not sample population data. Figure 4 illustrates the spatial distribution of the population. Table 1 presents population percent- ages by life history stage along with averages for height, basal caliper, tier number, capitulescence number, number of heads, and number of heads per capitulescence. Mean heights varied by life history stage and ranged from 0.38 m for juveniles to m for reproductive individuals. Similarly, mean basal cal- iper ranged from 3.36 mm to 12.83 mm. Mean tier number also ranged from one for vegetative individuals to nine for the reproductive stage. The reproductive individuals averaged 35 capitules- cences and 308 heads (capitula) per plant, leading to a mean of 9 heads per capitulescence. Figures 2, 6, and 7 illustrate the height class structure, basal caliper class structure, and tier number struc- ture, respectively, by life history stage for the entire population. Age class structures or distributions commonly have been used to interpret and predict population dynamics. Also, age distributions often have been inferred from plant size where age determination was either infeasible or impossible. In many cases it was assumed that size is correlated with age (Lorimer, 1985). However, if possible, this as- Volume 78, Number 4 Laven et a 1077 1991 Population Structure of Tetramolopium arenarium 25 — | Е 10 је. с s А ` s | x = Ya y Ы • о 0 “ly "+ NA * LZ = -5 . \, Е я 6 + -10 еа 1 е = ^ о 15 2 ° o . 20 А 25 ( 5 50 10 125 150 175 0 225 250 275 300 Plot Length (m) FIGURE 4. Coordinate map of Tet lop jum population ill ing spatial distributi f all individual sumption must be satisfied as prerequisite to the use of size as & surrogate for age. In our study it was not possible to determine plant age because this would entail destructive sampling of an ex- tremely small population without guarantee that a ring structure either exists or is related to age. Until aspects cf plant architecture can be related to age, we are compelled to use height, basal cal- iper, and tier structures in lieu of age structure to interpret population dynamics. Unimodal (bell-shaped or normal distribution) and descending monotonic (reverse J-shape) curves represent two contrasting size frequency distribu- tions. Unimodal size distributions may represent even-aged populations or multi-aged populations with relatively few seedlings or young individuals (Lorimer, 1985). This type of size distribution has TABLE 1. been interpreted to represent a senescent popula- tion that is not reproducing and is part of a seral community (Barbour et al., 1987). Reverse J-shaped size distributions typify all-age stands with rela- tively large numbers of seedlings of young individ- uals. Barbour et al. (1987) interpreted this struc- ture to indicate a high probability of a population maintaining itself in a relatively steady state com- munity. Figures 5, 6, and 7 clearly represent the bell- shaped size frequency distribution. These size struc- tures may indicate that the Tetramolopium pop- ulation is in jeopardy with a high probability of imminent extinction. There is, however, the pos- sibility that episodic establishment occurs when en- vironmental conditions are appropriate. Periodic establishment would be sufficient to maintain a Population characteristics of Tetramolopium arenarium. Life history characteristics Average + Standard deviation Number Height (m) Basal caliper (mm) Number of tiers Live plants Juvenile 7 0.34 + 0.11 3.16 + 1.06 0 Vegetative 17 0.59 + 0.10 5.51 + 1.59 1 + 0.70 Reproductive 110 1.00 + 0.25 12.83 + 4.65 9 + 4.65 Total live 134 0.91 + 0.30 11.39 + 5.27 7+ 5.14 Dead plants 25 0.97 + 0.36 11.08 + 4.80 7+ 3.91 Total for all plants 159 0.92 + 0.31 11.35 + 5.20 7 + 4.97 Maia chatacteristics Average + Standard deviation Capitulescences 34.81 + 43.52 eads 307.63 + 366.77 Heads per capitulescence 8.95 + 1.71 1078 Annals of the Missouri Botanical Garden АА Yw | AnA КС | КУ R۹۱١ КУ SSSSSSSSSSSN Number of Individuals o | RK ZIX Ф, E <> CX 7 Ф, е; + <] <Я | A хх? Росо хх = oY + ре = % <> Se хх 020202029. OOO 2252 ~ | Ф; + 9 bed >, eee 000, eee, SOKO 2 сг? <> SEK ооо MAAN | RR 53 653 VA, 0.1 0.2 0.3 04 0.5 06 07 0.8 0.9 10 11 12 13 14 15 16 17 1.8 Height (m) —— Size Classes ИШ e RSS Vegetative EZZA Reproductive С] Dead FiGURE 5. for each height class. population if the individuals are long-lived. Al- though life span for this species is not known, Lowrey (1986) considered individuals of the genus as probably not long-lived. The paradox, then, is how can a supposedly short-lived population such Size structure illustrating the number of individuals of Tetramolopium arenarium by life history stage as this persist in spite of low numbers of small (presumably young) individuals. In discussing the strategy of Leeuwenberg's ar- chitectural model, Hallé et al. (1978) suggested that species with this architecture may tend toward Number of Individuals 1 2 354 56 7 8 9 10 Basal Caliper (mm) RSS Vegetative B Juveniles 1 J 4 12 13 14 15 16 17 18 19 20 21 22 23 24 —— Size Classes EZA Reproductive Dead FIGURE 6. Size structure illustrating the number of individuals of Tetramolopium arenarium by life history stage for each basal caliper class. Volume 78, Number 4 1991 Laven et al. 1079 Population Structure of Tetramolopium arenarium 20 19 4 18 4 17 4 s] 1 С ü ©- a < L VOS Bea | GY 5 LIB 7 ub 1,2 EM [7 717 5 И 71717 4 + 99 ЛИН 2 E 222 1. 222 о 1 2 11 12 35 Number of Tiers —— Size Classes ШИШ Juveniles RRRA Vegetative Reproductive С] Dead RE 7. structure illustrating the number of individuals of Tetramolopium arenarium by life history stage Ficu for each tier clas an “r strategy” with rapid establishment rather coupled with Lowrey's (1986) estimates of floret numbers for this species (i.e., 22-45 ray florets and 4—9 disk florets per head), potential seed pro- duction may average approximately 10,000 per reproductive plant at any point in time. In light of the observation that flowering occurs in January, April, and August (and likely year-round), Теёга- output to qualify as an data on seed numbers, germination ео and percentage т which are necessary to sub- stantiate this claim, are yet to be obtained. Hallé et al. (1978) also suggested that species with this architecture may be associated with dis- turbed sites. Although fire history data do not exist, the study site apparently burned within the last decade. Tetramolopium has no obvious fire en- during or evading traits and is likely killed by fire, but perhaps fire produces a suitable, ephemeral regeneration environment that facilitates periodic establishment. If this is true, with the proper fire frequency, the size class distributions in Figures 5-7 may result from individuals establishing in the post-fire environment and acting as a relatively even-aged cohcrt over time. Consequently, periodic establishment raay be both sufficient and necessary to ensure the long-term persistence of this extreme- ly rare taxon. Caution should be exercised when interpreting these figures because of the relatively subjective way in which size classes were chosen. Because we do not know how well size is correlated with age, the possibility exists that individuals assigned to seedling or juvenile classes are simply old plants which, for unknown reasons, have not developed into reproducing individuals. If, with the passage of time, these individuals do not reproduce, our concerns about extinction would be strengthened. Alternatively, by assigning all individuals capable of flowering into either the vegetative or repro- ductive class (i.e., combining all age classes of branched individuals into only two life history stages), we may be mistakenly overstating our con- cerns about population decline. This overstatement may be true if individuals mature at different rates due to genetic or microenvironmental factors and continue to reproduce for long periods of time. However, regardless of which age surrogate is used, the population appears to be declining. This is true even if one of the measures (Figs. 5-7) happens to represent the true age of the population more accurately. ven if the size class structures suggested a vital, reproducing population, the uniqueness, isolation, and size of this population lead to the inescapable 1080 Annals of the Missouri Botanical Garden conclusion that the species is in danger of extinc- tion. Consequently, specific management plans must be developed to maximize the potential for popu- lation perpetuation. MANAGEMENT ÍMPLICATIONS AND RESEARCH NEEDS Franklin (1980) suggested an effective popu- lation size o as a minimum for maintaining long-term viability from a genetic viewpoint. Al- though this figure is an approximation, Lande (1988), and likely the biological conservation com- munity at large, does not feel justified in blanket application to species conservation. Regardless of the minimum number chosen, Tetramolopium is clearly in imminent danger of extinction. Several factors exacerbate the apparent tenuous existence of this species. Military operations, feral animals, exotic plants, and possibly fire should all be con- sidered when management plans are developed for the conservation of this species. Because of the immediacy of the threat to this species, we suggest the following changes be im- plemented to ensure its survival: (1) Military operations should be restricted, in- cluding reductions in foot and vehicular traffic to eliminate trampling, dust, and the potential use of Tetramolopium as camouflage. Accidental fire from ordnance also should be minimized. (2) The population should be fenced to exclude feral animals. The activities of feral hogs, sheep, and goats may have direct, deleterious impacts on this species. Additionally, rooting by feral hogs exposes bare mineral soil, a suitable establishment substrate for exotic plants. (3) This area is currently open to public hunt- ing, and until the effects of hunting activity (in- cluding accidental fire) can be determined, the area should be designated off-limits. Because so little is known about the biology and ecology of Tetramolopium, several research im- peratives exist. We are initiating studies on repro- duction biology (including pollination ecology, ger- mination and seedling establishment requirements, propagation also will be crucial to ensure the per- petuation of this species. We envision this study and the research imperatives described above as the first steps in a long-term monitoring program. The eventual creation of a suitable database will increase the likelihood of safeguarding this imper- iled species. LITERATURE CITED AYENSU, E. A. à R. A. DEFILIPPS. and Threatened Plants of the United States. Smith- sonian Institution and ús World Wildlife Fund, Inc., Washington, D. BARBOUR, M. G., J. H. Bunk & W. D. Pitts. 1987. oe Plant Ecology, 2nd edition. pe Cummings Publishing, Menlo Park, о V. 1978. Endangered narium (Asteraceae: Astereae) on the Pohakuloa "Training Area, Hawaii. Ann. Missouri Bot. Gard. 76: 1182-1185. И I. R. 198 Evolutionary change in small populations. Pp. 135-149 in Soulé & B. A. Wilcox (editors), Conservation Biology: An Evolu- tionary-Ecological Perspective. Sinauer Associates, Sunderland, Massachuset Harré, F., К. A. A. OLDEMAN & P. B. TOMLINSON. 1978. Tropical Trees and Forests. An Architectural Anal- ysis. Poun. Berlin. HERBST, D. R. & J. J. Fay. 1979. Endangered and threatened wildl that three Hawaiian plants are К ык species. Federal Register 44: 62468-62469. ГАМЕ, К. 1988. Genetics and о in biological conservation. Science 241: LerkovrrcH, L. P. 1965. The study of RR E in organisms grouped by stages. Biometrics 21: 18. > un P. W. LiPMAN, L. D. PETERSON & F. 1988. Generalized ages of surface lava flows of Mauna Loa Volcano, Hawaii. U.S. Geo- logical Survey Misc. Publ. Map I— 1908, U.S. Gov- ernment Printing Office, Washington, D.C. LORIMER, C. С. 1985. Methodological осе іп the analysis of forest disturbance history. Canad. J. For. Res. 15(1): 200-213. REY, T. К. 198 LOCKWOOD, J. P. R. Lowrey, T 6. A biosystematic revision of Ha- walian PONDUS (Compositae: Astereae). Al- на 4: 65. MENGES, E. S. 1986. Predicting the future of rare plant и demogr Аа e and modeling. Natural Areas J. 6(3): 1 `. GAWLER & D. M Waen. 198 5. Pop- up n of the endemic plant, Furbish's louse- UNITED STATES DEPARTMENT OF AGRICULTURE (USDA). 19 Soil Survey of Island of Hawaii, State of Hawaii. U.S. Government Printing Office, Washing- ton, D.C. m ED J. 8. y size in a stage projection matrix. anek 32: 79-8 NOTES A NEW SPECIES OF COSTUS FROM VERACRUZ, MEXICO Costus L. is a genus of about 90 species, with a pantropical distribution mostly in the Neotropics. The most recent taxonomic studies dealing with the genus are those by Maas (1972, 1977) and Maas & Maas (1990). Mexico represents the northernmost distribution limit of this genus in America. This genus is rep- resented in Mexico only by five species of subg. Costus: C. comosus (Jacquin) Roscoe var. bakeri (Schumann) Maas, C. scaber Ruiz & Pavón, and C. pulverulentus C. B. Presl within sect. Orni- thophilus; and C. pictus D. Don, C. guanaiensis var. tarmicus (Loesener) Maas, and C. guanaien- sis var. macrostrobilus (Schumann) Maas belonging to sect. Costus. Studies carried out during a comprehensive flo- ristic project of the Flora de la Estación de Biologia Los Tuxtlas, Veracruz, Mexico (Ibarra & Sinaca, 1987), have brought to light a distinctive new species. Costus dirzoi Garcia-Mendoza & Ibarra-Manri- TYPE: Mexico. Veracruz: Muni- cipio San Andrés Tuxtla, Estación de Biologia Tropical Los Tuxtlas, 18?34'-18?36'N, 95904' –9 5909 МУ, 200 m, 7 June 1989 (fl), Ibarra 3400 (holotype, MEXU; isotypes, BM, ENCB, К, LE, MO, U, US, ХАГ). Figure 1. A C. picto D. Don inflorescentia terminali in caule aphyllo (raro 1-2 foliis), callo conspicuo, bracteola puber- ula, ovario villoso et foliis supra glabris, infra dense ve- lutinis differt. quez, sp. nov. Plants 1-1.5(-2) m tall, pale green when dry, sheaths, ligules, and petioles sparsely puberulous. Sheaths (4-)4.5-7(-8) x 1-1.4(-2) cm, green; ligule (3-)4-9(-12) mm long, obliquely truncate, mostly dilacerating into fibers; petiole 3-10(-15) mm long. Leaves narrowly to broadly elliptic, (15-)20-27(-32) x 8-15 cm, (1.5-)2.1-2.5(- 2.8) times longer than wide, cuneate at the base, with a (1.5-)2-3 cm acumen at the tip, upper surface glabrous, lower surface densely veluti- nous. Inflorescence ovoid to fusiform, 3-6(-10) x (2.5-)3-4 cm, elongating to 16 х 5.5 cm in fruit, borne on a leafless shoot 20-40(-70) cm tall, or occasionally terminal on a leafy stem. Sheaths obliquely truncate, 4-5(-8) x 1.3-2 cm, minutely puberulous, sericeous at the insertion, reddish to green. Bracts broadly ovate, 2.5- cm, acute, coriaceous, exposed part green, glabrous, covered part reddish, puberulous; callus 5-6(-8) mm long, yellowish, rarely inconspicuous; bracteole 1.5-1.8(-2) ст long, sparsely puberulous; calyx _ 9 mm, sparsely puberulous, lobes deltate, (2- )3-4 mm long. Corolla 4.5-6 cm long, bright yel- low, lobes narrowly obovate, 4-5 x 1-1.8 cm, the dorsal one to 6 cm long, outer side minutely puberulous, tube 1-2 cm long. Labellum broadly obovate when spread out, 5-6.5 x 3-3.5 cm, yellow, margins crenulate, lateral lobes with a pu- bescent line inside, with dark red stripes, the middle lobe reflexed, irregularly 3—5 lobulate, tube 2-2.5 3.3-4.5 x 1-1.2 cm, yellowish white, apex dark red, rectangular, reflexed, anther 8-9(-10) mm, dorsifixed; ovary 4—6(-7) mm, densely villose; style filiform; stigma bilamellate with a 2-lobed appendage and one lu- nular, ciliate structure. Capsule broadly obovoid, 8-11 x 6-8 mm, densely villose, seeds 4-4.5 x 2.5-3 mm, black, 5-20 per fruit. Seedlings epi- geous. Common names. cm; stamen narrowly elliptic, Bordón, cana agria, caria de venado. Uses. The local people chew and suck on the stems to satisfy thirst. Distribution and habitat. known only from the Los Tuxtlas region, Veracruz, Costus dirzoi is Mexico, where it grows in primary tropical rain- forest but is more abundant in gaps produced by tree falls or edges of the forest and river banks, at elevations of 160-300(-700) m. The aerial parts of the plants show growth quickly after their stems are cut. The fruits and seeds are eaten and depredated (perhaps dispersed) by mice. Flowering occurs from May to June, rarely in August, and mature fruits are found from November to Decem- ANN. Missouri Вот. GARD. 78: 1081-1084. 1991. 1082 Annals of the Missouri Botanical Garden FIGURE 1. Costus dirzoi. А-1 from Ibarra 3400, J, К from Ibarra & Sinaca 2209, and L from Ibarra po A. Habit. — B. Part of a leafy stem. — C. Detail of leaf indument, cross section. — D. e —E. Bract and flower bud.—F. Flower.—G. Flower, showing three petals, labellum, and stamen.—H. Stamen and stigma о petaloid part.—I. Stigma seen from ventral surface. —]. Capsule, bracteole, mg bract.— К. See d.—L. Seedling Моште 78, Митбег 4 1991 Notes 1083 TABLE 1. Costus dirzoi and C. pictus Comparison of morphological features in C. dirzoi C. pictus Height 1-1.5(-2) m 1-4 m Sheath and ligule — Puberulous Glabrous or strigose Petiole Puberulous Glabrous Lower surface Velutinous Glabrous to of leaves densely puber- ulous, rarely strigose Length/width (1.5-)2.1-2.5 . (2.1-)3.7-4.6 ratio of leaves (-2.8) (-5.3) Inflorescence Leafless or Leafy shoot rarely leafy Callus of bracts Conspicuous Inconspicuous Bracteole Puberulous Glabrous to puberulous Anther 8-9(-10) mm 1-8 mm Ovary Villose Glabrous ber. The seeds germinate in 2-4 weeks, and seed- lings are particularly abundant in February and arch. Cultivated plants are found in the green- houses of Jardin Botánico Francisco Javier Clavi- јего, Xalapa, Veracruz (29 May 1985 (fl), /glesias 23 (XAL)), and Jardin Botánico de la Universidad Nacional Autónoma de México, Distrito Federal. Additional specimens examined. MEXICO. VERA- cruz: Municipo San Andrés Tuxtla, Estacion de Biologia Tropical Los Tuxtlas, 31 May 1983 (fl), Ibarra 641 (MEXU); 9 Nov. 1983 (fr), [barra 996 (MEXU); 25 Jan. 1984 (seedling), Ibarra 1219 (MEXU); 28 Feb. 1984 (seedling), Ibarra 1376 (MEXU); 3 July 1984 (seedling), Ibarra 1843, 1871 (MEXU); 28 Dec. 1984 ра > t ~ ~ = e ча ш o со Ne) E ра CN SF 8 5 в О [90 © л 5 = [т] > ms c LE 0 (fl), Martínez. Calderón 3103 (MEXU); 6 M, ENCB, K, MEXU, O, U, US, XAL) 15 Jan. 1990 (fr), Sinaca 1564 (ENCB, K, MEXU, MO, XAL); Montepio, some km fore Montepio, Z May 1980 (sterile), van Rooden 812 P deum Santiago m 6.5 km de Santiago Tux- .6 km a Cerro El Vigia, 12 Aug. 1960 p Ге e e 5599 (MEXU); Municipio de Catemaco, km 18 camino Las Palmas-Catemaco, 6 Ma 1960, González & Garza 3333 (MEXU); Municipio de Pajapan, 5 km NW of Pajapan, SE slopes of Cerro San Martin Pajapan, 3 Nov. 1981 (fr), Vee & Calzada 22759 XAL). Costus dirzoi belongs to subg. Costus because of its folded bracteole, bilamellate stigma and co- riaceous bracts. In addition, the labellum with a short tube and a distinct, exposed yellow limb, with lateral lobes striped with red, permit the placement of the new species in sect. Costus. The most closely related species is C. pictus, from which it differs by the features shown in Table 1. Costus pictus is a plant with a broad distribution range, from southeastern Veracruz to Costa Rica, whereas C. dirzoi seems endemic to the Los Tuxtlas region. Costus scaber and C. pulverulentus are sympatric with C. dirzoi, sharing the same habitat. Costus scaber, however, differs from the new species in having red-orange bracts and flowers. Costus pul- verulentus, on the other hand, differs from C. dirzoi by its pointed fusiform inflorescence and narrowly obovate leaves. In order to identify Costus dirzoi in the key of Maas & Maas (1990), the eighth couplet needs to be changed as follows and a new couplet, 8A, added. The Maases' use of х to identify second halves of couplets is preserved here. 8. Calyx 5-9 mm long 8A X Calyx 10-22 mm lon 8A. Lower side of leaf velutinous; capsule villose; plant green when dry (Mexico) X Lower side of leaf dede glabrous; capsule gla- brous; plant mostly brownish when dry (Central America and western South America) ..... C. laevis irzoi This species is named after Rodolfo Dirzo, who has been working intensively in forest conservation, especially in the Los Tuxtlas a We are grateful to i Cha for the Latin diagnosis and Paul Maas, Patricia Dávila, Fernando Chiang, and María de Lourdes Rico for critical reviews of the manuscript. We also thank Albino Luna for the original illustration and San- tiago Sinaca Colin for helping in fieldwork. LITERATURE CITED IBARRA M., С. & S. SINACA С. 1987. Listados rg de México VII. Estación de Biologia Tropical Los uxtlas, piede, Instituto de Biologia, Univ. Na- cp rag de México, México. Maas, P. J. 1972. Flora Neotropica Monograph No. 8. Orte dp c dira Hafner, New York. 1977. otropica paa o. 18. Renealmia ma aceae, Zi d Costoi- deae жекен (Zingiberaceae). Hafner, New York. H. Maas. 1990. Not otes on ke NS dp Zingiberaceae: IV. Some new species of Costus and Renealmia. Notes Roy. Bot. Gard. Edinburgh 463) 20. 1084 Annals of the Missouri Botanical Garden —Abisai Garcia-Mendoza, Jardin Botánico, lIn- tlas, Instituto de Biología, Universidad Nacional stituto de Biología, Universidad Nacional Au- | Autónoma de México, Apartado Postal 94, San tónoma de México, Apartado Postal 70-614, Andrés Tuxtla, Veracruz, México. Present ad- 04510 México, D.F.; and Guillermo Ibarra-Man- dress: Apartado Postal 22-733, México 22, D.F. riquez, Estación de Biología Tropical Los Tux- NOTICE THE 1991 Jesse M. GREENMAN AWARD The 1991 Jesse M. Greenman Award has been won by Scott Zona for his publication “А mo graph of Saba! (Arecaceae: Coryphoideae),” Subs lished in Aliso 12: 583-666. 1990. graphic study is part of a doctoral dissertation from the Claremont Graduate School, Claremont, Cali- fornia, under the direction of Sherwin Carlquist. The Greenman Award, a certificate and a cash prize of $500, is presented each year by the Mis- souri Botanical Garden. It recognizes the paper his mono- judged best in vascular plant or bryophyte system- atics based on a doctoral dissertation published during the previous year. Papers published during 199] are now being accepted for the 24th annual award, which will be presented in the summer of 1992. Reprints of such papers should be sent to P. Mick Richardson, Greenman Award Committee, Missouri Botanical Garden, P.O. Box 299, St. Lou- is, Missouri 63166-0299, U.S.A. In order to be considered for the 1992 award, reprints must be received by 1 June 1992. Volume 78, Number 2, pp. 273-538, of the ANNALS OF THE MISSOURI BOTANICAL GARDEN was published on ay 1, 1991 Volume 78, Number 3, pp. 539-855, of the ANNALS OF THE or BOTANICAL GARDEN was published on 19 ugust 19, Instructions for Authors, Annals of the Missouri Botanical Garden The Annals publishes и ا‎ in systematic botany and related fields. n English or Spanish. When in Spanish, an English е in addition scripts. Manuscripts sent already reviewed and revised are generally subject to the full normal review process initiated here. Manuscripts not properly prepared may be returned for revision prior to review. This includes man- uscripts with improper ios or with diacritical marks not indicated. Page costs are $80 per page, although charges may be reduced under certain circumstances. Corrections in proof will be billed nonnegotiably to the author at $2.50 per line Manuscripts must be typed (output from dot matrix printers must be entirely clear) on one side of nonglossy 14 x 11 inch paper. Leave at least one inch as margin all around, except on the first page. There three inches cluding abstract, notes, legends, tables, enn lists, Literature Cited, and footnotes. Do no stify the right-hand margin, and do not break ee there. Italics, boldface, or other special typefaces should not be used—they will appear later as communicated to the printer by the editor. Words to be set in italics should be underlined. Accents, umlauts, and other diacritical marks are the responsibility of the author. Footnotes should be typed double-spaced as paragraphs on the first page. The first footnote contains acknowledg- ments, including information on granting agencies, and herbaria that loaned specimens. The second footnote is the author’s address. Addresses for additional authors are each separate additional footnotes. No other footnotes are permitted, except footnotes in tables, and these are best avoide Long, complex tables are discouraged. Tables not neat and easily understood will be returned for improvement. Leave out vertical lines, and avoid horizontal lines. Lines are added as needed by the editor and/or printer. Type table captions double-spaced as paragraphs at the tops of the tables. Periods are used after all abbreviations (which should ompass direc- br viated, and cities are usually spelled out, although “St.” is acceptable. Abbreviate periodicals according to Gardens, Kew, and books according to Taxonomic Literature. Note, however, that book abbreviations are to be capitalized, and that books listed taxonomic authorities may be spelled out in major taxo- nomic citations such as the formal citation at the head of à taxonomic treatment. Except by prior arrangement with authors not having access to crucial references, the editor does not provide the service of looking up abbreviations. In the Literature Cited, periodicals should be listed as follows: author's last name, initials). Year. Full title of article. Journal abbreviated as in BPH. Volume: pages. ore than one author, follow this style: author's last name, initial(s), second author's initial(s), last name & third au- thor's initial(s), last name. Citations of books should appear as follows: author's last name, initials). Year. Full Title (edited by Editor), 3rd edition, Volume 2. Publisher, Place. Citations of work “‘in prep.” and similar references to inaccessible sources are strongly discouraged. Except by prior arrangement when necessary, the editor does not perform library research to fill in missing bibliographic information A brief Latin diagnosis for each new taxon is preferred to a complete Latin description е one paragraph per basionym as follows: Taxon author, literature citation, type citation, e.g., Bauhinia dipetala Hemsl., Diag. Pl. Nov. 1880. Casparia dipe- tala (Hemsl.) Britt. & Rose, М. Au Fl. 23: 216. 1930. TYPE: Mexico. Veracruz: Valle de Cordóba, Bourgeau 1713 (holotype, K; isotypes, F, GH, P, US). Exclamation points are not used for specimens examined. Indicate unseen types as not seen. Keys should be dichotomous, indented, and consistent with descriptions. Couplets should be numbered as 1a, lb, 2a, and so forth Cite specimens as follows: Additional specimens examined (or Selected speci- mens examined or Paratypes). MEXICO. OAXACA: Sierra San Pedro Nolesco, Talea, 3 Feb. 1987 (fl), Jergensen 865 (BM, С, К, US). Dates and reproductive status are optional Abbreviate herbaria according to Index Herbariorum, edition Send copies of illustrations; do not send originals until requested if the manuscript is accepted. Illustrations will be returned after publication. All illustrative materials should be mounted on stiff cardboard. The maximum size for printed illustrations is 5 x 8!4 inches. Presize illus- trations to fit either column width (2% inches or са. 68 m) or full page width (5% inches or са. 140 mm). Figures are numbered in Arabic numerals in the order of their citation in the text. Parts of figures are labeled with capital letters. Photographs should be black-and-white ographs. Do not use stripping between plate or figure segments. Make sure edges are squared. Maps should include reference to latitude and longitude. Scale bars should appear on illustrations, on photographs, and on maps. Magnifications/reductions indicated in cap- tions are unwelcome to the point of usually not being accepted. Scanning electron micrographs should be free of conspicuous charging. Axes on graphs are all labeled. No explanatory text is to appear in figures. Rather, provide clear captions in sequential paragraphs apart from the other text, one par agraph for each group of figures, and following the style in current issues of the Annals. For other aspects of style, consult a recent issue of the Annals; the latest edition of The Chicago Manual of Style; or write the editor. Communication with the editor at all points of the pubhesnon proce is ево raged. cronies uld be addressed the Editor, Annals of the Miscou Bo- tanical Garden, P.O. Box 299, St. Louis, Missouri 63166, U.S.A ANNALS OF THE MISSOURI BOTANICAL GARDEN VOLUME 78 1991 Colophon This volume of the ANNALS of the Missouri Botanical Garden has been set in APS Bodoni. The text is set in 9 point type while the figure legends and literature cited sections are set in 8 point type. The volume has been printed on 70# Vintage Gloss. This is an acid-free paper designed to have a shelf-life of over 100 years. Vintage Gloss is manufactured by the Potlatch Paper Company. Photographs used in the ANNALS are reproduced using 300 line screen halftones. The binding used in the production of the ANNALS is a proprietary method known as Permanent Binding. The ANNALS is printed and distributed by Allen Press, Inc. of Lawrence, Kansas 66044, U.S.A. (O Missouri Botanical Garden 1991 ISSN 0026-6493 VOLUME 78 1991 ADAMS, C. DENNIS. New Combinations and a New Variety in the Cyperaceae of Mesoamerica 254 ANDERBERG, ARNE A. & KARE BREMER. Parsimony Analysis and Cladistic Reclassification of the Relhania Generic Group (Asteraceae-Gnaphal- ieae) 1061 Bari, AMINA. (See Peter Goldblatt, John C. Manning & Amina Bari) .......... 950 BARNEBY, RUPERT C. Notes on Swartzia (Leguminosae: Swartzieae) Primary to the Flora of Venezuelan Guayana 177 BARRIE, FRED R., James L. REVEAL, CHARLES E. Jarvis & ALWYN Н. GENTRY. On the Typification of Bignonia crucigera (Bignoniaceae) 264 BARRINGER, KERRY & FRANCISCO GUANCHEZ. A New Species of Aristolochia (Aristolocheaeceae) from Venezuela 262 ВЕНМКЕ, H.-DIETMAR. Sieve-Element Characters of Ticodendron .................... 131 Bourronp, Davip Е. (See Ying Tsun-Shen, David E. Boufford & Tu Yuling) 338 BREMER, KARE. (See Arne A. Anderberg & Kare Bremer) 1061 Briccs, BARBARA. One Hundred Years of Plant Taxonomy 19 BURGER, W. С. (See Barry Hammel & W. С. Burger) 89 Bunr-UrLEY, KATHLEEN. (See John Е. Utley & Kathleen Burt-Utley) ....... 266 CARLQUIST, SHERWIN. Wood and Bark Anatomy of Ticodendron: Comments on Relationships 96 CARVALHO, L. D'A. FREIRE DE. New Таха of Solanum (Solanaceae) from Brazil, Colombia, Central America, and Venezuela 224 CHEN, YuNc-Kuaw. (See Ching-I Peng & Yung-Kuan Chen) 995 CLARK, LYNN С. (See Yvonne Widmer & Lynn С. Clark) . 164 CRAWFORD, D. J. (See J. Н. Hunziker, A. К. Wulff, T. К. Stuessy, D. J. Crawford & J. V. Crisci) 512 AE J.V. (See J. H. Hunziker, А. F. Wulff, Т. F. Stuessy, D. J. Crawford & J. V. Crisci) 512 Рату, DoucLas C. (See John D. Mitchell & Douglas С. Пау) 184 DIERSING, VICTOR E. (See Richard D. Laven, Robert B. Shaw, Patricia P. Douglas & Victor E. Diersing) 1073 Риснев, Davip L. (Зее Robert N. Schwarzwalder, Jr. & David L. Dilcher) 962 Dopson, C. H. & A. H. GENTRY. Biological Extinction in Western Ecuador 273 DoucLas, Patricia P. (See Richard D. Laven, Robert B. Shaw, Patricia P. Douglas & Victor E. Diersing) 1073 Dowp, J. M. (See P. G. Martin € J. М. Dowd) . 296 ENpREss, PETER К. (See Henk van der Werff & Peter К. Endress) ............ 401 Ewan, JOSEPH. Who Conquered the New World? or Four Centuries of Exploration in an Indehiscent Capsule FEUER, SYLVIA. Pollen Morphology and the Systematic Relationships of Ticodendron incognitum García-MENDOZA, ABISAÍ & GUILLERMO IBARRA-MANRÍQUEZ. А New Species of Costus (Costoideae, Zingerberaceae) from Veracruz, Mexico _____ | GENTRY, ALWYN H. (See Fred К. Barrie, James L. Reveal, Charles E. Jarvis & Alwyn H. Gentry) GENTRY, A. H. (See C. H. Dodson & A. H. Gentry) GOLDBLATT, PETER & JAMES E. HENRICH. С ‘alydorea Herbert (Iridaceae- Tigridieae): Notes on this New World Genus and Reduction to Synonymy of Salpingostylis, Cardiostigma, Itysa, and Саша GOLDBLATT, PETER & NEIL SNOW. Systematics and Chromosome Cytology of Eleutherine Herbert (Iridaceae) GOLDBLATT, PETER, JOHN C. MANNING & AMINA Bari. Sulcus and Operculum Structure in the Pollen Grains of Iridaceae Subfamily Ixioideae .......... GÓMEZ-LAURITO, JORGE & Luis D. Gómez P. Ticodendraceae: A New Family of Flowering Plants GONZALEZ G., Favio. Notes on the Systemtics of Aristolochia Subsect. Hexandrae GRAHAM, ALAN. Studies in Neotropical Paleobotany. VIII. The Pliocene Communities in Panama— Introduction and Ferns, Gymnosperms, An- giosperms (Monocots) GRAHAM, ALAN. Studies in Neotropical Paleobotany. IX. The Pliocene Com- munities of Panama—Angiosperms (Dicots) GRAHAM, ALAN. Studies in Neotropical Paleobotany. X. The Pliocene Com- munities of Panama—Composition, Numerical Representations, and Pa- leocommunity Paleoenvironmental Reconstructions GUANCHEZ, FRANCISCO. (See Kerry Barringer & Francisco Guánchez) ___ GUILIETTI, ANA Maria. (See Nancy Hensold & Ana Maria Guilietti) ____ GUILIETTI, ANA MARIA. Synonymization of the Genera Comanthera and Carptotepala with Syngonanthus (Eriocaulaceae) HAMMEL, Barry & W. C. BURGER. Neither Oak nor Alder, but Nearly: the History of Ticodendraceae HENRICH, JAMES E. (See Peter Goldblatt & James E. Henrich) ________ Немзогр, Nancy. Revisionary Studies in the Eriocaulaceae of Venezuela HENSOLD, NANCY & ANA MARIA GuiLIETTI. Revision and Redefinition of the Genus Rondonanthus Herzog (Eriocaulaceae) НЕРРЕВ, К. NIGEL. The Niger and the Nile: Botanical Exploration Around Two African Rivers Некзнком1т2, MARK А. Leaf Morphology of Cistanthe Spach (Portulaca- ceae) 57 143 81 1022 HERSHKOVITZ, MARK А. Phylogenetic Assessment and Revised Circumscrip- tion of Cistanthe Spach (Portulacaceae) Hickey, Leo J. & Рамр WINSHIP TAYLOR. The Leaf Architecture of Ti- condendron and the Application of Foliar Characters in Discerning Its Relationships Hunziker, J. H., А. Е. Wurrr, T. F. SruessY, D. J. CRAWFORD & J. V. CRISCI. Chromosome Numbers in the Genus Nassauvia (Compositae- Mutisieae) IBARRA-MANRÍQUEZ, GUILLERMO. (See Abisai Garcia-Mendoza & Guillermo barra-Manriquez) Janvis, CHARLES E. (See Fred R. Barrie, James L. Reveal, Charles E. Jarvis & Alwyn H. Gentry) Jiménez M., Quirico. — Peltostigma parviflorum (Rutaceae), Nueva Especie de Costa Rica y Colombia Jupziewicz, EMMET J. (See Fernando O. Zuloaga & Emmet J. Judziewicz) KEATING, RICHARD С. (See Carol A. Тойла & Richard C. Keating) .............. KraL, R. Erratum for Xyris (Xyridaceae) from the Planalto of Brazil ........ Копт, Јов. Panamanthus, А New Monotypic Genus of Neotropical Loran- thaceae LAFERRIÈRE, JOSEPH E. — Пазу топ wheeleri var. durangense: A New Com- bination in the Nolinaceae Laven, RICHARD D., ROBERT B. SHaw, Patricia P. DOUGLAS & VICTOR Е. IERSING. Population Structure of the Recently Rediscovered Hawai- ian Shrub Tetramolopium arenarium (Asteraceae) LUTHER, Harry. (See John F. Utley & Harry Luther) MANNING, JOHN С. (See Peter Goldblatt, John C. Manning & Amina Bari) MANNING, STEPHEN D. Intraspecific Variation in Pavetta rigida (Rubiaceae): Estimates of Reliability of Taxonomic Information MARTIN, P. С. & J. M. Похр. Studies of Angiosperm Phylogeny Using Protein Sequences MICKEL, JOHN T. Four New Species of Elaphoglossum (Elaphoglossaceae) from Venezuela MITCHELL, JOHN D. & Doucras C. DALY. Cyrtopcarpa Kunth (Anacardi- aceae) in South America Moran, Rossin C. Monograph of the Neotropical Fern Genus Stigmatop- teris (Dryopteridaceae) MORRONE, OsvaLDO. (See Fernando Zuloaga & Osvaldo Morrone) ................. MORRONE, OSVALDO. (See Fernando Zuloaga & Osvaldo Morrone) ................. MORRONE, OsvaLDO & FERNANDO ZULOAGA. Revisión del Género Strepto- stachys (Poaceae-Panicoideae), su Posición Sistemática dentro de la Tribu Paniceae NicoLsoN, DAN H. A History of Botanical Nomenclature 1009 Orozco, CLARA INÉS. Análisis Multivariado del Complejo de Tres Especies de Brunnellia (Brunelliaceae) PENG, CHING-1 & YUNG-KUAN CHEN. — Hybridity and Parentage of Begonia buimontana Yamamoto (Begoniaceae) from Taiwan PiPoLy, JOHN J. Ш. Ardisia lundelliana, a New Species of Myrsinaceae rom Panama REVEAL, JAMES L. Botanical Explorations in the American West— 1889- 1989: An Essay on the Last Century of a Floristic Frontier ...................... REVEAL, JAMES L. (See Fred R. Barrie, James L. Reveal, Charles E. Jarvis & Alwyn H. Gentry) ROHWER, JENS G., HANS GEORG RICHTER & HENK VAN DER WERFF. Two New Genera of Neotropical Lauraceae and Critical Remarks on the Generic Delimitation RUDOLPH, EMANUEL. One Hundred Years of the Missouri Botanical Garden SCHUBERT, BERNICE С. (See Oswaldo Téllez Valdes & Bernice С. Schubert) SCHWARZWALDER, ROBERT N., JR. & Davip L. DILCHER. Systematic Place- ment of the Platanaceae in the Hamamelidae SHAW, RoBERT B. (See Richard D. Laven, Robert B. Shaw, Patricia P. Douglas & Victor E. Diersing) Snow, NEIL. (See Peter Goldblatt & Neil Snow) Sorta, NERIDA. Two New Species of Trixis (Asteraceae-Mutisieae) from araguay STUESSY, T. Е. (See J. H. Hunziker, А. Е. Wulff, T. Е. Stuessy, D. J. Crawford & J. V. Crisci) TAYLOR, CHARLOTTE M. Two New Species of Hillia subg. Карта (Rubi- aceae) from Venezuela and Mexico TAYLOR, Davip WINSHIP. (See Leo J. Hickey & David Winship Taylor) ... Tose, HIROSHI. Reproduction Morphology, Anatomy and Relationships of Ticondendron Topzia, CAROL A. & RICHARD C. KEATING. Leaf Architecture of the Chlo- ranthaceae UTLEY, JOHN Е. & KATHLEEN BurT-UTLEY. Studies in Middle American Bromeliaceae I UTLEY, JOHN К. € HARRY LUTHER. Studies in Middle American Bromeli- aceae II VALDÉS, Oswarpo TÉLLEZ & BERNICE С. SCHUBERT. Especies Nuevas y Colecciones Notables de Dioscorea (Discoreaceae) en Mesoamérica ..... VAN DER WERFF, HENK. A Key to the Genera of Lauraceae in the New VAN DER WERFF, HENK. New Species of Lauraceae from Ecuador and Peru VAN DER WERFF, HENK & PETER К. ENpnEss. Gamanthera (Lauraceae), а New Genus from Costa Rica 1073 VASQUEZ M., RODOLFO. Caraipa (Guttiferae) del Peru WEBSTER, GRADY L. A New Brazilian Species of Dalechampia subsect. Triphyllae (Euphorbiaceae) WIDMER. YVONNE & LYNN С. CLARK. New Species of Chusquea (Poaceae: Bambusoideae) from Costa Rica WULFF, А. К. (See J. Н. Hunziker, A. F. Wulff, T. Е. Stuessy, D. J. Crawford & J. У. Crisci) YING, TSUN-SHEN, Davip E. BOUFFORD & YULING Tu. Phytogeographical Relationships of the Genera of Angiosperms in the Fanjing Shan Moun- tain Range, Northeastern Guizhou, China Упимс, Tu. (See Tsun-Shen Ying, David E. Boufford € Yuling Tu) ...... ZULOAGA, FERNANDO & Еммет J. JupziEWiCZ. А Revision of Raddiella (Poaceae: Bambusoideae: Olyreae) ZULOAGA, FERNANDO & OsvALDO Morrone. Estudios Morfológicos en el Subgénero Dichanthelium de Panicum (Poaceae: Panicoidae: Pani- ceae), con Especial Referencia a Panicum sabulorum ZULOAGA, FERNANDO & OsvaLDO Morrone. New Species of Panicum, sub- genus Dichanthelium (Poaceae: Paniceae) from South America ZULOAGA, FERNANDO. (See Osvaldo Morrone & Fernando Zuloaga) 1002 Monographs in Systematic Botany from the Missouri Botanical Garden Number 34. ‚ FLORA OF NIGERIA: CARYOPHYLLALES. Shahina A. Ghazanfar. ( The Caryophyllales are represented by 73 species in Nigeria. In this publication, description, distri- | bution, flowering period, and notes on habitat are given for each species. Information on medicinal and | ether uses and vernacular names are also given for most of the species. , 39 pp., illustrated. 1991. $10.00. { ‘Number 35 | THE ETHNOBOTANY OF AUBLET'S HISTOIRE DES PLANTES DE LA СЛАМЕ FRAN. COISE (1775). Mark J. Plotkin, Brian M. Boom, and Malorye Allison. Fusée Aublet's Histoire des Plantes de la Guiane Frangoise is best known for its contributions to \ systematics but also contains significant ethnobotanical information. This publication presents Aublet's ; Comments on 124 useful plant species, translanted into English, and includes updated nomenclature. , Reproductions of the original plates of these plants are included. 108 pp, illustrated. 1991. $12.00 ¿ | | m—————_ To place an к! ri check or money order in U.S. funds, payable through | а: US: AES U.S. shipments: add $2.00 for one book and $.75 for each additional book; non-U.S. shipments: add $3.00 for one book, and $.75 ir | ĉach additional book. ови should be prepaid; а $1.00 fee EA be added to orders requiring invoices. No shipme | зу made until payment is received. Mail form with your check or money order, preme to Missouri Botanical 1 Garden, | | Please send copy(ies) of Monograph No. 34. وروا چا‎ Eleven, Missouri Botanical Gard | | Р.О. Please send copiis рог Mencgreph No. 55. St. Louis, MO 63166-0299, U.S.A. | Send Tapes order to: | [1 Payment enclosed. | un ; PI Send Stoic ERE 00 fee will be | | - added to to з d Address س‎ Postal Code genas о А | Bem uar noc TOM. Prices ARE SUMECY To Chance V Witnour Norice СОМТЕМТ5 Monograph of the Neotropical Fern Genus Stigmatopteris (Dryopteridaceae) Robbin C. Moran : Estudios" Morfológicos en el Subgenero Dichanthelium de Panicum (Poaceae: Panicoidae: Paniceae), con Especial Referencia a Panicum sabulorum Osvaldo Morrone y Fernando Zuloaga | А Revision of Raddiella (Poaceae: Bambusoideae: Olyreae) Fernando O. Zuloaga & Emmet J. Judziewicz Systematics and Chromosome Cytology of Eleutherine Herbert acess) Peter Gold- blatt & Neil Snow : Sulcus. and Operculum | Structure in the Pollen Grains of lridaceae Subfamily | Ixioideae Peter Goldblatt, John С. Manning & Amina Bari .. SESER TEN: Systematic Placement of the Platanaceae i in the Hamamelidae Robert N. QUE | walder, Jr. & David E Dilcher .. 2 alisis Multivariado del Complejo de Tres Especies de Brunellia penne) Clara Inés Orozco с BAEC Уу E _ Hybridity and Parentage of Begonia buimontana Yamamoto НЕ from Nee Taiwan _ Ching-l Peng & Yung-Kuan Chen ;. | MGE RACER а — COS Caraipa. (Guttiferae) del Peru - Rodolfo Vasquez M.. E demi fs x Phylogenetic Assessment and Revised iesu of Cistanthe Spach (Portulaca NOE. . Mark A. Hershkovitz DSA pa > Leaf Morphology of Cistanthe Spach фый | Mark А. Hershkovitz . eat : Parsimony Analysis and Cladistic Reclassification of the R li ата С‹ ic Group p (A teraceae- _ Gnaphalieae) A Arne A. Anderberg & Kåre Bremer ...... 5 PEE rum and s А Population Structure of the оа Rediscovered Hawaiian Shrub Tetramolopium aren- : ааа P: Douglas (Costoideae, Zig t from Veracruz, Mexico | oe Species of Costus ( _ Abisat ‚ García Mendoza & Guillermo fi Mead л ae Denis me, 857 915 928 942 950 . 1022 | 1061 pup ——— _"_ВЖУЋ ЦРНУ У РУШИ НИОР ко | да аи "тл Ak quem rp t nme "^