JOURNAL OF THE ARNOLD ARBORETUM HARVARD UNIVERSITY VOLUME 71 NUMBER 1 ISSN 0004-2625 Journal of the Arnold Arboretum The Journal of the Arnold Arboretum (ISSN 0004-2625) is published quarterly in January, April, July, and October for $70.00 per year, plus $10.00 or $15.00 postage for addresses outside of the United States, by the Arnold Arboretum of Harvard University. It is printed and distributed by Allen Press, Inc., 1041 New Hampshire treet, Lawrence, Kansas 66044. Second-class postage paid at Lawrence, Kansas. POSTMASTER: send address changes to Journal of the Arnold Arboretum, % Allen Press, Inc., P.O. Box 368, Lawrence, Kansas 66044. Subscriptions and remittances should be sent to Journal of the Arnold Arboretum, 1041 New Hampshire Street, Lawrence, Kansas 66044, U.S.A. Claims will not be accepted after six months from the date of issue. Volumes I-51, reprinted, and some back numbers of volumes 52-56 are available from the Kraus Reprint Corporation, Route 100, Millwood, New York 10546, ULS.A. EDITORIAL COMMITTEE S. A. Spongberg, Editor E. B. Schmidt, Managing Editor P. F. Stevens, Book Review Editor P. S. Ashton K. S. Bawa C. E. Wood, Jr. Printed at Allen Press, Inc., Lawrence, Kansas COVER: The stylized design appearing on the Journal and the offprints was drawn by Karen Stoutsenberger. THIS PUBLICATION IS PRINTED ON ACID-FREE PAPER. IMPORTANT NOTICE After the final issue of volume 71, the Journal of the Arnold Arboretum will suspend publication pending a comprehensive review of its role in supporting the mission of the Arboretum. In 1988 the Arnold Arboretum underwent administrative review following the resignation of Dr. Peter Ashton as director. This led to a reaffirmation of its mission by the President and Fellows of Harvard College (the Corporation; see mission statement on facing page). It also led to changes in its administrative position within Harvard University and a sub- sequent search for a new director, who assumed the position in Jan- uary 1989. In addition, the Harvard University Herbaria (of which the collections of the Arnold Arboretum are a part) are presently going through a period of staff and faculty transition. Consequently it has seemed prudent to suspend publication of a journal that was originally intended to serve the mission of the Arnold Arboretum as an organ “to publish within a reasonable time the material which is gathered in its laboratories.””! During the next two years, the research mission of the Arnold Arboretum will be undergoing review and clarification. The shape of systematic and botanical scholarship within Harvard’s academic departments will become clear as new staff and faculty are hired. At this future point, the potential supporting role of an in-house journal will be assessed. If the Journa/ can further the scholarly mission of the Arboretum, publication will be resumed. During the suspension, the Arnold Arboretum will continue to support publication of the Generic Flora of the Southeastern United States. I will be happy to address the queries and concerns of subscribers about this suspension of publication. Robert E. Cook Director and Arnold Professor 'C. S. Sargent, Introduction, Journal of the Arnold Arboretum, Vol. 1, p. [i], 1919. THE MISSION OF THE ARNOLD ARBORETUM The Arnold Arboretum is held in trust by Harvard University according to the terms of the Indenture of Trust of 1872. This legal document provided for the creation of the living collections as a practical demonstration of the variety of plants that could be grown in this climate. It also mandated the appointment of the Arnold Professor to manage the Arboretum and to teach the knowledge of trees and related topics. Although there have been many significant changes in the fields of botany and horticulture and the Arnold Arboretum itself has fulfilled many different purposes during the past century, the basic premises of the Indenture still hold. It is most appropriate to reemphasize the traditional strengths of the Arnold Arboretum through a strong focus on botany and horticulture. The mission of the Arnold Arboretum may be outlined as follows: (1) to develop, curate, and maintain a well-documented collection of living woody plants from around the world that are hardy in the Boston area; (2) to study these plants and their relatives and associates in nature through the maintenance of a herbarium and library and through directly related research in botany and horticulture; (3) to provide instruction in botany, horticulture, dendrology, and other fields related to the living collections. As part of the City of Boston’s park system, the Arboretum’s Jamaica Plain site functions as an outdoor museum open to the public. The highest priority of the Arboretum’s administration is the proper curation and maintenance of these living collections. Proper curation includes acquisitions through field expeditions and exchanges with other institutions as well as cultivation and propagation of existing specimens to enhance and maintain the scientific and instructional value of the collections. The second priority of the Arboretum, the study of its collections and their relatives and associates in nature, directly benefits the curation of the collec- tions. The preserved collections and the library are indispensable tools for this research, and their curation and maintenance are therefore essential. All of the Arboretum’s resources— the living collections, the preserved collections, and the library—are available for scholarly research. The director of the Arnold Arboretum is responsible for ensuring that adequate and up-to-date materials are available for ane and future scholars to teach and to pursue research in the areas represen Educational ak are the Arboretum’s third-highest priority. The Arnold Arboretum offers a variety of programs for public instruction in horticulture, botany, dendrology, and landscape gardening and, in addition, disseminates knowledge of plants through its publications. The director and Arboretum professional staff may also offer courses, as appropriate, within other academic programs of Harvard University. —Document approved by the Harvard Corporation, February 1988. JOURNAL OF THE ARNOLD ARBORETUM VOLUME 71 JANUARY 1990 NUMBER | THE GENERA OF BETULACEAE IN THE SOUTHEASTERN UNITED STATES! JOHN J. FURLOW? BETULACEAE S. F. Gray, Nat. Arr. Brit. Pl. 2: 243. 1821, ‘““Betulideae,” nom. cons. (BIRCH FAMILY) Small to large, columnar or pyramidal to spreading dec1 s or shrubs; sap watery; branching excurrent to deliquescent; trunks and nee terete to irregularly longitudinally fluted, the branchlets terete, slender, often distichous, uniform or differentiated into long and short shoots. Bark close or exfoliating ae iy | C 4h c } 'Prepared fo United States, a long-term project made possible by grants ee the National Science Foundation and at this writing supported ra BSR-8415637 (Norton G. Miller, principal investigator), under which this account BSR-8415769 (Carroll E. Wood, Jr., principal investigator). This treatment, the 130th i in the series, follows the format established in the fete paper (Jour. Arnold Arb. 39: 296-346. 1958) and continued to the present. The area covered by the Generic Flora includes North and South Carolina, Georgia, Florida, Tennessee, Alabama, MissespD i pTeansas, and Pons The descrip mons: ate based primarily on the plants of this area g brackets. Those references I did not verify are marked with asterisks. The illustrations were prepared by Arnold D. Clapman under an earlier grant from the National the drawings. The materials used were from plants in the Arnold Arboretum, as well as from wild I extend my sincere thanks to Norton Miller and Carroll Wood for their help, encouragement, and patience throughout this study. I am especially grateful to Dr. Wood for overse: eing the preparation Miller for providing me with funds to travel useum to study herbanum material there s. Wood and Miller have both read the manuscript and have made many useful suggestions. Bruce P. Dancik, Peter R. Crane, Tho Lammers have also read and commented on the manuscript. I am thankful to them, and also to the many others with whom I have discussed aspects of this wor 2Department of Botany, Ohio State University, Columbus, Ohio 43210. © President and Fellows of Harvard College, 1990. Journal of the Arnold Arboretum 71; 1-67. January, 1990. 2 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 in thin layers, thin, smooth, often marked with prominent lenticels, sometimes becoming thick, corky, and rough, furrowed or scaly with age, often strongly tanniferous; young twigs glabrous to pubescent, sometimes covered with resi- nous glands, terminal buds lacking; leaf scars raised, with 3 vascular bundle scars; winter buds stipitate or sessile, narrowly to broadly ovoid, terete or sometimes angular in cross section, divergent, held parallel to the twig, or appressed, acute to rounded at apex, covered with 2 smooth, valvate (stipular) scales or few to many smooth or longitudinally striate imbricate ones [or occasionally naked]; nodes trilacunar; wood light brown to nearly white, diffuse porous, close grained, moderately soft to very hard; pith relatively small, ho- mogeneous, triangular to circular in cross section. Leaves simple, petiolate, alternate, spirally arranged, 3-ranked to distichous; blades ovate to deltoid, elliptic, obovate, or suborbicular, coarsely to finely toothed, glabrous to spar- ingly pubescent adaxially, glabrous to tomentose and sometimes covered with resinous glands abaxially,; venation pinnate, the secondary veins craspedod- romous [or semicraspedodromous], divergent and straight to strongly ascend- ing, the tertiary (cross) veins usually prominent; leaves conduplicate in bud, open and convex but becoming conduplicate as they expand, or open and concave; stipules present, free, broadly ovate to narrowly linear, deciduous. Plants monoecious, the flowers imperfect, anemophilous, much reduced. In- florescences consisting of pendulous or erect catkins of reduced, 3-flowered cymules, or reduced to compact clusters of several minute flowers (florets); staminate catkins terminal or lateral on the branchlets, borne on either long or short shoots, solitary or in small to large racemose clusters, precocious [or developing during the current season], when formed the previous season ex- posed or enclosed in buds during the winter, elongate, cylindrical, pendulous, conspicuously bracteate, the scales and flowers densely to loosely arranged, the scales consisting of [(1-)]3—5 fused bracts; carpellate inflorescences terminal or lateral on the branchlets, borne on either long or short shoots, solitary or in small [to large] racemose clusters, precocious or developing during the current season, when formed the previous season exposed or enclosed in buds during the winter, consisting of short to moderately long, erect to pendulous, bracteate catkins or of small, compact clusters of flowers subtended by leafy involucres; scales and flowers densely to loosely arranged, the bracts [(1-)]3—5 per scale, becoming variously fused and often large and subfoliaceous or woody in the infructescences. Staminate flowers small, the perianth lacking or of 1-4(-6) minute tepals; stamens (1-)4(-6), in | whorl, sometimes appearing to be more due to development of part or all of additional flowers of the reduced cymule: filaments short, separate or partially to wholly connate, or sometimes divided part or all the way to the base; anthers tetrasporangiate, 2-locular, dorsifixed, extrorse, entire or partially to wholly divided from the apex, opening by lon- gitudinal slits; pollen grains smooth or slightly granular, spheroidal to oblately flattened, angular, 15—45 um in diameter, aspidote, poroid, the apertures (3 or) 4—7, circular to elliptic, equatorial, evenly spaced; rudimentary gynoecium usually absent. Carpellate flowers small, perianth usually lacking or highly reduced and adnate to the ovary; ovary compound, of 2 (or 3) carpels, inferior or nude (i.e., apparently inferior on the basis of vascular traces, but lacking a 1990] FURLOW, BETULACEAE 3 perianth), 2- (or 3-)locular below, unilocular above, with 2 (or 3) linear styles, these stigmatic above; ovules axile, | or 2 per locule, pendulous, anatropous, crassinucellar, unitegmic or bitegmic; staminodes usually absent; fertilization chalazogamous; endosperm nuclear, becoming cellular. Infructescences woody and strobiluslike or consisting of elongate to compact or irregular foliaceous clusters, the scales (bracts) persistent or deciduous, variously lobed and toothed. Fruits nuts, nutlets, or 2-winged samaras (sometimes with the wings reduced), maturing and dispersed the same season as or the season following pollination. Seed single, pendulous; endosperm present but thin at maturity; embryo large, straight, as long as the seed, the cotyledons small and flat, or greatly thickened, plano-convex, and oily; radicle superior. Embryo-sac development of the Po- lygonum type. Germination epigeal or hypogeal. Base chromosome numbers 7, 8. (Including Corylaceae Mirbel, Elem. 2: 906. 1815, nom. cons., and Car- pinaceae Kuprianova, Taxon 12: 12. 1963.) Type GEeNus: Betula L. A family of six genera and about 150 species, primarily of the boreal and cool-temperate zones of the Northern Hemisphere, but also represented at high elevations from Mexico southward through Central America to northern Ar- gentina. Despite its small size, the family includes dominant trees of forests of temperate Asia, northern Europe, northwestern North America, mountainous parts of Mexico, Central America, and northern South America, and the cir- cumboreal region. Five of the genera occur in North America and are repre- sented by one or more species in the southeastern United States. The remaining genus, Ostryopsis Decne. (most closely related to Corylus L.), consists of two species of shrubs restricted to northern and western China The Betulaceae are woody plants easily distinguished by their simple, pin- nately veined, usually ovate, sharp-toothed leaves, their long, dense staminate catkins that often develop the season before anthesis, and (except in Corylus and Ostryopsis) their strobiluslike infructescences. The family is held together on the basis of many characters, including habit; leat structure, venation, and wood anatomy ovary "structure: ue morphology and position; pollen morphology; embryo structure; chalazogamous fertilization; germination pattern; and serological re- actions. The Betulaceae are usually treated as a single family closely allied with the Fagaceae and placed with it in the order Fagales, although Hjelmqvist (1957, 1960) and Takhtajan (1969, but not 1980) placed the family in its own order, Betulales, on the basis of embryological differences, especially the presence of endosperm in the Betulales (since then shown to exist in both groups, as dis- cussed below). The earliest natural post-Linnaean systems treated the genera of the Betulaceae, together with the elms, oaks, willows, and other amentiferous trees, as members of a single large assemblage (e.g., ““Castaneae” of Adanson, ““Amentaceae” of Jussieu). Adanson segregated the betulaceous and fagaceous genera into one of three “‘sections” of the Castanieae, corresponding to the modern Fagales. In most of the important subsequent work until that of Prantl (e.g., De Candolle; Regel (1861, 1868); and Spach (1841, 1842a—c)), the group was divided into two families, Betulaceae and Corylaceae Mirbel, following 4 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 the lead of Linnaeus, who had placed Betula (including A/nus Miller) in the Monoecia Tetrandia, and Carpinus (including Ostrya Scop.) and Corylus in the Monoecia Polyandria. De Candolle, in his monograph of the Corylaceae, further divided that family into two tribes, Coryleae Meisner and Carpineae D6Il. However, Prantl treated the entire group as a single family, Betulaceae, com- posed of two tribes, Coryleae Meisner and Betuleae, the system followed by Winkler in the most recent family monograph. Koehne also adopted this ar- rangement, but he elevated the tribes to subfamily status; this treatment has been accepted by a number of modern authors, including Jury, Rendle, Takh- tajan (1980), and Thorne (1983). Bentham & Hooker combined the Betulaceae and Fagaceae into the single family Cupuliferae, in part based on a misunder- standing of the nature of the bracts subtending the fruits in these families (Abbe, 1974), but few have followed this path. The family is still sometimes divided into two families, the Betulaceae and the Corylaceae, and Kuprianova has proposed the further splitting of the Coryloideae into two families, Corylaceae and Carpinaceae, following De Candolle’s tribes Coryleae and Carpineae. Par- ticularly in Europe, the two-family scheme has been adopted by many authors, including Airy Shaw in Willis, Dahlgren (1975, 1977, 1980, 1983), Hjelmqvist (1957, 1960), Hutchinson (1959, 1967), and Jury. In America, it has been used by Mohlenbrock & Thomson, Rehder (1940), and Small (1903, 1933). Some of these treatments based recognition of two families in part on the belief that a fundamental difference exists in the staminate inflorescences of the two groups (there being three flowers, each consisting of one to six stamens per bract in the Betulaceae, and one flower, consisting of many stamens per bract in the Corylaceae). However, it has now been established that no such difference in staminate inflorescence structure exists in these two groups (see below). Most modern authors (e.g., Cronquist, 1981, 1988; Melchior; Takhtajan (1969, 1980); and Thorne, 1968, 1973, 1983) maintain the family as a single group. When one family is recognized, the name Betulaceae is conserved against Corylaceae. The Betulaceae are treated here as consisting of a single family subdivided into two natural subfamilies, the Betuloideae and the Coryloideae (Regel) Koehne, these corresponding to the tribes Betuleae and Coryleae recognized by Prantl and Winkler. To reflect the substantial differentiation between C arpinus and Ostrya Scop. on the one hand, and Corylus and Ostryopsis on the other, the Coryloideae are further divided into two tribes, the Coryleae and the Car- pineae DOIl. In an early analysis of generic affinities within the family, Anderson & Abbe found that species differences are consistently smaller than generic differences and are approximately equal from genus to genus. These authors noted an exception, however, in the case of Carpinus and Ostrya, which they concluded show less divergence than occurs among some of the subgroups within the genus Betula. Where best to place the Betulaceae within the angiosperms has been a source of continuing disagreement, in part because of an unclear fossil record of the ancestors of the group. The Englerian tradition, following earlier authors such as Jussieu, placed the amentiferous families together in a relatively low position in the dicotyledons (see Stern). Bessey, perhaps in an overreaction to the ap- 1990] FURLOW, BETULACEAE 5 parent phylogenetic inaccuracy of the Englerian arrangement, gave the Betu- laceae an exaggeratedly advanced level (in the Sapindales). Cronquist (1981, 1988), Melchior, Takhtajan (1969, 1980), and Thorne (1973, 1983) all placed the Fagales, along with other “core” orders of the Amentiferae, in a position more advanced than the magnoliids, less advanced than the rosids, and with close ancestral ties to the Hamamelidales. Dahlgren, in his early treatments (1975, 1977), adopted a similar scheme but placed the Hamamelidanae some- what nearer the rosid groups. However, in the most recent versions of his system (1980, 1983), he joined the Fagales with the Juglandales, Saxifragales, and Rosales in the Rosiflorae. Taking a radically different view, Meeuse (1975a— c) maintained that the Amentiferae, including the Betulaceae, were not derived from a group having well-developed flowers but instead represent a funda- mentally distinct line of evolution in the flowering plants. However, this po- sition has received little acceptance. There is disagreement on the phylogenetic arrangement of genera in the family. While virtually all authors recognize a fundamental dichotomy between the Betuloideae and the Coryloideae, and most believe the Betuloideae to be the less specialized, the arrangement of the genera within these groups is un- settled. Within the Betuloideae, Betula has traditionally (e.g., by Bentham & Hooker; Komarov; Prantl; Regel (1861, 1868); Spach (1841); Winkler) been placed before A/nus. However, modern workers (e.g., Furlow, 1979, 1983a; Hall; Kikuzawa; Takhtajan, 1969, 1980) generally consider A/nus the less spe- cialized, and it has been treated accordingly in a number of recent floras (¢.g., those by Scoggan; Soper & Heimburger; Voss). In the Coryloideae there is more confusion, and all possible arrangements of the four genera are found in the modern literature. Winkler, in part following Prantl, viewed Ostryopsis as most primitive, Ostrya and Carpinus as intermediate, and Cory/us as most advanced. Bentham & Hooker and Hutchinson (1959, 1967), on the other hand, listed Carpinus as least specialized, followed by Ostryopsis, Ostrya, and Corylus. On the basis of wood anatomy and leaf and bud structure, respectively, Hall and Kikuzawa considered Corylus to be the most primitive and Carpinus the most advanced, a view first proposed by Tippo. Hjelmqvist (1948), from floral and fruit structure, and Hardin & Bell, on the basis of trichome morphology, viewed Carpinus as the most primitive and Corylus as the most advanced. Obviously, a detailed examination of this problem is needed. In a preliminary cladistic study using many features of a variety of types, Furlow (1983a) concluded that Carpinus was the least specialized, followed by Ostrya and Ostryopsis, with Corylus the most specialized. The leaves of the Betulaceae are simple and pinnately veined, with non- glandular teeth of a modified urticoid type (L. J. Hickey & Wolfe). The sec- ondary veins are generally characterized as uniformly craspedodromous (L. J. Hickey; L. J. Hickey & Wolfe). However, venation in members (mostly Asian) of Alnus subg. CLETHROPSIS (Spach) Regel is semicraspedodromous (Furlow, 1979), a unique condition in the Fagales, although the phylogenetic significance of this pattern has not been investigated. The basal secondary veins tend to be crowded in most genera of the family, especially in Corylus (Meyerhoff). Branches of the lower secondary veins (“‘outer secondaries”) appear in all genera and 6 JOURNAL OF THE ARNOLD ARBORETUM [vov. 71 run, like the secondaries, to teeth at the margin; regular and usually prominent tertiary veins connect the secondaries. The general structure of vegetative features in the family has been treated by Boubier, Metcalfe & Chalk, and Solereder. The leaves of all genera are pubescent abaxially, although in individual species or populations they vary from glabrous to densely tomentose. Most betulaceous leaves have at least some hair, especially along the major veins and in the axils of the secondary veins. Hardin & Bell have studied the foliage of the five North American genera of the family in detail and have identified six distinct trichome types, including unicellular and multicellular hairs of several kinds, stipitate glands, and peltate scales (sessile glands). The four hair types are found in all five of the genera, tying the family together as a unit (Hardin & Bell). Most of these trichome types are also present in the Fagaceae (Hardin & Johnson), suggesting a close relationship between the Betulaceae and the Fagaceae. Large glands are frequent on the leaves and twigs of A/nus and Betula (Bell et al.; Furlow, 1979). It has been shown by Dorman that in A/nus these glands secrete a high-molecular- weight polyterpene and by Wollenweber that their product includes flavonoid compounds Several authors have sought the origin and phylogeny of the Betulaceae in the structure of their flowers and inflorescences (Abbe, 1935, 1938; Hjelmaqvist, 1948; Korchagina) and in their embryology (Benson; Hjelmqvist, 1957; Na- waschin). In a comprehensive investigation of the floral and inflorescence anat- omy and morphology of the family, Abbe (1935, 1938) proposed that the inflorescences consist of systems of reduced three-flowered cymules and hy- pothesized the loss of various bracts, flowers, and flower parts in various genera, leading to the present patterns. He argued (1935), on the basis of the position of the tepals or their vestigial vascular traces, that the ovaries of all members of the family are inferior, and he proposed that the bicarpellate ovaries of the various genera have arisen in several different ways: those in A/nus, Betula, and Cory/lus from loss of the carpel in the radius of the adaxial tertiary bract; those in Carpinus and Ostrya from loss of the carpel in the radius of the secondary bract. Hjelmqvist (1948) disagreed with the latter interpretation, concluding that the different positions noted were probably due only to twisting of the original transverse carpels. In a third paper Abbe (1974) reviewed the floral structure of the entire Amentiferae and argued that the Betulaceae form a single evolutionary unit with three clearly divergent lines, which he assigned to the tribes Betuleae, Carpineae, and Coryleae. The distinctive ways in which individual catkins are clustered in various genera and infrageneric groups of the Betulaceae have been discussed by Furlow (1979), Hjelmqvist (1948), Jager, and Murai. Jager, expanding upon the ideas of Hjelmqvist (1948), proposed that the hypothetically ancestral synflorescence of the family consists of an axis bearing a terminal cluster of staminate catkins, with lateral clusters of carpellate catkins placed below it. This type resembles the form seen today in members of A/nus subg. CREMASTOGYNE Schneider in Sarg. and, in somewhat reduced form, A/nus subg. ALNUs. J ager has traced the evolution of various synflorescence types occurring in most extant betulaceous genera, explaining their progressions by means of translocations of the shoot 1990] FURLOW, BETULACEAE ‘3 innovations into the synflorescence itself, tendencies to monopodial or sym- podial proliferation of the synflorescence, reduction in size and number of catkins, and winter protection of carpellate catkins by bud scales. He has also shown that these changes correlate closely with vegetative adaptations related to severity of climate. In A/nus subg. ALNOBETULA Peterm., as well as in most subgroups of Betula, the tendency has been toward a monopodial form with a terminal staminate cluster, the carpellate clusters being held on lateral shoot innovations. However, in Betula sect. Humites W. D. Koch the axes have diversified sympodially and have been reduced to a form in which solitary staminate catkins occupy a position below the carpellate ones on lateral short shoots. Sympodial modifications and reductions are seen as well in Carpinus and Ostrya. In Carpinus the solitary or clustered staminate catkins are located laterally below the terminal carpellate ones, while in Corylus this trend reaches its ultimate configuration: the staminate clusters are positioned laterally below reduced solitary carpellate inflorescences. The Betulaceae are generally uniform in fruit structure, but each genus has distinctive modifications associated with dispersal by means of wind, water, or animals. The tiny, lateral-winged samaras of Alnus and Betula are carried great distances and in large numbers by air currents. In certain species of Alnus, these wings have been reduced or lost and the fruits are apparently dispersed mostly by water; they have been shown to float for long periods (McVean). The fruits of Carpinus and Ostrya are also scattered by wind, but with the aid of greatly expanded bracts rather than wings on the fruits themselves. The fruits of these genera are widely called both achenes and nutlets, but since their walls are quite bony and tightly attached to the enclosed seed, the latter is more appropriate (see Hjelmqvist, 1948). The most striking fruit modifications of the Betulaceae are seen in the dramatic adaptations of Corylus fruits for zo- dchory (Stebbins; Stone). The seeds of all the genera are similar in internal structure (with axile and investing embryos), a type regarded as advanced in the angiosperms (A. C, Martin). In the Betuloideae the embryos are somewhat less investing than in the Coryloideae, and thus they may be considered less specialized. In both groups several ovules are initially laid down (Hagerup), but only one of these develops. In the Coryloideae (as well as in the Fagaceae (Benson, ee 1948)), but not in the Betuloideae, several embryo sacs may deve The wood of the Betulaceae has been studied by Bailey (1910, 191 L, 1912), Forsaith, Hall, Hoar, and others. Bailey (1911) discussed the presence of ag- gregate rays in the wood of the Betulaceae and related families and proposed a phylogenetic series involving the development of large multiseriate rays from uniseriate ones. In a second paper (1912) he developed this concept further and demonstrated reversals to uniseriate rays in various species. Hoar con- cluded, on the basis of the presence of aggregate rays, that the family was extremely primitive in the dicots. A study of the wood anatomy of the Betu- laceae by Hall showed the family to be “anatomically natural and closely knit” (p. 262) but with clear distinctions between the Betuloideae and Coryloideae. However, Hall did not speculate on whether these groups should be considered one family or two. He found the Betuloideae to be more primitive than the 8 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 Coryloideae in having relatively large vessels lacking spiral thickenings and having scalariform perforation plates. He also concluded that true tracheids were present in the Betuloideae and absent in the Coryloideae. However, Ka- sapligil and Yagmaie & Catling have shown the wood of Carpinus and Corylus to contain tracheids. Within the Betuloideae A/nus was seen by Hall as less specialized than Betula in terms of the number and spacing of the bars of the perforation plates and in the presence of opposite, as well as alternate, inter- vascular pitting. Within the Coryloideae he noted a trend of specialization leading from Corvlus to Carpinus and then to Ostrya, this involving reduction in the number of bars of the perforation plates, an increased presence of spiral thickenings on vessel walls, and other characters. Brunner & Fairbrothers concluded from serological investigations of the six genera (but based on very limited sampling) that these groups held together well and should be treated as a single family. Petersen & Fairbrothers showed the Betulaceae as a whole to be closely related to members of the Fagaceae, Myricaceae, and Juglandaceae on serological grounds, with members of the Anacardiaceae, Aceraceae, Moraceae, Oleaceae, and other families of rosid affinity forming a very distant group. Other than this, there is little positive chemical evidence demonstrating relationships between the Betulaceae and other amentiferous families (Mears). The Betulaceae are well known cytologically (J aretzky; Wetzel, 1927, 1928, 1929; Woodworth, 1929a-c, 1930a, b, 1931). Chromosomally, the family con- sists of groups that do not correspond to the tribes or subfamilies (Raven). The base chromosome number of Betula and Corylus is 14, while that of Carpinus, Ostrya, and Ostryopsis is eight. Alnus, usually placed in the x = 14 group with Betula and Corylus, was suggested by Furlow (1979), after a count by Chiba, to have a base number of seven, the probable original base number of the Fagales (Raven). This number had previously been predicted for the family by Woodworth (1931) and by Wanscher. An allozyme study of A. viridis subsp. crispa (2n = 28) by Bousquet and colleagues (1987) indicated that this species could be treated either as a diploid or a diploidized autotetraploid. Additional study of 4. incana subsp. rugosa by these authors also revealed diploidlike expression for all polymorphic allozyme loci (Bousquet ef al. 1988), and ad- ditional indirect evidence for a base number of seven has been provided by Brown & Al-Dawoodie, who found that meiotic behavior in hybrid birches suggests that 2 = 42 trees actually represent hexaploids. The Betulaceae are an ancient group, extending back in the fossil record to the Upper Cretaceous on the basis of both leaves and pollen. Fossils assigned to the family become abundant in strata of Paleocene and Eocene age in both the New and the Old worlds. This record has been summarized by Crane (1981), Crane & Stockey, Crepet, and Wolfe (1973). Members of the Betuloideae first occur in the Upper Cretaceous (Maestrichtian); coryloid types in the lower Paleocene. Evidence that Betula had diverged from A/nus by the mid-Eocene is provided by Crane & Stockey, and fossil evidence from northwestern North America indicates that the subgenera of 4/nus had differentiated and were present in the New World during the Miocene if not before (Wolfe, 1969). The ancestor of the Betulaceae is not obvious from the features of other 1990] FURLOW, BETULACEAE -) extant families. On the basis of floral structure, Abbe (1938) tentatively sug- gested the Fagaceae as the most likely candidate, but in spite of the similarities, there are important differences between the two families, including the structure of the carpellate inflorescences and the presence of a “stem cupule” around the fruits. Takhtajan (1969) believed that the Betulaceae share a common ancestry with the Fagaceae but are not derived directly from them. Hjelmaqvist (1948) concluded that the Betulaceae are not closely related to the Fagales, but that they show significant connections to the Juglandaceae in floral coalescence, chalazogamy, and other basic features. He believed that the two major subgroups of the family are closely related, although one probably did not originate directly from the other: the distinctive lines of specialization within each group, such as differences in the fruits and the involucre, indicated rather that they had developed from a common ancestor. Tippo first suggested that the Betulaceae may have been derived from ha- mamelidaceous stock, and many modern workers have adopted this position. Endress (1967, 1977) emphasized that the Hamamelidales combine features of the mainly insect-pollinated Cunoniales and Rosales with those of the wind- pollinated Fagales and argued that the Betulaceae and Fagaceae may be derived from a Corylopsis-like hamamelid ancestor. Ehrendorfer concluded that the Hamamelidae, including the Hamamelidales and the Fagales, can be regarded as remnants of an ancient stock of dicots linking the Magnoliidae and the Rosidae-Dilleniidae, but with tendencies toward anemophily and floral reduc- tion. Doyle has suggested that the ancestor of the more advanced Hamamelidae (including the Betulaceae) may have been a member of the Normapolles com- plex, known from its psilate, complex-walled, tricolporate pollen, which first appeared during the Cretaceous and reached a peak in the Santonian, and which seems to have been adapted for wind pollination. However, the pollen of modern Betulaceae is much more specialized than that of the modern Ha- mamelidales (L. J. Hickey & Doyle; Walker & Doyle). This has been used to support the view that some groups of modern Amentiferae (e.g., the Juglan- daceae) may have a rosid, rather than a hamamelid, ancestor. There is fossil evidence that the unlobed, pinnately veined leaves of the Hamamelidae are of secondary derivation from a platanoid ancestor with palmately veined and lobed leaves (L. J. Hickey & Wolfe; Wolfe, 1973). In Corylus the basal secondary veins often tend toward an actinodromous con- dition, rising abruptly toward the apex, where there is a suggestion of lobing. This intriguing pattern (which can sometimes also be seen to varying degrees in other Betulaceae—e.g., in A/nus viridis subsp. sinuata (Rydb.) Love & Love) resembles that of the hamamelidaceous Corylopsis (Wolfe, 1973) and may represent a remnant of a primitive venation pattern. However, such venation is also explicable by relatively minor structural adjustments to the ordinary form of the extant Fagales (similar distortions to the apical parts of the leaves of several species of A/nus have been reported by Furlow, 1979). Kasapligil (p. 85) explained the pattern in Cory/us as “‘due to the auriculate condition of the cordate [base] of blades and the abrupt acuminate form of the leaf apices.” Wolfe (1973) pointed out that in toothed primitive Juglandaceae, the secondary 10 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 veins vary from a craspedodromous pattern, the teeth being entered along the apical side by a branch from the secondary vein. He concluded (p. 351) that these leaf features ““do not conform to any known specialization of venation elsewhere in the Hamamelididae but rather are highly similar to members of would be difficult to reconcile such an origin for the Fagales with a close relationship between the Fagales and the Hamamelidales, as has been implied by most modern classifications. He offered the alternate hypothesis that the Betulaceae and their closely related amentiferous allies may have converged on the Normapolles group in their pollen morphology. These patterns will become clearer only as additional paleobotanical evidence accumulates. Although the family is old, it is neither particularly unspecialized nor greatly advanced. On the basis of wood anatomy, Hall found the Betulaceae to be moderately specialized relative to other woody angiosperm families. Sporne, in reviewing the degree of specialization of the flowers of a number of amen- tiferous taxa, found that for the Betulaceae 60 percent of the included characters could be considered primitive, compared with values of 88, 70, and 46 percent for the Magnoliaceae, Fagaceae, and Hamamelidaceae, respectively. Moseley, in a similar comparison but using both floral and vegetative characters, arrived at a figure of 43 percent for the Betulaceae, with values ranging from 27 (UI- maceae) to 64 percent (Myricaceae). The most obvious evolutionary trends within the Betulaceae are those related to fruit dispersal, as discussed above, and those correlated with apparent ad- aptations for survival in cold climates. The latter include shrubby growth forms, small leaves with few lateral veins, protection of the carpellate catkins during the winter by bud scales, and presence of true bud scales on the winter buds (Furlow, 1979; Jager; Kikuzawa). These trends are identifiable in all of the genera of the family. Some of the related anatomical reductions involve reten- tion of juvenile characteristics (Forsaith; Hall). Members of the Betulaceae are economically important as timber trees, as the source of hazelnuts and filberts (Corylus), as ornamental trees and shrubs, and as an aid in soil nitrification and stabilization (A/nus). Some are important as causes of pollen allergies in regions where they grow abundantly (Dalen & Voorhorst; Lewis ef a/l.; Lowenstein et al.; Solomon & Durham; Wodehouse, 1945). The bark of some contains substances of medicinal value (Lewis & Elvin-Lewis; Moerman). Wood of Betula and Alnus is widely used in the manufacture of furniture, paneling, boxes, and small wooden objects; that of Carpinus and Ostrya is used for making wooden tools such as mallets. In the past the bark of Betula has served as a commercial source of oil and methyl salicylate. The wood of several of the genera is used in the manufacture of high-quality charcoal. REFERENCES: AsBE, E. Studies in the sles of ae Betulaceae. I. Floral and inflorescence anatomy and morphology. Bot. Gaz. 97: 1-67. 1935; II. Extremes in the range of variation of floral and ie morphology. /bid. 99: 431-469. 1938. . Flowers and inflorescences of the ““Amentiferae.”’ Bot. Rev. 40: 159-261. 1974. 1990] FURLOW, BETULACEAE 11 ApANnson, M. Familles des plantes. Paris. 1763. Vol. 2. 640 pp. [Castaneae, 366-377.] ANDERSON, E., & E. C. ABBE. A quantitative comparison of specific and generic differ- ences in the Betulaceae. Jour. Arnold Arb. 15: 43-49. 1934. BaILey, I. W. Notes on the wood structure of the Betulaceae and Fagaceae. Forestry Quart. 8(2): 3-10. 1910. . The relation of the leaf-trace to the formation of compound rays in the lower dicotyledons. Ann. Bot. 25: 225-241. 1911. —. The A ecauey history of the foliar ray. [bid. 26: 647-661. 1912 & E. W. SINNOTT. Investigations on the phylogeny of the angiosperms. Bot. Gaz. 58: 36-60. 1914, BaARABE, D., Y. BERGERON, & G. VINCENT. La répartition des caractéres dans la clas- sification des Hamamelididae (Angiospermae). (English summary.) Canad. Jour. Bot. 65: 1756-1767. 1987 BEHNKE, H.-D. Sieve-tube plastids of Hamamelididae. Electron microscopic investi- gations with special reference to the Urticales. Taxon 22: 205-210. 1973. [S-type sieve- ye plastids in the Betulaceae and most other Hamamelidae.] BELL, J. M., D. G. Personius, & J.D. Curtis. The occurrence and variation of secretory structures in the Betulaceae. (Abstract.) Proc. Iowa Acad. Sci. 89: 9. 1982. Benson, M. Contribution to the embryology of the Amentiferae. Part I. Trans. Linn. Soc. London, II. Bot. 3: 409-424. 1894. BENTHAM, G., & J. D. Hooker. Cupuliferae. Gen. Pl. 3: 402-410. 3. Bessey, C. The phylogenetic taxonomy of flowering plants. Ann. as Bot. Gard. : 109-117. 1915. Boone: A. M. Recherches sur l’anatomie systématique des Betulacées-Corylacées. pale 10: 349-436. 1896. BousquET, J., W. M. CHELIAK, & M. LaLtonpe. Allozyme variability in natural popu- lations of green alder (4/nus crispa) in Quebec. Genome 29: 345-352. 1987. llozyme variation within and among mature populations of speckled alder (Alnus rugosa) and relationships with green alder (A/nus crispa). Am. Jour. Bot. 75: 1678-1686. 1988. Britton, N. L. North American trees. xii + 896 pp. New York. 1908. [Betulaceae, 240-267.] & A. Brown. Illustrated flora of the northern United States and Canada. Vol. 1. x1 + 612 pp. New York. 1896. [Betulaceae, 506—-513.] & . Ibid. ed. 2. Vol. 1. xxix + 680 pp. New York. 1913. [Betulaceae, 605-— 614 Brown, H. P., & A. J. PANSHIN. Commercial timbers of the United States: their structure, ‘dentifieation, properties, and uses. xxi + 554 pp. New York. 1940. [Betulaceae, 438-446. Brown, I. R., & D. M. AL-DAwoopigE. Observations on meiosis in three cytotypes of Betula alba L. New Phytol. 83: 801-811. 1979. BRUNNER, F., & D. E. FArRBROTHERS. Serological investigations of the Corylaceae. Bull. Torrey Bot. Club 106: 97-103. 1979. CANDOLLE, A. DE. Corylaceae. DC. Prodr. 16(2): 124-133. 1864. CHAMBERLAIN, C, Winter characters of certain sporangia. Bot. Gaz. 25: 124-128. 1898. [Pollen mature in the anthers of A/nus and Corylus the autumn before anthesis.] Cu1BA, S. Studies on the breeding of Betu/a and A/nus species. (1) On the differences of CRANE, P. R. Betulaceous ree and fruits from the British upper Paleocene. Bot. Jour. Linn. Soc. 83: 103-136. . Early fossil history ie the Betulacexe, a preliminary report. (Abstract.) Am. Jour. Bot. 71(5, part taka 109. 1984. &R . A. Stoc Betula leaves and reproductive structures from the middle Eocene of British Pane Canad. Jour. Bot. 65: 2490-2500. 1987 Lz JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71] CrepeT, W. L. The status of certain families of the Amentiferae during the middle Eocene and some hypotheses regarding the evolution of wind pollination in dicot- yledonous angiosperms. Pp. 103-128 in K. J. NiKLas, ed., Paleobotany, paleoecol- ogy, and evolution. Vol. |. New York. 1981. Cronauist, A. An integrated system of classification of flowering plants. xviii + 1262 pp. New York. 1981. [Betulaceae, 225-228. The evolution and classification of flowering plants. ed. 2. viii + 555 pp. New York. 1988. [Fagales, 306-308. DaHLGREN, R. M. T. A system of classification of the angiosperms to be used to dem- onstrate the distribution of characters. Bot. Not. 128: 119-147. 1975. [Hamameli- — neha. angiospermschema och dess anvandning vid kartering av egenskaper. Sv. Bot. aie 71: 33-64. 1977. [Hamamelidanae, 47, 48. A revised system of classification of the angiosperms. Bot. Jour. Linn. Soc. 80: 91- pe 1980. [Rosiflorae, 95, 96.] eral aspects of angiosperm evolution and macrosystematics. Nordic Jour. Bot. 3 119 149. 1983. [Rosiflorae, 132, 133. DALEN, G. von, & R. Vooruorst. Allergen community in pollen from certain tree species. ree Allergy 46: 276-278. oe -reactivity of A/nus, Betula, and Corylus pollen allergens. ] Davis, G. L. Systematic embryology of the angiosperms. x + 528 pp. New York. 1966. [Betulaceae, 57, 58. Deevey, E. S., JR. Biogeography of the Pleistocene. Geol. Soc. Am. Bull. 60: 1315- 1416. 1949. DOLL, J.C. Zur Erklarung der Laubknospen der Amentaceen. 28 pp. Frankfurt. 1848. [Carpineae, 15.] DorRMAN, F. Zur Kenntnis der Hautdriisen und der epee von Alnus viridis. Sitz.-ber. Akad. Wiss. Math.-Naturw. Wien 133: 585-612. 1924. Doy eg, J. A. Cretaceous angiosperm pollen of the Atlantic es Plain and its evo- lutionary significance. Jour. Arnold Arb. 50: 1-35. 19 DRESSLER, R. Some floristic relationships between Mexico and the United States. Rhodora 56: 81-96. 1954. Duncan, W. H. Preliminary reports on the flora of Georgia. 2. Distribution of 87 trees. Am. Midl. Nat. 43: 742-761. 1950. [Alnus serrulata, Betula nigra, Carpinus caro- liniana, Ostrya virginiana. | EHRENDORFER, F. [New evidence for relationships and modern errr Gee In angio- sperms.] Summary statement. Nordic Jour. Bot. 3: 151-155. 1983. ENDLICHER, S. Gen. Pl. Suppl. 2. 114 pp. Vienna. 1842. a 28.] Suppl. 4. 95 pp. Vienna. 1847. [Betulaceae, 19-22; Cupuliferae, 22-24] Enpress, P. K. Systematische Studie iiber die verwandtschaftlichen Beziehungen zwi- schen den Hamamelidaceen und Betulaceen. Bot. Jahrb. 87: 431-525. 1967. . Evolutionary trends in the Hamamelidales-Fagales-Group. Pp. 321-347 in K. KuBITZKI, ed., Flowering So evolution and classification of higher categories. Pl. Syst. Evol. "Suppl. 1. 1977 ERDTMAN, G. Pollen morphology and plant taxonomy. Angiosperms. 539 pp. Uppsala. 1952. [Betulaceae, 71-73. FERNALD, M. L. Gray’s manual of botany. ed. 8. Ixiv + 1032 pp. New York. 1950a. sy ce 530-539.] . Kinsey, & R. C. Rotuins. Edible wild plants of eastern North America. xi + 452 pp. New York. 1958. [Betulaceae, 151-154.] Forsaitu, C. C. Anatomical reduction in some alpine plants. Ecology 1: 124-135. 1920. FoweE ts, H. A., compiler. Silvics of forest trees of the re States. U. S. ee aa Agr. Handb. 271. vi + 762 pp. Washington, D. C. 1965. [Maps by E. L. Jr.; Alnus rubra, 83-88; Betula, 92-118, including B. ara B. lenta, ; “alle ghaniensis. | 1990] FURLOW, BETULACEAE 13 Furtow, J. J. The systematics of the American species of A/nus (Betulaceae). Rhodora 81: 1-121, 151-248. 1979. . The phylogenetic Sea ere of the genera ae eae taxa of the Betula- ceae. (Abstract.) Am. Jour. Bot. 70(suppl.): 114. Gieason, H. A. The new Britton and Brown ae es of the northeastern United States and adjacent Canada. Vol. 2. iv + 655 pp. New York. 1952. [Betulaceae, 30- 36.] — & A. Cronaguist. Manual of vascular plants. ii + 810 pp. New York. 1963. fea daceas, 243-247. GOoLpBERG, A. Classification, secre and phylogeny of the families of dicotyledons. Smithson. Contr. Bot. 58: 1-314. 1986. [Betulaceae, 78, 79. GRAHAM, A. History of the sertaeitog temperate element in the northern Latin Amer- ican biota. Pp. 301-314 in A. GRAHAM, ed., Vegetation and vegetational history of 73a. m : pee vegetational history i in Latin America. Pp. 315-360 in A. GRAHAM, , Ibid. 19 GRAY, se . sate pollen genera in the Eocene (Claiborne) flora, Alabama. Science 132: 808, 809. 1960. [Betulaceous pollen absent.] Hacerup, O. The see and biology of the Cory/us fruit. Danske Vidensk. Selsk. Biol. Medd. 17: 1-32. HALL, J. W. The seer ee and phylogeny of the Betulaceae. Bot. Gaz. 113: 235-270. 1952. Harp, J. W. The Juglandaceae and Corylaceae of Tennessee. Castanea 17: 78-89. 1952 . Studies of the southeastern United States flora. I. Betulaceae. Jour. Elisha Mitch- ell Sci. Soc. 87: 39-41. 1971. . Bett. Atlas of foliar surface eve in woody plants, [X. Betulaceae of eastern United States. Brittonia 38: 133-14 & NsoNn. Atlas of foliar surface aa in woody plants, VIII. Fagus and C astanea (Fapaceae) of eastern North America. Bull. Torrey Bot. Club 112: 11- 985. Heptinc, G. H. Diseases of forest and shade trees of the United States. U. S. Dep. Agr. Forest Serv. Handb. 386. vii + 658 pp. Washington, D. C. 1971. [A/nus, 70-73, Betula, 78-93, Carpinus, 93-95; Ostrya, 251, 252.] HERNANDEz X., E., H. Crum, W. B. Fox, & A. J. SHARP. A unique vegetational area in Tamaulipas. Bull. Torrey Bot. Club 78: 458-463. 1951. Hickey, L. J. eee of the architecture of dicotyledonous leaves. Am. Jour. Bot. 60: 17-33. noe gl Cretaceous fossil evidence for angiosperm evolution. Bot. Rev. 43: 1- 104. 1977 . WOLFE. The bases of pee a phylogeny: vegetative morphology. Ann. Missouri Bot. Gard. 62: 538-589. Hickey, M., & C. Kinc. 100 families of ieee plants. ed. 2. xvi + 619 pp. New York. 1981. [Betulaceae, 60-69. ] Hyevmovist, H. Studies on the Bee morphology and phylogeny of the Amentiferae. Bot. Not. elas 2(1): 1-171. —. Some notes on the casper and embryo development in Fagales and related orders. Bot. Not. 110: 173-195. Notes on some names and cen within the Amentiferae. Bot. Not. 113: 373. 380. 1960. Hoar, C. S. The anatomy and phylogenetic position of the Betulaceae. Am. Jour. Bot. 3: 415-435. 1916. HorHAMMER, L., E. VoRDRAN, & H. WAGNER. Zur Kenntnis der Flavonglykoside aus den Betulaceen. Arch. Pharm. 26: 316-323. 1956. 14 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Hu ten, E. On the American component in the flora of eastern Siberia. Sv. Bot. Tidskr. 22: 220-229. 1928. Hutcuinson, J. The families of flowering plants. ed. 2. Vol. 1. x + 510 pp. Oxford, England. 1959. [Betulaceae, 191, 192; Corylaceae, 193, 194. e genera of flowering plants. Vol. 2. xi1 + 659 pp. Oxford, England. 1967. [Betulaceae, 124-126; Corylaceae, 132-134. JAGER, E. J. Progressionen im ec und in der Verbreitung bei den Betu- laceae. Flora 170: 91-113. 0. JARETzKY, R. Zur Zytologie i pane Planta 10: 120-137. 1930. — S. L. Betulaceae. Pp. 59, . H. HeEywoop, ed., Flowering plants of the orld. New York. 1978. eee: A. L. be. Genera plantarum. 498 pp. Paris. 1781. [Amentaceae, 407-411.] Kartesz, J. T., & R. Kartesz. A synonymized checklist of the vascular flora of the oe States, Canada. and Greenland. 500 pp. Chapel Hill. 1980. [Betulaceae, 111, 12. ] a B. A contribution to the histotaxonomy of Corylus (Betulaceae). Adan- sonia, I]. 4: 43-90. 1964. KikuzAwa, K. Leaf survival and evolution in Betulaceae. Ann. Bot. II. 50: 345-353. 1982. KLUCKING, E. P. The fossil Betulaceae of western North America. 166 pp. Unpubl. M.A. thesis, Univ. California, Berkeley Kocu, W. D. J. Synopsis florae Germanic Ae Hielveticas: Pt. 2. 844 + 1x pp. Frankfurt. 1837. [Cupuliferae, 638-641; Betulaceae, 661-663.] KoeunE, B. A. E. Deutsche Dendrologie. xvi + 601 pp. Stuttgart. 1893. [Betulaceae, 106-120. Komarov, V. L. Betulaceae. Fl. URSS. 5: 252-319. 1936a. KORCHAGINA, I. A. perods tsvetka berezovykh. Trans. Mosc. Soc. Nat. Biol. Ser. Bot. 51: 50-74. 1974. oo L. A. On a hitherto undescribed family belonging to the Amentiferae. axon 12: 12, 13. 1963. . Palynological data on the taxonomy of the order Betulales. Trans. Mosc. Soc. Nat. Biol. Ser. Bot. 13: 63-70. 1965.* Lawrence, G. H. M. Taxonomy of vascular plants. xiii + 823 pp. New York. 1965. [Betulaceae, 457-459.] Leg, S.-C. Forest botany of China. 991 pp. Shanghai. 1935. [Betulaceae, 239-287. Lewis, W. H., & M. P. F. E-vin-Lewis. Medical botany: plants affecting man’s health. xv + 515 pp. New York. 1977. [Many references to present and historical medicinal uses of members eee Betulaceae.] ——, P. Vinay, & V. E. ZENGeR. Airborne and eres pollen of North America. xi + xvi + 254 pp. Boston. 1983. [Betulaceae, 23-30, -187. Li, P. C., & S.-H. CHENG. Betulaceae. (In Chinese.) In: K. - KUuANG & P.-C. Li, eds., Fl. Reipubl. Pop. Sinicae 21: 44-137. 1979. Linnaeus, C. Species plantarum. Vol. 2. Pp. 561-1200. Stockholm. 1753. [Betula, 982, 983; Carpinus, 998; Corylus, 998, 999.] LITTLE, E. L. = Atlas - United ves Vol. 1. Conifers and important hardwoods. . Agr. Forest Serv. Misc. Publ. 1146. v + 9 pp. + 200 maps + indexes Q pp). 1971. finclades Alnus ae Betula pe Meaes B. lenta, B. papyrifera, B. populifolia, B. nigra; Carpinus caroliniana: Ostrya virginiana Atlas of United States trees. Vol. 4. Minor eastern hardwoods. Ibid. 1342. vi + 17 pp. + 166 maps + indexes (2 pp.). 1977. [Includes Alnus maritima, A. rugosa, 7 serrulata; Betula uber, Corylus cornuta.] Checklist of United States trees (native and naturalized). U. S. Dep. Agr. Forest Serv, Agr. Handb. 531. iv + 375 pp. 1979. 1990] FURLOW, BETULACEAE 15 North American trees with relationships in eastern Asia. Ann. Missouri Bot. Gard. 70: 605-615. 1983. Love, A., & D. Léve. The origin of the North Atlantic flora. Aquilo : ae 1967. LOWENSTEIN, H., N. PETERSEN, E. P. MUNCH, J.-A. WIHL, H.-H. Ips . JANNICKE, & H. BowaD Characterization of reaction to andim acne al parti identity between ens from alder, birch, and hazel. Allergy 37(suppl. 1): 21, 1981. MaAcDOoNALD, A. D. Floral development in the “Amentiferae.”’ 136 pp. en Ph.D. dissertation, eis Univ., Montreal. 1971. Martin, A. C. The comparative internal structure of seeds. Am. Mid]. Nat. 36: 513- 660. 1946. eee eee, 640, 641. Martin, P. S., & B. E. HARRELL. The Pleistocene sae - temperate biotas in Mexico and eastern United States. Ecology 38: 468-480. McKenna, M.C. Was te connected directly to caer one prior to the Eocene? Evol. Biol. 6: 178-189. 1972. McVEAn, D. N. Ecology of Ans glutinosa (L.) Gaertn. II. Seed distribution and ger- mination. Jour. Ecol. 43: 61- Mears, J. A. Chemical ee HES and systematics of the Amentiferae. Brittonia 25: 385-394. 1973. Meuse, A. D. J. Floral evolution of the Hamamelidae. I. General assessment of the probable phylogeny and taxonomic position of the group. Acta Bot. Neerl. 24: 155- 164. 1975a; II. Interpretive floral morphology of the Amentiferae. [bid. 165-179. 1975b; If. Hamamelidales and associated groups including Urticales, and final conclusions. [bid. 181-191. 1975c. Meisner, C. F. Plantarum vascularium genera. 442 pp. Leipzig. 1837-1843. [Coryleae, 346. 184 MELcHIor, H. oa Pp. 47-49 in A. Engler’s Syllabus der Pflanzenfamilien. Vol. 2. Berlin. METCALFE, C. * & L. CHALK. Betulaceae. Corylaceae. Pp. 1302-1309 in Anatomy of the dicotyledons. Vol. 2. Oxford, England. 1 1950. MEYERHOFF, A. A. study of leaf venation in the Betulaceae, with its application to paleobotany. 168 pp. Unpubl. Ph.D. dissertation, Stanford Univ. 1952. MicHAux, A. Flora Boreali-Americana. Vol. 2. 340 pp. Paris. 1803. [Betula, 180, 181; Corylus, 201; Carpinus, 201, 202. MIRANDA, F., & ‘A. J. SHarp. Characteristics of the vegetation in certain temperate regions of eastern Mexico. Ecology 31: 313-333. 1950. [Habitat of Carpinus tropicalis. ] Mrvosul, N., & S. UEYANA. Pollen morphology by means of scanning electron micro- sco n Japanese with English summary.) Jap. Jour Palynol. 27: “19-26. 1981. [Descriptions and photographs of representative betula- ceous pollen.] MoerMAN, R. I. Medicinal plants of Native Americans. Vol. 1. xix + 534 pp. Ann Arbor. 1986. [Alnus, 26-30; Betula, 92, 93; Carpinus, 102; Corylus, 138; Oe 318, 319. MoHLENBROCK, R. H., & P. M. THomson. The illustrated flora of Illinois: flowerin plants, smartweeds to hazelnuts. xiii + 228 pp. Carbondale. 1987. [Betulaceae, 180- 201; Corylaceae, 201-210.] MorvILLez, M. F. L’appareil libéroligneux foliaire des na Corylacées et Cas- tanéacées. Compt. Rend. Acad. Sci. Paris 170: 674-677 0. MosELEY, M. F., Jn. Vegetative anatomy and morphology af eee Brittonia 25: 356-370. 1973. Mural, S. Phytotaxonomical and geobotanical studies on gen. A/nus in Japan (III). Taxonomy of whole a and distribution of each sect. Bull. Gov. Forest Exper. Sta. 171: 1-107. NAWASCHIN, S. Kurzer aries meiner fortgesetzten Studien iiber die Embryologie der Betulineen. Ber. Deutsch. Bot. Ges. 12: 163-169. 1894. — 16 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 NeE, M. Betulaceae. Fl. Veracruz 20: 1-20. 1981. [A/nus, 2-11; Carpinus, 11-16: Ostrya, 16-20.] magerk F. P., & D. E. FArRBROTHERS. A serotaxonomic appraisal of the “Amentifer- ae.” Bull. Torrey Bot. Club 112: 43-52. 1985. PRANTL, K. Betulaceae. Jn; A. Engler & K. Prantl, Nat. Pllanzenfam. 3(1): 38-46. 1894. RApDForD, A. E. Betulaceae. Pp. 366-370 in A. E. Radford, H. E. Ahles, & C. R. Bell. Manual of the vascular flora of the Carolinas. Chapel Hill. 1968. RAFINESQUE, C. S. Florula Ludoviciana: or, a flora of the state of Louisiana. 178 pp. New York. 1817. [Carpinum virginianum (= Carpinus caroliniana), 159; Zugilus virginica (= Ostrya virginiana), 159.] Raven, P. H. The bases of angiosperm phylogeny: cytology. Ann. Missouri Bot. Gard. 62: 724-764. 1975. Recorp, S. J., & R. W. Hess. Timbers of the New World. xi + 640 pp. New Haven. 1943. [Betulaceae, 74-76; Corylaceae, 127, 138. REGEL, E. aor faa Bearbeitung der Betulaceen. Mem. Soc. Imp. Nat. Moscou 13(2): 59-187. 1861. . Betulaceae. ao Prodr. 16(2): 161-189. 1868. ReHDeER, A. Manual of cultivated trees and shrubs. ed. 2. xxx + 996 pp. New York. 1940. [Corylaceae, 124-146.] Bibliography of cultivated trees and shrubs. xl + 825 pp. Jamaica Plain, Mas- -112. RENDLE, A. B. The classification of flowering plants. Vol. 2. xix + 640 pp. Cambridge, England. 1925. [Betulaceae, 23-30.] SARGENT, C. S. Manual of ae trees of North America. ed. 2. xxvi + 910 pp. Boston. 1922. [Betulaceae, 200-22 ScoGGan, H.J. Flora a. 4 vols. xiii + 1711 pp. Ottawa. 1978-1979. [Betulaceae, 587-596. 1978.] SMALL, J. K. Flora of the southeastern United States. 1370 pp. New York. 1903. [Cor- ylaceae, Betulaceae, 343-346. ] anual of the southeastern flora. xxii + 1554 pp. New York. 1933. [Corylaceae, as 415-419. Situ, E. C., & C. NicHoLs, Jr. Species hybrids in forest trees. Jour. Arnold Arb. 22: 443-454, 1941. [Nine artificial hybrids. ] SOLEREDER, H. Systematic anatomy of the dicotyledons. Vol. 2. vi + 1182 pp. Oxford, England. rans pee daa 791-797, SoLomon, W. R., & O. C. DuRHAM. Soe I. Pp. 340-397 in J. M. SHELDON, . K. Mica eds., Pollens and the plants that produce them. Philadelphia. 1967. [Hay fever caused by Alnus i in the Pacific Northwest; cross- a of vate and Betula pollen allergens. Soper, J. H., & M. L. HemmBpurcer. Shrubs of Ontario. xxxi + 495 pp. Toronto. 1985. Sanne 93-107. Spacu, E. Revisio Betulacearum. Ann. Sci. Nat. II. 15: 182-212. 1841. ——.. Notes sur les Carpinus. Ibid. 16: 248-254. 1842a . Notes sur les Ostrya. Ibid. 243-248. 1842b. ——. Notes sur les Corylus. Ibid, 98-108. 1842c. SPORNE, K. R. A new approach to the problem of the primitive flower. New Phytol. 48: 259-276. 1949. STEBBINS, G. L. Adaptive aes of teat: characters in angiosperms. II. Seeds and seedlings. Ann. Rev. Ecol. Syst. 2: 237-260. 1971. STERN, W. L. Development ie amicntfecos concept. Brittonia 25: 316-333. 1983. Stone, D. E. Patterns in the evolution of amentiferous fruits. Brittonia 25: 371-384. 1973. TAKHTAJAN, A. Flowering plants: origin and dispersal. x + 310 pp. Edinburgh. 1969. 1990] FURLOW, BETULACEAE Ly —. Outline of the classification of flowering plants (Magnoliophyta). Bot. Rev. 46: 225-359. 1980. [Fagales, 267. J) THORNE, R. B. Synopsis of a putativ tic classificati f the flowering plants. Aliso 6: 57-66. 1968. (Hamameliilorae, 62, 03. J The ““Amentiferae”’ or ] g p statement. Brittonia 25: 395-405. 1973. Proposed new realignments in the angiosperms. Nordic Jour. Bot. 3: 85-117. TIPPO, O. Comparative anatomy of the Moraceae and their presumed allies. Bot. Gaz. 100: 1-99. 1938. Titmuss, F. H. Commercial timbers of the world. ed. 2. vii + 351 pp. Cleveland. 1971. [Alder, 43, 44; birch, 66, 69, 70; hazel, 141; hornbeam, 148; ironwood, 155.] TOURNEFORT, J. P. Institutiones rei herbariae. editio altera. Vol. 1. xx + 697 pp. Paris. 1700. Tutin, T. G. Corylaceae. In: T. G. Tutin et al., eds., Fl. Europaea 1: 60. 1964. Voss, E. G. Michigan flora. Part II. Dicots (Saururaceae—Cornaceae). xix + 724 pp. Bloomfield Hills, Michigan. 1985. [Betulaceae, 61-72.] WacEnitz, G. Betulaceae. Jn: G. Heai, Illus. Fl. Mittel-Eur. ed. 3. 3(1): 136-196. 1981. WALKER, J. W., & J. A. Dove. The bases of angiosperm phylogeny: palynology. Ann. Missouri Bot. Gard. 62: 664-723. 1975. Wa ters, S. M. Betulaceae. Jn: T. G. TuTIn et al., eds., Fl. Europaea 1: 57-59. 1964. [Betula by S. M. Watters, A/nus by P. W. BALL.] Wanscuer, J. H. The basic chromosome number of the higher plants. New Phytol. 33: 101-126. 1934. [Predicted 7 to be the basic chromosome number of the Betulaceae.] WELsH, S. L., N. D. Atwoop, S. Goopricn, & L. C. Hiccins. A Utah flora. 111 + 894 pp. Provo, Utah. 1987. [Betulaceae, 57, 58.] WetzeL, G. Chromosomenzahlen bei den Fagales. Ber. Deutsch. Bot. Ges. 45: 251, 252. 1927. hromosomenstudien bei den Fagales. [Part I.] [bid. 46: 212-214. 1928; [Part IL] Bot. Arch. 25: 257-283. 1929. Wiius, J. C. A dictionary of the flowering plants and ferns. ed. 8 (revised by H. K. Airy SHAW). xxii + 1245 + Ixvi pp. Cambridge, England. 1973. [A/nus, 44; Betula, 136; Betulaster, 137; Carpinaceae, 206; Carpinus, 206; Corylus, 295; Ostrya, 834.] WINKLER, H. Betulaceae. Pflanzenr. IV. 61: 1-149. 1904. WopeHouse, R. P. Pollen grains. xv + 574 pp. New York. 1935. [Betulaceae, 362- 371 . Hayfever plants. xx + 245 pp. Waltham, Massachusetts. 1945. [Betulaceae, 67— 75.] Wo Fe, J. Neogene floristic and vegetational history of the Pacific ee Madrofio 20: 83-110. 1969. [Miocene fossil records in western North America. . Fossil forms of Amentiferae. Brittonia 25: 334-355. 1973. WoLLENWEBER, E. Flavonoidmuster im Knospenexkret der Betulaceen. Biochem. Syst. Ecol. 3: 47-52. 1975. WoopwortH, R. H. Parthenogenesis and polyploidy in A/nus rugosa (Du Roi) Spreng. Science 70: 192, 193. 1929a. . Cytological studies in the Betulaceae. I. Betula. Bot. Gaz. 87: 331-363. 1929b; II. “ Corylus and Alnus. Ibid. 88: 383-399. 1929c; III. Parthenogenesis and polyem- bryony in Alnus rugosa. Ibid. 89: 402-409. 1930a; IV. Betula, Carpinus, Ostrya, ee Ibid. 90: 108-115. 1930b. Polyploidy in the Betulaceae. Jour. Arnold Arb. 12: 206-217. 1931. YAGMAIE, M., & D. CaTLING. The occurrence of vascular tracheids in Betula and som other betulaceous genera. Plant Syst. Evol. 147: 125-132. 1984. [True tracheids occur in the wood of Carpinus and Corylus.] 18 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 KEY TO THE SUBFAMILIES, TRIBES, AND GENERA OF BETULACEAE IN THE SOUTHEASTERN UNITED STATES General characters: monoecious, anemophilous, deciduous trees or shrubs: leaves simple, petiolate, pinnately veined, serrate or doubly serrate to subentire, ovate, elliptic, or obovate, alternately arranged, with deciduous stipules; sta- minate flowers reduced to (1-)4(-6) stamens and an equal number of tiny, scalelike tepals (or these absent), borne in scale-covered, pendulous catkins gen- erally formed the growing season before anthesis; carpellate flowers usually consisting of a single inferior ovary composed of 2 (or 3) carpels, adnate to several tepals (when present), containing 2 locules below and 1 above, with 2 (or 3) linear stigmatic style branches, the ovules 2, parietal, borne near the summit, (except in Corylus) in usually uncrowded catkinlike, bracteate clusters; infruc- tescences often strobiluslike, with large, conspicuous bracts; fruits 2-winged sa- maras, nutlets, or nuts. A. Staminate flowers with several scalelike ne carpellate flowers without a perianth; carpellate flowers 2 or 3 per bract in the orescence; ovules apparently unitegmic; deciduous with the fruits or persistent; fruits small and laterally winged (the wings sometimes reduced to ridges) (subfamily Betuloideae B. Stamens generally 4, entire; carpellate flowers 2 oe bract; infructescence scales with (4 or) 5 lobes, greatly thickened, woody, and persistent long after release of PIS U e g Seg ck See ety Bs 2 ee ts Sed es was ene daa Sead Alnus. B. Stamens 2, bifid below the anthers; carpellate flowers 3 per bract; infructescence scales with (1-)3 lobes, thickened but not woody, deciduous with the release of WG Fe ees cepa and ona bee Ghee een arden chieetees 2. Betula. . Staminate ee lacking a carpellate flowers bearing several scalelike tepals; carpellate flo owers 2 per bract in the inflorescence; ovules bitegmic; infructescences > these deciduous with the fruits; fruits tiny to moderately large nuts, not winged (subfamily ines Leaves narrowly ovate to st veins 10 or more; infructescences elongate, loosely arranged spikes of 3 or more pairs of leafy bracts, these each either subtending or enclosing a ie nutlet (tribe Carpineae D. Infructescence bracts flat, open, 1- to 3-lobed and variously toothed. ..... Goo eG as Sain Ged alta iN ee ate ee ae eee eee a ae ee eet eo Carpinus. D. Infructescence bracts forming inflated bladders, these completely enclosing the fruits. rya. C. Leaves broadly ovate to suborbicular, veins 8 or fewer: infructescences festa clusters of several small to moderately large nuts, these each surrounded by an involucre of several coarsely toothed teaflike bracts, the involucre sometimes long and tubular (tribe Coryleae). 5. Coryl Subfamily BETULOIDEAE Tribe BETULEAE 1. Alnus Miller, Gard. Dict. abr. ed. 4. [alph. ord.] 1754. Small to large shrubs [or small to medium-sized pyramidal to round-crowned trees], usually with several trunks; branching excurrent to deliquescent, when 1990] FURLOW, BETULACEAE 19 excurrent, often becoming deliquescent in age; trunks and branches terete, the branchlets and twigs subdistichous to diffuse; twigs sometimes differentiated into pronounced long and short shoots (subg. ALNOBETULA). Bark close, thin and smooth [to thick, furrowed, and corky], when smooth usually dark and marked with prominent pale lenticels, these sometimes becoming elongate horizontally; young twigs glabrous or sparingly pubescent, often covered with resinous glands; leaf scars triangular to crescent shaped, with 3 more or less equidistant, deeply crescent-shaped vascular bundle scars, winter buds long stalked or subsessile, narrowly to broadly ovoid or elliptic, terete, often held more or less parallel to the twig, the apex acute to rounded, with 2 valvate (stipular) or several imbricate scales [or sometimes naked]; wood fine grained, nearly white, turning reddish upon exposure to air, moderately soft, moderately light in weight; pith triangular in cross section. Leaves 3-ranked to subdisti- chous, borne on long [or short] shoots; blades thin [to very leathery], ovate to elliptic or obovate, doubly serrate, serrate, serrulate [or subentire], abaxially glabrous to tomentose, sometimes covered abaxially with resinous glands; sec- ondary venation craspedodromous [or semicraspedodromous], mostly diver- gent and straight; leaves open and convex in bud, becoming conduplicate and plicate upon expansion; stipules broadly ovate [to narrowly linear]. Staminate catkins terminal [or lateral in leaf axils near the ends of branchlets], [solitary or] in racemose clusters, formed during the previous growing season and ex- posed [or enclosed in buds] during the winter, expanding before or with the leaves [or (in subg. CLETHROPSIS) formed and expanding during the same grow- ing season], crowded, the scales ovate, consisting of 5 fused bracts; carpellate catkins lateral, below the staminate, either on short shoots or laterally in leaf axils on long shoots, [solitary or] in small [to large] racemose clusters, devel- oping and maturing at the same time as the staminate, exposed or enclosed within buds during the winter, short, ovoid to ellipsoid, firm and erect to subpendulous, crowded, the scales composed of 5 fused bracts. Staminate flow- ers 3 per scale in the catkins, each with (3 or) 4(-6) scalelike tepals and an equal number of stamens, these borne opposite the tepals, undivided; pollen grains flattened, 19-27 wm in diameter, strongly aspidote, with (3 or) 4 or 5 (or 6) elliptic equatorial apertures connected by conspicuous pairs of thickened bands (arci). Carpellate flowers sessile, normally 2 per scale, rarely with | or more acaeabae or vestigial tepals (the latter, when present, adnate to the Ovary); OV 1 by abortion, unitegmic. Infructescences ellipsoid, ovoid [or short- ey strobiloid, conelike, borne [singly or] in racemose clusters, erect and subsessile or pendulous on long, thin peduncles, the bracts connate into woody, 5-lobed scales, these persistent until after dispersal of the fruits. Fruits small, ellipsoid to ovoid, rostrate samaras, maturing and dispersed the same season as [or the season following] pollination, the styles persistent, the wings 2, lateral, membranaceous, reduced in some species, the pericarp thin. Seeds with membranaceous testa and flat cotyledons; germination epigeal. Chromosome numbers 27 = 14, 28, 42, 56. Lecrorype species: A/nus glutinosa (L.) Gaertner; see Furlow, Rhodora 81: 74. 1979. (The ancient Latin name for the alder, used by Virgil, Pliny, and others; derived from a/o, to nourish, in reference to its usual close association with water.)— ALDER. 20 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Figure |. Alnus. a— at upper left, carpellate as on catkin in “a,” primary, secondary, and tertiary bracts (1, 2, 3) visible along with tepals of 3 flowers (4) and tips of anthers of 2 of these (5), x8; c, staminate cymules of 1990] FURLOW, BETULACEAE 21 About 25 species of forested parts of the temperate and boreal Northern Hemisphere and Central America south to northern Argentina at high eleva- tions. The alders resemble the birches but are easily distinguished from them by their infructescences, which consist of persistent woody scales with five lobes (vs. thin, deciduous, three-lobed scales). Except in members of subg. ALNOBETULA (which have subsessile buds with several true scales), they are also distinctive in having stipitate buds with two stipular scales. The fruits, borne two to a scale, are laterally winged, although the wings are sometimes reduced (occasionally to mere ridges). The alders have been variously combined and split at the generic level by many authors. Linnaeus and his immediate followers combined Alnus, as used by Tournefort and Linnaeus himself in the first edition of Genera Plantarum, with Betula, while Czerepanov, Ledebour, Murai (1963), Spach (1841), and others have treated the present subgenera as genera. However, historically and currently the genus has most widely been held to constitute a single natural entity. In addition to the family monographs, taxonomic work in A/nus has included a series of descriptive papers by Callier (1892, 1911, 1918), synopses y ene and Murai (1964), and a revision of the American taxa by Furlow (1979 The genus is s diverse, including four distinct lines of specialization. These a ( ALNOBETULA, CLETHROPSIS (Spach) Endl., and CREMASTOGYNE Schneider in Sarg.). Subgenera ALNUS and ALNOBETULA are further divided into sections, detailed below. Seven native (and several nat g three of the four subgenera occur in North America north ‘of Mexico, with an ad- ditional two being distributed throughout the mountains of Mexico, Central America, and northern South America (Furlow, 1979). Species of subg. CREMASTOGYNE, characterized by stipitate two-scaled buds, solitary axillary staminate and carpellate catkins, long-pedunculate infructescences, and fruits with broad hyaline wings, are restricted to south-central Asia. Subgenus ALNus is characterized by a shrubby or arborescent habit, winter buds with long stalks and two valvate (stipular) scales, inflorescences borne in racemose clusters, and development of both carpellate and staminate inflores- cences (which are exposed during the winter) during the growing season prior to anthesis. Its most unspecialized segment (sect. PHYLLOTHYRSUS Spach) con- 3 flowers, adaxial view, central flower seen from above, its 4 tepals (4) and 4 nae (5) visible, bracts at top of illustration as in “b” (1, 2, 3), x8; d, staminate cym abaxial surface, 2 lateral Owens seen from side, tepals (4) and stamens (5) partly visible, primary bracts (1) and paired stigmas visible, x 10; h, adaxial view of carpellate cymule with 2 flowers (ovaries undeveloped), bracts partially visible, x 20; 1, scale with flowers removed to show primary, secondary, and tertiary bracts (1, 2, 3), x20; j, late-season branchlet with mature ener and next season’s carpellate catkins above and staminate ones pendent, x 2; k, infructescence with mature fruits (stippled), x 3; 1, mature infructescence scale, abaxial surface, showing 5 lobes derived from primary, se oe and tertiary bracts (1, 2, 3), x6; m, mature fruit, style remnants at summit, x 22 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 sists of large trees of western North America, Mexico, and Central America (Alnus acuminata HBK., A. jorullensis HBK., A. oblongifolia Nutt., and A. rubra Bong.), while sect. ALNUs (subg. Gymnothyrsus (Spach) Regel; sect. CLE- THRA W. D. Koch) includes shrubby, more northern forms. In the United States and Canada sect. ALNus is represented by four taxa. A/nus incana (L.) Moench subsp. rugosa (Du Roi) Clausen and subsp. tenuifolia (Nutt.) Breitung, large shrubs of riverbanks and marshy areas, occur throughout the cooler portion of the Temperate Zone across the continent. In the East subsp. rugosa, speckled or tag alder, characterized by dark, lenticel-speckled bark and ovate to elliptic, coarsely doubly serrate leaves, is more or less confined to the region north of the glacial boundary. Subspecies tenuifolia, thinleaf or white alder, distinctive in its lighter bark and smaller leaves with more evenly spaced and less acute teeth, is distributed across much of northwestern Canada and through the western mountains as far south as central California and New Mexico. Addi- tional subspecies occur at equivalent latitudes in Europe and Asia. Alnus ser- rulata (Aiton) Willd., smooth alder, occurs throughout the Atlantic and Gulf coastal plains, in the southern Appalachians and the Ozark Highlands, and northward in the Mississippi Embayment. Its leaves are large, somewhat leath- ery, elliptic to obovate, and serrulate or finely and irregularly doubly serrate. The distribution of this species is essentially coastal, but it reaches north to central Ohio and Indiana in the interior, and to New York and Massachusetts in the East. Disjunct populations occur along the St. Lawrence River and the lower Great Lakes to southern Lake Michigan. The remaining species, 4. rhom- bifolia Nutt., white alder, is a small tree of riverbanks and canyons in mountains of the western United States. Subgenus ALNOBETULA consists of shrubby species of regions with cold cli- mates. It has sometimes been segregated as a separate genus, A/naster Spach (Alnobetula Schur, Semidopsis Zumaglini, Duschekia Opiz; see Furlow, 1979). In this group the buds are subsessile and covered by several imbricate scales. Both staminate and carpellate catkins are formed the season before anthesis, but only the staminate ones are exposed during the winter. In North America it is represented by the circumpolar 4/nus viridis (Chaix) DC. in Lam. & DC. In the Northeast the transcontinental, far-northern subsp. crispa (Aiton) Turrill, green alder, occurs along the Appalachian Mountains at progressively higher elevations southward to central New York and Massachusetts, with disjunct populations in southern Pennsylvania (Wherry) and on the summit of Roan Mountain on the Tennessee—-North Carolina border (Brown; Clarkson; A. Gray, 1842). It grows along streambanks and also in rocky and drier sites in colder climates. This subspecies is recognizable by its medium-sized, ovate, serrulate or finely serrate, usually glutinous leaves and its subsessile buds with more than two scales. In western Canada, and southward in the mountains of the Northwest, subsp. crispa is replaced by subsp. sinuata (Regel) Live & Live, Sitka alder, a large shrub with larger, broader, thinner, more coarsely toothed leaves. This sub- species occurs along streams and frequently covers moist mountain slopes near the timberline, especially where landslides have created open areas. Subspecies fruticosa (Rupr.) Nyman is distributed from coastal Alaska to British Columbia, 1990] FURLOW, BETULACEAE 23 Washington, Oregon, and northern California, as well as across the Bering Strait in northeastern Asia (Furlow, 1983b). Its vegetative morphology some- what resembles that of subsp. crispa, with which it has sometimes been confused in the = Subspecies viridis is distributed throughout the mountains of west- ern Europ The ata Asian subg. CLETHROPSIS is represented in America by a single species, A/nus maritima Muhl. ex Nutt., a small tree of stream banks, marshes, and the shores of shallow lakes. Its distribution is limited to two widely disjunct populations, one on the Delaware-Maryland-Virginia peninsula and the other in south-central Oklahoma (Furlow, 1979; Stibolt). Members of subg. CLETHROPSIS are unique in that they bloom in autumn, rather than in spring. They also differ from other native species in having essentially naked buds, leaves with semicraspedodromous venation, and solitary carpellate in- florescences borne in the axils of foliage leaves. This group was considered to be the most primitive one in A/nus by Murai (1963) and Takhtajan (1969) on the bases of morphology and phytogeography. However, Furlow (1979) con- cluded that many of its distinctive structural and life-history features represent derived conditions, and he placed it in a moderately advanced position in the family. Van Steenis, following Regel, has treated populations of the Asian A. japonica (Thunb.) Steudel as conspecific with American A. maritima. However, no critical comparison of the two has been made to determine the soundness of that arrangement. Alnus serrulata, a common shrub along open streambanks throughout the region, was erroneously called 4. rugosa by Britton & Brown (1896, 1913), Robinson & Fernald, and Small (1903, 1933) (see also Fernald, 1945a), and misapplication of the name A. rugosa continues in a few floras and herbaria. Alnus incana subsp. rugosa, a related taxon reaching its southern limit in West Virginia in the mountains and in Maryland on the Coastal Plain, hybridizes with A. serrulata where their ranges overlap, and extensive and often unrec- ognized hybrid swarms are formed (Furlow, 1979; Steele). From the results of a serological study (based on limited material taken entirely from the region of geographic overlap), Villamil & Fairbrothers concluded that the two taxa constitute a single species. Woodworth (1929a, 1930a) described apomixis from populations of A. serrulata (which he called A. rugosa) taken from this swarm. However, in a later paper (1931), he showed that populations located away from the region of intergradation have normal cytological characteristics. The two species are not oe fo one eae in the field outside the region of overlap. A/nus vate leaves, often with rounded apices and with fet sinall teeth, while ve incana subsp. rugosa has ovate to elliptic leaves with acute apices and usually coarse, doubly serrate margins. The species are also distinguishable by the dark, rather shiny red-brown bark marked by prominent light-colored lenticels in 4. incana and the dull, uniform light gray or light brown-gray bark with inconspicuous lenticels in A. serrulata. The least-well-understood segments of the genus, with respect to the diver- sification and relationships of the genus as a whole and to the circumscription of individual species and infrageneric groups, are those occurring in Latin America and China. The Latin American species, all members of subg. ALNUS 24 JOURNAL OF THE ARNOLD ARBORETUM [vov. 71 sect. PHYLLOTHYRSUS, were mostly described by Bartlett and Fernald (1904a), who were later followed by Standley. Furlow’s (1977, 1979) study of this com- plex showed the many species and varieties to constitute two species consisting ofa number of somewhat incompletely differentiated geographic and ecological races. He concluded that this group demonstrated many of the most unspe- cialized characters in the genus and represented remnants of a very old intro- duction from Asia, possibly dating from the Miocene or earlier. The relation- ships of the Asian taxa, particularly of species of subgenera CLETHROPSIS and CREMASTOGYNE, to the better-known American and European species, have not been taken sufficiently into account by Western taxonomists. These species will need to be studied thoroughly before many questions regarding the origin and diversification of the genus can be answered. The shrubby habit of the northern and eastern North American alders is regarded as a specialized condition that evolved in response to harsh northern winters. The species of more equable regions (e.g., the Pacific Northwest, the mountains of Latin America, and the foothills of the Himalayas) become large trees and develop relatively thick, corky bark. It has been shown by both Forsaith and Hall that some shrubby species of A/nus retain juvenile wood characters, suggesting that the shrub habit was derived through neoteny. The existence of the shrub habit in several divergent lines within the genus suggests that it has arisen independently several times Leaves of species other than those in subg. CLeripopsis are rather uniform in morphology, varying mostly in size, general shape, and size of the serrations. Occasionally forms with deeply cut leaves occur naturally (see Hylander, 1957a), and these have attracted horticultural interest. In leaf venation and margins species of subg. CLETHROpsis differ both from other A/nus subgroups and from other Fagales. The teeth are small, distant, and single (although on the basis of secondary vein endings, they are apparently derived from a doubly serrate form). The secondary veins branch before reaching the tooth: one of the branch- es enters the tooth along its apical edge, and the other connects with another vein or ends in the adjacent tooth. The tertiary (cross) veins are poorly de- veloped in relation to those in other species. The leaves of A/nus, like those of the other Betulaceae, bear trichomes of various types, including simple hairs, which are sometimes extremely dense on the abaxial surface, as well as both stipitate and sessile glands (Furlow, 1979; Hardin & Bell). In some taxa (e.g., 4. jorullensis subsp. /utea Furlow and A. viridis subsp. crispa) these glands are large and conspicuous under magni- fication and have been given diagnostic status (e.g., by Standley). However, the great variability in the presence and prominence of the glands, like the variation in leaf pubescence, renders their use for identification largely inef- fective (Furlow, 1979). The wood anatomy of A/nus is the least specialized in the family (Hall). It is similar to that of Betula (e.g., some species have opposite intervascular pitting, perforation plates with many scalariform bars, and vessels usually frequent and small (Furlow, 1979: Hall)), although certain species are char- acterized by more advanced features. A/nus viridis appears to be the most specialized of the American species, while 4. incana and A. serrulata are of 1990] FURLOW, BETULACEAE 20 intermediate advancement (Furlow, 1979) and A. maritima contains a mixture of primitive and advanced features The structure of the flowers and inflorescences of A/nus, together with their various adaptive trends, have been elucidated by Abbe (1935, 1938) and Hjelmqvist (1948). The staminate catkins consist of helically arranged cymules of three sessile flowers. Subtending each cluster are one primary, two secondary, and (usually) two tertiary bracts. Each flower consists of from (one to) four to six tepals and stamens (usually four in subg. ALNUS), with the stamens opposite and adnate to the tepals at the base. Where six tepals are present, these are borne in two whorls of three (Abbe, 1935). The thecae are connate or slightly separated by a short, forked connective. The short carpellate catkins consist of extremely crowded cymules of two flowers (apparently reduced from three, as seen in Betula) and are subtended by a scale made up of the same five bracts that subtend the staminate cymules. The ovaries were characterized as “‘nude”’ (apparently inferior on the basis of vascular traces, but lacking a perianth) by Abbe (1935). The staminate inflorescences are comparable to those of Betula, except that the tertiary bracts have been retained. In comparison to Betula, the carpellate cymules are specialized in having lost the secondary (central) flower, but primitive in having retained their tertiary bracts. A simple progression of forms in the genus leads from racemose clusters of catkins (the carpellate borne below the staminate) to solitary axillary catkins by reduction (Furlow, 1979; Jager; Hjelmqvist, 1948; Murai, 1964). In sub- genera ALNuS and ALNOBETULA the staminate and carpellate catkins occur in separate clusters on different shoots; however, the carpellate catkins are pro- duced on new growth in the spring in subg. ALNOBETULA, while they are formed the season before anthesis in subg. ALNUS. A/nus maritima has clustered sta- minate and solitary carpellate catkins (occurring on the same shoot). Murai (1963) viewed this condition as primitive; however, Furlow (1979) argued that the solitary carpellate inflorescences of subg. CLETHROPSIS are most likely of secondary origin by reduction of flowering branch systems such as those seen in subg. ALNUS. The genus has been studied cytologically by Gram and colleagues, Jaretzky, Poucques, Wetzel (1927, 1928, 1929), and Woodworth (1929a, c; 1930a). All of the American species for which data are available have chromosome num- bers of 2n = 28. Other members of the genus form a polyploid series of 2” = 14, 28, and 56, with several counts of 2” = 42, these apparently having orig- inated through hybridization between 2 = 28 and 2n = 56 types (see Furlow, 1979). The base chromosome number of A/nus (x = 7) 1s based on a single report (Chiba), but this is also supported indirectly by other cytological evidence (Brown & Al-Dawoodie) and the results of allozyme studies (Bousquet et al., 1987a, 1988), which show plants with 2m = 28 to behave genetically as tet- raploids. The alders are anemophilous and produce abundant pollen at anthesis (Wodehouse, 1935), which in temperate North America occurs before (subg. ALNus) or at the same time as (subg. ALNOBETULA) the unfolding of new leaves, or in late summer (subg. CLETHROPSIS) just after the new catkins mature. The latter condition has been interpreted as precocious and therefore derived (Fur- 26 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 low, 1979). In Latin American populations (subg. ALNUs sect. PHYLLOTHYRSUS) anthesis occurs mostly in December and January (very early spring). The shrub- by northern species often form extensive thickets that give the impression of clonal growth. However, genotypes in such populations have been found to be randomly distributed in A/nus incana subsp. rugosa (Huenneke), suggesting that they result from seeding, rather than from vegetative reproduction. In all species the tiny fruits are abundantly produced and widely distributed. In some species, as in Betula, the fruits are winged and are carried by the wind. In others (e.g., A. serrulata) the wings have been reduced to ridges, in which case dispersal may be primarily by water currents. Allozyme studies by Bousquet and colleagues (1987a-c, 1988) have shown low inbreeding with high levels of gene flow within populations of both 4. incana subsp. rugosa and A. viridis subsp. crispa. Little population differentiation was noted in these studies (1987b, 1988), suggesting relatively high interpopulational gene flow as well. Members of the genus hybridize readily where species occur together. However, most species are separated by habitat or geography, and except at the intersections of the ranges of the various taxa, extensive genetic mixing does not occur. Lists of named hybrid taxa in the genus are provided by Murai (1964) and Winkler. Alnus (and Betula) appear earlier in the fossil record than the other Betulaceae (Crane & Stockey; L. J. Hickey & Doyle; Wolfe, 1973), but the precise time and place of origin of the genus, like those of the family, are a matter of speculation. Takhtajan (1969) believed that the group developed in south- western Asia, while Murai (1964) placed its origin in the area of present-day Japan. Furlow (1979) concluded that the alders most likely originated in tem- perate Asia, with diversification there followed by progressive migrations east and west into Europe and the New World. The species of A/nus currently inhabiting North America appear, from fossil and phytogeographic evidence, to have entered from both the east and the west at several different times (Furlow, 1979; McKenna; cf. Love & Léve). The ancestors of both A. maritima and the Latin American taxa may have entered in the early Tertiary from Asia. Using fossil pollen, Graham (1973a) concluded that A/nus and other woody mesophytic genera from northwestern North America migrated to southern Latin America during the Miocene. Graham (1973a) and Martin & Harrell have reviewed the evidence covering the introduction of this element into southern Mexico and Central America (cf. Deevey; Dressler; Miranda & Sharp). Fossils suggest that subgenera ALNUS and ALNOBETULA had already differ- entiated and were present in western North America by the Miocene (Wolfe, 1969). However, a recent study employing allozyme data (Bousquet et ai., 1988) placed the time of divergence of A/nus incana subsp. rugosa from A. viridis subsp. crispa populations at only about one million years ago. The American subspecies of A/nus viridis and (to a lesser extent) A. incana are only very slightly differentiated morphologically from their Eurasian races, and it seems likely that they may be of recent (possibly post-Pleistocene) introduction, especially in the West (cf. Hultén). However, as noted by Bousquet and col- leagues (1988), Furlow (1979), and others, it is possible that these species, at least in part, survived Pleistocene glaciations in refugia in northern North America. A/nus serrulata, closely related to A. incana on the basis of mor- 1990] FURLOW, BETULACEAE 2 phology (Furlow, 1979), may have entered from Europe at an earlier time. However, alder pollen is not known in the pre-Pleistocene sediments of the Atlantic and Gulf coastal plains where A. serrulata would be expected to have existed during the Pleistocene (J. Gray). The alders associate symbiotically with species of the actinomycete Frankia, which lead to the formation of nodules on the roots of the plants and fix atmospheric nitrogen (Bond; Bond et a/.; Dalton & Naylor; Hawker & Fray- mouth). The importance of alders in plant succession is well documented for different species and a variety of physical settings (e.g., Crocker & Major; Fremstad; Newton et al.; Reiners et al.; Tarrant, 1968; Ugolini). During the past two decades foresters and plant physiologists have shown great interest in the nitrogen-fixing ability of actinorhizal plants, and an extensive literature has developed related to research into details of the process and the biology of the organisms involved, including identification, isolation, cultivation, morpho- genesis, ultrastructure, ecology, nitrogen-fixing activity, inheritance, metabo- lism, chemosystematics, growth, reaction to various environmental factors, nutritional requirements, infection of hosts, and host-endophyte interactions. Some of this work represents biotechnological research aimed at the “genetic improvement” of alders and their symbionts (see Gordon et al.; Hall & May- nard; Hall, McNabb, Maynard, & Green; Hall, Miller, Robison, & Onokpise; Normand & Lalonde). Of special interest is the genetic recombination of large tree species, especially A/nus glutinosa, European black alder, A. rubra, red alder, A. cordata (Loisel.) Loisel., Italian alder, A. incana (L.) Moench subsp. incana, European white alder, and other species. This activity is in part due to a recent interest in the possible use of alders in intensive silviculture (see Dickman; Gordon & Dawson; Tarrant, 1983). Symposium papers dealing with this subject have been published by Gordon & Wheeler, Gordon, Wheeler, & Perry, and Torrey & Tjepkema (1979, 1983). The articles cited illustrate the range and scope of current work in this field. Alders are not seriously bothered by diseases or insect pests, although various insects feed on their foliage (Sargent, 1896). In the Southeast cottony scale insects are frequent parasites of A/nus serrulata. Hepting reviewed the many fungal diseases known to affect A/nus, but he concluded that most are of little or no economic importance. The most serious pathogen of tree-sized alders in North America is heart rot (Fomes ignarius (L.) Kickx), which usually appears only in trees over 40 years old (Hepting; Worthington et a/.). A species of Taphrina affects the carpellate catkins of many species, resulting in curled, straplike enlargements of the infructescence bracts. The shrubby species of A/nus in eastern North America are mostly oppor- tunistic plants that rapidly colonize disturbed habitats. In other parts of the world, the genus includes large trees that are important components of the mature natural vegetation. In the Pacific Northwest A. rubra is a dominant tree of floodplain forests, where it has considerable commercial value. Throughout the mountains of Mexico and Central America, A. acuminata and A. jorullensis become large trees and serve locally as a source of lumber. Alders have been put to a great many uses by many cultures throughout the centuries. Various groups of North American Indians, as well as white settlers 28 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 in the New World, utilized the astringent properties of alder bark for a wide variety of medicinal purposes, including the treatment of burns, infections, leukorrhea, toothaches, and indigestion (see Lewis & Elvin-Lewis; Moerman). The triterpenes betulin and lupeol, extracted from bark and wood of A/nus rubra, have recently been found to have antitumor activity in laboratory an- imals (Sheth et a/.). In regions where alders make up a significant part of the vegetation, their pollen is an important cause of hay-fever allergies (Cham- berlain, 1927; Florvaag & Elsayed; Florvaag, Elsayed, & Apold; Florvaag, Elsayed, & Hammer; Lewis ef a/.; Lowenstein et a/.; Solomon & Durham). In Europe and America the wood, which is fine grained, although rather soft and not very durable, has been used for beams and piles, shipbuilding, cabinetry, boxes, and the manufacture of a wide variety of small wooden objects, ranging from toys and tool handles to wooden shoes. The wood was formerly greatly valued for the production of high-quality charcoal for gunpowder manufacture. One of the most important present uses in the United States and Canada, especially in the Pacific Northwest, is as a source of pulpwood for making paper (Worthington ef a/.). The U. S. Forest Service has published two symposium volumes (Briggs et al.; Trappe et al.) dealing with aspects of alder taxonomy, ecology, and silviculture, with particular reference to A. rubra. Several species, especially the European A/nus glutinosa, A. incana subsp. incana, and A. cordata in the East and A. rubra in the Northwest, are occasionally cultivated as or- namentals. REFERENCES: Under family ate see sg dey 1938); eee et al. (1987, 1988); BRITTON & Brown (1896, 1913); I. R. BRown & AL-DawoopiE; CHIBA; CRANE & STOCKEY; DEEVEY; DRESSLER; FORSAITH: Furiow (1979): ares (1973a); J. GRAY; HALL; Har- DIN & Bett; HeptinG; L. J. Hickey & DoyLe; Hye-mevist (1948); HULTEN; JAGER; JARETZKY; Lewis & ELvin-Lewis; Lewis et al.; LINNAEUS; LOvE & LOveE; LOWENSTEIN ef al, MArtIN & HARRELL; MCKENNA; MIRANDA & SHARP; MOERMAN; Mural (1964); SMALL (1903, 1933); SoLOMON & DURHAM; SPACH (1841); TAKHTAJAN (1969): WETZEL (1927, 1928, 1929); WINKLER; WODEHOUSE (1935); WOLFE (1969, 1973); and WoopworTH (1929a, 1929c, 1930a, 1931). An, C. S., J. W. Wiis, W. S. Riccssy, & B. C. MULLIN. SA neue ET acid base composition of 12 Frankia isolates. Canad. Jour. Bot. 61: 2859-2862. 3. BARTLETT, H. H. Notes on Mexican and Central American alders. Proc. ree Acad. Arts Sci. 44: 609-612. 1909. Berry, A. M., & J.G. Torrey. Root hair deformation in the infection process of Alnus rubra. Canad. Jour. Bot. 61: 2863-2876. 1983. Bio, J., W. ROELOFSEN, & A. D. L. AKKERMANS. Assimilation of nitrogen in root nodules of alder, A/nus glutinosa. New Phytol. 89: 321-326. 1981. Bonp, G. Evidence for fixation of nitrogen by root nodules of alder (A/nus) under field conditions. New Phytol. 55: 147-153. 1956. . Nitrogen fixation in some non-legume root nodules. Phyton Buenos Aires 24: 57-66. 1967 , Ww. ETCHER, & T. P. FERGUSON. The development and function of the root nodules of Alnus, Myrica, and Hippophaé. P|. Soil 5: 309-323. 1954. [Nitrogen fixation in mae different endophytes are associated with each genus.] BousaquET, J., M. CHELIAK, & M. LALONDE. Genetic diversity within and among |] 1990] FURLOW, BETULACEAE 29 juvenile populations of green alder (A/nus crispa) in Canada. Physiol. Pl. 70: 311- 318. 1987b. ———_,, ———., & ———.. Genetic differentiation among 22 native populations of green alder (Alnus crispa) in central Quebec. Canad. Jour. Forest Res. 17: 219-227. 1987c. Bricos, D. G., D. S. DEBELL, & W. A. ATKINSON, eds. Utilization and management of alder. 379 pp. Portland, Oregon. 1978. [Symposium volume dealing with A/nus silviculture. ] Brown, D. M. The vegetation of ar ee a phytosociological and successional study. Ecol. Monogr. 11: 61-97. BurcarRAaF, A. J. P., A. QUISPEL, a es J. VALsTAR. Methods of isolation and cultivation of rae species from actinorhizas. Pl. Soil 61: 157-168. : Cacuer, A. Uber die in Schlesien vorkommenden Formen der Gattung A/nus. Jahresber. Schles. Ges. Vaterl. Cult. 69(2): 72-85. 1892. . Alnus. Pp. 119-136 in C. K. Schneider, Illustriertes Handbuch der Laubholz- kunde. Vol. 1. Jena. 1904. Diagnoses formarum novarum generis A/nus. Repert. Sp. Nov. 10: 225-237. 1911. Aln eae der europdischen Herbarien und Garten. Mitt. Deutsch. Dendr. Ges. 27: 39-185. 1918. CHABOUD, A., & M. 4 LONDE. Lectin binding on surfaces of Frankia strains. Canad. Jour. Bot. 61: 2889-2897. 1983. CHAMBERLAIN, C. Hayfever in the Pacific Northwest. Ann. Otol. Rhinol. Laryngol. 36: 1083-1092. 1927. [Hay fever caused by A/nus pollen.] CLARKSON, R. B. Notes : the distribution of A/nus crispa in eastern North America. Castanea 25: 83-86. Crocker, R. L., & ce Soil development in ae to vegetation and surface age in Glacier Bay, Alaska. Jour. Ecol. 43: 427-448. 1955. CZEREPANOV, S. Systema generis A/nus Mill. s. str. se affinium. Not. Syst. Herb. Inst. Bot. Komarovi Acad. Sci. U.R.S.S. 17: 90-105. 1955 Datton, D. A., & A. W. NAyLor. Studies on nitrogen fixation by Alnus crispa. Am. Jour. Bot. 62: 76-80. 1975. DickMaN, D. I. Plant materials ie for intensive culture of wood fiber in the North Central region. Iowa State Jour. Res. 49: 281-286. 1975. saa M. L. Synopsis of the eee and Central American species of A/nus. Proc. cad. Arts Sci. 40: 24-28. 1904a. ; The green alders of New England. Rhodora 6: 162, 163. 1904b. [Some North American Corylaceae (Betulaceae).] zi Eastern North American representatives of Alnus incana. Rhodora 47: 333-361. 1945a. FLORVAAG, E., & S. ELSAYED. Comparative studies on tree erie allergens. VIII. Im- munological properties of the alder, A/nus incana, pollen extract. Int. Arch. Allergy ppl. Immunol. 75: 300-308. 1984. [Isolation of A/nus pollen allergens.] ,& J. APpotp. Comparative studies on tree pollen allergens. II. Isolation of: alder (Alnus incana) pollen allergens: purification and some eas of the major allergen, p14.78. Int. oe Allergy Appl. Immunol. 67: 49-56. ,&A.S, ER. Comparative studies on tree pollen Pe XIII. Partial Characterization of the alder (A/nus) extract by two-dimensional IEF/SDS- PAG electrophoresis combined with electrophoretic transfer and immunoautora- diography. Int. Arch. Allergy Appl. Immunol. 80: 26-32. 1986. a E. Role of black alder (4/nus glutinosa) in vegetation dynamics in west rway. Nordic Jour. Bot. 3: 393-410. 1983. Furtow . J. Nomenclatural changes in A/nus (Betulaceae). Ann. Missouri Bot. Gard. 63: 380, 381. 1976 . Betulaceae. Jn: W. BurGer, ed., Flora Costaricensis. Fieldiana Bot. 40: 56-58. 77 30 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 ———. Evolutionary divergence and classification of the Alnus viridis complex (Betu- laceae). (Abstract.) Am. Jour. Bot. 70(suppl.): 114. 1983b. ORDON, J. C., & J. O. Dawson. Potential uses of nitrogen-fixing trees and shrubs in commercial forestry. Bot. Gaz. 140(suppl.): S88-S90. 1979. [Possible importance of Alnus glutinosa.] Cc WHEELER, eds. Biological nitrogen fixation in forest ecosystems: foun- dations and applications. 342 pp. The Hague. 1983. [Silvicultural practice and research involving nitrogen-fixing plants, especially Alnus. : , & D. A. Perry, eds. Symbiotic nitrogen fixation in the management of oo. forests. 344 pp. Proceedings of a workshop held April 2-5, 1979. Corvallis, Orego 79. GRAM, K., a M. Larsen, C. S. LARSEN, & M. WESTERGAARD. Contributions to the cytogenetics of forest trees. I]. A/nus studies. Kong. Vet. Landbohojsk. Aarsskr. 1941: 44-58. 1942. Gray, A. Notes ofa botanical excursion to the mountains of North Carolina. Am. Jour. Sci. 42: 1-49. 1842. [Disjunct population of A/nus viridis subsp. crispa on the summit of Roan Mountain.] HALL, R. B., & C. A. MAYNARD. Considerations in the genetic improvement of alder. Pp. 322-344 in J. C. Gorpon, C. T. WHEELER, & D. A. Perry, eds., Symbiotic nitrogen ote in the maa of temperate forests. Corvallis, Oregon. 1979. .H. BB, JR. A. MAynarbD, & T. L. GREEN. Toward development of opinal Ans ee ane Bot. Gaz. 140(suppl.): $120-S126. 1979. ——.,G. LER, T. L. Ropison, & O. U. ONoxpiseE. Developing Alnus for use in tensive au Pp. 36-42 in E. A. HANSEN, ed., Intensive plantation culture: 12 years research. Washington, D. C. 1983. Hawker, L. E., & J. FRAYMouTH. A re-investigation of the root nodules of species of Elaeagnus, Hippophaé, Alnus, and Myrica, with reference to the morphology and life history of the causative organism. Jour. Gen. Microbiol. 5: 369-386. 1951. Horrigre, F., M. P. LeCHEVALIER, & H. A. LECHEVALIER. Jn vitro morphogenesis and ultrastructure of a Frankia sp. Arl3 (Actinomycetales) from Alnus rubra and a morphologically similar isolate (AirI2) from Alnus incana subsp. rugosa. Canad. Jour. Bot. 61: 2843-2854. 1983. HUENNEKE, L. F. Spatial distribution of a individuals in thickets of A/nus incana ssp. rugosa, a clonal shrub. Am. Jour. Bot. 72: 152-158. 1985. HyLANDER, N. On cut-leaved and small- eee forms of A/nus glutinosa and A. incana. Sv. Bot. Tidskr. 51: 437-453. 1957a KIKUZAWA, K. Why do alder leaves fall in 2 summer? Jap. Jour. Ecol. 30: 359-368. 1980. KNOWLTON, S., A. Berry, & J. G. Torrey. Evidence that associated soil bacteria may influence root hair infection of actinorhizal plants by Frankia. Canad. Jour. Micro- biol. 26: 971-977. Komaroy, V. L. ne Gacrit: Fl. URSS. 5: 306-319. 1936b. Lepesour, C. F. A/nus. Pp. 655-658 in Flora Rossica sive enumeratio plantarum. Vol. 3. Stuttgart. 1850. McVean, D.N. Regional variation of A/nus glutinosa (L.) Gaertn. in Britain. Watsonia 3: 26-32. 1953. —.. Ecology of Alnus glutinosa (L.) Gaertn. III. Seedling establishment. Jour. Ecol. 44: 195- . ee IV. Root system. [bid. 195-225. 1956b. Mian, S., & G. B The onset of nitrogen fixation in young alder plants and its relation to aiferentiotion | in the nodular endophyte. New Phytol. 80: 187-192. 1978. Morita, Y., & N. Mrvosut. Sear study of the genus A/nus (Betulaceae) in Japan. Ecol. Rev. 21: 183-199. 1988. [Differences in exine thickness and arcus development in subgenera ALNUS ae ALNOBETULA. Mural, S. Phytotaxonomical and geobotanical studies on genus A/nus in Japan (1). 1990] FURLOW, BETULACEAE 31 Comparative studies on tree species. Bull. Gov. Forest Exper. Sta. 141: 141-166. 1962. Phytotaxonomical and geobotanical studies on so-called genus Alnus in Japan (11). Comparative studies on all species, including shrubby species. [bid. 154: 21- 72. 1963. Newton, M. B., A. Ex Hassan, & J. JAvitKovski. Role of red alder in western Oregon forest succession. Pp. 73-84 in J. M. Trappe, J. F. FRANKLIN, R. F. TARRANT, & G. M. Hansen, eds., Pee of alder. Portland: Oregon. 1968. NorMAnb, P., & M. LALONDE. The genetics of Frankia: a review. Pl. Soil 90: 429-453. 1986. [Recent rseath into the biology of nitrogen-fixing Frankia species. ] NorMAnbD, P., P. Smmonet, J. L. Burour, C. ROSENBERG, A. Morroup, & M. LALONDE. Plasmids in Frankia sp. Jour. Bact. 155: 32-35. 1983. PETERMANN, W.L. Deutschlands Flora mit Abbildungen. 668 pp. Leipzig. 1849. [Betula- 16.] 5 O16. Poucguss, M. L. pe. Etudes caryologiques sur les Fagales. I. Le genre A/nus. Bull. Mus. Hist. Nat. Paris, II. 21(2): 147-152. 1949. Reimers, W. A., I. A. Wortey, & D. B. Lawrence. Plant diversity in a chronosequence at Glacier ‘Bay, Alaska. Ecology 52: 55-69. 1971. Rosinson, B. L., & M. L. FERNALD. Gray’s new manual of botany. ed. 7. 926 pp. New York. 1908. [Betulaceae, 332-337. SARGENT, C. S. Alnus. Silva N. Am. 9: 67-82. 1896; 14: 61, 62, 104. 1902. ScHopMEYER, C. S. Alnus Ehrh. Alder. Pp. 206-211 in C. S. SCHOPMEYER, ed., Seeds of woody plants in the United States. U. S. Dep. Agr. Agr. Handb. 450. Washington, D. C. 1974. [Techniques for a Alnus from seed.] SHETH, K., E. BIANCHI, R. WiepHoprF, & J. R. Cote. 1973. Antitumor agents from A/nus oregona (Betulaceae). Jour. Pharm. Sci. 62: 139, 140. 1973. Suipton, W. A., & A. J. P. BurGGRAAF. A comparison of the requirements for various carbon and nitrogen sources and vitamins in some Frankia strains. Pl. Soil 69:149- 161. 1982. ae P.C. ees In: Trees and shrubs of Mexico, part 1. Contr. U. S. Natl. Herb. 23(1): 167, 168. 1920. STEELE, F. L. Introgression - Alnus serrulata anad Alnus rugosa. Rhodora 63: 297-304. 1961. STEENIS, C. G. G. J. VAN. Betulaceae. Fl. Males. 1(5): 207, 208. 1956. [Alnus japonica and A. maritima Conspec Cl el ee V.M. Th {/ 'tima Muhl. ex Nutt. (Betulaceae). Castanea 46: 195-200. 1981. TARRANT, R. F. Some effects of alder on the forest environment. P. 193 in J. M. TRAPP _ F. FRANKLIN, R. F. TARRANT, & G. M. Hansen, eds., Biology of alder. Portland, Oregon. 1968. . Nitrogen fixation in North American forestry: research and application. Pp. 261-277 in J. C. Gorpon & C. T. WHEELER, eds., Biological nitrogen fixation in forest ecosystems: foundations and applications. The Hague. 1983. [Possible use of Alnus glutinosa in intensive silviculture.] Torrey, J. G., & J. D. TyepKema, eds. Symbiotic nitrogen fixation in actinomycete- nodulated plants. Bot. Gaz. 140(suppl.): S1-S126. 1979. [Symposium papers con- cerning nitrogen fixation, especially in relation to silviculture, in A/nus and other woody plants. E, nternational conference on the biology of Frankia, August 4-6, 1982, Madison, Wisconsin. Introduction. Canad. Jour. Bot. 61: 2765-2767. 1983. [Intro- duction to a collection of symposium papers.] Trappe, J. M., J. F. FRANKLIN, R. F. TARRANT, & G. M. HANSEN, eds. Biology of alder. 292 pp. Portland, Oregon. 1968. 32 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Ucoutmt, F. C. Soil development and alder invasion in a recently deglaciated area of Glacier eon Alaska. Pp. 115-140 in J. M. TRAppE, J. F. FRANKLIN, R. F. TARRANT, NSEN, eds., Biology of alder. Portland, Oregon. 1968. VILLAMIL, GC. B. “YD D. E. FAIRBROTHERS. Comparative protein population investigation of the Wan Sis le complex. Biochem. Syst. Ecol. 2: 15-20. 1974. Wuerry, E. T. Intermediate occurrences of A/nus crispa. Castanea 25: 135. 1960. [Dis- junct Sop oas in Pennsylvania.] WocpertT, J. Vergleichende Anatomie und Entwicklungsgeschichte von A/nus, Alno- betula und Betula. Flora 100: 37-67. 1910. Wor THINGTON, N. P., R. H. Rutu, & E. E. Matson. Red alder: its management and utilization. U. S. Forest Serv. ae Publ. 881. 44 pp. Washington, D. C. 1962. YOUNGER, P. D., & L. A. Kapustka. N,(C,H,)ase activity by Alnus incana ssp. rugosa (Betulaceae) i in the northern hardwood forest. Am. Jour. Bot. 70: 30-39, 1983. 2. Betula Linnaeus, Sp. Pl. 2: 982. 1753; Gen. Pl. ed. 5. 422. 1754. Small to large, conical, pyramidal, or round-crowned trees [or small to large shrubs], often with several trunks; branching excurrent (becoming deliquescent in age) [or in shrubby forms mostly excurrent]; trunks and branches terete, the branchlets and twigs subdistichous; twigs usually differentiated into pro- nounced long and short shoots. Bark thin and smooth, dark brown to chalky white, often exfoliating in very thin layers, becoming thicker and scaly or furrowed in age, marked with prominent lenticels, these frequently becoming much elongated horizontally; young twigs glabrous or sparingly pubescent, often covered with resinous glands, sometimes aromatic when crushed; leaf scars crescent shaped to suboval, with 3 nearly equidistant circular to elliptic vascular bundle scars; winter buds sessile, slender, terete, divergent or appressed along the lower half, the apices acute, with several smooth imbricate scales, only the outer 3 generally visible; wood fine grained, nearly white to reddish brown, moderately hard, moderately heavy; pith circular or remotely triangular in cross section. Leaves subdistichous, usually borne on short shoots; blades thin, ovate to deltoid, elliptic [or suborbicular], doubly serrate [or serrate to shallowly lobed], glabrous to abaxially tomentose, sometimes covered abaxially with resinous glands; secondary venation craspedodromous, the veins mostly divergent and straight; leaves in bud open and convex, becoming conduplicate and plicate during expansion; stipules broadly ovate. Staminate catkins ter- minal [or lateral in leaf axils near the ends of branchlets], [solitary or] in small racemose clusters, formed the previous growing season and exposed [or en- closed in buds] during the winter, expanding with the leaves, densely arranged, the scales ovate, consisting of 3 fused bracts; carpellate catkins lateral on the branchlets, below the staminate, mostly borne on short shoots, usually solitary, developing at the same time as the staminate, enclosed within buds during the winter and expanding with the leaves, ovoid to cylindrical, firm and erect, scales and flowers crowded, the scales compact, consisting of 3 fused bracts. Staminate flowers 3 per scale in the catkin, consisting of [C1 or)] 2-4 scalelike tepals and [(1 or)] 2 or 3 [(or 4)] stamens, these divided nearly to the base (giving the impression of twice as many stamens with 1-locular anthers); pollen grains flattened, 15—30(—40) um in diameter, with 3(-7) elliptic equatorial ap- ertures. Carpellate flowers sessile, [(1—)]3 per scale, consisting of a single 2-locular 1990] FURLOW, BETULACEAE 33 ovary with 2 linear styles, sometimes with | or more staminodes; ovule | by abortion, unitegmic. Infructescences cylindrical to ovoid, strobiluslike, erect or pendulous on short peduncles, the bracts connate into coriaceous or some- what woody [(1- or)] 3-lobed scales, these usually readily deciduous with the fruits. Fruits small, ellipsoid to ovoid, rostrate samaras, maturing and dispersed the same season as pollination, styles persistent, the wings lateral, membran- aceous, the pericarp thin. Seeds with membranaceous testa and flat cotyledons; germination epigeal. Chromosome numbers 2n = 28, 42, 56, 70, 84. LEcTOTYPE species: Betula alba L.; see N. L. Britton, N. Am. Trees, 246. 1908. (The Latin name for birch used by Pliny; from batuere, “to beat,” for the birch rods used by Roman lictors to beat back crowds of people.)— BIRCH. About 35 species of small to large trees and shrubs of the Temperate and Boreal zones of the Northern Hemisphere. Like Alnus, the genus is highly diversified, especially in the Old World. In the United States and Canada it includes about 17 species, which occur in the area south to the Gulf Coastal Plain in the East and to Colorado and central California in the mountains of the West. The species hybridize freely; 16 named hybrids are listed by Kartesz & Kartesz. The birches occupy a variety of habitats, characteristically including peat lands; stream banks; lake shores; cool, damp woods; cool, moist slopes in coves; and (in cooler regions) drier, more open sites. Spach (1841) treated the birches as two genera, Betula and Betulaster, the latter an Asian group distinguished by many-veined, acuminate-toothed leaves, fruits with exceptionally wide wings, and carpellate inflorescences (and fruiting catkins) borne in racemose clusters. In his monograph of the Betulaceae, Regel (1861) recognized these two major groups as parts of Betu/a but did not clearly denote their rank or names. He indicated that these were subgenera in his subsequent (1865) revision of Betula and A/nus, but he again failed to provide a suitable name (erroneously referring to Spach’s genus A/naster, a segregate of A/nus). In his revision for De Candolle’s Prodromus, Regel (1868) named these taxa properly at the rank of section. In his 1865 revision, he further divided the two major groups into seven taxa bearing only the rank of “Gruppe.” Winkler (1904) treated the two major taxa as sections of Betula, subdividing these into four subsections corresponding to various of Regel’s subgroups, and W J. Koch, Koehne, Schneider, and others have since elevated most of them to sections, the arrangement adopted by Kuzeneva and the one used here.? Endlicher (1842, 1847) is frequently cited as the author of sections or subgenera in Betula, but he did not indicate ranks for his names either in the text or in his subsequent references to it (see Brizicky). In a recent synopsis of the genus, Fontaine recognized 52 species, 24 varieties, eight natural hybrid species, and 23 cultivars and artificial hybrids. Section CostTaTAE (Regel) Koehne consists of large, mesophytic trees, often with dark (close or exfoliating) bark; large, thin leaves; infructescence scales 3Sections CosTATAE (Regel) Koehne and Humites W. D. Koch have not yet been treated at the subgeneric level— perhaps a better course, considering their high degree of differentiation (equivalent to that of the subgenera of A/nus). 34 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 NEN p : aay an GS A Cet | ey Beeb, yA ) Ave Ficure 2. Betula. a—j, B. nigra: a, flowering branchlet with 2 ascending carpellate catkins and 3 pendent staminate ones, x 2; b, staminate cymule, oblique view of adaxial side, showing stamens of 3 flowers (tepals not visible) and tips of primary and secondary bracts (1, 2), portion of axis of catkin below, x8; c, same, side view, primary and secondary bracts at left, x 8; d, staminate cymule, seen as in “‘b,” anthers removed, to show primary (1) and secondary bracts (2), 1 tepal (4) of each of 3 flowers, and partial filaments of each of 6 stamens, x8; e, 2 views of stamens, showing half-anthers, x 12; 1990] FURLOW, BETULACEAE 35 with long, narrow lobes; and fruits with relatively narrow wings. Three species of this group occur in the southeastern United States. Betula nigra L., river birch, red birch, a large (to 30 m) tree usually with spreading clusters of trunks (each up to 1 m in diameter), distinctive rhombic-ovate leaves, and creamish to reddish exfoliating bark on young branches (dark, scaly bark on older trunks), is found throughout the region, except in peninsular Florida and certain areas of the Gulf Coastal Plain, including Alabama, Mississippi, Louisiana, and Arkansas (see Coyle et a/., 1983b; Duncan; Koevenig). This species grows on stream banks and on bottomlands. Cribben & Ungar, Fritts & Kirtland, and McClelland & Ungar have shown that in Illinois and Ohio river birch is pre- dictably present on acid soils, especially along streams heavily affected by coal- mine drainage, and largely absent from alkaline soils. However, Wolfe & Pittillo found no such relationship in western North Carolina and concluded that in their area, the availability of continuous moisture constituted the most im- portant limiting factor. Betula nigra is unique among our species in that its fruits mature, are released, and germinate in early summer, apparently an adaptation associated with the floodplain habitat (which is frequently inundated in the spring). Betula alleghaniensis Britton, yellow birch, and B. /enta L., cherry birch, sweet birch, black birch, are more northern, occurring in suitable habitats from southern Newfoundland to southeastern Manitoba and south- ward along the Appalachians to northern Georgia and northern Alabama (B. lenta only). Betula alleghaniensis (incorrectly spelled “alleghanensis” by Brit- ton & Brown, 1913) is a large forest tree, usually with a single trunk, reaching a height of about 35 m and a trunk diameter of 1.5 m. Its leaves are large, thin, ovate, and doubly serrate. The bark of young branches is usually yellowish and exfoliates in ragged curls, but Dancik and Dancik & Barnes (1971) have shown this character to be inconsistent. The bark of older trunks becomes dark and scaly. Betula lenta has somewhat similar characteristics, but it is smaller and its dark cherrylike bark does not exfoliate. The two species can be distin- guished by their infructescence scales, which in Betula alleghaniensis have pubescent, more elongate, and often more strongly ascending lateral lobes and in B. lenta glabrous, more expanded, more divergent ones. The bark and twigs of both species contain wintergreen oil (methyl salicylate), which can be de- tected by chewing fresh twigs. Betula alleghaniensis is an important constituent of the hemlock-hardwoods forest in the northern Appalachians, occurring on a variety of soil types and in various drainage conditions. In the southern Appalachians it occurs only at elevations over 1000 m. The yellow birch has been widely known in the past as B. /utea Michx. f., a superfluous and therefore illegitimate name (Michaux, after first misapplying the name B. excelsa to the f, adaxial side of eas cymule of 3 naked flowers, tips of primary and secondary bracts visible, x8; g, bract complex, abaxial side, the 2 secondary bracts (2) partially united with the ieee one ees style tips of cymule visible, x 8; h, branchlet with mature infructescences, x '4; i, abaxial side of 3-lobed bract complex of mature carpellate cymule (primary bract and 2 secondary bracts partially united (see g), x 6; j, mature achene with membranaceous lateral wings, x 6. k, |, B. /enta: k, branchlet with mature infructescences, x Y>: |, abaxial side of bract complex of mature fruiting cymule (cf. g, 1), 36 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 yellow birch, substituted the epithet /utea, which he considered more descrip- tive; see Brayshaw, 1966a). Betula uber (Ashe) Fern., known for many years only from herbarium ma- terial (Mazzeo), was rediscovered in 1975 in a single southwestern Virginia population (Ogle & Mazzeo; Reed). On the basis of leaf shape, the presence of wintergreen oil in its bark, and other characters, it had been speculated that this birch could represent a relative or hybrid of B. pumila L. (now occurring 500 miles to the north) or a variant or hybrid of B. lenta (see A. G. Johnson). However, wood anatomy clearly places B. wher in sect. COSTATAE, not with the dwarf birches, and additional evidence suggests that it is not closely related to other sympatric species of that section (e.g., B. lenta) (Hayden & Hayden). This view is supported by discriminant analysis (Sharik & Ford). The mostly circumboreal sect. BETULA (sect. Albae (Regel) Schneider) consists of small to medium-sized trees with rather large, thin leaves and fruits with relatively wide wings (wider than the body of the fruit). A characteristic feature of trees in this group is their white bark, which often peels apart in sheets due to its alternating layers of tabular cells with thick walls and larger ones with thin walls, the latter containing grains composed largely of the triterpenoid betulin, which also makes the bark waterproof (Metcalfe & Chalk). The birches of northern North America with white bark (including Betula papyrifera Marsh., B. populifolia Marsh., and B. cordifolia Regel) are often little differentiated from each other and from races of this complex in Europe and Asia, and they commonly hybridize in nature. Betula papyrifera, paper birch, canoe birch, is transcontinental in distribution across the Boreal Zone, extending south in cool forests at high elevations in the Appalachians. Although individuals of this species are relatively short lived (about 150 years), they sometimes reach a height of 30 m and a trunk diameter of nearly | m. Their distinctive features include pinkish to chalky-white exfoliating bark marked with dark, horizontal lenticels; ovate, doubly serrate, acute to acuminate leaves with rounded or cuneate bases; and infructescence scales having relatively wide, rather angular ascending lobes of about the same length. Betula populifolia, gray birch, occurs from Quebec to southwestern Ontario and south to Delaware, northern Penn- sylvania, and northern Indiana. It is distinguished from B. papyrifera by its close bark, deltoid to rhombic leaves with long-acuminate tips, and cone scales with very short central lobes. The only white-barked birch to enter our range is Betula cordifolia, heartleaf birch. This species, sometimes treated as B. papyrifera var. cordifolia (Regel) Fern., occurs from Labrador and central Ontario to northern New York, Mich- igan, Wisconsin, and northern Indiana, and south along the Appalachians as small disjunct populations as far as Mount Mitchell in the Black Mountains of North Carolina (listed as B. papyrifera in Radford). It is similar in aspect to B. papyrifera but differs in its cordate leaves with more lateral veins, reddish bark, narrower and longer infructescence-scale lobes, larger fruits, and other characters. It is found at higher elevations than B. papyrifera throughout its range. Its different chromosome number (27 = 28 or 56 in B. cordifolia, 2n = 56, 70, or 84 in B. papyrifera; Grant & Thompson; Live & Live, 1966), plus results from a discriminant analysis of morphological characters (Grant & 1990] FURLOW, BETULACEAE ai Thompson), support recognition of B. cordifolia at the species level. This in- terpretation is also supported by a study of betulin content of the bark of B. cordifolia (O’Connell et al.). Populations of a shrubby, small-leafed white birch, Betula minor (Tuckerman) Fern., occasionally occur with B. cordifolia and B. glandulosa (discussed below) from Labrador south to the Gaspé Peninsula and the Laurentian Mountains, Quebec, with disjunct populations in northern New England and the Adiron- dack Mountains. This birch was treated as a variety of B. papyrifera by Tuck- erman, asa subspeci ies of B. pubescens by Love & Léve (1966) and as conspecific with B. fontinalis by Scoggan. It has long been suspected of representing a hybrid. Lepage concluded that its holotype represents a hybrid plant, but that Canadian populations, which he named B. saxophila Lepage, constitute a nat- ural species. Love & Léve (1966) determined that plants near the summit of Mount Washington had a chromosome number of 2” = 56. However, at least one of the putative parents (B. glandulosa) has never been reported with a number higher than 2n = 28, and the other (B. minor) has usually been found also to be diploid. Therefore, if B. minor actually represents a hybrid, it may be of alloploid origin. From western Ontario to northeastern British Columbia and south in the mountains of the western United States to northern New Mexico and California, Betula papyrifera is replaced by B. fontinalis Sarg., water birch, a tall, shrubby race with darker, mostly nonexfoliating bark, smaller leaves, and cone scales with broad, ascending lateral lobes. The name B. occidentalis Hooker, often applied to this species, is illegitimate because both the original description and the specimens cited in the protologue are mixed (Dugle, 1969). A fifth species, B. resinifera (Regel) Britton, Alaska birch, occurs from central Canada to Alas- ka. This species resembles B. papyrifera, but it is smaller in stature, reaching a height of only about 12 m, and it differs in its more acuminate leaves and in details of the shape of its infructescence-scale lobes. Two additional species with white bark, Betula caerulea Blanch., blue birch, and B. caerulea-grandis Blanch., big blue birch, have been a source of confusion and controversy. Betula caerulea-grandis, which occurs from southern Quebec to Nova Scotia and in adjacent areas of New England and New York, resembles B. papyrifera in size, bark morphology, and general aspect, but its leaves are glabrous with more extended apices and more rounded or strongly cuneate bases, and its infructescence scales, like those of B. populifolia, have a short central lobe. Betula caerulea is similar in habit but smaller, reaching a height of only 8 or 9 m, and it has somewhat smaller and more sharply cuneate leaves. Sargent (1922) suggested that both of these forms are hybrids of B. papyrifera and B. populifolia, while Fernald (1922) concluded that B. caerulea-grandis was a “good” species, and that B. caerulea represented a hybrid between it and B. populifolia (cf. Fernald, 1950a). From a morphological analysis of the com- plex, Brayshaw (1966b) found that B. caerulea and B. caerulea-grandis fall between B. populifolia and B. papyrifera in many characters and concluded that the blue birches represent extremes of a hybrid swarm between those species. A paper and thin-layer chromatographic analysis of the northeastern white-barked birches by Koshy and colleagues demonstrated close flavonoid 38 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 relationships among B. caerulea, B. populifolia, B. caerulea-grandis, B. cor- difolia, and B. papyrifera and showed patterns supporting this conclusion. However, Brittain & Grant (1967a), Grant & Thompson, and Guerriero and co-workers, in further morphological and cytological studies, concluded that B. caerulea and B. caerulea-grandis represent hybrids between B. populifolia and B. cordifolia. Dehond & Campbell’s recent multivariate analysis of a single community in Maine containing B. papyrifera, B. cordifolia, B. populifolia, and B. caerulea-grandis suggested that B. caerulea-grandis represents a hybrid be- tween B. populifolia and B. cordifolia, with B. papyrifera apparently not entering into the hybridization. These results have been substantiated by a study of betulin content in the bark of trees of the same population (O’Connell et al.). Section Humites W. D. Koch (subsect. Nanae (Regel) Winkler), the “dwarf birches,” are shrubs of the cold circumpolar region that are characterized by small, rounded leaves with few veins and by staminate catkins that are borne laterally and (usually) singly, enclosed in buds during the winter prior to an- thesis. The usually solitary carpellate catkins emerge with new growth from the apical buds of short shoots. Betula pumila L., bog birch, an upright spreading shrub to ca. 4 m in height with leaves to ca. 7 cm long, is a common and variable species throughout bogs and fens of cool northeastern North America. A scarcely distinct more northern variety, B. pumila var. glandulifera Regel (B. glandulifera (Regel) Dugle), is marked by pubescent, somewhat gland-dotted branchlets and often smaller, more glandular leaves. This variety occurs from Newfoundland to the Yukon, extending southward in the western mountains to Oregon. A second dwarf species, B. glandulosa Mich., is found from Green- land and Labrador to western Canada and south in the Rocky Mountains. It is distinguished by its much smaller (to 3 cm long) leaves, its stems that are warty with large resinous glands, and its ascending lateral infructescence-scale lobes. This species reaches its southernmost limit in the East on the summits of high peaks, including Mount Washington (New Hampshire) and Mount Marcy (New York). A third member of this group, B. nana L., usually a prostrate shrub with tiny leaves, is circumpolar across the high latitudes of Europe, Asia, and North America. A similar species, B. Michauxii Spach, occurs in Nova Scotia and Newfoundland. It differs from B. nana primarily in the shape of the infructescence bracts (often lacking the side lobes) and in its wingless fruits (Fernald, 1950b; Rousseau & Raymond). Ina preliminary multivariate analysis of these species, Furlow (1984) found that B. Michauxii differs very little from B. nana and concluded that it did not deserve specific status. Betula rupestris, an intriguing birch apparently related to this complex, was described by Ra- finesque in 1819 (p. 229) from northern Kentucky on “‘the cliffs and on the sandstone rocks of the Kentucky river in Estill County.” Although no specimen of this record exists today, Rafinesque’s description agrees almost perfectly with that of B. pumila, not presently known farther south than central Ohio. Birches and alders share many features, but they are easily distinguished by the bracts of their infructescences, which are three-lobed and deciduous in Betula and five-lobed and persistent in A/nus. In vegetative morphology, in- cluding the structure of their leaves, buds, shoots, and bark, and their broad- winged fruits, species of Betula resemble those of A/nus subg. ALNOBETULA. 1990] FURLOW, BETULACEAE 39 However, the leaves are distinct in that they lack uniseriate-stalked glandular trichomes (Hardin & Bell). Like the alders, some birches have cut-leaved forms (Hylander, 1957b). In pollen morphology the two genera are distinct, with grains of A/nus normally bearing four or five apertures and those of Betula most frequently having three and lacking the prominent arci characteristic of alder pollen. Overall, the genus is much more homogeneous morphologically than A/nus is. The birches are a difficult group taxonomically because of their high vege- tative variability and frequent hybridization. Particular confusion has centered around the variable white-barked birches of the circumpolar Betula alba com- plex, the North American representatives being considered geographic races of a single species, B. pubescens Ehrh. (B. alba L.) (e.g., by Fernald, 1902), or separate species or hybrids (see Fernald, 1945b; Grant & Thompson; Hitch- cock; Hultén). The response of several American authors (e.g., Britton; Butler) to the observed diversity was to name numerous new species (see Dugle, 1966). Others (e.g., Fernald, 1902) have recognized the American forms as varieties of the European species. Gleason and Gleason & Cronquist (1963) suggested that B. papyrifera and B. pubescens might better be considered parts of a single circumboreal species, but they, as well as Fernald (1950a) and most other modern authors, have maintained the American plants as separate species. Recent cytological research has begun to elucidate some of the subtle relation- ships of American representatives of the complex (see the review of Dugle, 1966). Several of the races differ in chromosome number and on the basis of meiotic irregularities (Woodworth, 1931) are possibly of hybrid origin. How- ever, the American B. papyrifera is interfertile with both of the European white- barked species, B. pubescens Ehrh. (8. alba) and B. pendula Roth (B. verrucosa Ehrh.), even though a sterility barrier exists between the two European species (Johnsson, 1949). Many morphological and cytological studies have dealt with variation within and among separate and (mostly) mixed populations of the European white- barked birches, Betula pubescens and B. pendula. The most comprehensive reviews of this work are those of Natho (1959, 1964). Jentys-Szaferowa (1949, 1950, 1952), using simple statistical and graphic methods, analyzed morpho- logical variation in Polish populations, and Gardiner & Jeffers and Gardiner & Pearce, employing multivariate statistics, examined leaf-shape variation in populations in Scotland. These studies have shown both species to be extremely variable and suggest that they hybridize whenever they occur together. Jentys- Szaferowa (1950) noted that, because of this high variability, B. pubescens and B. pendula cannot be separated on the basis of any single character, but that each one is held together on the basis of combinations of characters and rep- resents a natural group. The cytogenetics of the European white-barked birches has been studied extensively (see the reviews of Brown & Al-Dawoodie, 1979; Gardiner, 1984; and Johnsson, 1974). Helms & Jorgensen first pointed out that the chromosome number of Betula pubescens is 2n = 56, while that of B. pendula is 2n = 28. This fact was discussed further by Woodworth (1931), and Johnsson (1945) noted the presence of nonchromosomal sterility barriers between the species. 40 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 Nevertheless, triploid (2n = 42) plants have been widely reported in mixed populations (see Brown & Al-Dawoodie, 1977, 1979; Brown & Williams; Eifler, 1956, 1958; Gardiner & Pearce; Helms & Jorgensen). Lindquist has noted that some of the triploid plants described by Helms & Jérgensen lack intermediate characters and thus might represent autopolyploids. A result of some of this work has been the formal recognition (e.g., by Gunnarsson) of many simple variants and putative hybrids as separate species or varieties. In a more moderate treatment of the European birches, Lindquist consolidated the variants of Betula pendula into three varieties that correspond with major phytogeographic regions. From a study of B. pubescens in relation to the subspecies recognized in Flora Europaea (Walters), Gardiner (1984) concluded that two main races of that species occur in Europe, an arctic and southern montane group corresponding to subsp. tortuosa (Ledeb.) Nyman, and a common lowland form corresponding to subsp. pubescens. A third sub- species recognized in Flora Europaea, subsp. carpathica (Willd.) Ascherson & Graebner, differs little from subsp. tortuosa. The taxonomy of the Betula alba complex has long been complicated by disagreement over the correct name of B. alba itself, as well as those of the other white-barked birches with which it occurs. According to Winkler (1904, 1930), Linnaeus circumscribed B. a/ba in such a way that he included both of the white-barked species of northern and central Europe in his concept (al- though Linnaeus’s name “‘Betula foliis ovatis acuminatis serratis” and the listed synonym from Flora Lapponica, “Betula foliis cordatis serratis,” together with his herbarium material, seem in fact to apply well to only one element of the complex). The major components of B. alba were separated by Roth as B. alba and B. pendula Roth. However, European authors have since mostly used the later name B. pubescens Ehrh. for the species with pubescent leaves and upright branches (B. alba as interpreted by Roth) (see Fernald, 1945b, p. 309, who condemned “the very doubtful Germanic practice of rejecting all Linnaean names of European species if they included what are now considered two or more species ...”). Many nineteenth-century workers (e.g., W. D. J. Koch; Lamarck & De Candolle) at the same time incorrectly applied the name B. alba to what should have been called B. pendula, and many (although not all) recent systems substitute B. verrucosa Ehrh. for B. pendula (see Fernald, 1902, 1945b). Recent European authors have mostly used B. a/ba in the sense of a “collective species” or Grossart (a named species complex; cf. Natho, 1964: Winkler, 1930), ignoring its nomenclatural priority for one of the elements of that complex. In the present treatment, the name B. pubescens has been em- ployed to follow prevalent current usage, pending final clarification of the issue. Taxonomic confusion also exists with regard to the dwarf birches (see Furlow, 1984; Lepage). The various North American taxa of this complex have been combined and split into a large number of species and infraspecific taxa. The most comprehensive recent analysis of these problems is found in the work of Dugle (1966), who studied the relationships and hybridization patterns among the various taxa occurring in western Canada. Using statistical analyses of morphological characters in combination with chromatographic and cytological procedures, Dugle recognized and described the variation patterns of two species 1990] FURLOW, BETULACEAE 4] (Betula glandulosa and B. glandulifera) and four hybrids of dwarf birches. Similar work is needed for the eastern American and European segments of the complex, followed by a comprehensive taxonomic revision of the entire group Several studies have been made of the vegetative variability of the three southeastern species. Coyle and colleagues (1983a) have described clinal vari- ation and population differentiation in Betula nigra based on measurements of leaf characters. Dancik and Dancik & Barnes (1971) have shown that the bark of B. alleghaniensis varies from light colored and exfoliating to dark and close in certain populations. Trees exhibiting the latter characters were at first thought to represent hybrids between B. alleghaniensis and B. lenta but after study were judged to be dark-barked variants of B. alleghaniensis. Further work (Dancik & Barnes, 1975; Sharik & Barnes, 1979) has shown that the two species vary considerably, both within and among populations (often more so within populations), but with discernible trends for many characters over latitudinal and altitudinal gradients. Wood of the dwarf northern birches, like that of the shrubby alders, exhibits primitive (juvenile) characters (e.g., many small vessels and numerous tra- cheids), while that of species of sects. CosrATAE and BETULA is more Sculie (Hall). The most specialized wood is present in members of sect. BETULASTER Regel (cf. Roskam). The staminate inflorescences of Betula are similar to those of Alnus except that they lack the two tertiary bracts subtending the cymules (Abbe, 1935, 1938, 1974). As in Alnus, the number of stamens and tepals in each flower differs among the species—i.e., generally three or four in members of sect. CosTATAE, two or three in sect. BETULA, and one or two in sect. HuMILEs. In sect. BETULASTER, the number of stamens has been reduced to two, but four tepals have been retained, a condition seen also in A/nus but not elsewhere in Betula (Abbe, 1935). The carpellate cymules of Betula differ from those of all other Betulaceae in that they usually retain all three flowers; the secondary one is absent in the other genera (Abbe, 1935). As in Alnus, various lines of Beuna ae Decome specialized 1 mn the grouping, number, and position of tl t, 19 Jager). The staminate catkins are produced the season before blooming in all sections except sect. HUMILES. The carpellate ones develop with the new growth in all sections. The number of both staminate and carpellate catkins in each cluster has been reduced from four or more in subg. BETULASTER to one in sect. Humites (cf. Jager). Accompanying this reduction are alterations in branching that place the staminate catkins (which occur near the ends of branches and above the carpellate clusters in subg. BETULASTER and sects. COSTATAE and BETULA of subg. BETULA) below the terminal carpellate ones on short shoots in sect. Humives (Jager). These changes parallel modifications, interpreted as adaptations to cold climates, seen in A/nus subg. ALNOBETULA (Furlow, 1979). Little chemosystematic work has been attempted with woody plants in gen- eral in comparison with herbaceous groups. However, a surprising number of studies have been undertaken in Betu/a. An early flavonoid study was con- ducted by K. E. Clausen (1960b) to identify hybridization between B. papyrifera 42 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 and B. pumila. Other work, in addition to that of Koshy and colleagues (de- scribed above), has included an extensive series of investigations of northern European birches by Pawlowska (1980a-c, 1982a, 1982b, 1983a, b) to dem- onstrate relationships of flavonoid occurrence among various species, species segregates, and putative hybrids. An electrophoretic analysis of pollen proteins in B. populifolia by Payne & Fairbrothers showed a high level of variation in proteins among populations in that species and suggested that local population differentiation was occurring. In a study of ten Betula species of the Soviet Far East, Baranov and co-workers found triterpene data to be taxonomically useful in the identification and separation of groups and subgroups of various species. Species of Betula form a polyploid series, with chromosome numbers of 27 = 28, 56, 70, 84, and 112, plus dysploid numbers in some hybrids (Woodworth, 1929b; Dugle, 1966; Poucques; Wetzel, 1927, 1928, 1929; Barnes & Dancik). Of the southeastern species, B. nigra and B. alleghaniensis are diploids (2n = 28), while B. alleghaniensis is hexaploid (2n = 84). Meiosis is normal in the diploids and somewhat abnormal in B. alleghaniensis (Woodworth, 1929b). Although the European B. pubescens (B. alba) and two of its northeastern American races, B. papyrifera and B. fontinalis, form a circumpolar complex having distributional and morphological patterns similar to those of Alnus incana, these birches represent different polyploidy levels (2n = 56, 70, and 84, respectively), lending support to their continued treatment as separate species. The European species, B. pendula Roth, and its American and Asian counterparts, B. populifolia and B. japonica Sieb., are diploids (2n = 28), but these forms are more differentiated morphologically than are the members of the B. pubescens group. Consequently, there has been little tendency to treat them as conspecific. In both complexes the segments should be examined in relation to modern species concepts to determine whether they might better be treated as a single species. The birches are anemophilous and produce large quantities of pollen (Wode- house, 1935). In all subgroups of the genus, the carpellate catkins appear with the new growth, and anthesis occurs as the leaves unfold. Achenes are produced in large numbers and are carried for considerable distances by the wind. Natural hybridization is common (Alam & Grant; Johnsson, 1945), and many of the resulting hybrids have been named (see L. P. V. Johnson: Kartesz & Kartesz; Winkler, 1904). In eastern North America Betula alleghaniensis and B. lenta have been shown to hybridize (Sharik & Barnes, 1971). These species, as well as B. papyrifera and B. populifolia, also hybridize with B. pumila in the North where their ranges overlap (Cousins). Through artificial crosses hybridization between B. papyrifera and B. populifolia has been studied by Alam & Grant, who found the progeny to resemble B. papyrifera more closely than B. populifolia in juvenile leaf characters. Seeds from single trees of various taxa of Betula often give rise to offspring of two or more ploidy levels, and there “appears to be little barrier to cross fertilization between Betula plants with different levels of polyploidy” (Grant, 1969, p. 81). He suggested that this feature may have permitted the genus to take advantage of new ecological niches that opened up following the Pleistocene. The earliest pollen and leaf material of Betula is from the Upper Cretaceous, 1990] FURLOW, BETULACEAE 43 and fossils of Betula are widespread and highly diversified by the Middle Eocene (Crane & Stockey). Differentiation of subgenera and sections appears to have occurred largely in response to major climatic differences (Jager; Kikuzawa). A cladistic analysis of the birches by Roskam, done as part of a study of coevolutionary patterns in the birches and their gall-midge parasites, indicated that sect. CosTATAE is the most plesiomorphic subgroup and the sister group of Alnus subg. ALNOBETULA, with which the tree birches share many characters. However, it seems unlikely that Betula sect. CosTATAE is closely related to Alnus subg. ALNOBETULA. If the shrubby growth form indeed represents ad- aptation to cold climates, as is strongly suggested by morphological and phy- togeographic patterns in both genera, it is improbable that unspecialized birches could have evolved directly from one of the most highly specialized groups of alders, or vice versa. In a preliminary cladistic analysis by Furlow (1983), all of A/nus and all of Betula appear as sister groups. Within Betula, sect. COSTATAE is most closely related to sect. BETULASTER. Birches serve as important sources of food for browsing animals (LeResche & Davis; Oldemeyer). Palo, Pehrson, & Knutsson and Palo, Sunnerheim, & Theander have shown that phenolic compounds become much more concen- trated in the twigs and bark of white-barked birches during the winter and have correlated this fact with striking examples of weight loss and reduced food consumption in vertebrate herbivores feeding on birch twigs and branches in the winter. They suggested that phenolic compounds may constitute a major chemical defense in birches against browsing animals. In areas recently exposed by logging or natural causes, Betula papyrifera and B. alleghaniensis often exhibit symptoms of distress and gradually die from the top downward. This occurrence, known as ‘“‘decadence,”’ has generally been attributed to suddenly changing environmental conditions. During the 1930's, a disease with similar symptoms (termed “dieback”) appeared in New Bruns- wick and rapidly spread throughout the Northeast, although there had been no alteration of the surrounding forest. By 1905 at least 80 percent of the mer- chantable birch had been killed in the Maritime region of Canada to as far west as New Hampshire (Clark; Clark & Barter). Dieback, which affects B. alleghaniensis more severely than B. populifolia, and which in Europe and in ornamental plantings also affects especially B. pendula, has since spread west- ward through New York, Ontario, Michigan, and Minnesota. Various expla- nations have been proposed, but thus far no single climatic or biotic cause has been identified. It has been suggested by Ball & Simmons and Houston that a population is first weakened by adverse environmental conditions, such as a period of drought, then is invaded and eventually decimated by the bronze birch borer, Agrilus anxius Gory, a native buprestid beetle. Others (e.g., Berbee; Cooper & Massalsk) believe, on the basis of the nature of dieback symptoms, that the initial causative agent of decline may be a virus, with borer invasion following in the weakened trees. Many insect species feed on or otherwise affect birches, the most detrimental being the bronze birch borer. Others include the gypsy moth, tent caterpillars, leaf miners, and scale insects. Fungal diseases result in the destruction of large numbers of trees and are therefore of considerable economic consequence 44 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 (Hepting). The most important of these include heartwood rots caused by various species of Fomes and Poria, especially Fomes ignarius (L.) Kickx (Bash- am & Morawski; Campbell & Davidson), and nectria canker (Nectria galligena Bres.), the most damaging external stem disease of yellow birch and paper birch (Hepting). The pollen of birches, in regions where they are numerous, is a significant cause of hay fever (Lewis & Imber; Lewis er a/.; Lowenstein et al.; Wodehouse, 1945). During the past two decades considerable progress has been made in Europe to elucidate the basis of this ailment (Apold et a/.; Berlin: Vik & Elsayed; Vik et al.). This research has revealed that the allergenic reaction to Betula pollen is related to that caused by 4/nus and Corylus pollen. The responsible allergens have been partly identified (Dalen & Voorhorst; Lowenstein ef al.). It has been shown that in children, birch-pollen allergies are sometimes related to food allergies (Dreborg & Foucard; Halmepuro et a/.: Lahti er al.; Lowenstein & Eriksson). Birch sap has also been shown to cause a contact dermatitis in persons sensitive to birch pollen (Lahti & Hannuksela). The wood of the birches has many uses (reviewed by Lines). In eastern North America Betula alleghaniensis and B. lenta are important sources of hardwood timber employed in the manufacture of doors and windows, flooring, cabinetry, interior molding, wood paneling, barrels, shoe lasts, furniture, and plywood. These and other species, especially B. papyrifera, are also widely used for making small specialty products, including wooden toys, athletic equipment, broom handles, clothespins, ice-cream sticks, spools, bobbins, and toothpicks. Wood of various species has long been utilized to make charcoal for gunpowder and for filtration purposes. In the northern Appalachians Betula /enta is sometimes tapped during the spring in the same way sugar maple trees are tapped, the collected sap being fermented to produce a naturally carbonated birch beer. A tea is made from the bark and twigs of this species by steeping them in hot water (Fernald et al.; Sargent, 1896). Betula lenta was formerly a major commercial source of methyl salicylate, the chief constituent of wintergreen oil, widely used as a flavoring and as a component of pharmaceuticals, including aromatic cascara sagrada fluid extract (sweet cascara). Its chief medicinal use has been as a rubefacient and, in the past, as an antirheumatic (Lewis & Elvin-Lewis). Today, methyl salicylate is largely produced synthetically. Infusions of the bark of various birch species were widely used by North American Indians as treat- ments for infections, colds, pulmonary problems, burns, leukorrhea, and other ailments (Lewis & Elvin-Lewis; Moerman). Twigs of Betula lenta and B. al- leghaniensis have been used in modern times as chewing sticks for cleaning the teeth (Lewis & Elvin-Lewis). A pyroligneous oil is obtained by distillation from the bark and wood of Betula pendula and other species. This material has been widely used in north- eastern Europe in the preparation of leather and in the manufacture of lotions, ointments, and medicines. Birch wood is a common source of high-quality firewood and of pulp for manufacturing paper in regions where the trees are plentiful. Birch bark, rich in oil and starch, has been used for centuries by people in times of famine as a source of food. Many species, especially those 1990] FURLOW, BETULACEAE 45 with white bark (B. pendula, B. pubescens, RB. populifolia, and B. ele as well as B. nigra, are utilized horticulturally (of these, B. pendula is by fa the most widely used in the United States). The bark of B. papyrifera, ee is waterproof and easily workable because of its betulin and oil content, was extensively employed by northern North American Indians as a covering ma- terial for canoes, houses, and bundles, and as a material for making various articles of clothing. REFERENCES: Under family references see ABBE (1935, 1938, 1974); BRITTON; BRITTON & BROWN (1913); I. R. BRown & AL-DAWOODIE (1979); CRANE & STOCKEY; DALEN & VOORHORST; ENDLICHER (1842, 1847); FERNALD (1950a); FERNALD ef al.; FUuRLow (1979, 1983a); GLEASON; GLEASON & CRONQUIST; HALL; HARDIN & BELL; HEPTING; HJELMQvisT (1948); JAGER; KarTeEsz & KARTESZ; KikuZAWA; W. D. J. KocH; KOEHNE; LEwis & ELvIN-LeEwIs; Lewis et al.; LINNAEUS; LOWENSTEIN ef al.; METCALFE & CHALK; MICHAUX; MOERMAN; PRANTL; RADFORD; REGEL (1861, 1868); Scoc GGAN; SPACH (1841); WALTERS: WETZEL an 1928, 1929); WINKLER (1904); WoDEHOUSE (1935, 1945); and W. (1929b, poly: AcaAM, M. T., & W. F. Grant. Interspecific hybridization in birch (Betula). Nat. Canad. Rake te ae D, J., E. VAAG, & S. ELSAYED. Comparative studies on tree pollen allergens. “tL ie and partial characterization of a major allergen from birch pollen, Betula verrucosa. Int. Arch. Allergy Appl. Immunol. 64: 439-447. 1981. BALL, J., & G. Simmons. The ete te between bronze birch borer and birch dieback. Jour. Arboric. 6: 4 314. 1980. BarRANoV, VY. I., G. MALINovskKalIA, N. D. PokHILo, V. V. MAKHAN’KOV, N. I. OuvaAROVA, & P a Gorovoy. Khemotaksonomicheskoie issledovanie vidov Bet- ula L. oases Dal’nego Vostoka. Rast. Res. 19: 159-166. 1984. [Reviewed in Taxon 33: 359. BARNES, B. V., & B. : DANCIK. oct ee aus origin of a new birch species, Betula Murrayana, from southeastern Michigan. Canad. Jour. Bot. 63: 223-226. 1985. , & T. L. SHARIK. Natural eRe of yellow birch and paper birch. Forest Sci. 20: 215-221. 1974. BASHAM, J. T., & Z. J. R. MoRAwSKI. Cull studies, the ee and associated basidio- mycete fungi i in the heart wood of living trees in the forests of Ontario. Canad. Dep. Forest. Publ. 1079. 69 pp. Ottawa. 1964. [Photographs Sere fungal damage to birch woods. ] BERBEE, J. G. Virus symptoms associated with birch dieback. Canad. Dep. Agr. Sci. Serv. Forest Biol. Div. Bimonthly Progress Rep. 13(1): 1. 1957. BERLIN, C. Immunological analysis of birch pollen antigens, “with special reference to the allergenic components. Int. Arch. Allergy Appl. Immunol. 42: 300-322. 1972. BEVINGTON, J. Geographic differences in the seed germination of paper birch, Betula papyrifera. Am, Jour. Bot. 73: 564-573. 1986. BJORKSTEN, F., I. SUONIEMI, & V. Koski. Neo-natal birch ai contact and subsequent allergy to birch pollen. Clin. Allergy 10: 585-592. 19 BLANCHARD, W. H. New white birches. Betula 1: 1. 1904. BLANCHETTE, R. A. Progressive stages of discoloration and decay associated with the canker rot fungus, Jnonotus obliquus, in birch, Betula papyrifera. Phytopathology 72: 1272-1277. 1982. BRAYSHAW, T. C. The names of yellow birch and two of its varieties. Canad. Field-Nat. 80: 160-161. 1966a. [Status of the name Betula lutea Michx. f.] . What are the blue birches? /bid. 187-194. 1966b. 46 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 BRINKMAN, K. A. Betula L. Birch. Pp. 252-257 in C. S. ScHOPMEYER, ed., Seeds of woody plants in the United States. U. S. Dep. Agr. Agr. Handb. 450. Washington, D. C. 1974. [Techniques for propagating Betula from seed.] BRITTAIN, W. H., & W. F. Grant. Observations on Canadian birch (Betula) collections at the Morgan Arboretum. I. Bp papyrifera in eastern Canada. Canad. Field-Nat 189-197. 1965a; II. B. papyrifera var. oo. Ibid. 253-257. 1965b; Il. B. papyrifera from British Columbia. [bid. 80: 147- 157. IV. B. caerulea- grandis and hybrids. /bid. 81: 116-127. 1967a; V es eastern Canada. Ibid. 251-262. 1967b; VI. B. papyrifera from the ae Mountains. [bid. 82: 44-48. 1968a; VII. B. papyrifera and B. resinifera from northwestern Canada. Ibid. 82: 185- ae as VUI. Betula from Grand Manan Island, New Brunswick. Thid. 361-383. & : oo ee on the Betula caerulea complex. Nat. Canad. 98: 48- 58. 1971. Brizicky, G. K. Subgeneric and sectional names: their ene. aa and early sources. Taxon 18: 643-660. 1969. [Names attributed to ENDLICH Brown, I. R., & D. M. AL- DAWoopIE. Cytotype diversity in a ! population of Betula alba L. New Phytol. 79: 441-453. 1977. [Chromosome numbers of 1 = 28, 42, and 56 in a mixed population of B. pubescens and B. pendula.] Prag Cytology of Betula alba L. complex. Proc. Roy. Soc. Edinb. B. 85: 49- 64. BuTLeER, B. T. es American birches. Bull. Torrey Bot. Club 36: 421-440. 1909. [Recognized 17 species. Caesar, J. C., & A.D. MAcDona.p. Shoot development in Betula papyrifera 2. Com- parison of vegetative and reproductive short-shoot growth. Canad. Jour. Bot. 61: 3066-3071. 1983; 4. Comparisons between growth characteristics and expression of vegetative long and short shoots. /bid. 62: 446-453. 1984a: 5. Effect of male inflorescence formation and flowering on long shoot development. Ibid. 1708-1713. 1984b CAMPBELL, W. A., & R. W. Davipson. Cankers and decay of yellow oe associated wit mes ignarius var. laevigatus. Jour. Forestry 39: 559, 560. 194 CLarRK, J. Birch dieback. Pp. 1551-1555 in Recent advances in botan — lectures and symposia presented to the IX International Botanical Congress, Montreal, 1959. Toronto. 196 . W. BARTER. ie and climate in relation to dieback of yellow birch. Forest Sci. 4: 343-364. 1958. CLAUSEN, J. J., & T. T. Kosiowsx Heterophyllous shoots in Betula papyrifera. Nature 205: 1030, 1031. CLAUSEN, K. E. A survey variation in pollen size within individual plants and catkins of three taxa of Betula. Pollen Spores 2: 299-304. 1960a. The a hybrid of paper birch and bog birch. Proc. Minn. Acad. Sci. 25/ 26: 98-100. Ob. ee of a hybrid birch and its parent species. Canad. Jour. Bot. 41: 441-458. 1963. Cooper, J. 1., & P. R. MASSALSK. Viruses and virus-like diseases affecting Betula spp. Proc. Roy. Soc. Edinb. 85: 183. 1984. Cousins, S. M. The comparative anatomy of the stems of Betula pumila, Betula lenta, and the hybrid Betula Jackii. Jour. Arnold Arb. 14: 351-355. 1933. fer B. F., T. L. SHARIK, & P. P. Ferner. Variation in leaf morphology among disjunct and continuous populations of river birch, Betula nigra. Silvae Genet. 31: 122-125. 1983a. ,& . The utility of range-wide maps for identifying disjunct pop- aetione Ai river birch (Betula nigra L.). Castanea 48: 285-288. 1983b. CriBBEN, L. D., & I. A. UNGAR. River birch (Betula ais ria communities of south- eastern Ohio. Ohio Biol. Surv. Biol. Notes 8: 1-37. 1990] FURLOW, BETULACEAE 47 — M., J. W. HiccrnsoTHaM, & C. R. Parks. Morphological and aa ariation over an elevational gradient in rae eens an U.S. A. (eta) An Jour. Bot. 71(5, suppl. 2): 162. Dancik, B. P. Dark-barked birches of southern ae. Mich. Bot. 8: 38-41. 1969. & B. V. BARNES. Meer in bark morphology of yellow birch in an even- aged stand. Ibid. 10: 34-38. 197 & ———. Leaf variability in otis birch (Betula oT in relation to environment. Canad. Jour. Forestry Res. 5: 149-159. 1975 Davy, A. J., & J. A. Git. Variation due to environment and heredity in birch trans- planted between heath and bog. New Phytol. 97: 489-506. DEHOND, P. A., & C. S. CAMPBELL. Natural hybridization aii Betula sea Cee and B. populifolia Voges in Maine. (Abstract.) Am. Jour. Bot. 74: 7 987 Downs, R. J., & J. M. Bevincton. Effect of temperature and ese on growth and dormancy of Betula papyrifera. Am. Jour. Bot. 68: 795-800. 1981. DresorG, S., & T. FoucarD. Allergy to apple, carrot, and potato in children with birch pollen allergy. Allergy 38: 167-172. 1983. [Skin sensitivity. ] Dua te, J. R. taxonomic study of western Canadian species in the genus Betula. Canad. Jour. Bot. 44: 929-1007. 1966. ome sae eas problems in North American Betula. Canad. Field-Nat. 83: 251-253. EIFLER, I. Arreurange der Betula. Ziichter 26: 342-346. 1956. zungen zwischen Betula verrucosa und B. se ene Ibid. 28: 331-336. ELKINGTON, T. T. Introgressive hybridization between Betula nana L. and B. pubescens Ebrh. in north-west Ic eland. New Phytol. 67: 109-118. 1968 FERNALD, M. ps of some American and Old World birches. Am. Jour. Sci. 169: ion 194. 1902. . Notes on the flora of Nova Scotia. Rhodora 24: 165-180. 1922. [Predicted Betula caerulea is a hybrid between B. caerulea-grandis and B. populifolia.] ome North American Corylaceae (Betulaceae). I. Notes on Betula in eastern North America. Ibid. 47: 303-329. 1945b Betula Michauxii, a brief symposium. |. Introductory note. Ibid. 52: 25-27. 1950b. FonTAINE, F. J. The genus Betula (contribution to a monograph). (In Dutch.) Misc. Pap. Landbouwhogeschool Wageningen 6: 99-180. 1970. Fritts, H. C., & B. J. KirtLanp. The distribution of river birch in Cumberland County, Illinois. Trans. Illinois Acad. Sci. 53: 68-70. 1960. Furow, J. J. The evolution and eae of the Betula nana complex. (Abstract.) Am. Jour. Bot. 71(5, suppl. 2): 166. Garpiner, A. S. A biometric study of oe tee in some British birch populations. Forestry 45: 37-47. 1972. . Taxonomy of infraspecific variation in Betula pubescens Ehrh., with particular references to the Scottish Highlands. Proc. Roy. Soc. Edinb. B. 85: 13-26. 1984. _R. Jerrers. Analysis of the collective species Betula alba L. on the basis of leaf igs eas Silvae Genet. 11: 156-1 N. J. RcE. Leaf ae as an indicator of introgression between Betula pendula and B pubescens. Trans. Bot. Soc. Edinb. 43: 91-103. 1979. GitperT, A. M. Silvical aEeae of yellow birch. U. S. Forest Serv. N. E. Forest aN Sta. Pap. 134. 18 p Gut, J. A., & A. J. Davy. Varia tion and polyploidy within lowland populations of the Betula pendula/B. pubescens complex. New Phytol. 94: 433-451. 3. Gittrap, N. J. Hebeloma spp. as mycorrhizal associates in birch. Trans. Brit. Mycol. Soc. 79: 157-160. 1982. Grant, W. F. Decreased DNA content of birch (Betula) chromosomes at high ploidy as determined by cytophotometry. Chromosoma 26: 326-336. 1969. 48 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 ———.. The evolution of karyotype and polyploidy in arboreal plants. Taxon 25: 75- 84. 1976. [Plants with different euploid chromosome numbers obtained from seeds ofa ae parent tree.] & B. K. THompson. Observations on Canadian birches, Betula cordifolia, B. mae B. papyrifera, and B. x caerulea. Canad. Jour. Bot. 53: 1478-1490. 1975. GREENIDGE, K. N. H. Further studies of birch dieback in Nova Scotia. Canad. Jour. Bot. 31: 548-559. 1953. GUERRIERO, A. G., W. F. GRANT, & W. H. BritTAIN. Interspecific hybridization between Betula ae and B. populifolia at Valcartier, Quebec. Canad. Jour. Bot. 48: 2241-2247. 1970. GUNNARSSON, J. aC Monografi over Skandinaviens Betula. xi + 136 pp. Malmé. 1925. Haxkansson, A. Notes on endosperm formation in Betula. Bot. Not. 110: 201-204. 1957. HALMEPuRO, L., K. VUONTELA, K. KALIMO, & F. BIORKSTEN. Cross reactivity of immuno- globulin antibodies with allergens in birch pollen, fruits, and vegetables. Int. Arch. Allergy Appl. Immunol. 74: 235-240. 1984. [Allergy to fruits and vegetables in persons suffering from birch-pollen allergy.] HAwso_pt, L. S. Aspects of yellow birch dieback in Nova Scotia. Jour. Forestry 45: Haypen, W. J., & S. M. Haypen. Wood anatomy and relationships of Betula uber. Castanea 49: oe 30. 1984. Hevms, A., & C. A. JoRGENSEN. Birkene paa Maglemose. Bot. Tidsskr. 39: 57-133. 7 rn, C.L. Betula. Pp. 76-83 in C. L. Hitcucock, A. CRONQUIST, M. OWNBE & Ws TROMESON, cular plants of the Pacific Northwest. Vol. 2. Seattle. 1964. Houston, D.R. Die d by stress, including defoliation. Arb. News ae 73-77. 1973. Hu ten, E. Flora of Alaska and Yukon. Lunds Univ. Arsskr. II. Sect. 2. 40(1): 1-1902. 1940-1950. [Betula, 572-585. 1944 HuTnik, R. J., & F. E. CUNNINGHAM. Silvical characteristics of paper birch. U. S. Forest Serv. N. E. Forest Exper. Sta. Pap. 141. 1961. Hy tanper, N. On cut-leaved and small-leaved forms of Scandinavian birches. Sv. Bot. Tidskr. 51: eee 1957b. Ipsen, H., & H. Lowenstern. Isolation and immunochemical characterization of the major cs of birch pollen, Betula verrucosa. Jour. Allergy Clin. Immunol. 72: 150-159. 1983. Ivarson, K. C., & H. KATZNELSON. Studies on the rhizosphere microflora of yellow birch seedlings. Pl. Soil 12: 30-40. 1960. JENTYS-SZAFEROWA, J. Analysis of the collective see Betula alba L. on the basis of leaf measurements. I: Aim and method of the work on the example of Betula verrucosa Ehrh. Bull. Acad. ee Sci. Lett. a Math. Nat. BI. 1949: 175-214. 1949; I: Betula pubescens Ehrh., B. tortuosa Ledeb., B. carpatica Waldst. et Kit. Ibid. 1950: 1-63. 1950; III. i oycoviensis Bess. and Betula obscura Kotula. Determination on the basis of a single leaf. Zbid. 1951: 1-40. 1952. Jounson, A. G. Betula lenta var. uber Ashe. Rhodora 56: 129-131. 1954. JoHNsoN, L. P. V. A descriptive list of natural and artificial interspecific hybrids in North American forest-tree genera. Canad. Jour. Res. C. 17: 411-444. 1939, [Betula, hybrids. ] JOHNssON, H. Interspecific hybridization within the genus Betula. Hereditas 31: 163- 176. 1945, . Studies on birch species hybrids. I. Betula verrucosa x B. japonica, B. verrucosa x B. papyrifera, and B. pubescens x B. papyrifera. Ibid. 35: 115-135. 1949. [Artificial hybrids. ] Genetic characteristics of Betula verrucosa Ehrh. and B. pubescens Ehrh. Ann. Forestry 6: 91-127. 1974. 1990] FURLOW, BETULACEAE 49 JosepH, H. C. Germination and vitality of birch seeds. Bot. Gaz. 87: 127-151. 1929. Kawase, M. Dormancy in Betu/a as a quantitative state. Pl. Physiol. 36: 643-649. 1961. KINKEAD, E. Our footloose correspondents: the search for Betula uber. New Yorker 51(47): 58-69. 1976. [Popular account of the search for B. uber, cf. OGLE & Mazzeo. ] KirKPATRICK, M. Spatial and age dependent patterns of growth in New England, U.S. A., black birch, Betula lenta. Am. Jour. Bot. 68: 535-545. 1981. Kocu, K. H. E. Dendrologie. Vol. 2, part 1. i + 665 pp. Erlangen. 1872. [Betulaceae subfam. Betuleae, 622-662.] Ibid. Vol. 2, part 2. 1 + 424 pp. Erlangen. 1873. [Cupuliferae subfam. Coryleae, 1-1 KoeveniG, J. B. Distribution of river birch, Betula nigra, in the United States. Florida Sci. 38: 13-19. 1975. [Northern distributional limit coincides with the Wisconsin glacial ea ] Kosny, T. K., W. F. Grant, & W. H. Brittain. Numerical chemotaxonomy of the Betula cere Rana ee Symp. Biol. Hungar. 12: 201-211. 1972. KUuzeEneEVA, O. I. tula L. Fl. URSS. 5: 269-305. 1936. Lantl, A., F. aera & M. HANNUKSELA. Allergy to birch pollen and apple cross reactivity of the allergens studied with the radio allergo sorbent test. ee 35: 297-300. 1980. HANNUKSELA. Immediate contact allergy to birch (Betula verrucosa) leaves and sap. Contact Dermatitis 6: 464, LAMARCK, J. B. A. P. M. De, & A. P. D CANDOLLE. Flore francaise. ed. 3. Vol. 4. 111 + 944 pp. Paris. 1805. [A/nus, Betula, 304.] Leak, W. B. Silvical characteristics of the sweet birch (Betula lenta). U.S. Forest Serv. N. E. Forest Exper. Sta. Pap. 113. 14 pp. 1958. LepAGE, E. Les bouleaux arbustifs du Canada et de I’Alaska. Nat. Canad. 103: 215-233. 1976. [Eight species and 16 interspecific hybrid taxa.] LERESCHE, R. E., & J. L. Davis. Importance of nonbrowse Sars to moose on the Kenai Peninsula, Alaska. Jour. Wildl. Managem. 37: 279-287. Lewis, W. H., & W. E. Imper. Allergy epidemiology in the x ie Missouri, area. III. Trees. Allergy 35: 113-119. [Hay fever caused by Betula pollen.] Linpqutst, B. On the variation in Scandinavian Betula verrucosa Ehrh. with some notes on the Betula series Verrucosae Sukacz. Sv. Bot. Tidskr. 41: 45-71. 1947. Lines, R. Man’s use of birch—past and present. Proc. Roy. Soc. Edinb. B. 85: 203-213. 1984. Love, A., & D. LOvE. eee ar of the alpine plants of Mount Washington. Univ. Colorado Stud. Biol. 24: 1-74. 1966. [Betula, 30-32.] Lowenstein, H., & N. E. Eriksson. Hypersensitivity to foods among birch pollen allergic patients: immunochemical inhibition studies for evaluation of possible mechanisms. Allergy 38: 577-588. 1983. MacDOonaLb, A. D., & D. H. MoTHERSILL. Shoot development in wee papyrifera. |. Short shoot organogenesis. Canad. Jour. Bot. 61: 3049-3065. 1983. & J.C. CAESAR. Shoot yg ere in Betula papyrifera. Long shoot organ- ogenesis. per ee Bot. 62: 437-445. 1984. Mason, P. A., J. ON, & F. T. LAST. Mycorrhizal fungi of Betula spp.—factors affecting ane occurrence. Proc. Roy. Soc. Edinb. 85: 141-152. 1984. Mazzeo, P. M. Betula uber—what is it and where is it? Castanea 39: 273-278. 1974. McCLELLAND, M. K., & I. A. UNGAR. The influence of owe factors on Betula nigra L. distribution in southeastern Ohio. Castanea 35: 99-11 oa, G. Variationsbreite und Bastardbildung bei se Birkensippen. Feddes Repert. Sp. Nov. 61: 211-273. 1959. Status and problematics of Betula taxonomy in central Europe. Biol. Zentralbl. 83: 197-230. 1964. [Review of past and present taxonomic and cytological research involving the B. alba complex.] O’ConNELL, M. M., M. D. BENTLEY, C. S. CAMPBELL, & B. J. W. Cote. Betulin variation in four white-barked birches in Maine. (Abstract.) Am. Jour. Bot. 74: 703. 1987. 50 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 OctE, D. W., & P.M. Mazzeo. Betula uber, the Virginia round-leaf birch, rediscovered in southwest Virginia. Castanea 41: 248-256. 1976. OLpeMeYeER, J. L. Estimating production of paper birch and utilization by browsers. Canad. Jour. Sees Res. 12: 52-57. 1982. Pato, R. T., A. PEHRSON, & P. -G. Knutsson. Can birch phenolics be of importance in the defense ae browsing vertebrates? Finn. Game Res. 41: 75-80. 1983. [Phe- nolic content of birch twigs Habeas le weight loss and reduced food consump- tion in the mountain hare after brow , K. SUNNERHEIM, & O. cet i cone variation of phenols, crude protein and cell wall content of birch (Betula pendula Roth) in relation to ruminant in vitro digestibility. Oecologia a 314-318. 1985. [Phenolic content of birch twigs increases uring the winter seaso PAWLOWSKA, L. Flavonoids i in the leaves of Polish species of the genus Betula. 1. The flavonoids of Betula pendula and Betula obscura leaves. Acta Soc. Bot. Polon. 49: 281-296. 1980a; 2. The flavonoids of Betula nova and Betula humilis leaves. Ibid. 297-310. 1980b; 3. The flavonoids of Betula oycoviensis leaves. Ibid. 311-320. 1980c; 4. The flavonoids of Betula pubescens, Betula carpatica, Betula tortuosa, and Betula nana. Ibid. 51: 403-412. 1982a. 5. The taxonomic position on the basis of flavonoid composition. /bid. 415-422. 1982b. . Flavonoids from the leaves of some American species of the genus Betula L. Ibid. 52: 295 300. 1983a. . Biochemical and systematic study of the genus Betula. Ibid. 301-314. 1983b. PAYNE, R. C., & D. E. FAIRBROTHERS. Disc electrophoretic study of pollen proteins from natural populations of Betula populifolia in New Jersey. Am. Jour. Bot. 60: 182— 9, 1973. Poucques, H. L. Etudes caryologiques sur les Fagales. II. Le genre Betula. Bull. Soc. Sci. Nancy 8: 1-5. 1949. Preston, D. J. The rediscovery of Betula uber. Am. a 16-20. 1976. [Account of the rediscovery of B. uber; cf. OGLE & Mazz RAFINESQUE, C. 8S. Descriptions of two new shrubs Fae Kentucky, &c. West. Rev. Misc. Mag. 1: 228-230. 1819. Reep, C. F. Betula uber (Ashe) sales rediscovered in Virginia. Phytologia 32: 305- 311. 1975. [cf. OGLE & Mazz REGEL, E. Bemerkungen neers aes Betula und Alnus nebst Beschreibung einig REICHARDT, P. B., J. P. BRYANT. T. P. CLAUSEN, & G. D. WIELAND. Defense of winter dormant Alaska U.S.A. paper birch (Betula g , Lepus americanus. Oecologia 65: 58-69. 1985. RicG, G. B. Birch succession in sphagnum bogs. Jour. Forestry 20: 848-850. 1922. ROSENDAHL, C. O. Evidence of the hybrid nature of Betula Sandbergi. Rhodora 30: 8. RoskaM, J. C. Biosystematics of insects living in female birch catkins. II. Inquiline and predaceous a midges belonging to various genera. Netherlands Jour. Zool. 29: 283-351. Rotu, A. W. tamen florae germanicae. Vol. 1. xvi + 560 pp. Leipzig. 1788. [Betula alba, B. pendula, 404, 405. Rot, P. L. Phe eoreae variation in river birch (Betula nigra L.). Proc. Indiana Acad. Sci. 80: 225-229. 1971. RoussEAu, J., & M. ecu Betula Michauxii, a brief symposium. 2. Betula Mi- chauxii in northeastern America. Rhodora 52: 27-32. 1950. [See FERNALD, 1950b.] SARGENT, C. S. Betula. Silva N. Am. 9: 45-66. 1896; 14: 53-60, 104. 1902. SCHNEIDER, C. K. Illustriertes Handbuch der Laubholzkunde. Vol. 1. 810 pp. Jena. 1904-1906. [Betulaceae, 96-150, 1904.) SHARIK, T. L., & B. V. BARNES. Hybridization in Betula alleghaniensis Britton and B. lenta L.: a comparative ha of controlled crosses. Forest Sci. 17: 415-424. 1971. 1990] FURLOW, BETULACEAE 51 aes ogy of shoot growth among di ti f yellow birch ag emis and sweet birch (B. lenta). Canad. rae Bot. 54: 2122-2129. ural variation in mor phol 0 ie (Betula eho and sweet birch (B. maa Ibid. 57: 1932- 1939. 1979. Variation and taxonomy of Betula uber, B. lenta, and B. alle- ae Sree 36: 307-316. 1984. SoLomon, D. S., & K. W. KENLAN. Discriminant analysis of interspecific hybridization in Betula. Silvae Genet. 31: 136-149. 1983. Taper, L. J., & W. F. Grant. The relationship between chromosome size and DNA content in birch (Betula) species. Caryologia 26: 263-273. 1973. TUCKERMAN, E. Observations on some oo plants of New England. Am. Jour. Sci. 45: 27-49. 1843. [B. minor, sp. no VAARTAJA, O. Photoperiodic response in a aation of seed of certain trees. Canad. Jour. Bot. 34: 377-388. 1956. [Germination of Betula pubescens and B. pendula seeds requires light and low to moderate temperatures. otoperiodic responses in seedlings of northern tree species. [bid. 35: 133- 138. 1957. lege of Betula alleghaniensis affected by photoperiod.] Vik, H., & S. ELSAYED. Raye studies on tree pollen allergens. XIII. Further fee and N-terminal amino acid sequence analysis of the major allergen of birch pollen (Betula me. Int. Arch. Allergy Appl. Immunol. 80: 17-25. 1986. art a testing and characterization of the allergen isolated in Vik et al., J ——, p, & B.S. PAULSEN. Comparative studies on tree pollen allergens. IIT. A purified cg pene weight major allergen from birch pollen a ver- rucosa) isolated by gel germination chromatography. /bid. 68: 70-78. Weiss, F. Seed germination in the gray birch, Betula populifolia. Am. ie Bot. 13: 737-742. 1926. WINKLER, H. Der Gegenwartige Stand der Betula- Systematik. we ae Dendr. Ges. 42: 36-39. 1930. [Review of alba complex.] Wo Le, C. B., Jr., & J. D. Pirt1tto. Some ecological factors funcing distribution of Betula nigra L. in western North Carolina. Castanea 42: 18-30. Th Wo pert, J. Vergleichende Anatomie und Entwicklungsgeschichte von Alnus, Alno- betula und Betula. Flora 100: 37-67. 1910. Subfamily COR YLOIDEAE (Regel) Koehne Tribe CARPINEAE D6ll 3. Carpinus Linnaeus, Sp. Pl. 2: 998. 1753; Gen. PI. ed. 5. 432. 1754. Small [to large] usually spreading trees, mostly with a single trunk; branching mostly deliquescent; trunk and branches irregularly longitudinally ridged and fluted, the branchlets and twigs conspicuously distichous; twigs differentiated into long and short shoots. Bark close, thin, smooth, bluish- to brownish-gray, becoming thicker and scaly or furrowed in age, the lenticels generally incon- spicuous; young twigs glabrous or sparingly pubescent; leaf scars narrowly crescent shaped to suboval, with 3 circular to elliptic vascular bundle scars; winter buds sessile, ovoid, 4-angled in cross section, usually appressed, the apices acute, with many smooth, imbricate scales; wood fine grained, nearly white to light brown, extremely hard, very heavy; pith circular to slightly angular in cross section. Leaves distichous, borne on long and short shoots; blades thin, narrowly ovate to ovate, elliptic, or obovate, doubly serrate [to 52 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 FiGure 3. Ostrya and Carpinus. a-j, O. virginiana: a, flowering branchlet, pendent staminate catkins above, carpellate catkin partly hidden by expanding leaves below, with 2 secondary ones) visible behind, portion of axis of catkin below, <8; c, d, 2 views of stamens, showing division of anther and upper part of filament into halves, x15; e, 1990] FURLOW, BETULACEAE 53 serrulate], glabrous to tomentose abaxially, sometimes with small glands abax- kins lateral, from axillary buds on short shoots, solitary [or in small racemose clusters], borne below the carpellate catkins, formed the previous growing season and [exposed or] enclosed in buds during the winter, expanding with the leaves, the scales broadly ovate [to elliptic], relatively uncrowded [to crowd- ed], each consisting of 3 fused bracts; carpellate catkins terminal on short shoots from leafy new growth, solitary, developing at the same time as the staminate ones, enclosed within buds during the winter and expanding with the leaves, more or less erect, uncrowded [to crowded], with paired flowers subtended by a primary scale and each surrounded by a 3-lobed scale consisting of 3 fused bracts. Staminate flowers 3 per scale, each flower consisting of 3(-6) stamens, several such clusters crowded together on a pilose torus at the base of the scale; stamens short, the anthers divided into 2 1-locular parts, pilose at the apex the filaments often divided partway to the base; pollen grains spheroidal to slightly flattened, 20-45 um in diameter, slightly aspidote, with 3(-6) circular to slightly elliptic equatorial apertures. Carpellate flowers sessile, 2 per primary scale; ovary 1, 2-locular, with 2 linear styles; perianth of several scalelike tepals, these adnate to the ovary and apparent as a membranaceous or short-fringed margin at the apex; sometimes with | or more staminodes; ovule | by abortion, bitegmic. Infructescences elongate, pendulous, consisting of a loose racemose [to densely imbricate] cluster of pairs of expanded, [(1- to)] 3-lobed and var- iously toothed foliaceous bracts, each bract subtending a single fruit, splitting away with the adnate fruit. Fruits small, ovoid, longitudinally ribbed nutlets, usually crowned with the persistent tepals and styles, maturing and dispersed the same season as pollination: pericarp relatively thick and bony; seeds with membranaceous testa and somewhat thickened cotyledons; germination epi- geal. Chromosome numbers 2 = 16, 32, 64. LECTOTYPE SPECIES: Carpinus Betulus L.; see N. L. Britton, N. Am. Trees, 241. 1908; N. L. Britton & A. daxial side of carpellate cymule, showing primary bract (1) and 2 flowers (only styles visible), each with sheath composed of secondary bract united with 2 tertiary ones, x 15; f, carpellate flower at anthesis, showing 2 receptive styles and hardly developed ovary crowned by rudimentary perianth, x15; g, carpellate cymule in young fruit, inflated surrounding bracts (secondary united with tertiary) removed from eae ae fruit at right to show persistent styles and collarlike perianth topping ovary, x5; h, branchlet with nearly mature infructescence, each fruit surrounded by inflated bracts (cf. e, g), x; i, mature fruit with persistent rudimentary perianth and styles at top, x4; ds seed: vat aborted ovule at upper left, x 4. k-g, C. caroliniana: k, adaxial side of sta showing stamens of 3 flowers with bract (composed of united primary and secondary bracts) behind, portion of axis of catkin below, x8; 1, stamen, showing half-anthers and partly divided filament, x15; m, carpellate cymule, from adaxial side, showing primary bract (1) and 2 flowers (only styles visible), each surrounded by 3-lobed bract composed of a secondary bract (2) united with 2 tertiary ones (3), x15; n, carpellate flower, the hardly developed ovary with minute perianth, x15; 0, branchlet with nearly mature infructescences, each fruit subtended by 3-lobed bract, x '/; p, mature fruit adnate to 3- lobed bract (united secondary (2) and tertiary (3) bracts), portion of axis splitting away, x 14; q, mature fruit topped by accrescent perianth and persistent styles, x 54 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Brown, Illus. Fl. No. U. S. & Canada, ed. 2. 1: 606. 1913. (The Latin name used by Pliny and other ancient writers for the hornbeam; possibly derived from carpentum, the name of a horse-drawn vehicle made from its wood.)— HORNBEAM, IRONWOOD. A genus of about 25 species of small to large trees mostly of the North Temperate Zone, but with a few extending into Central America along the Sierra Madre and in the Old World in the mountains from the North Temperate Zone to India and Iran. Carpinus Betulus is a large and important forest tree throughout much of Europe (where it attains trunk diameters of up to 4 m). In mountainous Mexico and Central America C. tropicalis (J. D. Sm.) Lundell forms a dominant canopy component. Some of the Asian species also become large trees. However, C. caroliniana Walter subsp. virginiana (Fern.) Furlow, of the northeastern United States and adjacent Canada, and subsp. caroliniana, found throughout the Coastal Plain (Fernald, 1935; Furlow, 1987b), consist of smaller forms of the forest understory, often near streams, where they occupy a subdominant position. Carpinus was treated taxonomically by Spach, De Candolle, and Winkler (1904). De Candolle divided the family Corylaceae into tribes CARPINEAE and CorYLEAE, the latter containing only Cory/us (as followed in the present treat- ment). He further divided Carpinus into two genera, Carpinus and Distego- carpus Sieb. & Zucc., the latter an Asian group (D. japonica Blume, D. cordatus Blume) characterized by elongate, stipitate, more densely imbricate staminate floral bracts and crowded infructescences composed of numerous broad un- lobed scales (as opposed to broadly ovate, subsessile, more or less uncrowded staminate bracts and open infructescences of relatively few distinctly three- lobed scales in Carpinus). Winkler (1904) treated these segregates as sections of Carpinus, and this remains the most frequently used treatment today. In a further revision of the genus (1914), he named a number of new species and varieties, based largely on the shape and size of the leaves and the infructescence bracts. Rafinesque modified the name of the genus, which ne considered too similar to Pinus, to Carpinum, and this variant i Additional study is needed to determine whether sect. DisTEeGOcARPUS (Sieb. & Zucc.) Sarg. is distinct enough to warrant continued recognit Numerous Asiatic species of Carpinus have been described in aa decades. In an early enumeration of Chinese Carpinus, Hu (1933) reported 23 species. Lee (1935) listed 24 species in Forest Botany of China, and 52 in the supplement to this work (1973). In Flora Reipublicae Popularis Sinicae Li & Cheng listed 25 species, together with an additional 15 infraspecific taxa. Although it is doubtful that all of these taxa deserve formal recognition, some of them appear to represent good species. The genus as a whole is in need of a comprehensive taxonomic revision. In North America Carpinus consists of two species, C. caroliniana Walter and C. tropicalis J. D. Smith, each with several geographic races (Furlow, 1987a). Fernald (1935) first distinguished an Atlantic and Gulf Coastal Plain race, with small, blunt-toothed leaves, from the widespread Appalachian and continental form. Furlow (1987a) analyzed this complex using multivariate 1990] FURLOW, BETULACEAE 55 statistics and concluded that the Latin American hornbeams constitute a di- vergent group, most likely not derived from the species in the eastern United States, and recognizable as a separate species. The Coastal Plain populations of the United States were shown to form a distinctive and cohesive subgroup of C. caroliniana. This and the Appalachian race were recognized formally at the level of subspecies. Carpinus caroliniana, American hornbeam, ironwood, blue beech, is easily recognized by its smooth, gray, often fluted stems, normally ovate to elliptic sharp-toothed leaves, and racemose infructescences of pairs of uncrowded, leaflike, three-lobed bracts, each subtending a small triangular nutlet. The staminate (but not the carpellate) catkins develop in the autumn, although they are enclosed in buds throughout the winter prior to anthesis. The carpellate catkins are produced on the first new growth in the spring. Both the staminate and carpellate catkins (except in sect. DisteEGOCARPUS) are much more un- crowded than those of A/nus or Betula. Leaves of Carpinus closely resemble those of Ostrya. Both lack peltate scales (sessile glands), and they have similar kinds of trichomes. However, they differ in the structure of their stipitate glands: in Carpinus the stalks are uniseriate, rather than multiseriate, and the heads are more globose (Bell et a/.; Hardin & Bell). The color and degree of development of these glands in C. caroliniana were shown by Furlow (1987a, b) to be of value in characterizing the subspecies. The wood, which overall has been regarded as rather advanced in the Betula- ceae, has both primitive features (e.g., numerous vessels of small diameter) and advanced ones (e.g., spiral thickenings on the vessels, homogeneous rays) (Hall). In some characters (e.g., structure of the perforation plates), relatively primitive states are present in some species and more advanced ones in others. Hall concluded that true tracheids were absent in Carpinus, although he noted fiber tracheids in all genera of the family. Recently, Yagmaie & Catling reported the presence of true tracheids in the wood of Carpinus. Staminate inflorescences of Carpinus and Ostrya are much more difficult to interpret than those of members of the Betuloideae because the flowers lack tepals and the cymules lack tertiary bracts (see Abbe, 1935, 1974). In Carpinus the catkins consist of clusters (“‘partial inflorescences”) of about 18 stamens. From the patterns observed in the Betuloideae and in Corylus, and from MacDonald’s anatomical observations in Ostrya (see below under Corylus), such clusters have been interpreted as highly reduced cymules of three flowers, each consisting of six stamens (Abbe, 1974). The carpellate cymules consist of two ovaries subtended by one primary, two secondary, and four tertiary bracts (Abbe, 1935). As the infructescence develops, the primary bract abscises and the united secondary and two tertiary bracts associated with each fruit develop into a characteristic wing by which the fruits are dispersed (see FiGuRE 3p). Species of Carpinus form a straightforward polyploid series of 2n = 16, 32, and 64 (in C. caroliniana, 2n = 16). In most species both the staminate and carpellate catkins are produced along with growth of the new leaves. The fruits, attached to expanded winglike bracts that dehisce from the infructescences with them, are dispersed by the wind. The paleobotanical history of the genus has been reviewed by Crane (1981) 56 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 and by Berger. Fossils referred to the Coryloideae first appear in the Paleocene. The genus Paleocarpinus Crane, from the upper Paleocene, morphologically links Carpinus with Corylus on the basis of fruit and bract characters (Crane, 1981, p. 131). Crane proposed that this fossil ‘may approach the generalized Carpinus form envisaged by Hjelmqvist as having given rise to the extant genera of Coryleae.” He further showed that modern betulaceous characters began to differentiate in the Upper Cretaceous, with genera such as C. arpinus greatly diversifying in the late Paleogene and early Neogene, and concluded (p. 131) that this shows that the ‘‘strong morphological adaptation for dispersal exhibited by most extant species” had not developed before that time. He proposed that the primary diversification took place in Eurasia, perhaps in relation to vegetation changes following climatic deterioration during the Eocene and Oligocene (cf. Wolfe, 1973). The origin of populations of Carpinus in the mountains of Mexico and Central America (Hernandez X. et al.) has been the subject of considerable speculation. Some workers (Deevey; Dressler; Miranda & Sharp) considered these populations to be closely related to those of the eastern United States, while others (e.g., Martin & Harrell) emphasized obstacles to the dispersal of mesophytic plants between these areas. Furlow (1987a, b) has shown by means of multivariate analyses that the taxa in Latin America are distinct morpho- logically and concluded on this basis, as well as on that of phytogeographic evidence, that Latin American Carpinus has more likely been derived from an extinct western taxon. Clinal variation and population differentiation have been demonstrated for several characters of C. caroliniana (Wardell & Winstead; Winstead er a/.). Furlow (1987a) showed that these and other characters vary geographically in complex ways, and that the patterns are related to climatic factors—in different ways in different regions. There are no serious insect pests or fungal pathogens associated with Car- pinus, although many fungi including mildews and rusts attack the leaves (Hepting). Sargent (1896) listed a variety of insects known to feed on the leaves of members of the genus. The very hard wood of Carpinus has been used, especially in Europe (where the trees are larger), for making mallet heads, tool handles, levers, and other small, hard, wooden objects. It has also been employed to make high-quality charcoal for use in manufacturing gunpowder. The branches are utilized ex- tensively in Europe for fuel. Carpinus Betulus and (less frequently) C. carolinia- na are cultivated as ornamentals, the former being available in a number of cultivars. REFERENCES: Under atte references see ABBE (1935, 1974); BELL et al.; CANDOLLE; CRANE (1981); DEEVEY; DReESSLER; HALL; HARDIN & BELL; HEPTING; HERNANDEZ X. et al.; LEE (1935, WINKLER (1904); WOLFE (1973); and YAGMAIE & CATLI ANDERSON, E. The European hornbeam, Carpinus Betulus. Missouri Bot. Gard. Bull. 52(10): 13, 14, 1964. 1990] FURLOW, BETULACEAE a1 Benson, M., E. SANDAY, & E. BERRIDGE. Contribution to the embryology of the Amen- tiferae. Part II. Carpinus Betulus. Trans. Linn. Soc. London, II. Bot. 7: 37-44. 1906. Bercer, W. Studien zur Systematik und Geschichte der Gattung Carpinus. Bot. Not. 106: 1-47. 1953. Bosrov, E. G. Carpinus L. (In Russian.) Fl. URSS. 5: 254-262. 1936. FERNALD, M. L. Midsummer vascular plants of southeastern Virginia. Rhodora 37: 378- 413, 423-454. 1935. Furow, J. J. The Carpinus caroliniana complex in North America. I. A multivariate analysis a variation. Syst. Bot. 12: 21-40. 1987a; II. Systematics. [bid. 416-434. Hu, H. H. Carns in China. Sunyatsenia 1: 103-120. 1933. materials on the monography of genus Carpinus Linn. of China. Acta Phy cee ae 9: 281-298. 1964. seas SZAFEROWA, J. The genus Carpinus in Europe in the palaeobotanical literature. nogr. Bot. 5. 1-59. 1958. —.. Pisaidies on the epidermis of recent and fossil fruits of Carpinus and Ostrya and its significance in the systematics and history of these genera. Acta Palaeobot. 16: 3-70. 1975 JouHNsson, H. Die Chromosomenzahl von Carpinus Betulus L. Hereditas 28: 228-230. 1942 Maekawa, F. A Japanese fossil Carpinus and its living allies. Jour. Jap. Bot. 26: 357, 358. 1951. Rappe-Fomin, O. Beitrage zur Systematik der Gattung Carpinus in Russland. Mem. cad. Sci. Ukraine 15: 51-107. 1929. ees P. O., & H. Piers. Carpinus L. Hornbeam. Pp. 266-268 in C. S. SCHOPMEYER, ed., Seeds of woody plants in the United States. U. S. Dep. Agr. Agr. Handb. 450. Washington, D. C. 1974. [Techniques for propagating eee from seed.] SARGENT, C. S. Carpinus. Silva N. Am. 9: 39-43. 1896; 14: 104. SpAcu, E. Notes sur les Carpinus. Ann. Sci. Nat. Bot. II. 16: 248- nr 184 2. WARDELL, G. I., & J. E. WINSTEAD. Populational differences in bud bursting of Carpinus caroliniana Walter. Trans. Kentucky Acad. Sci. 39: 127-130. 1978. WINKLER, H. Neue Revision der Gattung Carpinus. Bot. Jahrb. 15(suppl.): 488-508. 1914. WINSTEAD, J. E., B. J. Siro, & G. I. WARDELL. Fruit weight clines in populations of ash, ironwood, cherry, dogwood, and maple. Castanea 42: 56-60. 1977. 4. Ostrya Scopoli, Fl. Carniolica, ed. 2. 2: 243. 1772, nom. cons. Small to medium-sized, usually spreading trees, mostly with a single trunk; branching mostly deliquescent; trunks and branches terete, the branchlets and twigs conspicuously distichous; twigs differentiated into long and short shoots. Bark thin, light brownish gray to light brown, breaking into slender, shaggy vertical shreds, the lenticels generally inconspicuous; young twigs glabrous or sparingly pubescent; leaf scars narrowly crescent shaped to suboval, with 3 circular to elliptic vascular bundle scars; winter buds sessile, ovoid, somewhat laterally compressed, divergent, the apices acute, with many longitudinally striate imbricate scales; wood fine grained, nearly white to light brown, ex- tremely hard, very heavy; pith circular to slightly angular in cross section. Leaves distichous, borne on long and short shoots; blades thin, narrowly ovate to ovate, elliptic, or obovate, doubly serrate [to serrulate], glabrous to abaxially tomentose; secondary venation craspedodromous, the veins divergent and 58 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 straight; new leaves in bud concave, plicate, not conduplicate; stipules broadly ovate. Staminate catkins terminal on branchlets, [solitary or] in small racemose clusters, formed the previous growing season and exposed during the winter, expanding with the leaves, the scales relatively uncrowded, broadly ovate, each consisting of 3 fused bracts; carpellate catkins terminal on short shoots from leafy new growth, below the staminate, solitary, developing at the same time as the staminate, enclosed within buds during the winter and expanding with the leaves, more or less erect and uncrowded, the scales 3 fused bracts, these later completely fusing so as to enclose the young fruits completely in blad- derlike involucres. Staminate flowers 3 per scale, each consisting of 3(-6) stamens, several such clusters crowded together on a pilose torus at the base of the scale; stamens short, the anthers divided into 2 1-locular parts, pilose at the apex, the filaments often divided partway to the base; pollen grains spheroidal to slightly flattened, 20-45 um in diameter, slightly aspidote, with 3(-6) circular to slightly elliptic equatorial apertures. Carpellate flowers sessile, 2 per scale; ovary 1, 2-locular, with 2 linear styles; perianth of several scalelike tepals, these adnate to the ovary and apparent as a membranaceous or short- fringed margin at the apex; sometimes with | or more staminodes; ovule 1 by abortion, bitegmic. Infructescences consisting of loosely imbricate, pendulous, strobiloid clusters of closed bladderlike involucres derived from the encircling bracts of each flower in the catkins, each bract enclosing and deciduous with a single fruit. Fruits small, ovoid, longitudinally ribbed nutlets, maturing and dispersed during the same season as pollination, often crowned with the per- sistent tepals and styles; pericarp relatively thick and bony; seeds with mem- branaceous testa and somewhat thickened cotyledons; germination epigeal. Chromosome number 2n = 16. Type species: Ostrya carpinifolia Scop. (The Greek name used by Theophrastus for a tree with very hard wood; from the Greek ostryos, “a scale,” in reference to the scaly catkins.)— Hop HORNBEAM, IRONWOOD. About five species of small trees of the North Temperate Zone. Ostrya carpinifolia isa common and important forest tree throughout southern Europe. In North America the genus consists of small trees of the northeastern deciduous forest and the mountains of the southwestern United States and adjacent Mex- ico, south to northern Central America. Ostrya was included as a single species of Carpinus (C. Ostrya) by Linnaeus. Miller accepted this generic concept, but he separated the American species (as C. virginiana Miller) from the European; Michaux treated it as C. Ostrya americana. The genus was segregated from Carpinus in 1772 by Scopoli, who named the common European tree Ostrya carpinifolia. Ostrya has since mostly stood as a separate genus, yet on the basis of its inflorescences, infructescences, and vegetative features, the two genera are closely allied. Willdenow, in the fourth edition of Species Plantarum, named the American species O. virginica. Spach (1842b), in a revision of the genus, recognized two species, O. italica Micheli (including all the European forms) and O. virginica Willd. De Candolle also recognized these species but correctly selected the earlier name, O. car- pinifolia, to designate the former. In 1873 K. Koch transferred Miller’s name to Ostrya. Winkler submerged both O. virginiana (Miller) K. Koch and QO. 1990] FURLOW, BETULACEAE 59 carpinifolia as subspecies of O. italica but recognized the western North Amer- ican O. Knowltonii Cov. as separate. Rafinesque, in Florula Ludoviciana, sub- stituted the name Zugilus for Ostrya because he believed the name to be too similar to Ostrea. Four species are listed for China by Lee (1973) and Li & Cheng, although some of these may be found to be too indistinct to deserve specific status. Ostrya virginiana is a common tree in North America from Nova Scotia to eastern Manitoba, south to Virginia, northern Georgia, Tennessee, and Okla- homa, with a disjunct population in the Black Hills (South Dakota). Although frequent in the Northeast, O. virginiana is seldom a major forest component. There, like Carpinus caroliniana, it usually occupies a subdominant position in the understory (although Greenidge has reported that the species is nearly absent from closed old-growth forests in Nova Scotia). Unlike Carpinus, it is characteristic of drier or better-drained, more upland sites. It is seldom seen in wet areas. Ostrya virginiana is much less abundant in the Southeast than farther north (Duncan, Radford), occurring—when present at all—mostly in the mountains and Piedmont. The leaves of Ostrya virginiana are similar to those of Carpinus caroliniana, as are its infructescences and fruits, except that the infructescences are some- what more compact, with the bracts fused into bladders that completely enclose the fruits. As in Carpinus, only the staminate catkins develop in the autumn, although in Ostrya these occur in small clusters and are exposed during the winter near the tips of lateral branchlets (short shoots). Also as in Carpinus, the carpellate and staminate catkins are loosely arranged at anthesis. One of the most characteristic field characters of O. virginiana is its light brownish- gray bark, which shreds into thin, narrow vertical strips. In the winter the trees are distinctive in their numerous small terminal clusters of dormant catkins (absent in Carpinus caroliniana). Coastal Plain populations of Ostrya virginiana are represented _ a small- leaved and somewhat pubescent geographic race (var. /asia Fern.). However, the O. virginiana complex has not been studied in detail. Two shrubbier species occur in the Southwest: O. Knowi/tonii Cov. is found in mountains and canyons from southwestern Texas to southeastern Utah (including both rims of the Grand Canyon), and O. chisosensis Correll occurs in the Chisos Mountains in Big Bend National Park in southwestern Texas. These two species differ some- what from each other and from O. virginiana in characters of leaf shape, leaf margin, and plant pubescence. However, no comprehensive study has consid- ered the distinctness of these species in terms of their variation patterns or their relationships to other North American taxa. Additional populations of Ostrya occur in the eastern and western mountains of Mexico and northern Central America. Rose believed these segments to differ significantly from O. virginiana and named the Mexican group, char- acterized by more narrowly lanceolate and more gradually acuminate leaves, O. mexicana Rose. He called the southern group, with similar features but somewhat broader and more pubescent leaves, O. guatemalensis Rose. How- ever, in current work (e.g., Nee) these forms are usually treated as conspecific with O. virginiana. The Latin American representatives are especially in need of critical taxonomic examination. The morphological differentiation and pa- 60 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 leoecology of these segments, as well as of O. Knowltonii and O. chisosensis, need to be examined in relation to the complex as a whole Although Ostrya shares many vegetative features with Carpinus, its habit is more treelike. As in Corylus but not Carpinus, the stipitate glands of the leaves have multiseriate rather than uniseriate stalks, and these bear more elongate heads, features not seen in other Betulaceae (Hardin & Bell). The wood is similar to that of Carpinus, but the vessels are of a more specialized type, with largely simple perforation plates (Hall). The genus is seen by Hall as the most advanced of the family on the basis of wood structure. The inflorescence and flower structure of Ostrya is also similar to that of Carpinus, except that in the infructescences the secondary and two tertiary bracts of each floret are fused into a sac that envelops the fruit, rather than a flat wing that subtends it (Abbe, 1935, 1974). Even though direct evidence was lacking, Abbe (1935) hypothesized that the clusters of stamens found in sta- minate catkins of Carpinus and Ostrya represented the three reduced florets of a cymule comparable to those seen in the Betuloideae. This position was recently supported by the work of MacDonald, who demonstrated three growth areas in the primordia of the staminal groups in developing catkins. All species of Ostrya for which counts have been made have a chromosome number of 2 = 16 Unlike Carpinus, in most species of Ostrya the staminate catkins are pro- duced the season before anthesis and exposed during the winter. The carpellate catkins develop in the spring with the new shoots, with anthesis occurring as the leaves are forming. Dispersal is as in Carpinus, except that the bracts form closed bladderlike structures rather than flat wings. The evolution of Ostrya parallels that of Carpinus, but the genus first appears somewhat later in the fossil record (Miocene), and fossils of Ostrya are not nearly so well represented (Crane). Although the disjunct populations of Ostrya in the mountains of Mexico have generally been considered to be conspecific with the species of the eastern United States (Miranda & Sharp; Nee), the same phytogeographic evidence cited in connection with Latin American Carpinus suggests that these populations may have been derived not from O. virginiana, but rather from an earlier and more western species. Ostrya, like Carpinus, suffers from few insect pests or diseases, and none of these is regarded to be of economic importance. Sargent (1896) listed a number of insects that feed on or otherwise affect hornbeams. Hepting discussed various parasitic fungi, mostly found on Ostrya leaves. The wood of Ostrya is employed for fuel, fence posts, and other utility purposes. It was formerly used for making items subject to prolonged friction, including sleigh runners, wagon tongues, wheel rims, spokes, windmill vanes, and airplane propellers. Because of its density, it has been used for tool handles, mallet heads, and other hard wooden objects. Millspaugh listed a tincture of the heartwood of Ostrya virginiana as a treatment for intermittent fever. Ostrva virginiana, and sometimes O. carpinifolia, are occasionally cultivated in eastern North America. 1990] FURLOW, BETULACEAE 61 REFERENCES: Under family references see ABBE (1935, 1974); CANDOLLE; DUNCAN; HALL; HARDIN & Bett; HEPTING; Kocn; LEE (1973); Li & CHENG; MACDONALD; MICHAUX; MIRANDA & SHARP; RADFORD; ocr ed SARGENT (1896); SPACH eee and WINKLER. Bosrov, E. G. Ostrya Scop. (In Russian.) Fl. URSS. 5: 253, 254. ELuison, L. A. Occurrence of hop ee (Ostrya virginiana) on aa Island. Proc. Staten Island Inst. Arts Sci. 13: 70, 71. 1951. FERNALD, M. L. Plants from the outer i Plain of — Rhodora 38: 376-404, 414-452. 1936. [Geographic variation in O. virginia FiicHe, M. Notes sur les formes du genre Ostrya. Bull. Soc. Bot. France 34: 462-473. 1890. GREENIDGE, K. N. H. Distribution and ecological characteristics of ironwood, Ostrya virginiana (Miller) K. Koch, in northeastern Nova Scotia. Rhodora 86: 139-149. 1984. 98 Mace, R. F. Hophornbeam (Ostrya virginiana) for handles. New Hampshire Dep. Forest. Notes 36: 1. 1948. [Use of hop hornbeam for tool handles.] MILLER, P. The gardener’s dictionary. ed. 8. London. 1768. MILLspAuGH, C. F. American medicinal plants. 2 vols. 806 pp. Philadelphia. 1884— 1887. [Ostrya virginiana, 637-639. Rose, J. N. Notes on Ostryva, with two new species. Contr. U. S. Natl. Herb. 8: 291- 293. 1905. SARGENT, C. S. Ostrya. Silva N. Am. 9: 31-38. 1896; 14: 104. 1902. SCHOPMEYER, C. S., & W. B. LEAK. Ostrya virginiana (Mill.) K. Koch. Eastern hornbeam Pp. 264, 265 in C. S. SCHOPMEYER, ed., Seeds of woody plants in the United States. U. S. Dep. Agr. Agr. Handb. 450. Washington, D. C. 1974. [Techniques for prop- agating Ostrya from seed.] WILLDENOW, K. L. Ostrya. Sp. Pl. ed. 4. 4(1): 469. 1805. Tribe CoryLEAE Meisner 5. Corylus Linnaeus, Sp. Pl. 2: 998. 1753; Gen. Pl. ed. 5. 433. 1754. Medium-sized to large shrubs [or small to medium-sized, usually spreading trees]; branching mostly deliquescent; trunks and branches terete, the branchlets and twigs subdistichous to diffuse; twigs differentiated into long and short shoots. Bark close, thin, smooth, grayish brown, breaking into vertical strips and scales with age, lenticels inconspicuous; young twigs glabrous or sparingly pubescent, sometimes with resinous glands; leaf scars suboval to triangular, with 3 nearly equidistant circular to elliptic vascular bundle scars; winter buds sessile, broadly ovoid, terete, divergent, the apices acute, with several smooth, imbricate scales; wood fine grained, nearly white to light brown, moderately hard, moderately heavy; pith circular to slightly angular in cross section. Leaves distichous, borne on long and short shoots; blades thin, broadly ovate, the base often cordate, occasionally somewhat lobed above, doubly serrate, usually pu- bescent and sometimes glandular abaxially; secondary venation craspedodro- mous, the lowest veins sometimes crowded at the base of the midrib and rising abruptly toward the apex; leaves in bud conduplicate and plicate; stipules broadly ovate. Staminate catkins lateral in bud axils on short shoots, in nu- merous racemose clusters of (2 or) 3-5, formed the previous growing season 62 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 and exposed during the winter, expanding long before the leaves, the scales broadly ovate, relatively uncrowded, consisting of 3 fused bracts; carpellate inflorescences lateral, borne near the tips of the branchlets producing staminate catkins, developing at the same time as the staminate, enclosed within buds during the winter and expanding long before the leaves, consisting of a small cluster of flowers and bracts, only the styles protruding from the buds at an- thesis, the scales 3 fused bracts. Staminate flowers 3 per scale in the catkin, congested, the tepals lacking [(or 1|—-4)], the stamens 4, divided nearly or entirely to the base to form 8 half-stamens, the filaments very short, fused along with 2 bractlets to the scale; pollen grains flattened, 12-30 um in diameter, slightly to moderately aspidote, with (2 or) 3 (or 4-6) slightly elliptic equatorial ap- ertures. Carpellate flowers sessile, 2 per scale, with 4 extremely reduced tepals (displayed as a thin irregular fringe on the ovary); ovary 1, 2-locular, with 2 linear styles; ovule | by abortion, bitegmic. Infructescences consisting of com- pact clusters of several fruits, each subtended and surrounded by an involucre of 2 hairy [or spiny] expanded foliaceous bracts, these sometimes fused into a [short to] elongate tube. Fruits relatively thin-walled, subglobose to ovoid, somewhat laterally compressed, longitudinally ribbed nuts; pericarp bony; seed with membranaceous testa, the cotyledons thick and oily; germination hypo- geal, the seed being raised to the surface but remaining in the fruit. Chromosome number 2” = 28. Lecrotype species: Corylus Avellana L.; see N. L. Britton, N. Am. Trees, 246. 1908; N. L. Britton & A. Brown, Illus. Fl. No. U.S. & Canada, ed. 2. 1: 607. 1913. (The Latin name used by Virgil, Pliny, and other ancient writers for the European hop hornbeam; from korus, “helmet,” for the shape of the shells of the nuts.) —Haze-. About 15 species of trees and shrubs of the North Temperate Zone. Corylus Colurna L., Turkish filbert or hazel, is a medium-sized tree of southeastern Europe and Asia Minor. The other species of Europe and North America are small to large shrubs or small trees. Two species, C. americana Walter, Amer- ican hazel, and C. cornuta Marsh. (C. rostrata Aiton), beaked hazel, occur throughout much of the northeastern United States and adjacent Canada. In the Southeast these are mostly confined to the mountains southward to northern Alabama and Georgia. Several varieties of C. cornuta, including the tree-sized var. californica (A. DC.) Sharp, occur to the west. Corylus was treated as a genus by Linnaeus and his predecessors. Spach (1842c) divided the genus into three sections, Ave/lana Spach (C. Avellana and C. Colurna), Tubo-Avellana Spach (C. tubulosa Willd. and C. rostrata), and Acanthochlamnys Spach (C. ferox Wall.). The second of these groups is char- acterized by an elongate tubular involucral beak, and the last by densely spiny bracts. De Candolle modified this scheme, making the first two groups sub- sections of sect. Avellana. Winkler recognized no infrageneric categories of Corylus in his monograph of the Betulaceae. In his synopsis, Beijerinck de- scribed 32 species, varieties, and cultivars of the genus. Lee (1973) listed 15 Chinese species, while Li & Cheng recognized seven. As in the other genera of the Betulaceae, the relatively poorly known Asian species need to be examined in relation to the genus as a whole. No infrageneric taxa are recognized here, pending detailed study of subgroups of the genus. FURLOW, BETULACEAE 63 Ficure 4. Cory C. americana: a, branchlet with qe anne the 4 eee ones 2a 2 3 carpellate ones ascending, the styles visible, x 1; b, adaxial side of staminate cymule with 8 half-anthers, primary bract and | secondary bract (at right) visible behind them, <8; c, same, 8 half-anthers removed, bases of nee a shown, primary (1) and 2 secondary (2) bracts visible, x8; d, carpellate catkin, x 4; adaxial side of carpellat le of 2 flowers, styles receptive, ovaries hardly eae bra each ovary with delicate, rudime entaey perianth, primary bract (1) and parts of tertiary bracts (3) visible behind flowers, secondary bracts absent, x 13. f, branchlet with 3 nearly mature fruits with their accrescent paired tertiary bracts, <1; g, 2 mature fruits, each surrounded by tertiary bracts, x1; h, mature nut with 1 tertiary pleas (3) behind, the other removed, x 1;1, nut to show seed with ha right and aborted vule at upper left, x2. j, C. cornuta: — eat with paired pear bracts and (at eee left) undeveloped fruits i bracts, 64 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Corylus americana 1s a shrub to about 3 m tall, occurring mostly in thickets, open woods, fence rows, and forest edges, especially on well-drained soils, from Maine to Missouri and south to Georgia and Oklahoma. Its broadly ovate or roundish, doubly serrate leaves are distinctive in that they are often expanded apically to give a squarish appearance. Corylus cornuta is similar in habitat and distribution, but it extends farther north and west (from Newfoundland to British Columbia and south to Georgia, eastern Kansas, Colorado, and California). It is a larger plant than C. americana, reaching a height of about 5 m, and its leaves are narrower and more ovate. Drumke concluded that C. cornuta var. californica (A. DC.) Sharp, which becomes a small tree, is suffi- ciently distinct to warrant varietal status. He noted that this form grades clinally into var. cornuta to the north in Oregon and Washington. Useful field characters for separating the two species include the presence of reddish stipitate glands on the petioles and young twigs of C. americana and their absence on the petioles of C. cornuta (see Wiegand), and more rounded bud apices in C. americana. Corylus cornuta differs most noticeably from C. americana in the narrow, extended, tubular involucres surrounding its fruits, those of C. amer- icana being short and leaflike. Drumke examined populations of these two species in their region of overlap and found them to be morphologically distinct, with little or no evidence of hybridization. Although clearly related at the family level, the hazels are morphologically distinct both from the Betuloideae and from Carpinus and Ostrya. Their most distinctive features lie in their infructescences, which consist of a small cluster of small to moderately large nuts, each enclosed by a loose involucre of leaflike bracts. As in Ostrya, staminate catkins are formed during the summer and are exposed through the winter prior to anthesis. However, these are represented by numerous clusters of catkins borne on short shoots arranged evenly along the branches. The carpellate catkins develop at the same time as the staminate and consist of only a few flowers protected by the scales of special buds. The leaves of Cory/us resemble those of the other Betulaceae in overall aspect, but they are modified in shape and venation. As noted by L. J. Hickey & Wolfe and Wolfe (1973), the blades are frequently broader, and the lowest secondary veins, which are congested at the base of the midrib in some species, rise sharply toward the apex, a pattern also seen in Corylopsis of the Hamameli- daceae (see further discussion above under the family treatment). The indu- mentum on the leaves of Corylus is very similar to that of Ostrya and (to a slightly lesser degree) Carpinus (Hardin & Bell). All three genera lack peltate scales but have the five other trichome types described by Hardin & Bell. However, the stipitate glands of Corylus and Ostrya have multiseriate stalks, while those of Carpinus are more primitive in their uniseriate stalks (Hardin & Bell). The genus is the most specialized of the family in its fruit type (well- developed nuts) and the accompanying involucre (Stebbins; Stone). Corylus stands apart from the remainder of the family in terms of flower and inflorescence morphology. The staminate inflorescences are similar in structure to those of the other Coryloideae except that up to four tepals are occasionally present, clearly defining the three individual flowers that make up each cymule. The carpellate catkins are much modified. The inflorescence itself 1990] FURLOW, BETULACEAE 65 is reduced to a small cluster of flowers, only one or two of which develop further. Present in each partial inflorescence are two flowers, as in Carpinus and Ostrya, plus one primary and several additional bracts. Abbe (1935) in- terpreted the latter to represent various of the four tertiary bracts of a cymule (the secondary bracts not developing), but Hjelmqvist (1948) believed the two secondary bracts to be present, each fused to one of the two tertiary bracts associated with every flower (cf. Abbe, 1974). The two resulting foliaceous bracts grow around the developing fruit, either free from each other, as in Corylus americana, or fused into a tube, as in C. cornuta (see Abbe, 1974). Abbe (1974) reviewed the development of present concepts regarding the nature of the involucre of Cory/us and various misconceptions that have involved its structure and development. All investigated species of Corylus have chromosome numbers of 2” = 28. The inflorescences are produced the season before flowering, with the staminate being exposed during the winter and the carpellate enclosed in buds. Anthesis usually occurs ext ly early (January or February), even in northern areas, and well before production of the new leaves. The plants are anemoph- ilous, producing large quantities of pollen (Wodehouse, 1935). Dispersal is (apparently) by means of small mammals that carry the nuts away. The wood of Corylus, like that of Carpinus and Ostrya, contains numerous small vessels with spiral secondary thickenings (Hall). However, its vessels, like those of A/nus and Betula, have scalariform perforation plates and are thus regarded as more primitive than those of the other Coryloideae (Hall). Hall concluded that true tracheids are absent from all of the Coryloideae, although fiber tracheids are present. However, Kasapligil (1964) and Yagmaie & Catling have reported tracheids in Cory/us wood. Kasapligil (1964) noted two distinct subgroups of the genus on the basis of wood anatomy, with one, including C. Avellana, mostly lacking true tracheids and having fewer and wider bars in the perforation plates of its vessels, and a second, more primitive group (including C. Colurna) with wood composed of both tracheids and vessels, the latter with numerous narrow scalariform perforations. He suggested, on the basis of in- dumentum and other characters, that this second group might also include C. ferox, but that a formal assignment of C. Colurna to sect. Acanthochlamnys would require additional study. According to Hall, the wood of Corylus is anatomically indistinguishable from that of Ostryopsis. Fossil leaves from the Late Cretaceous and early Paleocene have been iden- tified as Corylus, but fossil fruits of the Corylus type are scarce during this period (Crane, 1981). A probable small Corylus nut was discovered in the Danian of Greenland by Koch. By the Pliocene, the genus had become well established in North America and Europe. Zoéchory, a novelty in the Betu- laceae, has been regarded by both Stebbins and Stone as a specialized condition; all of the other genera (except Ostryopsis) rely on wind or water for dispersal. Crane (1981) pointed out that zodchory is related to more stable K-selective environments than is anemochory, and its development may have paralleled the evolution of suitable animal dispersal vectors during the Paleocene, these permitting the genus to diversify into new niches. Corylus is the source of hazelnuts and filberts. The commercial filbert (C. 66 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Avellana and C. maxima Miller) and the Turkey nut (C. Co/urna) are cultivated as crop plants in various parts of the world, particularly Turkey, Italy, Spain, China, and Japan. In the United States these species are grown commercially in the Pacific Northwest, where they produce over 10,000 metric tons of nuts annually, about five percent of the world crop (Schery). The fruits are used mostly as dessert nuts, but they are rich in oil (up to 68 percent) and serve as a commercial source for cooking and salad oil in Europe (Eckey; Vaughn). The kernels are sometimes ground into meal used to make a sweet, cakelike bread (Fernald et al.). Wild hazelnuts are gathered locally in both America and Eu- rope. The pollen of hazels causes hay-fever allergies in regions where they occur in abundance (Dalen & Voorhorst; Lewis et al.; Lowenstein et al.; Wodehouse, 1945). The wood of the tree-sized species, which is similar in structure to that of birches, is used in limited amounts for pipe stems, hoops, tool handles, carved items, molding, and boxes. Cory/us cornuta spreads aggressively and is considered a weedy pest in northern forest plantations (Tappeiner). Cultivars of a number of species, especially the shrubby C. Avellana and the arborescent C. Colurna, are widely planted as ornamental shrubs. REFERENCES: Under family cals see ABBE (1935, 1974); CANDOLLE; CRANE (1981); DALEN & VOORHORST; FERNALD et al.; HALL; HARDIN & BELL; L. - janet KEY & WOLFE; HJELMQVIST (1948); KASAPLIGIL (1964. Lee (1973); Lewis et al.; Li & CHENG; LOWENSTEIN ef al.; SPACH (1842c); Pike STONE; WINKLER; a (1935, 1945); WoLFE (1973); and YAGMAIE & CATL BARNOLA, P. Recherches sur la croissance et la ramification du noisetier (Corylus Avella- na L.). Ann. Sci. Nat. Bot. XII. 17: 223-257. 1976. BAUCKMANN, M. Morphologische Untersuchungen in Corylus . olurna (Baumhasel) Blattern und Fruchtstaénden. Mitt. Rebe Wein 26: 255-260. 1977. ear W. The distribution of the species and forms of ce genus Corylus. (In Dutch.) Nederl. Dendrol. Ver. a 17: 67-107. 1949. [Synopsis of the genus; includes a bibliography of 48 works.] Bosrov, E. G. Corylus L. (In Russian.) Fl URSS. 5: 262-268. 1935. . Histoire et systématique du genre Cory/us. (In Russian.) Sovetsk. Bot. 1936(1): 11-51. 1936. [Includes a phylogenetic analysis of the genus.] Borcu, S. M., & B. A. SkKALHEGG. Some characteristics of allergenic components in an aqueous extract from hazel pollen, Corylus Avellana. Allergy 35: 194, 195. 1980. BRINKMAN, K. A. Corylus L. Hazel, filbert. Pp. 343-345 in C. S. SCHOPMEYER, ed., Seeds of woody plants in the United States. U. S. Dep. Agr. Agr. Handb. 450. Washington, D. C. 1974. [Description of Corylus fruit structure and procedures for propagating hazels from seed Drumkg, J.S. A systematic survey of Cory/us in North America. (Abstract.) Diss. Abstr. 25: 4925, 4926. 1965. Eckey, E. W. Vegetable fats and oils. ix + 836 pp. New York. 1954. [Corylaceae (Betulaceae), 383, 384; composition and characteristics of hazelnut oil.] GoESCHKE, F. Die Haselnuss, ihre Arten und ihre Kultur. 100 pp. Berlin. 1887. [Com- prehensive treatment of cultivated Corylus.] GOTTWALD, T. R., & H. R. CAMERON. Disease increase and the dynamics of spread of canker caused by Anisogramma anomala in European filbert, ee Avellana, in the Pacific Northwest, U. S. A. Phytopathology 70: 1087-1092. 1990] FURLOW, BETULACEAE 67 Jeavons, R. A., & B. C. Jarvis. The breaking of dormancy in hazel seed by pretreatment with ethanol and mercuric chloride. New Phytol. 96: 551-554. 1984. KanieEwskI, K. Development of the pericarp of the fruit of Corylus Avellana L. Bull. Acad. Polon. Sci. Biol. 12: 215-226. 1964. Kasapuicit, B. A. A bibliography on Corylus (Betulaceae) with annotations. No. Nut Growers Assoc. Rep. 63: 107-162. 1972. [Over 400 titles, mostly dealing with horticultural topics.] Kocu, B. E. Fossil plants from the lower Paleocene of the Agatdalen (Angmartussut) area, central Nugssuaq Peninsula, northwest Greenland. Medd. Grenl. 172(5): 1- 120. 1963. Kova, G. K. Differentiation of male and female flowers of Corylus L. (In Ukrainian with English summary.) Ukr. Bot. Zhur. 28: 199-203. 1971. ScHERY, R. W. Plants for man. ed. 2. viii + 657 pp. Englewood Cliffs, New Jersey. 1972. [Hazelnut, 465.] SHANNON, P. R. M., R. A. Jeavons, & B. C. Jarvis. Light sensitivity of hazel, Corylus Avellana, seeds with respect to the breaking of dormancy. Pl. Cell Physiol. 24: 933- 936. 1983. TappEIner, J. C., II. Invasion and development of beaked hazel in red pine stands in northern Minnesota Ecology 52: 514-519. 1971. VAUGHN, J. G. e structure and utilization of oil seeds. xv + 279 pp. London. 1970. [Hazelnuts, rae WIEGAND, K. M. Recognition of Corylus rostrata and Corylus americana. ae 11: 107. 1909. [Vegetative characters for distinguishing the native s ZIELINSKI, Q. B. Techniques for collecting, handling, germinating, a Sone pollen of the filbert (Corylus spp.). Euphytica 17: 121-125. 1968. PRICE, TAXACEAE 69 THE GENERA OF TAXACEAE IN THE SOUTHEASTERN UNITED STATES' RoBeERT A. PRICE? TAXACEAE S. F. Gray, Nat. Arr. Brit. Pl. 2: 222. 1821, “Taxideae,” nom. cons. (YEW FAMILY) Evergreen dioecious (or rarely monoecious) trees or shrubs. Foliage leaves entire, linear to linear-lanceolate, spirally arranged (often apparently 2-ranked) [or opposite in Amentotaxus], short-petiolate, with 2 stomatal bands on the abaxial surface; resin canal single, abaxial to the vascular bundle (or absent in Taxus, Pseudotaxus, and Austrotaxus). Wood without resin canals, axial wood parenchyma present or absent, helical thickenings present on the tracheid walls [apparently absent in Austrotaxus]. Pollen cones (microsporangiate strobil1) apparently simple, borne singly in the axils of foliage leaves [or compound and ‘Prepare Fl fthe Southeastern United States, a long-term project made possible through ie see of National Science Foundation Grant BSR-87 16834 (Norton G. Miller, principal investigator), under which this account was prepared, and BSR-8717333 (Carroll E. Wood, Jr. ants o ee ake ue members of a family or genus in brackets. The references that I have rified a rked with asterisks. a (eae il ae and Norton Miller for the opportunities afforded by participation in the Generic Flora Project and for their guidance in the study. I am grateful to Rudolf Schmid for bibliographic assistance, to Paul Groff for critical reading of the manuscript, and to Robert Godfrey herbarium collections associated with Harvard University and the University oe Berkeley, were at in this study, and 1 wish to thank the ate of these institution of ae taxifolia are based on living material from the University of California Botanical Garden, R. K. Godfrey 55449 (uc). The Shee and cross sections of ihe seed of Torreya are ei Contribution 616 from the New York State Science Ser 2University Herbarium, University of California, sa eeN fiona 94720. Please address reprint requests to the Biological Survey, New York State Museum, The State Education Department, Albany New York 12230 © President and Fellows of Harvard College, 1990. Journal of the Arnold Arboretum 71: 69-91. January, 1990. 70 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 aggregated in Amentotaxus and Austrotaxus]; microsporophylls several per strobilus, more or less whorled, each with [2 or] 3-9 microsporangia borne in a radial arrangement (Taxus, Pseudotaxus) or developing fully only on the abaxial side (Torreya, Amentotaxus); pollen grains nonsaccate, lacking pro- thallial cells. Ovules arillate, borne singly at the ends of short axillary shoots? bearing decussate [or spirally arranged] scale leaves; archegonia usually few per ovule, not clustered. Seeds with a stony coat, largely surrounded by the fleshy aril; embryo straight, cotyledons 2 (or occasionally | or 3). Chromosome num- ber 2m = 22 or 24 [14 in Amentotaxus]. (Including Amentotaxaceae Kudo & Yamamoto, Austrotaxaceae Nakai, Torreyaceae Nakai: Taxineae L. & A. Rich- ard.) Type GENUS: Taxus L. A small family of five genera and perhaps 20 species, the Taxaceae are widely distributed in the Northern Hemisphere in moist, forested habitats from the subarctic of Eurasia and North America to subtropical or even tropical areas of Central America and southeastern Asia. The monotypic Austrotaxus Comp- ton is endemic to New Caledonia, while Taxus L. ranges south of the equator on Sumatra and Celebes (Florin, 1963). Only the two most widespread genera, Yorreya Arnott and Taxus, occur in North America, including the southeastern United States. The monotypic Pseudotaxus Cheng (Nothotaxus Florin) is en- demic to eastern China, while Amentotaxus Pilger has four species in south- eastern Asia, occurring from central and southern China, including Taiwan and Hongkong, to southern Vietnam and extreme eastern India (Alvin et al; Ferguson, 1985; Florin, 1963). Amentotaxus is also known from fossils of Tertiary age in western North America and Europe, but it had disappeared from these areas by the Late Miocene or Early Pliocene (Alvin et al.: Florin, 1963). The Taxaceae, both fossil and extant, are unique among the conifers in having arillate ovules borne singly at the ends of lateral short shoots, but with no evidence of a biaxial ovulate cone or ovuliferous scales. Thus Florin (1938- 1945, 1948a) proposed that they form an evolutionary lineage separate from the conifers. However, the Taxaceae have the specialized pattern of proem- bryogeny typical of the modern conifers (Dogra; Doyle, 1963), and they share derived features of wood anatomy and pollen and leaf morphology with other modern families of conifers, particularly the Cephalotaxaceae, a monogeneric family native to eastern Asia. On the bases of overall similarity and preliminary phylogenetic analyses (Hart), it seems most likely that the Taxaceae are conifers, most closely related to Cephalotaxus Sieb. & Zucc. ex Endl., that have sec- ondarily lost the coniferous ovulate cone organization, although there are no intermediates indicating how the ovules have come to be terminal in position. Cephalotaxus (plum-yew, with nine species in Fu’s treatment) is very similar to the Taxaceae in appearance and in a number of morphological features. It differs most prominently from the Taxaceae in having its young ovules borne ‘In Taxus and Torreva the ovule-bearing short shoots are axillary to scale leaves borne on very short axillary shoots, while in the other genera they arise from the axils of foliage leaves (Florin, 1948a). In Torreya there are usually two ovuliferous shoots on each dwarf axillary shoot, but only one ovule generally matures. 1990] PRICE, TAXACEAE 71 in pairs along a cone axis. Each pair of ovules is subtended by a bract and is associated with a small outgrowth that has been interpreted as an extremely reduced ovuliferous scale (Florin, 1951; Singh, 1961). Generally only one or two ovules mature per cone. The ovular integument differentiates into an inner stony layer and an outer fleshy layer, making the seed similar in structure and appearance to those of Torreya and Amentotaxus, but there is no evidence of a separate aril in the development of the ovule (Singh, 1961). Until the early 1900’s the Taxaceae were often treated as including both the Cephalotaxaceae and the Podocarpaceae (Pilger, 1903, 1916b), primarily on the grounds that these groups have fleshy structures surrounding the seed and often have the ovuliferous scales greatly reduced. The Podocarpaceae are very diverse in their ovulate-cone morphology. Cones of some species are highly reduced and bear only a single ovule, while the more primitive members of the family have prominently biaxial ones with a number of ovules (Quinn; Sporne). The Podocarpaceae have a unique binucleate cellular stage in their early embryogeny, and all but two of the genera have an epimatium (an unusual structure generally interpreted as a modified ovuliferous scale that partially folds around the ovule) as the fleshy structure associated with the seed (Florin, 1958b; Quinn). The aril surrounding the seed of the Taxaceae arises aS an outgrowth at or just below the base of the ovular integument (Coulter & Land; Keng; Loze). Elsewhere among the conifers, a comparable structure occurs only in the podocarpaceous genus Phyllocladus L. & A. Rich. (Sporne), a highly derived group otherwise very dissimilar from the Taxaceae. The Taxaceae and Podocarpaceae also differ in wood anatomy, pollen structure, and chemistry, and there is thus little evidence of a close relationship between the two families (see Florin, 1958b; Hegnauer, 1962, 1986; Phillips). Two tribes are often recognized in the Taxaceae, following the treatment of Janchen. In plants of tribe Torreyeae Janchen, including Torreya and Amen- totaxus, the microsporangia are borne abaxially on the microsporophylls (vs. radially arranged in tribe Taxeae) and the aril tightly invests the mature seed and is largely adnate to the seed coat (Florin, 1948a; Keng). In these two genera the free portion of the aril is displaced toward the apex of the ovule by inter- calary growth, while in the three genera of tribe Taxeae (Taxus, Pseudotaxus, and Austrotaxus) the entire aril remains free from the ovular integument and forms a cuplike structure around the seed (Florin, 1948a). The genera of tribe Taxeae also lack the foliar resin canal present in Torreya, Amentotaxus, and the outgroup Cephalotaxus (Gaussen). Austrotaxus is unique in the family in having the microsporangia partially connate (Saxton; Wilde) and in apparently lacking helical thickenings on the tracheid walls (Greguss, 1955; Phillips). It has recently been found, however, to produce taxane alkaloids, which had previously been known only from 7axus (Guéritte-Voegelein et a/.). Chromosome numbers have been reported for three of the genera of Taxa- ceae. Counts of 2n = 24 have been obtained for four species of Taxus (Dark; Sax & Sax: Sugihara, 1946b). This is presumably the primitive chromosome number for the family, since it has also been obtained for several species of the outgroup genus Cephalotaxus and is apparently the primitive number for the conifers generally (Ehrendorfer; Khoshoo, 1961, 1962). Eleven of the chro- ie) JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 mosomes in 7axus are metacentric or submetacentric, while the shortest one is subtelocentric (Dark; Sax & Sax). The chromosome number 27 = 22 has been obtained from two species of Torreya, T. nucifera (L.) Sieb. & Zucc. (see, for example, Tahara, Terasaka) and 7. californica Torrey (see Gaussen). All of the chromosomes of 7. nucifera are apparently metacentric or submetacentric (Terasaka). The count of 2n = 22 has also been reported for Amentotaxus argotaenia (Hance) Pilger by Su- gihara (1946a), but his illustration is unclear. In contrast, Chuang & Hu pro- vided an excellent illustration of the chromosomes of A. formosana Li (as A. argotaenia), with 2n = 14, equaling the lowest chromosome number known for any gymnosperm (see Khoshoo, 1961). All of the chromosomes in this species are clearly heterobrachial, and two are markedly shorter than the rest (Chuang & Hu). Counts for other populations of 4mentotaxus and for Austro- taxus and Pseudotaxus, as yet unstudied, are much needed. The proembryogeny of the Taxaceae is similar to that of other groups of conifers, including Cephalotaxus and members of the Podocarpaceae (see re- views in Dogra; Doyle, 1963; Doyle & Brennan; Singh, 1978). After cell walls are formed, there is a single upper tier of cells that are open at the top, with an irregularly storied group of cells below. There is a trend among the gymno- sperms toward reduction in the number of free-nuclear mitotic divisions in the proembryo prior to cell wall formation. Eight sets of free-nuclear mitoses are found in Ginkgo, and there are still higher numbers in the cycads, while six sets or fewer occur in the conifers (Dogra; Singh, 1978). There are usually four sets of free-nuclear mitoses in Cephalotaxus, Taxus, Pseudotaxus, Amen- totaxus, and Austrotaxus (Buchholz, 1940: Chen & Wang, 1978, 1984a, b; Saxton; Singh, 1961, 1978; Sterling, 1948; Sugihara, 1946a, b), while in Torreya the number of such divisions has been reduced to two or three sets, yielding a four- or eight-celled proembryo (see review in Doyle & Brennan). Each of these genera frequently exhibits simple polyembryony. Cleavage polyembryony has been documented in Torreya (Chen & Wang, 1984a: Doyle & Brennan), but not in the other genera of Taxaceae and only as a rare event in species of Cephalotaxus (Singh, 1961). The microsporangiate strobili of the Taxaceae are diverse in structure and have been the subject of varied morphological interpretations (see, for example, Florin, 1948a; Sporne; Wilde). In Amentotaxus the strobili are compound structures with some 20 to 30 small strobilar units arranged in a more or less decussate fashion along the axis (Keng; Wilde). The outgroup Cephalotaxus also has compound microsporangiate branches, with each lateral strobilus in the axil of a bract, indicating that a compound strobilus may be primitive in the Taxaceae (Wilde). No bracts are evident subtending the lateral units in Amen- totaxus, each of which bears eight to 12 microsporophylls in A. argotaenia. Austrotaxus also has an unusual spikelike strobilar structure, in which the very reduced sporangia-bearing structures occur in the axils of spirally arranged bracts along the cone axis. Based on the pattern of vasculature, Wilde has interpreted the axillary structures as highly reduced lateral cones in which the sporophyll stalks are virtually absent and the sporophylls and sporangia are partially fused. Pseudotaxus seemingly has simple pollen cones in the axils 1990] PRICE, TAXACEAE 73 of foliage leaves, but two scale leaves are positioned on the axis between whorls of sporophylls (Florin, 1948b; Wilde), possibly indicating reduction from a more complex structure. Taxus has pollen cones rather similar to those of Pseudotaxus (but lacking sterile scales between the sporophylls), while Torreya has simple axillary pollen cones without sterile scales. The radial arrangement of microsporangia in Taxus and Pseudotaxus is very unusual among the conifers and is somewhat reminiscent of the arrangement of sporangia in Equisetum L. (Sporne). Radially arranged microsporangia are also sometimes found at the apices of strobili in Cephalotaxus and Amento- taxus, in which the sporangia are usually fully developed only on the abaxial surface (Wilde). By analogy, Wilde has thus suggested that the sporangia-bearing structures of Taxus and Pseudotaxus may be equivalent to reduced lateral cones. In Torreya, however, the sporangial initials are radially arranged, and those on the adaxial side abort early in development to give the asymmetric arrangement seen at maturity (Coulter & Land). The bark of Taxus and Torreya is unusual in having fibers with prominent crystals of calcium oxalate on the outer cell walls (Chang; Lotova). Crystalli- ferous fibers apparently are also present in Amentotaxus but are absent in Austrotaxus and in the outgroup Cephalotaxus (Outer & Toes). Resin canals are absent in both the bark and the wood of Taxaceae and Cephalotaxaceae (Chang; Phillips; Suzuki, 1979a), although they are present in the leaves and arils of Torreva and Amentotaxus, as well as in the leaves and fleshy seed coat of Cephalotaxus (Keng; Singh, 1961). Individual resinifer- ous parenchyma cells are apparently present in the wood of Cephalotaxus (Greguss, 1972) and also in that of Torreya (Bliss), in accord with the distinctive odor of the wood in several species of the latter. Resiniferous parenchyma cells are evidently absent from the stem wood of Taxus but may be present in the root wood (Bliss). Helical thickenings on the secondary walls of the axial tracheids are a notable feature of the wood in Taxaceae and Cephalotaxus (Greguss, 1955; Penhallow; Phillips). Apparently they are usually absent in Austrotaxus (Gaussen; Greguss, 1955), although according to Phillips they were reported for this genus by Prince. The form and distribution of helical thickenings on the tracheids of Taxaceae are strikingly similar to those of Cephalotaxus but differ from those found in some of the Pinaceae (Penhallow; Yoshizawa et al.). Leaf anatomy has proved particularly useful in differentiating the genera of Taxaceae (Ferguson, 1978; Florin, 1931, 1948b, 1951) and has often allowed unequivocal identification of fossil leaves. Amentotaxus 1s notable for the star- like arrangement of subsidiary cells around the guard cells, with some of the subsidiary cells shared by adjacent stomata. The outer surface of the subsidiary cells is prominently thickened and papillate in Taxus but unthickened in Pseu- dotaxus. The stomata are monocyclic in Torreya and Amentotaxus and are usually so in Pseudotaxus, while they are amphicyclic in Taxus and Austrotaxus. The pollen grains of Taxaceae are nonsaccate and lack prothallial cells, as is also the case in the Cephalotaxaceae, Sciadopityaceae, and Cupressaceae (in- cluding Taxodiaceae) (Erdtman, 1957, 1965; Singh, 1961, 1978; Ueno, 1959, 1960; Wodehouse). Presence of pollen saccae (as in most Pinaceae and Po- 74 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 docarpaceae and several extinct groups of conifers) and two prothallial cells (as in Pinaceae and Ginkgo, with secondarily higher numbers in Araucariaceae and most Podocarpaceae) are apparently the primitive states among the conifers (see Florin, 1951; Millay & Taylor; Singh, 1978). The sculpture and structure of the pollen grains of the Taxaceae are similar in general features to those of the Cephalotaxaceae and Cupressaceae (Owens & Simpson; Ueno, 1959), al- though unique features have recently been ascribed to Amentotaxus (Xi, 1986a, The Taxaceae are wind pollinated, as are virtually all gymnosperms, and have a pollination-drop mechanism of pollen capture, as is the case in Ginkgo and a number of groups of conifers, including the Cephalotaxaceae and Cu- pressaceae (Doyle, 1945; Singh, 1978). The Taxaceae are one of several gymnospermous groups in which dioecy is coupled with a seed-dispersal syndrome involving mammals or birds as vectors (Givnish). Taxus exhibits a typical pattern of characters related to bird dis- persal, with the attracting aril becoming sweet and red when the seeds ripen, while the seed itself is toxic and protected by a hard coat. The large, edible seeds of Vorreya, with their dull purplish, resinous aril, are well suited to mammal dispersal. The seeds of Torreva taxifolia are highly sought after by squirrels (U. S. Fish & Wildlife Service). Field studies are needed to determine the dispersal mechanisms in the other genera, although preliminary morpho- logical and chemical data tend to suggest bird dispersal in the other two genera of Taxeae. Studies of the natural-product chemistry of the Taxaceae have been largely restricted to the genera Taxus and Torreya, with the exception of preliminary studies of the biflavonoid composition of Amentotaxus and Pseudotaxus (see reviews in Geiger & Quinn, 1975, 1982; Hegnauer, 1986) and of a few of the compounds extractable from the wood of Pseudotaxus Chienii (Cheng) Cheng (Ma et al., 1982) and the leaves of Amentotaxus argotaenia (Ma et al., 1986). What is known of the chemistry of the Taxaceae does not provide evidence favoring its treatment as a group separate from the other families of conifers (Hegnauer, 1962, 1986). Only the amentoflavone series of biflavonoids appears to be present in the Taxaceae and Cephalotaxaceae, in contrast to the much greater diversity of biflavonoid structures found in the Araucariaceae, Cupressaceae (including Taxodiaceae), and Podocarpaceae (Geiger & Quinn, 1975: Hegnauer, 1962, 1986). Kayaflavone, the sole biflavonoid reported from four species of Torreya (see Geiger & Quinn, 1975; He et al., 1983: Ma et al., 1985), has not been found in Amentotaxus, Pseudotaxus, or Taxus, while the parent compound amentoflavone is the only biflavonoid reported to date from Amentotaxus. Mono-, di-, and trimethyl ethers of the series, including sciadopitysin, gink- getin, and sequoiaflavone have been isolated from Taxus baccata (M.S. Y. Khan er a/.; Morelli), while only sciadopitysin was reported from several other species of Taxus by Ma and colleagues (1985). Leaf oils have apparently been investigated only in Jorreya, in large part because foliar resin canals are absent in the other genera except for Amento- faxus. The major monoterpene components of Torreya leaf resin, including limonene, a-pinene, and myrcene, are also of wide occurrence elsewhere among 1990] PRICE, TAXACEAE 75 the conifers (He et al., 1986; Hegnauer, 1962, 1986; Yatagai & Sato). A series of more unusual resin sesquiterpenes has also been isolated from the stem wood of Torreya nucifera (see reviews by Burke; Hegnauer, 1986), but only limited comparative studies have been conducted within the genus. Taxus is notable for the highly poisonous taxane alkaloids, an unusual class of diterpene alkaloids that is characteristic of its leaves, stems, and seeds (Lyth- goe; R. W. Miller). Some of these alkaloids, most notably taxol, have been the subject of considerable interest because of their potent antimitotic activity and potential utility as anticancer chemotherapeutic agents (Guéritte-Voegelein ef al.. R. W. Miller). Taxane alkaloids had been considered to be unique to Taxus (Hegnauer, 1986, 1988) but have recently also been found to occur in Austro- taxus (Guéritte-Voegelein er a/.). They are not present in Torreya but need to be thoroughly sought after in Pseudotaxus. A biosynthetically unrelated class of cytotoxic alkaloids is characteristic of Cephalotaxus, and these compounds are also under investigation for their anticancer activity (Chu; Huang & Xue). Alkaloids are of quite restricted distribution among conifers outside of the Taxaceae and Cephalotaxaceae, having been found only in a few members of the Pinaceae and in Athrotaxis D. Don (Cupressaceae sensu lato) (Hegnauer, 1988). Taxus, which is widely grown as an ornamental hedge or tree in North America and Europe, is the only genus of Taxaceae with major economic importance. Torreya is of lesser importance as an ornamental, but it is valued in Asia for its wood, edible seeds, and seed oil. Amentotaxus and Pseudotaxus are of potential horticultural interest, but only 4. argotaenia has been intro- duced into cultivation outside of China (Rushforth). REFERENCES: A.vin, K. L., D. K. FeERGuson, & H. JAHNICHEN. Amentotaxus Pilger from the European Tertiary. Feddes Repert. 86: 379-410. pls. 24-43. 1978. [One extinct species; leaf anatomy of fossil oe extant species compared; range map of extant species from herbarium-label dat ANDRE, D. een one a étude morphologique du cone femelle de quelques gym- nospermes (Cephalotaxées, Juniperoidées, Taxacées). Nat. Monspel. Bot. 8: 3-35 1956.* BaiLey, L. H. The cultivated conifers in North America. ix + 404 pp. 158 pls. New York. 1933. BAILLON, H. Coniféres. Hist. Pl. 12: 1-45. 1894. [Taxaceae, 1-3, 31, 32.] BEAN, W. J. Trees and shrubs hardy in the British Isles. ed. 8. Vol. 4. xv + 808 pp. 108 pls. London. 1980. [Taxus, 560-570; accounts of six cultivated species; Torreya, 611-613, pl. 85.] BENTHAM, G., & J. D. Hooker. Taxaceae. Gen. Pl. 3(1): 431, 432. 1880. Buss, M. C. Interrelationships of the Taxineae. Bot. Gaz. 66: 54-60. pls. 1, 2. 1918 [Purportedly resiniferous axial parenchyma cells in the stem wood of Cephalotaxus and Torreya but only in the root wood of Taxus. ] Britton, N. L. North American trees. x + 894 pp. New York. 1908. [Taxaceae, 122- 128, figs. 95-98. ] Bucuuotz, J. T. The embryogeny of Torreva, with a note on Austrotaxus. Bull. Torrey Bot. Club 67: 731-754. 1940 _ Generic and subgeneric distribution of the Coniferales. Bot. Gaz. 110: 80-91. 1948. 76 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 BuRKE, J. G. Human use of the California nutmeg tree Torreya californica and other members of the genus. Econ. Bot. 29: 127-139. 1975. Butts, D., & J. T. BucHHOoLz. Cotyledon number in conifers. Trans. Illinois Acad. Sci. 33: 58-62. 1940. [Taxus and Torreya usually with two cotyledons; includes original counts and literature summary. } CARRIERE, E. A. Traité général des coniféres. ed. 2. xii + 910 pp. Paris. 1867. [Taxaceae, 723-750; includes account of toxic effects of Taxus. CHAMBERLAIN, C. J. Gymnosperms: structure and evolution. xi + 484 pp. Chicago. 1935. CHANG, Y. P. Bark structure of North American conifers. U. §. Dep. Agr. Tech. Bull. 1095. 11 + 86 pp. 1954. [Taxaceae, 8-12, figs. 1, 2; photomicrographs of Torreya: Taxus and Torreya compared to other North American genera. & F CHEN, Z. K.., . X. Wanc. Embryogeny of Pseudotaxus Chienii in relation to its systematic position. (In Chinese; English summary.) Acta Phytotax. Sinica 16(2): 1-10. pis. J, 978 . oe gametophytes and fertilization in Pseudotaxus Chienii. (In Chinese: English summary.) Acta Bot. Sinica 21: 19-29. pis. J, 2. 1979, & n the systematic position of Amentotaxus from its embryological investigation. (In Chinese; English summary.) Acta Phytotax. Sinica 22: 269-276. pls, 1-3. 1984a & Early embryogeny and its starch distribution in Amentotaxus. (In Chinese; English cea cta Bot. Sinica 26: 359-364. pls. 1, 2. 1984b. ovule structure and development of male and female Go a in. Amentotoxus. ts Chinese; English summary.) /bid. 27: 19-25. pls. 1, 2. 1985a. & ntribution to pollination and fertilization o entra ar- pe es Chinese: English summary.) bid. 239-245. pis. 1, CHENG, W. C., & L. K. Fu, eds. Gymnospermae. Flora Reipublicae ae ees (In ae nomenclatural portions primarily in English and Latin.) Vol. 7. xiv + 544 pp. 1978. [Taxaceae, 437-467; illustrations of Chinese species eae ae Pseudotaxus, Taxus, Torreya. CHowpuury, C. R. The embryogeny of conifers: a review. Phytomorphology 12: 313- 338. 1962. Cuu, T. P. A study of the alkaloids in Cephalotaxus and their bearing on the chemo- taxonomic Os cae a genus. (In Chinese; English summary.) Acta Phytotax. Sinica 17(4): 7-20. CHUANG, T. I, & W. W. ‘ os Study of Amentotaxus argotaenia (Hance) Pilger. Bot. Bull. Acad. Sinica, I. 4: 10-14. 1963. [Chromosome number, n = 7, and karyotype of A. formosana, here included in A. argotaenia.] Coker, W. C., & H. R. Torren. Trees of the southeastern states. ed. 3. viii + 419 pp. 7 pls. Chapel Hill. 1945. [Taxaceae, 52-55, p/. [2].] Compton, R. H. Gymnosperms. Jn: A ena account of the plants collected in New Caledonia and the Isle of Pines. Bot. fee Linn. Soc. 45: 421-434, pis. 26, 27. 1922. ena gen. nov., 427, 428, pi. 2 Coucter, J. M., & C. J. CHAMBERLAIN. Spain of gymnosperms. ed. 2. xi + 466 pp. Chicago. 1917. & W. J AND. Gametophytes and embryo of Torreya taxifolia. Bot. Gaz 39: 161-178. p/. A, I-//T. 1905. [Important account of ovular and microsporangial development, notes on human utilization of the species.] DALLIMoRE, W., & A. B. Jackson. A handbook of Coniferae and Ginkgoaceae. ed. 4 (revised by S. G. HARRISON). xix + 729 pp. 46 pls. London. 196 Dark, S. O. S. Chromosomes of Taxus, Sequoia, Cryptomeria, and Thuya. Ann. Bot. Il. 46: 965-977. 1932. De LAusenFeLs, D. J. The external morphology of coniferous leaves. Phytomorphology -20. 1953 1990] PRICE, TAXACEAE a . Gymnospermes. Flore de la Nouvelle Calédonie et dépendances 4: 1-168. 1972. [Taxaceae, | 1-14; range map and original illustrations, . Coniferales. Fl. Malesiana, I. 10(3): 337-453. 1988. [Taxaceae, 347-351; world- wide range map for Taxus and detail map for the Malesian 7. sumatrana, nomen- clature of 7. swmatrana clarified. ] Docra, P. D. Embryogeny of gymnosperms and taxonomy—an assessment. Glimpses Pl. Res. 5: 114-128. 1980. Doy te, J. Developmental lines in pollination mechanisms in the Coniferales. Sci. Proc. Roy. oes Soc. II. 24: 43-62. 1945. Proembryogeny in Pinus in ag to that in other conifers—a survey. Proc. Roy. Irish Acad. B. 62: 181-216. 3. & M. BRENNAN. Cleavage pee in conifers and taxads—a survey. I. Podocarps, taxads & taxodioids. Sci. Proc. Roy. Dublin Soc. A. 4: 57-88. 1971. [See also ibid. 4: 137-158. 1972.] EHRENDORFER, F. Evolutionary significance of chromosomal se ee patterns in gymnosperms and primitive angiosperms. Pp. 220-240 in C. B. K, ed., Origin and early evolution of angiosperms. iii + 341 pp. New York. crm EICHLER, A. W. Coniferae. Jn: A. ENGLER & K. PRANTL, Nat. Pflanzenfam. II. 1: 28- 116. 1889. anes 111-113. EMBERGER, L. Gymnospermes. Pp. 383-459 in M. CHADEFAUD & L. EMBERGER, Traité de botanique systématique. Vol. 2(1). Paris. 1960. [Taxaceae, 452-459, placed in the order Taxales along with the Cephalotaxaceae. ErpDTMAN, G. Pollen and spore morphology/plant taxonomy. Vol. 2. Gymnospermae, ee and Bryophyta (illustrations). Frontisp. + iv + 151 pp. 5 pls. Stock- holm. 1957. [Taxaceae, 8, 9, 31, 42, 43.] Vol. 3. eee her ee a 191 pp. oy pls. Stockholm. 1965. [Taxaceae, 17, 21, 70, 75, 78; pls. 24.] FERGUSON, D. K. Some current research on fossil and recent phon a ot Palynol. 26: 213-226. 1978. [Leaf anatomy of genera of Taxaceae plus Cephalotaxus, new i on leaves of Amentotaxus, range map of North American species of Taxus. ] w species of Amentotaxus (Taxaceae) from northeastern India. Kew Bull. 40: Is es 1985. [4. assamica. Ferré, Y. DE, M. L. Rouane, & P. Woxtz. Systématique et anatomie comparée des feuilles de Taxaceae, Podocarpaceae, Cupressaceae de Nouvelle-Calédonie. Cahiers Pacifique 20: 241-266. 1978. [Leaf anatomy of Austrotaxus spicata.] FitscHEN, J. Handbuch der Nadelholzkunde. xv + 765 pp. Berlin. 1930. [Revised third edition of Beissner’s Handbuch der Nadelholzkunde.] Florin, R. Untersuchungen zur Stammesgeschichte der Coniferales und Cordaitales. Sv. Vet.-akad. Handl. III. 10: 1-588. p/s. 1-58. 1931. [Classic treatment of leaf anatomy of the conifers; Taxaceae, 68-74, 235-242, 304-306, 465-468, 474-476, pls. 10, 21, 24-26, 33, 34.] . Die Koniferen des Oberkarbons und des unteren Perms. I-VIII. Palaeonto- graphica 85B: 1-729. p/s. 1-186, table 1. 1938-1945. [Taxaceae, 515-518, 618-632, 657-662, 668, pls. 185, 186; morphology of fossil and extant taxa compared. ] . On the morphology and relationships of the Taxaceae. Bot. Gaz. 110: 31-39. 1948a. n Nothotaxus, a new genus of the Taxaceae, from eastern China. Acta Horti Berg. 14: 385-395. pls. 1-3. 1948b. [Pseudotaxus. | . Nothotaxus or Pseudotaxus? Bot. Not. 1948: 270-272. 1948c. . Evolution in cordaites and conifers. Acta Horti Berg. 15: 285-388. 1951. [Leaf- surface anatomy of Taxus, ee xus, Amentotaxus, and Cephalotaxus compared; Taxaceae excluded from the conife The female reproductive a ions and taxads. Biol. Rev. 29: 367-389. 1954. The systematics of the gymnosperms. Pp. 323-403 in E. L. KesseL, ed., A 78 JOURNAL OF THE ARNOLD ARBORETUM [vov. 71 century of progress in the natural oo 1853-1953. San Francisco. 1955. [His- torical review, mostly of morpho n Jurassic taxads and ae from northwestern Europe and eastern Green- land. Acta Horti Berg. 17: 257-402. a I-56. 1958a. [Several extinct species of Torreya compared to modern specie Notes on the systematics of the i clocussecse Ibid. 403-411. 1958b. [Podo- carpaceae compared to Taxaceae and Cephalotaxaceae The distribution of conifer and taxad genera in time and space. /bid. 20(4): 121-312. 1963. [Maps of current and fossil distribution for all genera of Taxaceae.] Fu, L. K. A study on the genus Cephalotaxus Sieb. et Zucc. (In Chinese; English sum- mary.) Acta Phytotax. Sinica 22: 227-288. 1984. [Revision of the genus; two sections and nine species recognized. ] Gaussen, H. Les Taxines. /n: Les gymnospermes actuelles et fossiles. Trav. Lab. Forest. Toulouse, Tome II, Vol. I, Partie 1(3), Fasc. XV. Chap. XXIII: 1-24. pls. 98-100; Chap. XXIV: I-10; Chap. XXV: I-82. p/. 101. 1979. [Worldwide treatment of the Taxaceae.] Bibliographie. /bid. Tome II, Vol. I, Partie III: 1-180. 1979. [Bibliography for the entire work.] GelGcer, H., & C. QuINN. Biflavonoids. Pp. 692-742 in J. B. HARBORNE, T. J. MABRY, & H. Masry, eds., The flavonoids. Vol. 2. London. 1975. [Taxaceae, 723, 732, 733. See also updated cases J. B. HARBORNE & T. J. MAsBry, eds., The flavonoids: advances in research. Vol. 1. pp. 505-534. London and New York. 1982. Vol. 2. Chap. 4. 1988.* GIvNIsH, T. I. Ecological constraints on the evolution of breeding systems in seed plants: dioecy and dispersal in gymnosperms. Evolution 34: 959-972. 1980. [Dioecy coupled with animal dispersal of seeds in wind-pollinated plants.] Grecuss, P. Identification of living gymnosperms on the basis of xylotomy. 263 pp. 350 pls., 8 tables. Budapest. 1955. [Taxaceae, 147, 148, 150-157, pls. 73, 78-89. | . Xylotomy of the living conifers. 172 + 329 pp. Budapest. 1972. [Taxaceae, 18, 19, 85; pls. 7, — ood of Taxus floridana illustrated and discussed. ] GUERITTE-VOEGELEIN, F., D. GUENARD, & P. PoTieR. Taxol and derivatives: a biogenetic hypothesis. ie Nat. Prod. 50: 9-18. 1987. [Possible eae pathways to the taxane structure; first report of taxane alkaloids from Austrotaxus spicata. ] Harpin, J. W. Studies of the southeastern United States flora. II. The gymnosperms. Jour. Elisha Mitchell Sci. Soc. 87: 43-50. 1971. [Taxaceae, 44.] Harris, T. M. Two neglected aspects of fossil conifers. Am. Jour. Bot. 63: 902-910. 1976a. [Abscission of foliage and phyllotaxis in modern Taxaceae compared to fossil material. ] T he ozoic gymnosperms. Rev. Palaeobot. Palynol. 21: 119-134. 1976b. ee of Jurassic fossil Taxaceae reviewed; possible mechanisms of shift from lateral to terminal ovules discussed. Hart, J. A. A cladistic analysis of conifers: preliminary results. Jour. Arnold Arb. 68: 269-307. 1987. [Both intergeneric and interfamilial cladograms; character states tabulated for the genera of Taxace He, G. F., Z. W. Ma, & W. F. Yin. Biflavone of Torreya Jackii Chun and its taxonomical significance. (In Chinese; English summary.) Acta Phytotax. Sinica 21: 433-435. 1983. [Kayaflavone found in several species of Torreya; apparently absent in Amen- totaxus, Po i a XUS. Zs, , J. G. PAN, . a C. ZHu. Studies on essential oil composition in leaves = sae grandis cv. Merrillii and chemotaxonomy. (In Chinese; English summary.) Acta ioe ee 24: 454-457. 1986. [Main com- ponents are limonene, a-pinene, 5-3-ca HEGNAUER, R. Chemotaxonomie der Pflanzen, Bd. 1. Thallophyten, Bryophyten, Pteri- dophyten und Gymnospermen. 517 pp. Basel & Stuttgart. 1962. — 430- 440, 482.] Bd. 7. Nachtrage zu Band | und Band 2. xii + 804 pp. 1986. [Taxaceae, 540-546, 802; extensive literature review. ] 1990] PRICE, TAXACEAE 19 _ Biochemistry, distribution and taxonomic relevance of higher plant alkaloids. Jour. Nat. Prod. 27: 2423- ee 1988. [Diterpene alkaloids not known from gym- nosperms outside of the Taxace HICKEL, R. Graines a plantules ae conifres. Bull. Soc. Dendrol. France 19: 13-115; 20: 134-204. 19 Hitt, T. G., & E. DE ae On the seedling structure of gymnosperms. I. Ann. Bot. 22: 689- 712. pl. 35. 1908. [Species of Taxus and Torreya compared to members of Cephalotaxaceae, Podocarpaceae, and Cupressaceae. ] 5 Neti _Yao. Transfusion tissue of gymnosperm leaves. Jour. Linn. Soc. Bot. 83: 263- 272. 1981. [Similar in Taxus, Pseudotaxus, Torreya, Amentotaxus, and Cephalotaxus.] _A., H. X. W x Cc <0 iS ~ > Z fo OQ — 5 u, & Z. X. QIAN. Biochemical systematics of gymno- 260-263. pl. 1. 1986. [Major seed protein bands of Taxaceae and Cephalotaxaceae are quite similar.] Huan, L., & Z. Xue. Cephalotaxus alkaloids. Alkaloids 23: 157-226. 1984 ee G. R., & E. H. Rep. The leaf lipids of some conifer species. Phytochemistry 1: 269-275. 1972. [Fatty-acid profile of Taxus baccata relatively similar to that in a genera of Cupressaceae. ] JANCHEN, E. Das System der Koniferen. aes ber. Akad. Wiss. Math.-Naturw. Wien 158(1): 155-262. 1949 [195 0]. [Taxacea ided into tribes Torreyeae and Taxeae. | Keno, H. Aspects of morphology of Amentotaxus formosana with a note on the taxo- nomic position of the genus. Jour. Arnold Arb. 50: 432-448. 1969. [Leaves, mi- crosporangiate strobili, ovules, seeds, and seedlings illustrated.] Kuan, A. G., & P. G. VALDER. Occurrence of root nodules in Ginkgoales, Taxales, and Coniferales. Proc. Linn. Soc. New S. Wales 97: 35-41. 1972. [Endophytic mycor- rhizae but no nodules in Taxus baccata; no other genera of Taxaceae examined.] Kuan, M. S. Y., I. Kumar, J. S. PRASAD, G. R. NAGARAJAN, M. R. PARTHASARATHY, & H. G. KRISHNAM uRrTY. Phenolic constituents of Taxus baccata leaves. P|. Med. 30: 82-85. 1976. [Betuloside and three biflavonoids isolated.] KuosHoo, T. N. Chromosome numbers in gymnosperms. Silvae Genet. 10: 1-9. 1961. [Comprehensive summary of the literature. ] —. Cytogenetical evolution in the gymnosperms—karyotype. Pp. 119-135 in P. MAHESHWARI, B. M. Jourt, & I. K. VAsIL, eds., Proceedings of the Summer School of Botany, Darjeeling. New Delhi. 1962. KRAL, R. A report on some rare, threatened, or endangered forest-related vascular plants of the South. Vol. 1. Isoetaceae through Euphorbiaceae. U.S. Dep. Agr. Forest ae Southern Region Tech. Publ. R8-TP2. x + 718 pp. 1983. [Taxaceae, 24—- 30; m phology, ecology, distribution of Taxus floridana and Torreya taxifolia.] KRrussMANN, G. Handbuch der Nadelgehélze. ed. 2 (revised by H. D. Warpa). [6] + 396 pp. 160 pls. Berlin and Hamburg. 1983. [Also published in English as KrUss- MANN, G., Manual of cultivated conifers. (M. E. Epp, translator.) 361 pp. /60 pls. Portland, Oregon. 1985. Li, H. L. The genus 4mentotaxus. Jour. Arnold Arb. 33: 192-198. 1952. [Three new species described. LITTLE, E. ~ Jr. Atlas of United States trees. Vol. 1. Conifers and important hardwoods. Dep. Agr. Misc. Publ. 1146. v + 9 pp. 2 base maps, 200 species maps, 9 overlay maps. 1971. Vol. 5. Florida. vi + 22 pp. 6 general maps, 256 species maps. 1978. [Detailed distributional maps for Taxus and /orreya The Audubon Society field guide to North America trees. Eastern region. 714 p. New York. 1980. [Taxaceae, 273-275.] tera L. I. The bark structure of Taxaceae, Cephalotaxaceae, and Podocarpaceae. a oe English summary.) Vestn. Moskov. Univ. Ser. 6. Biol. 1979(3): 3-13. 80 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Loze, J.C. Etude de l’ontogénie de l'appareil reproducteur femelle de l’if, Taxus baccata L. Revue Cytol. Biol. Veg. 28: 211-256. pls. I-3. 1965. [Detailed study of the development of the ovule, aril, and ovuliferous shoot.] Lytucog, B. The Taxus alkaloids. Alkaloids 10: 597-626. 1968. Yin. Study of major chemical components of Pseudotaxus " Chienii (In ‘Chinese: Fnglich summary.) Acta Bot. Sinica 24: 554-557. 1982. [Three crystalline compounds extracted from the wood, including 6-sitosterol and tsuga- lactone. ] ——.. The distribution of bisflavones in Taxaceae. (In Chinese; English summary.) Act Phytotax. Sinica 23: 192-195. 1985. [Kayaflavone found in several s es of Torreya and sciadopitysin in several species of Taxus: no biflavonoids ee in Pseudotaxus or Amentotaxus; see also GEIGER & QUINN.] ,& tudy on chemical components in leaves of Amentotaxus argotaenia (Hance Pilger native to China. (In Chinese; English summary.) /bid. 24: 443-446, MAHESHWARI, “i : H. SinGH. The female gametophyte of gymnosperms. Biol. Rev. 42: 88-130. 1967. Martin, A.C. The comparative internal morphology of seeds. Am. Midl. Nat. 36: 531— 660. 1946. Heeree es 530. McKNIGHT, B. A. Refugees from the ice age. Am. Forests 53: 543, 566. 1947. [Taxus Noridana, Torreya taxifolia; includes historical account and bark and habit photo- graphs of the latter.] Meura, P. N., & T. N. KHosHoo. Cytology of conifers. II. Jour. Genet. 54: 181-185. 1956. [Chromosome counts for Cephalotaxus drupacea. Taxaceae and Cephalotax- aceae compared. a Mitiay, M. A., & T. N. Taytor. Evolutionary trends in fossil gymnosperm pollen. Rev. Palaeobot. aon 21: 65-91. 1976. MILter, C. N., JR. Mesozoic conifers. Bot. Rev. 43: 217-280. 1977. [Taxaceae, 230- 8.] Micter, H. J. The wood of Amentotaxus. Jour. Arnold Arb, 54: 111-119. 1973. Miter, R. W. A brief survey of Taxus alkaloids and other taxane derivatives. Jour. Nat. Prod. 43: 425-437. 1980. More.ut, I. Constituenti di Taxus baccata L. Fitoterapia 47: 31-38. 1976.* [Chemical constituents of 7. baccata. Oxtver, F. W. The ovules of the older gymnosperms. Ann. Bot. 17: 451-476. pl. 24. 1903. [Ovule anatomy of Torreya californica, 466-473, figs. 10-14.] OupEN, P. DEN, & B. K. Boom. Manual of cultivated conifers. x + [2] + 526 pp. The Outer, R. W. pen, & E. Toes. The secondary phloem of Amentotaxus. Jour. Arnold Arb. 55: 119-122. 1974, [Phloem anatomy of the genera of Taxaceae (except Pseu- ara and see compared. Owens, J. N., & S. Simpson. Pollen from conifers native to British Columbia. Canad. Jour. Forest R Res. 16: 955-967. 1986. [SEM and LM photographs and descriptions: Taxus brevifolia included.] PARLATORE, P. Coniferae. Jn’ DC. Prodr. 16(2): 361-521. 1868. [Taxaceae, 499-502, 504-506. ] PENHALLOW, D. P. The generic characters of the North American Taxaceae and Conif- erae. Trans. Roy. Soc. Canada, II. 2(4): 33-57. pls. 1-6. 1896. [Wood anatomy of ‘axus and Torreya PHILLIPS, E. W. J. iden neation of softwoods by their microscopic structure. Forest Prod. Res. Bull. 22. iii + 56 pp. 3 pls. 1948. [Wood anatomy for all genera except Pseudotaxus compared; a similar account appeared in Jour. Linn. Soc. Bot. 52: 59— 320. 1941.] PiGeR, R. Taxaceae. Pflanzenr. IV. 5: 1-124. 1903. [Including Cephalotaxaceae and Sie eae the species of a us treated as geographic subspecies of 7. baccata.] 1990] PRICE, TAXACEAE 81 . Die Taxales. Mitt. Deutsch. Dendrol. Ges. 25: 1-28. 1916a. . Kritische Ubersicht iiber die neuere Literatur betreffend die Familie der Tax- aceac. Bot. Jahrb. 54: 1-43. 1916b. [Amentotaxus separated from Cephalotaxus as a new sas ] —. Taxaceae. Jn: A. ENGLER, Nat. Pfla sea ed. 2. 13: 199-211. 1926. [Amen- totaxus maa under Cephalotaxaceae, 270, 271.] . MELCHIOR. Gymnospermae. /n: H. Sao A. Engler’s Syllabus Pflan- zenfam. ed. 12. 1: 312-344. 1954. [Taxaceae, 340, 341. Prince, J. B. Stem-wood in the eymnosperms: 96 pp. Unpubl. Master’s Thesis, Univ. New Brunswick. Fredericton. 1938.* Quinn, C. J. Generic nee in the Podocarpaceae. Proc. Linn. Soc. New S. Wales 94: 166-172. 197 ReHper, A. Manual - cultivated trees and shrubs hardy in North America, exclusive of the subtropical and warmer temperate regions. ed. 2. xxx + 996 pp. New York. 1940. [Taxaceae, 1-5. Bibliography of cultivated trees and shrubs hardy in the cooler temperate regions of the Northern Hemisphere. xl + 825 pp. Jamaica Plain, Massachusetts. 1949. [Detailed nomenclatural treatment; Taxaceae, 1—-4.] RusHFortH, K. D. Conifers. 232 pp. 20 pls. London. 1987. SAHNI, B. On certain archaic features . the seed of Taxus baccata, with remarks on the antiquity of the Taxineae. Ann. Bot. II. 34: 117-133. 1920. SARGENT, C. S. Silva N. Am. 10. vii + . pp. pls. 497-537. 1896. [Taxaceae, 55-68, pls. 512-515; sea illustrations of Taxus and Torreya; uses and lore Sax, K., & H. J. SAx. Chromosome number and morphology in the conifers. Jour. Arnold Arb. 14: 356-374. pis. 75, 76. 1933. [Chromosome numbers and karyotypes for species of Taxus and Cephalotaxus.] SAXTON, W. T. Notes on conifers. VIII. The morphology of Austrotaxus spicata Comp- ton. Ann. Bot. 48: 411-427. 1934. [Embryogeny; ovule development; microspor- angiate strobilar morphology. ] SincH, H. The life history and systematic position of Cephalotaxus drupacea Sieb. et Zucc. Phytomorphology 11: 153-197. 1961. [Important account of the morphology of Cephalotaxus, Taxaceae and Se aaa compared _ Embryology of gymnosperms. Band 10, Teil 2 in K. LinspAuer, Handbuch der Pflanzenanatomie. ed. 2. Berlin. 1978. [Male and female gametophytes; pollination mechanisms; embryogeny; extensive bibliography. ] SMALL, J. K. Man ep of the southeastern flora. xxii + 1554 pp. New York. 1933. [Taxaceae, 11, 12.] Sporn_E, K. R. The i sheies of gymnosperms. ed. 2. 216 pp. London. 1974. [Tax- STERLING, C. Pro mbryogeny and early embryogeny in Taxus cuspidata. Bull. Torrey Bot. Club 75: 469-485. 1948. . Structure of the male gametophyte in gymnosperms. Biol. Rev. 38: 167-203. 1963 SucimarA, Y. Morphologische ae, liber Amentotaxus argotaenia Pilger. 8 Japanese; German summary.) Bot. Mag. Tokyo 59: 61-67. 1946a. [Chromosome ount, n = 11; illustration ee unclear.] e€ proembryoge ny of Taxus Wallichiana Zuccarini. (In Japanese; English summary. ) Ibid. 96-98. 1946b. [Proembryogeny and chromosome complement (71 = ]2) illustrated.] Suzuki, M. The course of resin canals in the shoots of conifers. 1. Taxaceae, Cephalo- taxaceae, and Podocarpaceae. Bot. Mag. Tokyo 92: 235-251. 1979a. [Resin canals absent in Taxus, associated with the leaf traces in Torreya and Cephalotaxus, see also ibid. 92: 253-274. 1979b; ibid. 92: 333-353. 1979c.] Tanara, M. Embryogeny of Torreya nucifera S. et Z. Sci. Rep. Tohoku Imp. Univ. IV. 15: 419-426. pls. 1, 2. 1940. [See also ibid. 17: 9-16. pl. 1. 1942.] 82 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 TERASAKA, O. Nuclear differentiation of male gametophytes in gymnosperms. Cytologia 47: 27-46. 1982. [Chromosome number, 1 = 11, and gametophytic development in Torreya nucifera. | TIEGHEM, P. vAN. Anatomie comparée de la fleur femelle et du fruit des Cycadées, des Coniféres et des Gnetacées. Ann. Sci. Nat. Bot. V. 10: 269-304. pls. 13-16. 1869. Ueno, J. Some palynological observations of Taxaceae, Cupressaceae, and Araucari- aceac. Jour. Inst. Polytech. Osaka Univ. D. 10: 75-87. 1959. [LM and TEM com- parisons of pollen structure.] . Studies on pollen grains of Gymnospermae: concluding remarks to the rela- Hoshi between Coniferae. /bid. 11: 109-126. pl. 1, foldout table. 1960. U. S. H & WILDLIFE Service. Florida torreya (Torreya taxifolia) recovery plan. 42 pp. Sree Georgia. 1986 VIERHAPPER, F. Entwurf eines neuen Systems der Coniferen. Abh. Zool.-Bot. Ges. Wien 5(4): 1-56. 1910. WANG, F. X., Z. K. CHen, & Y. S. Hu. On the systematic position of Taxaceae from the embryological and uerora studies. (In Chinese; English summary.) Acta Phytotax. Sinica 17(3): 1-7. 1979. Witpe, M. H. A new is of microsporangiate cones in Cephalotaxaceae and Taxaceae. Phytomorphology 25: 434-450. 1975. [Perisporangiate microsporophyll of Taxus may result from microstrobilar reduction.] WobeHousE, R. P. Pollen grains. xv + 574 pp. New York. 1935. [Taxaceae, 249-252, 282, 283. Wo tz, P., & Y. BAILLY. Austrotaxus spicata Compton, Taxacée de Nouvelle-Calédonie: aspects anatomiques et évolution de l'appareil conducteur de la plantule. Bull. Soc. Bot. France 129: 223-230. 1982. [Includes photographs of seed and seedling, illus- trations of leaf cross-sectional anatomy.] X1, Y. Z. Studies on pollen morphology of Taxaceae of China. (In Chinese; English summary.) Acta Phytotax. Sinica 24: 247-252. pls. 1-4. 1986a. cee is Pseu- dotaxus, Taxus, and Torreya compared at LM, SEM, and TEM levels.] . Ultrastructure of pollen exine in Amentotaxus Pilger and =e significance in taxonomy. (In Chinese; English summary.) /bid. 439-442. pls. 1, 2. 1986b. [Treat- ment as monogeneric family suggested based on unusual features of pollen structure. ] YATAGAI, M., & T. Sato. Terpenes of leaf oils from conifers. Biochem. Syst. Ecol. 14: 469-478. 1986. [Includes Torreya nucifera.] Yosuizawa, N., T. Iron, & K. SHimasi. Helical thickenings in normal and compression wood of some softwoods. IAWA Bull. 6: 131-138. 1985. [Pattern of helical thick- enings on tracheid walls of Taxus, Torreya, and Cephalotaxus differs from that in the Pinaceae. ] KEY TO THE GENERA OF TAXACEAE IN THE SOUTHEASTERN UNITED STATES General characters: dioecious (rarely monoecious) evergreen shrubs or trees; foliage leaves alternate, appearing 2-ranked [or less oon opposite], linear to linear-lan- ceolate, short-petiolate; pollen cones with whorled o r tightly clustered microsporophylls; each sporophyll with [2 or] 3-9 abaxial or ae rranged microsporangia; pollen nonsaccate, lacking prothallial cells; ovules bor nearae terminating short axillary shoots; seeds arillate, with hard coat; cotyledons 2 (rarely 1 or 3). A. Seed ca. 2-5 cm long, completely surrounded by a greenish or purplish aril: leaves sharply pointed aromatic when bruised, with a resin canal. .......... 1. Torreya. A. Seed < m long, partially enclosed by a cup-shaped ae aril; leaves not sharply pointed or oa aromatic, without a resin canal. ................... 2. Taxus. 1990] PRICE, TAXACEAE 83 Tribe TORREYEAE Janchen 1. Torreya Arnott, Ann. Nat. Hist. 1: 130. 1838, nom. cons. Evergreen dioecious (rarely monoecious) trees with branches seemingly in whorls. Bark fissured. Wood with little or no axial parenchyma, persistently aromatic. Foliage leaves spirally arranged (appearing 2-ranked due to twisting of the leaf bases), linear to linear-lanceolate, sharply pointed at apex, strongly aromatic when bruised [or slightly so in 7. grandis]; abaxial surface with 2 narrow, whitened [or sometimes brownish] stomatal bands with sunken sto- mata; resin canal single, abaxial to the vascular bundle. Pollen cones short- stalked, borne singly in leaf axils of the current year; microsporophylls in ca. 6-8 whorls of 4 per strobilus, distal to a series of tightly clustered decussate scale leaves: sporangia abaxial, usually 3-5 per sporophyll. Ovules borne singly, terminating short secondary lateral shoots, generally with 2 such ovuliferous shoots borne in the axils of scale leaves on a primary dwarf shoot axillary to a foliage leaf, usually only 1 ovule maturing per associated foliage leaf. Seeds ca. 2~5 cm long, surrounded by the resinous greenish or purplish aril, maturing the second year; seed coat heavily sclerified, adnate to the surrounding aril except at the distal tip; gametophytic storage tissue irregularly channeled (“ru- minate”); embryo distal, very small when seed is shed. Chromosome number 2n = 22. (Tumion Raf.) Type species: Jorreya taxifolia Arnott. (After John Torrey, 1796-1875, noted American botanist.)— TORREYA, STINKING CEDAR. A genus of seven species in moist temperate areas of eastern Asia and North America, with two native to the United States (Yorreya californica Torrey in California and 7. taxifolia Arnott in Florida and adjacent Georgia), one to Japan (T. nucifera (L.) Sieb. & Zucc.), and four to central and southern China (see the recent treatment in Cheng & Fu) and adjacent Burma (Florin, 1963). Torreya had a more extensive Northern Hemisphere range during Mesozoic and Tertiary times. The oldest fossils of the genus, from Great Britain and southern Scandinavia, are of middle Jurassic age, and the genus persisted in central and southern Europe until the Pliocene (Florin, 1958a, 1963). Torreya most closely resembles 4mentotaxus (and Cephalotaxus) in having large, drupelike seeds and elongate leaves with a single resin canal. Torreya differs from these genera and the remainder of the Taxaceae in having narrower stomatal bands with sunken stomata on the abaxial leaf surface (Ferguson, 1978; Florin, 1931), a pungent aromatic odor when its foliage is bruised, simple rather than compound microsporangiate strobili, and highly channeled (“rumi- nate”) gametophytic storage tissue in its seeds. The last feature is unique among the gymnosperms and is caused by the irregular growth of megagametophytic tissue into the surrounding nucellar and integumentary tissues (Coulter & Land). It has given rise to the name ‘‘California nutmeg” for Torreya californica. Torreya is apparently not divisible into clearcut infrageneric groups sup- ported by multiple characters but has been divided into sect. NUCIFERAE Hu, with only slightly channeled seed-storage tissue (7. nucifera and T. grandis Fortune only), and sect. TORREYA (sect. Ruminatae Hu), with prominent chan- neling of the seed tissue (Cheng & Fu). JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 tue Ge as uw OA —< ASS erik 1989 Wr / Me A ont Tribe Torreyeae. a-m, Torreya taxifolia: a, leafy shoot, x 4; b, detail of , shoots bearing microsporangiate strobili, x 4; d, detail of x 53e, f, microsporophyll i in abaxial 1, dry seed with aril, FIGURE 1, leaf in abaxial view, x 14; ¢ microsporangiate strobilus with subtending bracts and adaxial views, 20; g, shoot bearing matur re seed, x 4h longitudinal and radial views, x '4; j, k, seed with aril a longitudinal and radial x 4: 1, longitudinal section of mature seed 5 “ruminate” gametophytic tissue, x %; m, cross section of moist arillate seed, peels resin canals in the aril and “ruminate” gametophytic tissue, x 3/. Our sole species, Torreya taxifolia Arnott (Tumion taxifolium (Arnott) Greene), Florida torreya, stinking cedar, gopherwood, is a very narrow endemic restricted to moist, wooded slopes and limestone bluffs in the vicinity of the Apalachicola River in three counties (Gadsden, Liberty, Jackson) of north- 1990] PRICE, TAXACEAE 85 western Florida and in adjacent Decatur County, Georgia (Kral; Little, 1978; Stalter & Dial). The species was already significantly diminished in population size by the early 1900’s because of cutting for lumber and fenceposts (Britton; Coulter & Land) and is now critically endangered because of the effects of a fungal blight. It is a listed endangered species at both the state and federal levels (U. S. Fish & Wildlife Service). At some point between 1954 and 1962, fungal disease(s) of the leaves and stems began to attack the native populations (Godfrey & Kurz). Several fungal pathogens have been isolated from diseased trees (Alfieri ef a/.; U. S. Fish & Wildlife Service), but the causative agent is still unknown. Drought and habitat degradation may have played a significant role in making the wild trees less resistant to fungal attack (Savage; Stalter & Dial). By 1981, essentially all wild individuals had been infected and all of the wild trees over 3 m tall had died back except for stump sprouting (Stalter & Dial). Only approximately 100 wild individuals were alive as of 1981. Further propagation of cultivated trees outside the native range will be needed to avoid the extinction of the species. Torreya taxifolia differs from the other North American species, 7. califor- nica, in having yellowish brown rather than reddish brown second-year twigs and generally shorter leaves (2.5-4 cm vs. 3-8 cm) that are less flat on the adaxial surface. The stomatal bands are also less deeply sunken into the leaf in T. taxifolia than in T. californica and the other species of the genus (Kriiss- mann). The two species also evidently differ in their volatile oil chemistry, as indicated by differences in the odor of the bruised stems and foliage, described as pungently aromatic in T. californica and foul-smelling in 7. taxifolia (Kriiss- mann; Sargent). No systematic investigation of crossability in Jorreya has been conducted. An apparently seta hybrid of T. californica and T. nucifera was men- tioned by Kriis The various species on Torreya are attractive ornamentals in cultivation. Torreya californica is the hardiest and most widely cultivated species of the genus in North America and Great Britain, while 7. nucifera and T. grandis are grown primarily in their native countries in Asia, where several cultivars have been selected (Bean; Kriissmann). The large seeds of several species are highly esteemed food items. Seeds of 7. californica were gathered by a number of tribes of native Americans, while those of 7. nucifera are both eaten and used as a commercial source of cooking oil in eastern Asia (Burke). The very durable wood of Torreva has been valued for furniture, cabinetry, and fence posts (Burke; Dallimore & Jackson), and thus few large trees remain outside of cultivation. REFERENCES: Under family references see BEAN; BRITTON; BUCHHOLZ, 1940; BURKE; BuTTs & BUCHHOLZ, CHANG; CHENG & Fu; CHOWDHURY; COKER & TOTTEN; COULTER & LAND; DALLIMORE & JACKSON; DoyLe, 1963; ERDTMAN, 1957, 1965; FERGUSON, 1978; FLOR, 1931, 1948a, 1954, 1958a, 1963: GAUSSEN; GREGusS, 1955, 1972: He et al., 1983, 1986; HEGNAUER, 1962, 1986; KHosHoo, 1961; KRAL; KRUSSMANN; LiTTLe, 1971, 1978, 1980; Ma et al., 1985; MAHESHWARI & SINGH; MCKNIGHT; PHILLIPS; PILGER, 1903, 1916a, 86 JOURNAL OF THE ARNOLD ARBORETUM [vov. 71 1916b, 1926; PILGER & MELCHIOR; REHDER, 1940, 1949: oe TAHARA; TERASAKA; U.S. Fish & WILDLIFE SERVICE; WILDE; and YATAGAI & SAT AFier!, S. A., JR., A. P. MARTINEz, & C. WEHLBURG. Stem and needle blight of Florida torreya (Torreva taxifolia Arn.). Proc. Florida State Hort. Soc. 80: 428-431. 1967.* [First account of fungal pathogens putatively involved in the blight of Torreya.] Barnes, L. Micro-propagation of endangered native conifers Torreya taxifolia and Taxus floridana. Hort. Sci. 18: 617. 1983. CHAPMAN, A. W. Torreya taxifolia Arnott; a reminiscence. Bot. Gaz. 10: 251-254. 1885. [Including distribution map. Cuick, E. The seedling of Torreya myristica. New Phytol. 2: 83-91. pls. 7, 2. 1903. [T. californica. FAvRE-DucuHartreE, M. Contribution 4 l'étude de la reproduction sexuée chez le Torreya californica. Compt. Rend. Acad. Sci. Paris, D. 255: 3208-3210. 1962. [See also ibid. 258: 661-664. 1964; ibid. 264: 574-576. 1967.] Goprrey, R. K., & H. Kurz. The Florida torreya destined for extinction. Science, II. 136: 900- 902. 1962. [T. taxifolia.] He, G. F., Z. W. Ma, & W. F. Yin. A new diterpenoid component torreyagrandate from leaves of Torreya grandis Fort. endemic in China. (In Chinese; English summary.) Acta Bot. Sinica 27: 300-303. 1 Hu, Y. S., L. Q. Guan, & Z. X. TANG. Anatomy of the secondary phloem and the crystalliferous fibers in the stem of Hae grandis. (In Chinese; English summary.) Acta Bot. Sinica 27: 569-575. pls. 1-3. 1985. Jou, Y. G., K. S. Bark, & T. M. ane es on the composition and chemical structure of desmethylsterols from Torreya nucifera seeds. Korean Jour. Food Sci. Tech. 13: 127-132. 1981.* Kemp, M. Morphological and ontogenetic studies on Torreya californica Torr. 1. The vegetative apex of the megasporangiate tree. Am. Jour. Bot. 30: 504-517. 1943: II. Development of the megasporangiate shoot prior to pollination. /bid. 46: 249-261. 959. Lotti, G., & R. Izzo. Seed oils of Torreya nucifera and Cephalotaxus drupacea. (In Italian; English summary.) Agr. Ital. 39: 163-173. 1974.* Rosertson, A. Spore formation in Jorreya californica. New Phytol. 3: 133-148. pls. 4. nen in the morphology of oe californica Torrey. II. The sexual organs and furtilication. Ibid. 3: 205-216. pls. 7-9. 1904b. SAVAGE, T. A Georgia station for Toreva taxifolia Arn. survives. Florida Sci. 46: 62— 64. 1983 Suin, C., & T. CHAo. A oe on seed dormancy of Torreya grandis. (In Chinese.) Acta Bot. Sinica 15: 279, 280. 1973. STALTER, R., & S. DIAL. cca status of the stinking cedar, Torreya taxifolia. Bartonia 50: 40-42. 1984. [Detailed review of status of the native populations.] TANG, S. H. The embryogeny of Torreya grandis. Bot. Bull. Acad. Sinica 2: 269-275. 1948 TANG, Z. X., Z. K. CHEN, & F. X. WANG. The development and structure of the late embryo in Jorreya grandis. (In Chinese; English summary.) Acta Bot. Sinica 27: 582-588. pls. 1, 2. 1985. & Investigation on sexual reproductive cycle in Torreya grandis. dn Chinese: English summary.) Acta Phytotax. Sinica 24: 447-4 86. Toops, C. The “stinking cedar” is in big trouble. Am. Forests 87(7): 46-49, 51. 1981. [Color plates of plant habit, seeds, and leaves attacked by blight.] Upuor, J.C. T. Protection of the 7wmion in Florida. Science, II. 64: 405. 1926. [Torreva taxifolia.] 1990] PRICE, TAXACEAE 87 Tribe TAXEAE 2. Taxus Linnaeus, Sp. Pl. 2: 1040. 1753; Gen. Pl. ed. 5. 462. 1754. Dioecious (rarely monoecious) shrubs or trees. Bark reddish brown, becom- ing scaly with age. Wood without axial parenchyma. Foliage leaves spirally arranged (often appearing 2-ranked), linear, with 2 yellowish or grayish, abaxial stomatal bands; resin canals absent. Pollen cones short-stalked, borne singly in axils of foliage leaves; microsporophylls densely clustered, ca. S—15 per strobilus, each with 4-9 radially arranged sporangia. Ovules borne singly in the axils of foliage leaves, each terminating an ovuliferous dwarf shoot borne laterally on a very short vegetative shoot in the axil of the foliage leaf. Seeds ca. 0.5[-1] cm long, largely surrounded by the cuplike reddish [rarely yellow] aril, which is not adnate to the highly sclerified seed coat; cotyledons 2 (rarely 3). Chromosome number 2n = 24. Lecroryre species: Taxus baccata L.* (Clas- sical Latin name for yew.)— YEw. A genus of perhaps eight or nine poorly differentiated allopatric taxa, treated most often as separate species but sometimes as geographic subspecies (Pilger, 1903, 1926). The species of Taxus are very difficult to distinguish by either gross morphology or leaf anatomy (Florin, 1931; Kwei & Hu; Pilger, 1903). A thorough revision making use of comparative biochemistry as well as mor- phology is much needed. Four species are native to North America, of which two, T. canadensis Marsh. and T. floridana Chapman, occur in our region. Taxus brevifolia Nutt. is widely distributed but relatively uncommon in the extreme western United States and Canada, while the poorly known 7. globosa Schlecht. is endemic to Mexico and Guatemala. Taxus baccata L. (English yew, Irish yew) is native to Europe, Asia Minor, and north Africa, while 7. cuspidata Sieb. & Zucc. (Japanese yew) and two or three other species are native to eastern Asia and the Himalayas (Cheng & Fu; De Laubenfels, 1988). Given the fact that the genus apparently dates back to the Jurassic (Florin, 1958a; Harris, 1976b), it is remarkable how little morphological differentiation is seen among its extant members. Taxus canadensis (T. baccata subsp. canadensis (Marsh.) Pilger; 7. baccata var. canadensis (Marsh.) A. Gray, 1856; 7. baccata var. minor Michx., 1803; T. minor (Michx.) Britton), Canada yew, American yew, ground hemlock, 2” = 24, is native to an area extending from Newfoundland to Manitoba, south to portions of Iowa, Illinois, Indiana, eastern Kentucky, Virginia, and extreme northern Tennessee and North Carolina. It is usually found under cool, moist conditions on well-drained soil as an understory plant in coniferous or mixed- mesophytic forest. In the southern part of its range, it almost always occurs on upland sites. In our area it has only recently been found in Pickett County, Tennessee, on protected northern slopes on the western edge of the Cumberland Plateau (Gonsoulin) and under montane forest conditions in Ashe and Watauga ‘The genus was effectively lectotypified by Siebold & Zuccarini, Fl. Japon. Fam. Nat. 2: 108. 1846, when they transferred the only other Linnaean species, Taxus nucifera, to the newly described genus Torreya. 88 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Ki orobik 1989 FiGurRE 2. Tribe Taxeae. a-g, Taxus floridana: a, leafy shoot with microsporangiate strobili at time of pollen release, x 4; b, detail of leaf in abaxial view, x 2%; c, d, e, microsporangiate strobili before, during, and after shedding of pollen, x 5; f, detail of microsporophyll, x 20; g, shoot with arillate ovule, x 14. counties in extreme northwestern North Carolina (McDowell). Despite the fact that the foliage of the plant is toxic to most animals, Canada yew is a favored browse plant of deer and has been greatly reduced in numbers in much of the northeastern United States by the large resident deer populations (Martell). Taxus floridana Chapman, Florida yew, savin, is a very rare species native only to areas of Gadsden and Liberty counties in northwestern Florida, largely in the vicinity of the Apalachicola River in the same area where Torreya taxifolia is native (Little, 1978). It occurs primarily on moist ravine slopes and sometimes bluffs in mixed deciduous forest, although one population has been found in an acidic bog (Kurz). Taxus canadensis is usually a low shrub, occasionally becoming 2-3 m tall, while 7. floridana is a small tree up to 10 m in height. The leaves of 7. canadensis are yellowish green above, relatively flat, and ca. 1.3-2 cm long, while those of 7. floridana are dark green above, usually falcate, and ca. 2-2.5 cm long (Rehder, 1940; Sargent). Taxus floridana could easily be treated as a subspecies of 7. canadensis (or perhaps of T. baccata) on the basis of its morphology, but its ecological tolerances are clearly different. It would be instructive to compare their chemistry in detail to assess the extent of other genetic differences between them. The genus Taxus is most similar to the monotypic Chinese Pseudotaxus, which differs in having a white rather than a reddish aril, ovuliferous shoots apparently borne directly in the axils of foliage leaves (vs. on dwarf axillary shoots), sterile scales present (vs. absent) between the sporophylls in the mi- crosporangiate strobilus, and epidermal papillae only on the margins of the leaves (vs. on the subsidiary cells of the stomata) (Florin, 1948b). The New 1990] PRICE, TAXACEAE 89 Caledonian Austrotaxus is similar 10 Taxus in the general features of its ovular development (Florin, 1948a) and in the presence of taxane alkaloids (Guéritte- Voegelein ef a/.), but it differs substantially in the structure of its microsporangi- ate strobilus and in having much larger leaves and seeds. Crossability among species of Taxus has not been investigated in a systematic manner, but hybrids between 7. baccata and T. cuspidata (T. x media Rehder) and 7. cuspidata and T. canadensis (T. x Hunnewelliana Rehder) have origi- nated spontaneously in cultivation (Dallimore & Jackson; Rehder, 1923) and are now commonly planted in the United States. Taxus is most notable chemically for the apparently ubiquitous presence of taxane alkaloids, an unusual class of diterpene alkaloids (Hegnauer, 1988, Lythgoe; R. W. Miller). Taxol, an alkaloid of this group that was isolated first from T. brevifolia and later from T. baccata, T. cuspidata, and T. Wallichiana Zucc., has been of particular interest as an antimitotic agent with activity against a number of types of cancer cells (Guéritte-Voegelein ef al.; Kingston ef al.; Wani et al.). Taxol and related compounds bind to tubulin and promote un- usually rapid microtubule assembly and are thus also useful in studying the mitotic process (Bajer et al.; Schiff et al.). The well-known toxicity of virtually all parts of the yew plant (only the aril is edible) is due in large part to more immediate physiological effects of the various taxane alkaloids, although other biologically active compounds are also present. The degree of toxicity of the foliage appears to vary considerably within species. Cyanogenic glycosides have been found in the leaves of Taxus baccata, T. canadensis, and T. cuspidata (Hegnauer, 1986) and presumably contribute to the toxicity of the plants. Biflavonoids and other types of glycosides may also have important pharmacological effects and are probably the active ingredients in nonalkaloidal leaf extracts used in traditional medicine in India (M. S. Y. Khan et al.; Vohora). The ecdysterones, biologically active compounds related to insect-molting hormones, are another interesting class of compounds found in the foliage of at least Taxus baccata and T. cuspidata (Hegnauer, 1986). Yews (Taxus baccata) have long been important plants in the history and folklore of Great Britain and Europe (Bean; Bialobok; Voliotis). The wood of Taxus is very strong and durable and was heavily utilized for the manufacture of bows for archery before the advent of gunpowder. The early Celts, who considered yew trees to be sacred, built temples near them in Britain and Ireland. These temples were often perpetuated by later Christian churches (Bean). Today the main economic importance of Taxus is as an ornamental hedge, shrub, or tree. The most widely used horticultural yews are forms of T. baccata, T. cuspidata (which is more cold tolerant), and their hybrid, 7. < me- dia. REFERENCES: Under family Sew see BAILEY; seen Butts & BUCHHOLZ; CARRIERE; CHANG; CHENG & Fu; CHOWDHURY; COKER & TOTTEN; DALLIMORE & JACKSON; DARK; DE LAu- BENFELS, 1988; aoe 1945, 1963; eae 1957, 1965; Florin, 1931, 1948a, 1948b, 1954, 1958a, 1963: GAUSSEN; GreGuss, 1955, 1972; Gutrirre- VOEGELEIN et al.; HARRIS, 90 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 1976a, 1976b; HEGNAUER, 1962, 1986, 1988; HickeL; M. S. Y. KHAN et al.; KHOSHOO, 1961, 1962; KRAL; KRUsSSMANN; LITTLE, 1971, 1978, 1980; Loze; LYTHGOE; Ma et al., 1985; MAHESHWARI & SINGH; C. N. MILLER; R. W. MILLeR; MORELLI; OUDEN & Boom; PHILLIPS; PILGER, 1903, 19 16a, 1916b, 1926; REHDER, 1940, 1949: SARGENT: SAX & SAX; SINGH, 1961, 1978; STERLING, 1948; UENo, 1959, 1960; WILDE; WODEHOUSE: and XI, 1986a Bager, A. S., C. CypHER, J. MOLE-BAsER, & H. M. Howarp. Taxol-induced anaphase reversal: evidence that elongating microtubules can exert a pushing force in cells. Proc. Natl. Acad. Sci. U. S. A. 79: 6569-6573. | BIALOBOK, S.,ed. The yew— Taxus baccata L. 183 pp. Warsaw. 1978. Sealine of Cis polity— Taxus baccata L.; a monographic treatment of the history, m phology, ecology, and physiology of the species. Camerort, H. Evolution de la structure des plastes pendant la maturation de l’arille de if (Taxus baccata L.). Compt. Rend. Acad. Sci. Paris 258: 1017-1020. pls. 1-4. 1964. DISTELBARTH, H., & U. KuLi. Physiological investigations of leaf mucilages. II. The mucilage of Taxus baccata and of Thuja occidentalis. Israel Jour. Bot. 34; 113-128. 1985.* Duper, A. W. The gametophytes of Taxus canadensis Marsh. Bot. Gaz. 64: 115-136. pls. 11-14. 1917 ——. Staminate strobilus of Taxus canadensis. [bid. 68: 345-366. pls. 24-26. 1920a. Ovuliferous structures of Taxus canadensis. Ibid. 69: 492-520. pl. 23. 1920b. [Imp ortant account of the anatomy and development of the ovuliferous shoot.] FAVRE-DUCHARTRE, M. Contribution a l’étude de la reproduction sexuée chez Taxus baccata. Compt. Rend. Acad. Sci. Paris 246: 979-982. 1958. Contribution a l’étude des spermatozoides de Taxus baccata. Revue Cytol. Biol. Veg. 21: 329-337. pl. 7. 1960. [Two approximately equal sperm nuclei produced.] GONSOULIN, G. Taxus canadensis Marsh: a new state record for Tennessee. Castanea 40: 253-255. 1975. HOLLWartTu, M. The development of heavy metal contents in needles of Taxus baccata L. at urban habitats from 1975 to 1982. Ang. Bot. 58: 21-30. 1984. [Taxus used to monitor heavy- metal pollution in Germany.] JAGER, L. Beitrage zur k d Embryologie von Taxus accaie t oe 86: 241-288. pls. 15-19. 1899. Kincston, D.G.1., D. R. HAwkins, & L. OviNGTon. New taxanes from Taxus brevifolia. Jour. Nat. Prod. 45: 466-470. 1982. Kurz, H. A new and remarkable habitat for the endemic in tae yew. Torreya 27: 90- 92. 1927. po in an acidic swamp in Liberty County. Kwel, Y. L., & Y. S. Hu. Epidermal feature of the ee ves of 7axus in relation to taxonomy. (In Chinese; English summary.) Acta Phytotax. Sinica 12: 329-334. p/. 67. 1974. [No caer dea ou differences seen among species MarTELL, A. M. Canada yew: Taxus canadensis Marsh. Pp. 158-160 ial. D. GILL & W. M. HEALY, eee Shrubs and vines for northeastern wildlife. U. S. Dep. 4. McDowe Lt, G. W. American yew in North Carolina. Jour. Elisha Mitchell Sci. Soc. 85: 16, 17. [7. ae Ashe and Watauga counties. ] MILLER, R. W., J. L. MCLAUGHLIN, R. G. PowELL, R. D. PLATTNER, D. WEISLEDER, & C.R. Sura, Ir. Lignans from 7axus W a Jour. Nat. Prod. 45: 78-82. 1982. [Lignan constituents of sun review . PowELi, C. R. SMITH, mel oe ia ote. & J. CLarpy. Antileukemic alkaloids Bs Taxus Wallichiana Zuce. Jour. rare Chem. 46: 1469-1474. 1981. Nik.as, K. J. ind pollination of Taxus cuspidata. Am. Jour. Bot. 72: 1-13. 1985. Saas e data compared to results from computer models.] 1990] PRICE, TAXACEAE 91 PENNELL, R. I., & P. R. BELL. Microsporogenesis in ee baccata: the development of the archaesporium. Ann. Bot. II. 56: 415-428. 1985. & . Microsporogenesis in Taxus pen the formation of the tetrad and development of the microspores. [bid. 57: 545-556. 1986a. The development of the male en ee and spermatogenesis in Taxus baccata. Proc. Roy. Soc. London, B. 228: 85-96. 1986b & egas sigs op and the subsequent cell ees within the ovule of Taxus haccata L. a Bot 59: 693-704. 1987. nation ~ the archegonium and fertilization in Taxus baccata L. Jour. Cell Sci. 89: 551 —560. 1988. ReHDER, A. New species, varieties and combinations from the herbarium and t collections at the Arnold aun Jour. Arnold Arb. 4: 107-116. 1923. Taxus x media, = T. baccata x cuspidat ScuiFF, P. B., J. FANT, & Horwitz. peenactea of microtubule assembly in vitro by taxol. Nature 277: a 667. 1979, [Taxol acts by promoting microtubule poly- merization and tending to suppress subsequent depolymerizatio STERLING, C. Embryonic differentiation in Taxus cuspidata. Bull. Torrey Bot. Club 76: 116-133. 1949. Vouora, S. B. Studies on Taxus baccata. 11. Pharmacological investigation of total extract of leaves. Pl. Med. 22: 59-65. 1972. [Aqueous leaf extracts, which test negative for alkaloids, ei a depressant effect on cardiovascular muscles; see also ibid. 20: 100-107. 197 VouioTis, D. Historical ae environmental significance of the yew (Taxus baccata L.). Israel Jour. Bot. 35: 47-52. 1987. Want, M.C., H. L. Taytor, M. E. WALL, P. Coccon, & A. T. McPHaIL. Plant antitumor agents. VI. The isolation and structure of taxol, a novel antileukemic and antitumor agent from Taxus brevifolia. Jour. Am. Chem. Soc. 93: 2325-2327. 1971. [Initial report of taxol and its biological activity; also found by the authors in 7. baccata and 7. cuspidata.] AL-SHEHBAZ, BRAYOPSIS AND EUDEMA 93 GENERIC LIMITS AND TAXONOMY OF BRAYOPSIS AND EUDEMA (BRASSICACEAE) IHSAN A. AL-SHEHBAZ! Brayopsis colombiana subsp. colombiana, B. colombiana subsp. ecuador- iana, and Eudema incurva are described as new. They represent first generic given. The sectional ‘classification of Eudema is discussed, and its value is questioned. In his treatment of the Brassicaceae (Cruciferae) for the Flora of Peru, Mac- bride (1938) followed Baehni and Macbride (1937) in uniting Brayopsis Gilg & Muschler with the earlier-published E-ng/erocharis Muschler. Evidence sup- porting the maintenance of both genera was given in Al-Shehbaz (1989a); there I pointed out that Brayopsis and Englerocharis are unrelated to each other, and that their nearest relatives are Eudema Humb. & Bonpl. and Catadysia O. E. Schulz, respectively. During the study of some Colombian species of Draba L. (Al-Shehbaz, 1989b), I discovered an undescribed species of Brayopsis among undetermined mate- rial. The new species, hereafter Brayopsis colombiana Al-Shehbaz, raised some taxonomic and phytogeographic questions. First, the genus had not previously been reported from Colombia; second, B. colombiana is disjunct from the nearest range of Brayopsis in Peru by at least 1600 air kilometers; and third, one of the close relatives of this plant is a Bolivian species that Schulz (1924) treated as Eudema diapensioides (Wedd.) O. E. Schulz (= B. diapensioides (Wedd.) Gilg & Muschler). The phytogeographic disjunctions of Brayopsis are not surprising, and several other South American genera of the Brassicaceae show even more dramatic distributional gaps; these will be dealt with in a forthcoming paper. Adequate material of this complex was not available to me 18 months ago when I submitted a revision of Brayopsis for publication (Al-Shehbaz, 1989a). The subsequent discovery of B. colombiana and E. in- curva and the study of almost all the types of Eudema have prompted a critical reevaluation of the boundaries between these two genera. Brayopsis resembles Eudema in habit and in having few to several solitary flowers borne on peduncles that arise from the center of a well-developed rosette. The maturation of these flowers follows a racemose, centripetal pattern; the direction of anthesis is from the periphery to the center. No other South ‘Arnold Arboretum, Harvard University, 22 Divinity Avenue, Cambridge, Massachusetts 02138. © President and Fellows of Harvard College, 1 Journal of the Arnold Arboretum 71: 93-109. tae. 1990. 94 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 American genera have such a synapomorphy, and therefore Brayopsis and Eudema are sister groups. The somewhat remotely related Dactylocardamum Al-Shehbaz and Xerodraba Skottsb. have only one pedunculate or subsessile flower from the center of the leaf cluster terminating a given branch (AI- Shehbaz, 1989c). Schulz (1924, 1936) separated Brayopsis from Eudema solely on the basis of its having siliquose instead of siliculose fruits. Additional morphological differences between the two genera, particularly in funicle length and seed-coat sculpture, are also found (see TABLE, FiGuRE 1). These differences strongly support the maintenance of both genera. Brayopsis colombiana is a perfectly good representative of the genus, and characters of the fruits, as well as the filiform funicles and the seed sculpture (see Figures 1, 2; TABLE), strongly define its generic identity. Schulz’s (1924) initial assignment of B. diapensioides to Eudema was based solely on his study ofa fragmentary isotype at Berlin. Although I have seen only three additional specimens of the type collection, the fruit shape and dimensions, funicle type, and seed-coat sculpture are characteristic of Brayopsis and not Eudema (see TABLE). Therefore, I believe that Gilg & Muschler’s (1909) placement of B. diapensioides in Brayopsis is more acceptable than its assignment by Schulz (1924) to Eudema. BRAYOPSIS The taxonomy of Brayopsis alpaminae Gilg & Muschler, B. calycina (Desv.) Gilg & Muschler, and B. monimocalyx O. E. Schulz has been studied in some detail (Al-Shehbaz, 1989a), and therefore an account need not be repeated here. The addition of B. diapensioides has expanded the morphological diversity within the genus only a little. A key to the five species of Brayopsis and a taxonomic account of B. colombiana and B. diapensioides are given below. A. Styles in fruit 1-3 mm long; sepals erect, somewhat coherent; fruit valves densely covered with long, straight trichomes. ............. 00.00.0000 eee . alpaminae. . Styles in fruit obsolete or 0.2-0.8(-—1) mm long; sepals ascending, free; fruit valves glabrous or rarely with short, crisped trichomes. B. Leaves 0.4-0.8(-1) mm wide, ciliate with trichomes 0.2-0.3 mm long; fruits stipitate; gynophores |.2—2.5 mm long, thinner than fruiting ee 16s aeeecus ah ond tee deibvg isitiel Deltas ae mnie apne een Wa ane ed a ae B. diapensioides. B. Leaves |.5-4 mm wide, glabrous or ciliate with trichomes (0.8- y\- 1.5 mm long; fruits sessile, or if stipitate, then gynophores up to 0.5 mm long and thicker than fruiting peduncle. : meee persistent even after fruit dehiscence; leaves ovate, densely desea n upper surface. 2.0... cece eee . Monimocalyx. Seaals usually caducous shortly after anthesis; leaves oblong, ee or linear, glabrous or pubescent on upper surface. D. Fruits short-stipitate, fruiting peduncles 3-15 mm long; leaves ue to oblanceolate or rarely lanceolate; styles conspicuous, to 1.5 mm long; Colombia and Ecuador. .......................0005 B. colombiana. Fruits sessile; fruiting peduncles (9-)12-42(-55) mm long; leaves linear to lanceolate; styles obsolete or very rarely to 0.8 mm long; Argentina, BOW. POs. 1.2 nheexedeeceunnionarsasy eeea sto aiewas B. calycina. > i u Comparison between Brayopsis and Eudema. TAXON CHARACTER Brayopsis B. colombiana B. diapensioides Eudema Fruits Shape Linear, lanceolate, Lanceolate Lanceolate Oblong, ovoid, suborbicu- or oblong-linear lar, obovo id, o or pyri- orm Length/width quotient (3.5-)5-8(-10) : 3. 1-1.5(-2.5) Funicle Long, filiform Long, filiform Long, ao Short, thick Septum Complete Complete Complet Complete or reduced to rim Seed-coat sculpture Colliculate-reticulate Colliculate-reticulate Colliculate-reticulate Coarsely reticulate* *Mature seeds of only E. incurva, E. nubigena, and £. rupestris were available. [0661 VWACNA ANY SISCOAVUE “ZVAHAHS-1V 96 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 é i i S” ~ + a7 7 ‘ FiGurRE |. Scanning electron eae awl of seeds of Brayvopsis and Eudema: a, B. alpaminae (Macbride & Featherstone 887, GH), b, B. colombiana subsp. colombiana H),c, E. ber a subsp. nubige na eds 7652, F), d, E. rupestris (Jameson ., US). Scale bars = 10 1990] AL-SHEHBAZ, BRAYOPSIS AND EUDEMA 97 Brayopsis colombiana Al-Shehbaz, sp. nov. FIGURE 2. Herba perenna dense caespitosa, caudicibus ramosis, foliis emortuis persis- tentibus vestitis. Folia rosulata, spathulata vel oblanceolata, ciliata, 4-12 mm longa, 1.5—4 mm lata, pilis simplicibus (0.8—-)1-1.5 mm longis. Pedunculi fruc- tiferi tenui, 3-15 mm longi. Petala alba, spathulata 2.2-2.5 mm longa, 0.7-0.9 mm lata. Siliqua lanceolata, subtereta, breve stipitata, 6-8 mm longa, 1.2-1.5 mm lata; septum hyalinum completum; locula 4-6 sperma; stylus 0.4-1 mm longus. Semina oblonga notorrhiza, brunnea, 1.4-1.6 mm longa, 0.7-0.9 mm lata; funiculi filiformi, 1.2-2 mm longi. Cespitose, cushion-forming perennials; caudices woody, branched, covered with whole leaves of previous years. Leaves forming dense rosettes, spatulate to oblanceolate, 4-12 mm long, 1.5—4 mm wide, obtuse to subacute at apex, ciliate with trichomes (0.8-)1—1.5 mm long, cuneate at base. Sepals erect, ovate, ca. 2 mm long. Petals white, spatulate, 2.2-2.5 mm long, 0.7-0.9 mm wide. Filaments 1.2-1.8 mm long; anthers ovate, ca. 0.6 mm long. Fruiting peduncles erect, slender, 3-15 mm long, sparsely pubescent. Fruits lanceolate, 6-8 mm long, 1.2-1.5 mm wide, dehiscent, subterete, exserted well above leaf rosette; gynophores stout, to 0.5 mm long; valves glabrous, acute at both ends, incon- spicuously veined; septa hyaline, complete; styles 0.4-1 mm long. Seeds 4 to 6 per locule, oblong, |.4-1.6 mm long, 0. ae 9 mm wide, brown, pendulous; cotyledons incumbent; funicles filiform, 1.2—-2 mm long. Type. Colombia, Depto. Boyaca, Cordillera Oriental, Sierra Nevada del Cocuy, Alto Ritacuva, 11 April 1959, 4600 m alt., H. G. Barclay & P. Juajibioy 7355 (holotype, GH; isotype, MO). KEY TO THE SUBSPECIES Leaves ciliate; sepals ca. 2 mm long; petals 2.2-2.5 mm long; styles in fruit 0.4-1 mm 2 WMDIias hoe we duieseanoee tee eee sein a. subsp. colombiana. Leaves not ciliate; sepals 3-4.5 mm long; petals 4-5.5 mm long; styles in fruit 1-1.5 m lONG HE CNACOM. 2242550 toad dows ceed pecan aay aoaan tela 2b. subsp. ecuadoriana. 2a. Brayopsis colombiana subsp. colombiana ADDITIONAL SPECIMENS EXAMINED. Colombia. Depto. BoyAca: Cordillera Oriental, Sierra Nevada del Cocuy, Alto Ritacuva, 4400 m alt., Barclay & Juajibioy 7410 (GH); same locality, 4600 m alt., Grubb, Curry, & Ferndndez- Pérez 304 (k, US Brayopsis colombiana subsp. colombiana is the sole representative of the genus in Colombia. It grows on unstabilized, fine, sandy soil in crevices of bedrock above moraines of the Ritacuva glacier at an altitude of ca. 4400- 4600 m 2b. Brayopsis colombiana subsp. ecuadoriana Al-Shehbaz, subsp. nov. Folia glabra, nonciliata; sepala 3-4.5 mm longa; petala 4-5.5 mm longa, ca. 1.5 mm lata; stylus 1-1.5 mm longus. 98 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Ficure 2. Brayopsis colombiana subsp. colombiana (holotype): a, plant: b, fruit with valves removed. Scale bars = 1 cm (a), | mm (b). Leaves glabrous, lanceolate to oblanceolate, not ciliate. Sepals oblong, 3—4.5 mm long. Petals white, spatulate, 4-5.5 mm long, ca. 1.5 mm wide. Fruits 5— 6 mm long; styles 1-1.5 mm long. Type. Ecuador, Prov. Azuay, paramo de las Cajas, Lagunas Suerococha, 35 km on Cuenca-Molleturo road; mountain 1 km NW of pass, 79°14’W, 2°48’S, 4200-4400 m alt., 30 Jan. 1988, U. Molau & B. Eriksen 2787 (holotype, GB). The type collection of Brayopsis colombiana subsp. ecuadoriana represents 1990] AL-SHEHBAZ, BRAYOPSIS AND EUDEMA 99 the first record of Brayopsis from Ecuador, and the type locality somewhat bridges the gap between the main range of the genus and that of subsp. colom- biana. The differences in flower size between the two subspecies is significant. In the absence of additional material, and because of the evident similarity in all other morphological characters, I have refrained from recognizing these taxa as separate species. The presence vs. absence of leaf trichomes, which is a good distinguishing character, should not be overemphasized: the same pattern oc- curs within B. calycina. Brayopsis colombiana resembles B. diapensioides in several features, partic- ularly in having lanceolate, stipitate fruits, short styles, and filiform funicles. It is easily distinguished, however, by its spatulate to oblanceolate or lanceolate leaves 1.5-4 mm wide, its stout gynophores to 0.5 mm long, and its either glabrous or ciliate leaves with fine trichomes (0.8-)1-1.5 mm long. In contrast, B. diapensioides has narrowly lanceolate to linear leaves 0.4-0.8(-1) mm wide, slender gynophores 1.2—2.5 mm long, and stouter trichomes 0.2-0.3 mm long. Perhaps the nearest relative of B. colombiana is B. calycina, from which it is easily distinguished by several characters given in the key above. Brayopsis diapensioides (Wedd.) Gilg & Muschler, Bot. Jahrb. Syst. 42: 484. 1909 Draba diapensioides Wedd. Ann. Sci. Nat. Bot. 5(1): 285. 1864. Eudema he lg ae . Schulz, Pflanzenr. IV. 105(Heft 86): 245. 1924. Type: Bolivia, Prov. Larecaja, vicinity of Sorata, between Ancouma and Turelague in Cerro de ee. 4500 m alt., November 1860, Mandon 894 (holotype, P!; isotypes, B! (photos, FI, GH!), BM!, G!). Densely cespitose perennials, with thick, much-branched caudices covered with persistent petiolar remains of previous years. Leaves narrowly lanceolate to linear, forming well-developed rosettes, 4-8 mm long, 0.4-0.8(-1) mm wide, subacute, ciliate with unbranched trichomes 0.2-0.3 mm long; blades thin, shorter than persistent, conspicuously thickened petioles. Sepals erect, ovate, 2-2.5 mm long, |.2-1.3 mm wide, membranaceous at margin, glabrous. Petals obovate, 2.5-3.5 mm long, |.2-1.5 mm wide, cuneate at base. Filaments 2- 2.5 mm long; anthers ovate, 0.5-0.6 mm long. Fruiting peduncles 5-8 mm long, thick, up to 1 mm wide, sparsely pubescent with trichomes to 0.3 mm long. Fruits lanceolate, subterete, 3.5-5 mm long (excluding style and gyno- phore), stipitate, glabrous; styles 0.4-0.7 mm long; gynophores 1.2-2.5 mm long, slender, about half as wide as fruiting peduncle. Seeds oblong, 1.2-1.4 mm long, 0.7-0.8 mm wide, borne on filiform funicles ca. 1.2 mm long. Hardly anything can be said about the variation and biology of this species, which is apparently known only from the type collection. Brayopsis diapen- sioides can easily be distinguished from the remainder of the genus in having slender gynophores much narrower than the thickened fruiting pedicels. EUDEMA As interpreted here, Eudema consists of six species, including only five of the eight recognized by Schulz (1924, 1936). One species, £. remyana (Wedd.) 100 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 O. E. Schulz, 1s reduced here to a subspecies of E. nubigena Humb. & Bonpl. Another, £. diapensioides, is treated above as a species of Brayopsis. The third species, F. colobanthoides (Skottsb.) O. E. Schulz, is reassigned to Xerodraba. The sixth species, FE. incurva, is described below as new. Schulz (1924, 1936) recognized the heterogeneity of Eudema and divided the genus into four sections. The monotypic sect. Gynophoridium O. E. Schulz was defined solely on the basis of having stipitate instead of sessile fruit. Its single species is recognized here as Brayopsis diapensioides. Section Lepteudema O. E. Schulz is separated from the rest of the genus mainly on the basis of having slender, scaly, loosely branched rhizomes instead of a thick caudex. Two species of this section, Eudema friesii O. E. Schulz and E. werdermannii O. E. Schulz, have dendritic leaf trichomes that are lacking in the rest of the genus. This section might justifiably be raised to a separate genus, but in the absence of adequate material at my disposal, I refrain from taking such an action. The monotypic sect. Xereudema O. E. Schulz is based on a species that was originally described as Yerodraba colobanthoides Skottsb. Schulz (1924) trans- ferred the species to Eudema because it allegedly lacks the expanded leaf bases, subsaccate sepals, and somewhat fleshy petals characteristic of Yerodraba. However, these features are unreliable in this complex, and Yerodraba is easily separated from Eudema by its solitary terminal flowers, usually complete sep- tum, and densely imbricate, minute, fleshy, wholly persistent leaves. Eudema has several flowers at the branch apex, a usually rudimentary septum, and nonimbricate, large, unfleshy leaves with persistent petioles. Boelcke (1982) transferred Yerodraba colobanthoides to the later homonym Skottsbergiella Boelcke and then (in Boelcke & Romanczuk, 1984) to Skotts- bergianthus Boelcke. However, the single character on which the latter genus is based, the presence of a minute, scalelike appendage at the base of each petal, might not justify its recognition as distinct from Xerodraba. I have not seen adequate material of Y. colobanthoides, but the species does not belong to Eudema. 1 have concluded that Schulz’s (1924, 1936) sectional classification of Eudema is artificial and does not improve the taxonomy of the genus. Eudema Humb. & Bonpl. Pl. Aequinoct. 2: 133. 1813. LEcrotTyPE (here des- ignated): £. rupestris Humb. & Bonpl. Cespitose, scapose perennials with thick, branched caudices or slender rhi- zomes, the branches covered with petiolar remains of previous years and ter- minated by rosettes; trichomes simple or dendritic, sometimes absent. Leaves rosulate, petiolate, somewhat fleshy, entire or rarely pinnately lobed: petioles persistent, thick or flattened at base. Flowers solitary, borne on peduncles originating from center of rosette, maturing centripetally. Sepals erect, ovate, nonsaccate at base, caducous. Petals white to creamy white, obovate, spatulate, oblanceolate, or sublinear, usually not differentiated into blade and claw. Sta- mens 6; anthers oblong to ovate. Nectar glands 4, | on each side of a lateral stamen. Fruits dehiscent, suborbicular, ovoid, obovoid, pyriform, or rarely oblong, subterete to conspicuously flattened parallel or rarely perpendicular to 1990] AL-SHEHBAZ, BRAYOPSIS AND EUDEMA 101 replum, glabrous; septa reduced to narrow rim or rarely complete; stigmas subentire. Seeds few to many, coarsely reticulate; cotyledons incumbent; fu- nicles short, thick. KEY TO THE SPECIES OF EUDEMA A. Leaves with dendritic trichomes; plants with slender rhizomes. B. Leaves pinnately lobed, rarely coarsely toothed, upper surface sparsely pubescent. SA rae cist ht Bes dances has bpetie te eae ted ile tai CRE bp a aie ee es 6. E. friesii. B. Leaves entire, upper surface densely pubescent. .......... 5. E. werdermannii. A. Leaves glabrous or with simple trichomes; are with thick ae (if with rhi- zomes, then leaves glabrous and fruits flattened). C. Fruits eg to pyriform, conspicuously flattened parallel to ae plants PROMO Se tn trea ant cui awe deed pad neta esas 4. E. hauthalii. C. Fruits ablone or ovoid, subterete or slightly flattened; plants with thick caudex. D. Fruits oblong; seeds 10 to 30 on each ee petioles flattened at base; leaves linear to narrowly oblanceolate, glabro E. Leaf blades ciliate; septa complete; fruits sol incurved; seeds plump, m lon 1. 15 to 30 on each placenta, 0.7-0.8 mm long. .......... . Incurva. E. Leaf blades not ciliate; septa reduced to rim; ae straight; seeds flattened, 10 to 13 on see placent ee 2-1.7 mm long. ......... rupestris. D. Fruits ovoid; seeds 1 to 5 on each sea petioles thick at base: leaves spatulate to ovate or ee ciliate or glabrous. ..... 3. E. nubigena. 1. Eudema incurva Al-Shehbaz, sp. nov. FIGURE 3. Herba perenna caespitosa scaposa, caudicibus tenuibus. Folia rosulata lineari vel lineari-lanceolata, ciliata, 5-15 mm longa, 0.6-1.4 mm lata, petiolis per- sistentibus, ad basin complanatibus. Pedunculi fructiferi glabri, 4-10 mm longi. Siliqua oblonga, incurva, angustiseptata, 4-8 mm longa, 2.5-3.5 mm lata; septum hyalinum completum; locula 15—30-sperma; stylus 0.4-0.7 mm longus. Semina late ovoidea, notorrhiza, biseriata, reticulata, 0.7-0.8 mm longa, 0.5- 0.6 mm lata Cespitose, perennial herbs with slender, simple, or few-branched caudices covered with petiolar remains of previous years. Leaves rosulate, petiolate, linear to linear-lanceolate, 5-15 mm long, 0.6-1.4 mm wide, ciliate with un- branched, straight trichomes to 2 mm long; petioles persistent, not ciliate, flattened at base, 2-7 mm long. Flowers not seen. Fruiting peduncles glabrous, 4-10 mm long. Fruits oblong, strongly incurved, glabrous, 4-8 mm long, 2.5- 3.5 mm wide, flattened contrary to septum, rounded at apex and base; septa complete, hyaline; styles 0.4-0.7 mm long. Seeds 15 to 30 per locule, broadly ovoid, plump, 0.7-0.8 mm long, 0.5-0.6 mm wide, dark brown, conspicuously reticulate, biseriately arranged in each locule; funicles 0.2-0.5 mm long; cot- yledons incumbent. Type. Peru, Depto. Ancash, Prov. Huaraz, Punta Callan and above, summit area of the Huaraz-Casma road, 30 km from Huaraz, 1-3 km N of road, heavily grazed puna with rock outcrops, 77°37'W, 9°32'S, 4200-4400 m alt., 7 April 1988, U. Molau & B. Eriksen 3512 (holotype, GB (photocopy, GH)). 102 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Figure 3. Eudema incurva (holotype): a, plant; b, leaf; c, fruit; d, valve; e, replum and septum; f, seed. Scale bars = 1 cm (a), | mm (b-f). Eudema incurva is closely related to the Ecuadorian endemic E. rupestris. It is easily distinguished from the remainder of the genus in having angusti- septate, strongly incurved fruits, a complete septum, many (15 to 30 per locule) biseriately arranged, small seeds to 0.8 mm long, and conspicuously ciliate leaf blades with trichomes to 2 mm long. No flowers are found on any of the six plants of the holotype. However, floral remains on a young fruit reveal that the sepals are oblong, pubescent, and ca. 2.5 mm long, and that the filaments and the oblong anthers are ca. 2 and 0.8 mm long, respectively. Eudema incurva is somewhat anomalous in the genus because it has a com- 1990] AL-SHEHBAZ, BRAYOPSIS AND EUDEMA 103 plete instead of a rimlike septum and numerous seeds per locule. In these characters it resembles Brayopsis, but the overall morphology strongly supports its placement in Eudema. 2. Eudema rupestris Humb. & Bonpl. Pl. Aequinoct. 2: 133. 1813. Sisymbrium rupestre (Humb. & Bonpl.) Wedd. Ann. Sci. Nat. Bot. 5(1): 290. 1864; Hesperis rupestris (Humb. & Bonpl.) Kuntze, Revis. Gen. Pl. 2: 935. 1891; Draba humboldtii Desv. J. Bot. (Desvaux) 3: 171. 1815; Eutrema Kelle ae ar gel, Syst. Veg. 2: 880. 1825. Type: Ecuador, between Quito and Cuenca, Montana - Assuay, Humboldt & Bonpland s.n. (holotype, Pp (IDC 6209. 119: 1 6): isotype, Ben grandiflora Planchon in W. J. Hooker, London J. Bot. 3: 620. 1844; Sisymbri- um grandiflorum (Planchon) Wedd. Ann. Sci. Nat. Bot. 5(1): 290. 1864; Hesperis planchoniana Kuntze, Revis. Gen. Pl. 2: 936. 1891; Brayopsis grandiflora (Planchon) Gilg & Muschler, Bot. Jahrb. Syst. 42: 482. 1909. Type: Ecuador, Andes, monte Assuay, 15,000 ft [ca. 4570 m] alt., Jameson s.n. (holotype, MPU ve NA; isotype, NY!). Cespitose, glabrous perennials with much-branched caudices covered with petiolar remains of previous years. Leaves numerous, rosulate, petiolate; blades linear to narrowly oblanceolate, rounded at apex, entire, attenuate at base, 7— 17 mm long, 1-3 mm wide; petioles conspicuously flattened at base, 7-10 mm long. Sepals erect, narrowly oblong to sublanceolate, 6-7 mm long, ca. 2 mm wide, scarious at margin, with few subapical trichomes. Petals oblong to sub- linear, 11-12 mm long, ca. 2 mm wide. Filaments erect, slender, 5.5-6.5 mm long; anthers narrowly oblong, 1.2-1.5 mm long. Fruiting peduncles glabrous, 10-14 mm long. Fruits oblong, subterete to slightly compressed contrary to replum; valves thin, glabrous, with conspicuous midvein; septa reduced to rim, styles slender, 3-3.5 mm long. Seeds about 10 to 13 on each placenta, ovoid, slightly compressed, |.2-1.7 mm long, 0.8-1.1 mm wide, dark brown to black- ish, conspicuously reticulate; funicles thick, ca. 0.2-0.5 mm long. REPRESENTATIVE SPECIMENS. Ecuador: Andes above Quito, 14,000 ft [ca. 4267 m] allt., Jameson s.n. (US, 2 sheets) Eudema rupestris is apparently a very rare species. I have examined only a few specimens, which were collected around the middle of the last century. It is easily distinguished from its nearest relative, E. nubigena, in having oblong fruits with 10 to 13 seeds on each placenta, larger flowers with longer sepals, petals, and anthers, and linear to narrowly oblanceolate leaves with flattened petioles. In FE. nubigena the ovoid fruits have | to 5 seeds on each placenta, the flowers are smaller, and the leaves are usually spatulate to ovate with thick petioles. 3. Eudema nubigena Humb. & Bonpl. Pl. Aequinoct. 2: 136. 1813. Densely cespitose perennials forming well-developed cushions. Caudices woody, much branched, densely covered with persistent petioles of previous years. Leaves numerous, forming dense rosettes, somewhat fleshy; blades spat- ulate to oblanceolate or ovate, (3-)4-8(-9) mm long, 1.5—4(-5) mm wide, entire, 104 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 obtuse to somewhat subacute, nonciliate or ciliate throughout, with unbranched trichomes to 1.2 mm long; petioles persistent, thick, (2-)3-10 mm long. Sepals ovate to oblong, (2—)2.5—3.5(-6) mm long, (1.2—)1.5—2 mm wide, scarious at margin. Petals broadly obovate, 3-5 mm long, 1.5—2.5(-—2.8) mm wide, rounded alt apex, not clawed, white to creamy white. Filaments 2—3(-3.5) mm long: anthers oblong, 0.7-1 mm long, usually violet. Fruiting peduncles 2—5(-10) mm long. Fruits broadly to narrowly ovoid, subterete, (2—)3—5(-6) mm long, 2-3 mm wide; septa usually reduced to narrow rim; styles 0.4—2.5(-3) mm long. Seeds (1 or) 2 to 4 (or 5) on each placenta, oblong-ovate, (1.4-)1.6-2 mm long, (0.8—)1—1.3(-1.4) mm wide; dark brown, conspicuously reticulate; funicles thick, to 0.6 mm long. KEY TO THE SUBSPECIES Leaf blades conspicuously ciliate; fruits usually broadly ovoid, — 0.4—-1.5(-3) long. fia es Greet are ay te deve hated vice at desoutes Saws uti As ew stags eee a subsp. nubigena. i a usually not ciliate; fruits narrowly to broadly ovoid, ses (i. 2-)1.5-2.5(-3) MEAG Ge ph cers Seg aie aid aid sein ead hah hata ta ae Sy he oks . subsp. remyana. 3a. Eudema nubigena subsp. nubigena Draba nubigena (Humb. & Bonpl.) Desv. J. Bot. (Desvaux) 3: 171. 1815: eal plandii ‘Sprengel, Syst. Veg. 2: 880. 1825. Type: Ecuador, between Quito and Cuenca, Montana del Assuay, Humboldt & Bonpland s.n. (holotype, p (IDC 6209. 119: 1. 7!)). Leaf blades ciliate throughout, with trichomes to 1.2 mm long. Fruits broadly or rarely narrowly ovoid, usually plump; styles stout to rarely slender, 0.4— 1.5(-3) mm long. REPRESENTATIVE SPECIMENS. Ecuador. Prov. IMBABURA: volcano Cotacachi, Boysen Lar- sen, Eriksen, Kvist, Nissen, & Korning 45635 (AAv); Imbabura, Hirsch 155 (GH); Cayambe Mt., Cazalet & Ale a 5735 (uc). PRov. PICHINCHA: Cerro Antisana, Grubb, Lloyd, Pennington, & Whitmore 632 (Ny), Pichincha and Antisana, Sodiro 52 (B); Loma Pilongo, NE slope of Nevado linia, Molau & Eriksen 2269 (GB); Cerro Uiniza, near border to Cotopaxi, Harling 11175 (Gs); W Rucu Pinchincha, Padre Encantado, Molau, Eriksen, & Klitgaard 2407, 2408 (GB), Mt. Corazon, Asplund 17499 (G, Ny, s); Volcan Iliniza, agile 17457 (s); summit of Cerro Pichincha, Holmgren & Heilborn 175 (s, 2 sheets); Pichincha, near Quito, Hitchcock 21064 (GH, Ny, us); Rucu Pichincha, Asplund a Nae ); Volcan de Pichincha, Bel/ 362 (am, 2 sheets), Mexia 7652 (F, UC, US); Andes 'e Quito, Jameson . oe 33 (G, GH, P), 752 (BM, G, LE); on Chimborazo and sr cae sm, Sept. 1824 Subspecies nubigena grows on consolidated moraine and sandy soil between rocks or in crevices, as well as in grass paramo at altitudes of 4000-4820 m. Plants of this subspecies vary in flower color from white (Boysen Larsen et al. 45635) to greenish yellow (Grubb et al. 632). Style length is also variable: styles of most plants that I have examined are usually less than 1.5 mm long, while those of plants comprising Sparre 17457 are up to 3 mm. Although the 1990] AL-SHEHBAZ, BRAYOPSIS AND EUDEMA 105 septum in almost all collections of subsp. nubigena is reduced to a narrow rim, it is unusually complete in Boysen Larsen et al. 45635. 3b. Eudema nubigena subsp. remyana (Wedd.) Al-Shehbaz, comb. nov. Based on Sisymbrium remyanum Wedd. Ann. Sci. Nat. Bot. 5(1): 290. 1864; Hesperis remyana (Wedd.) Kuntze, Revis. Gen. Pl. 2: 935. 1891; Brayopsis remyana es d.) Gilg & Muschler, Bot. Jahrb. Syst. 42: 482. 1909; Eudema remyana (Wed pe. Pflanzenr. IV. 105(Heft 86): 244. 1924. piv eae pated oe , 3 Nov. 1856 (holotype, Pp! (photo and fragmen Aschersoniodoe chimborazensis Gilg & Muschler, a ie Syst. 42: 470. 1909. pe: Ecuador, Chimborazo, Hall s.n., 1853 (holotype, B!). Leaf blades glabrous throughout, rarely sparsely ciliate. Fruits narrowly ovoid, rarely broadly so; styles slender to rarely stout, (1.2-)1.5—2.5(-3) mm long. REPRESENTATIVE SPECIMENS. Ecuador. Prov. BoLivar: W base of Volcan Chimborazo, Maguire & Maguire 61738 (Ny), 61746 (Ny, US); ca. 33 km N of Guaranda, W of Volcan Chimborazo, Luteyn & Cotton 11076 (GH). Prov. CHIMBoRAzO: E slope of Mt. Chim- 8392 (s); WSW slope of Mt. Chimborazo, below Whymper refuge, Molau & Eriksen 2991 (GB); Chimborazo, Hirsch E330 (GH), Whymper s.n., Jan. 1880 (BM); slope of Chimborazo, above Tortorillas, Penland & Summers 702 (GH); between Urbina and Mt. Chimborazo, Hitchcock 21976 (GH, Ny, US); Riobamba, Primbach 219 (us). mae TuNGuRAHUA: Las Minas, SE of Volcan Carihuayrazo, Brandbyge 42383 (AaAu); Tun- gurahua volcano, Korning & Thomsen 47335 (AAu); paramo of Minza, Penland & es mers 387 (GH, US). Subspecies nubigena and remyana are remarkably similar in almost all as- pects of habit, leaf morphology, floral dimensions, and size and sculpture of seeds, as well as the reduction of the septum to a narrow rim. They are somewhat different in leaf pubescence, style length, fruit shape, and geographic distri- bution. Schulz (1924) separated the two subspecies (as species) solely on the basis of the leaves being ciliate in subsp. mubigena and nonciliate in subsp. remyana. This difference holds for the majority of collections that belong to the latter taxon. However, in Luteyn & Cotton 11076 leaf blades and/or petioles vary continuously from completely glabrous to sparsely pubescent. Ciliate and non- ciliate petioles have also been observed in Asplund 8384 (Ny). The fruits of subsp. nuhigena are usually broadly ovoid, while those of subsp. remyana tend to be narrower. However, some collections of the latter (e.g., Asplund 8384; 8) have broadly ovate fruits. Styles of subsp. remyana are usually longer than those of subsp. nubigena. In Sparre 17457 and Asplund 17499 (Ny), both of subsp. nubigena, style length is 2-3 and 1.2-1.8 mm, respectively. Similarly, in a few plants of subsp. remyana (e.g., Asplund 8392) the styles are 1.2-1.6 mm long. A critical evaluation of these slight differences in style length, fruit shape, and leaf pubescence reveals that there is no justification for recognizing Eudema remyana as a species distinct from the earlier-published £. nubigena. The taxonomy of the group is best served by reducing the former to a subordinate taxon of the latter. 106 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Plants of subsp. remyana apparently occupy similar habitats at about the same altitude as those of subsp. nubigena. 4. Eudema hauthalii Gilg & Muschler, Bot. Jahrb. Syst. 42: 471. 1909, — oa (Gilg & Muschler) Skottsb. Kongl. Svenska Vetenskapsakad. Hand. 56(5): 2 1916. Type: Argentina, [Prov.] Santa Cruz, Rio Gallegos, Cerro Buitres, R. a 10618 (holotype, B! (photo, F!)). Brayopsis skottsbergeri Gilg in Skottsberg, Kongl. Svenska Vetenskapsakad. Hand. 56(5): 236. 1916. Type: Patagonia, Sierra de los Baguales, Paso Centinela—Baguales, Skottsberg 5.n., 1909 (not seen). Prostrate, glabrous perennials with slender, branched rhizomes, the branches terminated by rosettes and covered with petiolar remains of previous years. Leaves rosulate, petiolate, oblanceolate-spatulate to narrowly oblanceolate or linear, (6-)9-15(-22) mm long, (0.7-)l-1.5(-2) mm wide, entire, rounded at apex, attenuate at base; petioles glabrous, as long as or longer than blades, conspicuously flattened at base. Sepals erect, ovate, 2.5-3.5 mm long, glabrous, rounded at apex. Petals oblanceolate, (4.5-)5.5-6.5 mm long, (1.2-)1.5-1.8 mm wide, rounded at apex, attenuate at base, creamy white. Filaments slender, 2.5-3.5 mm long; anthers 0.5-0.6 mm long. Fruiting peduncles (3-)5—10 mm long, few to several from center of rosette. Fruits sessile, obovate to pyriform, rarely ovate, flattened parallel to replum, (4-)5—11(-12) mm long, 4-7 mm wide, dehiscing from apex downward; valves thin, glabrous, inconspicuously nerved; septa hyaline, reduced to narrow rim; styles 0.5-1 mm long. Seeds 2 to 5 on each placenta, oblong, 2—2.5(-3) mm long, 1.1-1.3(-1.8) mm wide, light brown, obscurely flattened at distal end; funicles thick. REPRESENTATIVE SPECIMENS. Argentina. Prov. SANTA CRuz: Depto. Lago Argentino, Ea. Pérez, Rio de las Vueltas, Meseta Quemada, Sleumer 1386 (G, us); Cerro Corona, La Victorina, Lago Paine, Pisano & Pisano 5603 (Gu, Hip), Giier Aike, Ea. Las Viscachas, Ensenada de Riques, 7BPA 2679 (Hip); Co. Sin Nombre, TBPA 2674 (uIP); Co. Pto. la Piedra, T7BPA 2542 (nip). Chile. MAGALLANES: Depto. Ultima Esperanza, Sierra de los Baguales, Ea. La Cumbre, Co. Sin Nombre, TBPA 757 (nip). Eudema hauthalii is most closely related to E. werdermannii, from which it is easily distinguished in having glabrous leaves, obovate to pyriform fruits (4-)5-11(-12) mm long, and seeds 2-2.5(-3) mm long. In contrast, E. werder- mannii has densely pubescent leaves with dendritic trichomes, suborbicular fruits 2.5-3 mm long, and seeds 0.9-1.2 mm long. The two species resemble each other in habit, leaf morphology, flower size and color, flattening of the fruit, reduction of the septum to a rim, and number of seeds on each placenta. Eudema hauthalii and E. werdermannii form dense cushions in moist seepage areas of rocky steppes. They often produce a tangled growth among fine gravel, particularly at altitudes of 700-1500 m. 5. Eudema werdermannii O. E. Schulz, Notizbl. Bot. Gart. Berlin-Dahlem 10: 462. 1928. Type: Chile, Prov. Atacama, Depto. Copiap6, Cord. Rio Figueroa, Co. Paredones, ca. 4300 m alt., Jan. 1926, E. Werdermann 974 a ae B! (photos, F!, GH!, Ny!); isotypes, F!, GH!, M!, Mo!, NY!, s!, uc!, us!, z!). 1990] AL-SHEHBAZ, BRAYOPSIS AND EUDEMA 107 Cespitose, rhizomatous perennials forming dense cushions, rhizomes branched, 1-2.5 mm wide, with sessile, ovate to lanceolate scales at nodes. Leaves rosulate, somewhat fleshy, narrowly oblanceolate to spatulate, (3—-)8- 20(-30) mm long, (0.7-)1-2(-3.5) mm wide, obtuse to rounded at apex, entire, attenuate at base, glabrous on lower surface, usually densely pubescent on upper surface with dendritic trichomes; petioles ciliate with simple or furcate tri- chomes, somewhat expanded at base. Sepals erect, oblong, 3-4 mm long, 1.4— 1.8 mm wide, obtuse at apex, scarious at margin, sparsely pubescent on back with dendritic trichomes. Petals oblanceolate, (4.5—)5-7 mm long, 1.5-2.2 mm wide, rounded at apex, cuneate at base, white to creamy white. Filaments 2.5— 3.5 mm long; anthers oblong, 0.8-1 mm long. Fruiting peduncles glabrous, 3.5-20 cm long. Fruits sessile, suborbicular, strongly compressed parallel to replum, 2.5-3 mm in diameter; septa reduced to narrow rim; styles 0.2-0.8 mm long; stigmas capitate, slightly 2-lobed. Seeds 3 to 5 on each placenta, ovoid, 0.9-1.2 mm long, 0.6-0.8 mm wide. REPRESENTATIVE SPECIMENS. Argentina: Prov. SAN JUAN: Quebrada Ortiga, E part of Cordillera de la Ortiga, Johnston 6174 (BA, GH, S), 6290 (GH); Quebrada de Coucouta, between Las Vicuftitas and el Portezuelo de Coucouta, Moreau 30/87 (BA); Iglesia, Rio Blanco cerca del Bafio del Bollete, Castellanos 15523 (us). Eudema werdermannii forms prostrate mats in wet gravel at altitudes of 3800 to 4300 m. Field notes on Johnston 6174 (Gu) indicate that the plant produces much-branched rhizomes that form a tangled growth in the gravel, and that the plants are smaller and less branched in drier sites than in wetter areas. Eudema werdermannii is highly variable in habit and leaf size and shape. Plants of the type collection have much-shortened rhizomes that terminate in a compact, few-branched caudex. In contrast, those of Johnston 6174 are rather lax and have long, slender rhizomes with a conspicuous, sessile scale at each node. The variation in leaf size and shape is evident within each population, and apparently plants of the drier sites have smaller leaves with shorter petioles than those of the wetter ones. 6. Eudema friesii O. E. Schulz, Pflanzenr. IV. 105(Heft 86): 245. 1924. Type: rgentina, Prov. Salta, Incachuli bei San Antonio de las Cobras, ca. 5000 m alt., 30 Oct. 1901, Fries 703 (holotype, B! (photos, F!, GH!, NY!); isotype, UPs!). Prostrate rhizomatous perennials; rhizomes slender, branched, with con- spicuous, sessile scale at each node; branches terminated by rosettes and cov- ered below leaves with petiolar remains of previous years. Leaves rosulate, petiolate, lanceolate to linear, (5-)9-16 mm long, (1.5—)2-3 mm wide, obtuse at apex, pinnately lobed to rarely dentate at margin, glabrous on lower surface, sparsely pubescent on upper surface and margin with dendritic trichomes. Flowering peduncles originating from center of rosette, 2-4 mm long. Sepals erect, ovate, 2.5—3.5 mm long, ca. 1.5 mm wide, glabrous or sparsely pubescent below apex, scarious at margin. Petals spatulate, 5—6.5 mm long, 1.2-1. wide, creamy white, rounded at apex, attenuate to clawlike base. Filaments 3— 3.5 mm long; anthers oblong, 0.9-1 mm long. Young fruits ovate, apparently 108 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 71 compressed parallel to replum, glabrous; septa reduced to narrow rim; styles ca. 0.5 mm long. Seeds not seen. ADDITIONAL SPECIMEN EXAMINED. Bolivia. DEPTO. ORURO: Prov. Carangas, Sajama, 4200 m alt., Asplund 3988 (s). — Eudema friesii is a very rare species that apparently grows on fine gravel. Its nearest relative is &. werdermannii, which it resembles in having branched trichomes, slender rhizomes, and creamy white flowers. It is easily distinguished from the remaining species of Eudema by its pinnately lobed instead of entire leaves. EXCLUDED NAMES Eudema colobanthoides (Skottsb.) O. E. Schulz, Pflanzenr. IV. 105(Heft 86): 246. 1924, = Xerodraba colobanthoides Skottsb. Kongl. Svenska Ve- tenskapsakad. Handl. 56(5): 234. 1916. Eudema diapensioides (Wedd.) O. E. Schulz, Pflanzenr. IV. 105(Heft 86): 245. 24. = Brayopsis diapensioides (Wedd.) Gilg & Muschler, Bot. Jahrb. Syst. 42: 484. 1909. Eudema glebaria (Speg.) Gilg & Muschler, Bot. Jahrb. Syst. 42: 472. 1909. = Xerodraba age ee ) Skottsb. Kongl. Svenska Vetenskapsakad. Handl. 56(5): 362. 19 Eudema ey a & Muschler, Bot. Jahrb. Syst. 42: 471. 1909. = Xerodraba lycopodioides (Speg.) Skottsb. Kongl. Svenska Vetenskap- sakad. Hand]. 56(5): 362. 1916. Eudema microphylla (Gilg) Gilg & Muschler, Bot. Jahrb. Syst. 42: 472. 1909. = odraba pycnophylloides (Speg.) Skottsb. var. microphylla (Gilg) O. E. Schulz, Pflanzenr. IV. 105(Heft 86): 250. 1924. Eudema monantha (Gilg) Gilg & Muschler, Bot. Jahrb. Syst. 42: 472. 1909. = Xerodraba monantha (Gilg) Skottsb. Kongl. Svenska Vetenskapsakad. Handl. 56(5): 362. 1916. Eudema patagonica (Speg.) Gilg & Muschler, Bot. Jahrb. Syst. 42: 471. 1909. = Xerodraba ave ae ) Skottsb. Kongl. Svenska Vetenskapsakad. Hand. 56(5): 362 Eudema pectinata (Speg.) ce & Muschler, Bot. Jahrb. Syst. 42: 471. 1909. = Xerodraba sores ote Kongl. Svenska Vetenskapsakad. Hand. 56(5): 23 Eudema pn oo Gilg & Muschler, Bot. Jahrb. Syst. 42: 471. ce = Xerodraba pycnophylloides (Speg.) Skottsb. Kongl. Svenska enskapsakad. Handl. 56(5): 362. 1916. Eudema es ia Philippi, Anales Univ. Chile 41: 675. 1872. = Draba hilippii O. E. Schulz, Notizbl. Bot. Gart. Berlin-Dahlem 10: 559. 1929. Eudema trichocarpa Muschler, Bot. Jahrb. Syst. 40: 276. 1908. = Weberbauera trichocarpa (Muschler) J. F. Macbr. Candollea 5: 356. 1934 ACKNOWLEDGMENTS Iam most grateful to Reed C. Rollins for his critical review of the manuscript and to Donald H. Pfister for obtaining funds from the Harvard University 1990] AL-SHEHBAZ, BRAYOPSIS AND EUDEMA 109 Herbaria that supported the SEM portion of the research. Many thanks are due Elizabeth A. Shaw for correcting the Latin diagnoses, Trisha Rice for the SEM work, Barbara Nimblett for typing the manuscript, Gustavo Romero for his help with the Spanish, Michael A. Canoso for obtaining the loans, my wife, Mona, for her continuous support, and Elizabeth B. Schmidt and Stephen A. Spongberg for their editorial advice. I am thankful to the directors and curators of the herbaria (abbreviations follow Holmgren et a/., 1981), who kindly sent the material for this study. LITERATURE CITED AL-SHEHBAZ, I. A. 1989a. The South American genera Brayopsis and Englerocharis (Brassicaceae). Nord. J. Bot. 8: 619-625. b. New or noteworthy Draba (Brassicaceae) from South America. J. Arnold Arbor. 70: 427-437. 1989c. Dactylocardamum (Brassicaceae), a remarkable new genus from Peru. Ibid 515-521. BAEHNI, C., & J. F. Macsripe. 1937. Remarques sur les Cruciferae-Sisymbrieae. Can- BoELcKE, O. 1982. Un nuevo género de la Patagonia Argentina, Skottsbergiella (Cru- ciferae). Hickenia 1: 305-310. & M.C. Romanczuk. 1984. Cruciferae. Fl. Patagonica 4A: 373-5 Giue, E., & R. Muscuier. 1909. Aufzahlung aller zur Zeit bekannten Seen ischen Cruciferen. Bot. Jahrb. Syst. 42: 437-487. HoLmarEN, P. K., W. KEUKEN, & E. K. SCHOFIELD. 1981. Index herbariorum. ed. 7. Regnum Veg. 106: 1-452. 13(2): 937-983. 1936. —— In: H. Harms, ed., Nat. Pflanzenfam. ed. 2. 17B: 227-658. BOGLE, MYTILARIA 111 MULTILACUNAR NODAL ANATOMY IN MYTILARIA (HAMAMELIDACEAE)! A. LINN BOGLE? Multilacunar nodal anatomy associated with unique sheathing stipules in the genus Mytilaria (subfam. Exbucklandioideae) is reported for the first time in the Hamamelidaceae, a family usually characterized by trilacunar nodes. The number of gaps and traces ranged from eight to 12 in the nodes analyzed. The significance of anatomical data from the stem-node-leaf continuum for systematic and phylogenetic studies is now generally recognized. The literature and the development of ideas in this field have been well reviewed by Howard (1979) and others (Carlquist, 1961; Radford et a/., 1974; Takhtajan, 1980). One of the major results of early studies of nodal anatomy in dicotyledons (Sinnott, 1914; Sinnott & Bailey, 1914) was the recognition of three major nodal types: trilacunar three-trace, unilacunar one-trace, and multilacunar mul- titrace. It was also established that the presence of stipules was usually cor- related with trilacunar nodes, the absence of stipules with unilacunar nodes, and the presence of sheathing leaf bases or stipular structures with multilacunar nodes. The purpose of this paper is to report the nodal structure of the unusual unistipulate leaves of the monotypic genus Mytilaria Lecomte. The nodes display a vascular pattern not previously known in the Hamamelidaceae but of apparent taxonomic significance. In the Hamamelidaceae all of the approximately 30 currently recognized genera are stipulate. In most cases the stipules are paired, small, membranous, and ephemeral. However, the stipules of the three genera usually grouped in subfam. Exbucklandioideae are among the most unusual in the family. Those of Exbucklandia R. W. Brown and Chunia H. T. Chang are large, leathery, oblong to orbicular, and connate in the bud, completely enclosing the younger leaves and the shoot apex (Griffith, 1836; Lubbock, 1899; Chang, 1948). Their broadly attached bases leave a girdling scar at the node, imposing a jointed appearance on the stem. In contrast, the leaves of Mytilaria laosensis Lecomte are unique in the family in having a single, long, narrowly conical, sheathing stipule, which Lecomte (1924) thought consisted of two stipules united into a conical sheath. It attaches all the way around the stem, opens by a single 'This paper represents scientific contribution no. 1604 from the New Hampshire Agricultural Experiment Station. 2Botany and Plant Pathology Department, University of New Hampshire, Durham, New Hamp- shire 03824. © President and Fellows of Harvard College, | Journal of the Arnold Arboretum 71: 111-118. rca 1990. £42 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 FETT SS @ b® Nodal anatomy of Mytilaria laosensis: A, stem tip, showing young leaf prior to un- folding, with stipule still attached, and slightly older, expanding leaf with stipule scar at ase (arrows indicate positions of sections in C-H); B, diagram of transverse section through bud, showing succession of 4 petiole-stipule sheath complexes in different stages of development (arrow indicates axillary bud; C, transverse section of internode, aes differentiation of 12 leaf traces (blank circles = = gum ducts of ground tissue in C-H); D opposite leaf base, and horizontal course of some leaf traces on side of stem opposite 1990] BOGLE, MYTILARIA 115 longitudinal suture that faces the petiole of the associated leaf, and at maturity is circumscissile caducous, also leaving the stem with a girdling scar and a jointed appearance (see FiGure, A, B). In general, anatomical data on the stem-node-leaf continuum in the Hama- melidaceae are fragmentary and scattered in a number of publications. Some of these studies (Sinnott, 1914; Sinnott & Bailey, 1914; Skvortsova, 1960; Bisht et al., 1983) have emphasized the number of gaps in the stem stele and the number of traces departing from them to enter the leaf base. Others (Thou- venin, 1890; Morvillez, 1919; Tong, 1930; Chang, 1948; Covin, 1959; Skvort- sova, 1960; Harjal et al., 1984) have emphasized the number of vascular strands entering the leaf base (without reference to the structure of the node), the pattern formed by the vascular tissue at different levels in the petiole and midrib, and whether the vascular bundles are collateral or concentric in form. With regard to the subfamily Exbucklandioideae, several of these authors (Mor- villez, 1919; Tong, 1930; Skvortsova, 1960) have included observations on the vascular structure of the petiole and midvein of one or another of the three genera of the subfamily in the context of broader studies of petiolar anatomy, but not on their nodal anatomy. Among authors considering the anatomy of the node in the Hamamelidaceae, Sinnott (1914) described the Hamamelidaceae as being typically trilacunar. Sinnott and Bailey (1914, p/. 44, fig. 4) studied at length the origin of the vascular supply to the stipules, including those of Hamamelis L., which they designated (p. 453) as the “‘typical trilacunar dicotyledon,” and indicated that in trilacunar nodes the stipular supply is typically derived from the lateral traces. Skvortsova (1960, p. 126), as part of a broader study of petiole structure in the Hamamelidaceae and the Altingiaceae, investigated the nodal structure of leaves in which the vasculature “in the lower part of the petiole was established as an arc or semicircle” and concluded that a correlation existed between trilacunar nodal structure, three traces entering the leaf base, and a vascular pattern in the form of an arc or semicircle in the lower part of the petiole. Bisht and colleagues (1983), on the other hand, reported unilacunar nodal structure for four genera of Hamamelidaceae (Corylopsis Sieb. & Zucc., Di- stylium Sieb. & Zucc., Hamamelis, Parrotia C. Meyer). the authors who have concentrated on the vascular pattern of the petiole, Morvillez (1919, fig. 5) described three traces entering the leaf base in species of Hamamelis, Parrotia, Fothergilla L., and Liquidambar L. and im- plied a similar condition in Disanthus Maxim., Exbucklandia (as Bucklandia leaf base; F, transverse section through base of petiole, showing vascular cylinder of 8 bundles; G, transverse section at mid-level of petiole, bundles forming vascular cylinder; H, transverse section through top of petiole, vascular cylinder broken up into discrete bundles, each of which will form leaf vein; I, diagrammatic representation of vascular pattern in node with 8 gaps and traces (dashed lines indicate branching patterns of lateral traces supplying petiole and stipule sheath in I, J); J, diagrammatic representation 2) of vascular pattern in node with 9 gaps and traces. (Abbreviations: gd = gum duct, pe petiole, stip = stipule.) 114 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 R. Br.), and A/tingia Noronha. He described the structure of the vascular system at different levels in the petiole and the derivation of the major leaf veins from the petiole vasculature. Tong (1930) described the vascular condition in the lower and upper parts of the petiole and the base, middle, and tip of the midvein for 20 species of 14 genera of Hamamelidaceae, including E-xbucklandia (as Bucklandia). He described three collateral bundles in the lower part of the petiole for 17 of the species, one collateral bundle in two species of Fothergilla, and one concentric bundle in the lower part of the petiole of Exbucklandia, this giving rise to three concentric and five collateral bundles in the upper part of the petiole. Covin (1959, fig. 6) illustrated the petiole vasculature of Mytilaria as being “polystelic” at mid-petiole level, while at the base of the midvein it formed a “simple fold” in Mytilaria and a “simple are” in Exbucklandia (as Bucklandia). Skvortsova (1960) examined the petiolar vascular system at the low, middle, and high levels in 22 genera of the family, including Exbucklandia and My- tilaria, and emphasized the configuration of the vasculature at mid-level. She found that all genera except Liguidambar, Altingia, Mytilaria, and Embolanthe- ra Merr. have a cylinder of vasculature at mid-level. Mytilaria is illustrated (fig. 1-22) as having four arcuate bundles and an adaxial vascular band at mid- level, : ae to a complete vascular cylinder at that level in Exbucklandia (fig. I- ae Aon. p. 67), in his description of the new genus Chunia, provided the only direct comparative data for all three genera of Exbucklandioideae, stating that in the petiole of Chunia the vascular tissue “unites into a cylinder with prominent ridges,” while in Exbucklandia (as Bucklandia) it is “horseshoe shaped,” and in Myrilaria it is “arranged in discrete bundles.” MATERIALS AND METHODS The observations reported here are based on hand sections, clearings, and serial sections of young, alcohol-preserved nodes and buds of Mytilaria laos- ensis obtained from the Biological Laboratory-Factory, Zhaoqing Training Col- lege, Zhaoqing, Guangdong, People’s Republic of China (no collection number or data provided); on herbarium specimens of Chunia (Wang 36075, NYBG); and on personal field collections of Exh Inea (Bogle 1379, 1387). All voucher materials are on deposit in the Botany and Plant Pathology De- partment of the University of New Hampshire (NHA). For clearing, young nodes were bleached in 5% NaOH, washed in tap water, and cleared in 5% chloral hydrate. Transverse sections of nodes and petioles were made by standard paraffin-embedding techniques and stained with saf- ranin-fast green (Sass, 1958). Hand sections were stained with phloroglucinol- 50% HCl and mounted in glycerine. OBSERVATIONS Transverse sections of a young bud of Mytilaria laosensis revealed a con- densed series of concentric stipular sheaths with their associated petioles. A 1990] BOGLE, MYTILARIA 115 single suture was seen as a thin area in the stipule sheath opposite the petiole. There was no evidence of a second suture on the opposite side of the sheath to support an interpretation of two stipules united into a conical sheath. The stipule sheath encloses the related axillary bud (see Ficure, B, arrow). In sections of the stem of Myvtilaria taken below the node of a young leaf, a number of leaf traces and gaps were clearly differentiated (see Ficure, C). The gaps were more or less equally spaced around the circumference of the vascular cylinder. The number of gaps and traces ranged from eight to 12 in the nodes examined (see FiGure, C, I, J). The three traces originating nearest the leaf base moved directly and inde- pendently into the leaf base to form the three major abaxial bundles of the petiole vascular system. The five to nine bundles originating farther around the stem from the leaf base bent toward the leaf and ran more or less hori- zontally around the stem in the ground tissue, toward the leaf base, often anastomosing with adjacent traces to form an incomplete vascular ring around the node. As many as 18 to 24 minor branches diverge from the horizontal traces to provide the vascular supply to the base of the sheathing stipule (see Ficure, I, J). The trace farthest from the leaf base bifurcates after leaving the gap, sending branches in each direction around the node, as well as to the stipule (see FIGURE, D, I, J). Occasionally a leaf trace provided traces to the stipule but not to the leaf base (see Ficure, J, arrow). In the base of the petiole, seven or eight major bundles were observed, sometimes accompanied by minor cortical strands (see FiGurE, F). The three large abaxial bundles were each derived from single leaf traces originating from the three gaps directly opposite the leaf base, as described above. The two larger bundles in each angle of the petiole cross section resulted from the branching of a single bundle derived on each side of the node from fusion of the horizontal trace system (see Ficure, I, J). Additional, smaller bundles may appear on the adaxial side of the petiole vascular cylinder, or in the central ground tissue, as derivatives of minor branches of the major bundles (see FIGURE, F, arrows). At mid-petiole level the major bundles of the petiole base had fused into a fluted vascular cylinder (see Figure, G), which continued almost to the tip of the petiole. Just prior to joining the blade, the vascular cylinder again broke up into distinct bundles, each of which formed one of the major veins of the palmately veined leaf (see FiGure, H). In contrast to the multilacunar nodal structure of Mytilaria, the nodes in my material of Exbucklandia exhibit trilacunar, three-trace structure, with vascular supply to the large stipules derived from the lateral traces. DISCUSSION Nodal anatomy is generally considered to be a conservative feature of plant structure and to be a useful character at various levels in phylogenetic and taxonomic studies. However, Carlquist (1961, p. 87) warned that nodal anat- omy can vary within a single plant, and that “there is no single point along 116 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 the petiole that could be called ‘characteristic’ or ‘typical,’ or be used to the exclusion of other sections for comparative purposes.” With regard to the Hamamelidaceae, Sinnott (1914) listed the family as trilacunar, while Skvortsova (1960) reported trilacunar nodes in those taxa of the family in which the vascular pattern in the lower part of the petiole is arc shaped or semicircular. Unfortunately, Skvortsova did not comment on the nodal structure of those taxa having a different vascular pattern in the lower petiole, and neither author was able to study all genera of the family. My observations (unpublished) on the nodes of foliage leaves of various genera of Hamamelidaceae (A/tingia, Liquidambar, Rhodoleia Champ., Exbucklandia, Disanthus, Hamamelis, Eustigma Gardner & Champ., Corylopsis Sieb. & Zucc., Sinowilsonia Hemsley, Parrotia, Parrotiopsis Schneider, Distylium, Matudaea Lundell) indicate that all have trilacunar, three-trace nodes. There are, however, reports in the literature that present conflicting infor- mation. Bisht and colleagues (1983, fig. 7) reported unilacunar, one-trace nodes in Distylium racemosum Sieb. & Zucc., Hamamelis japonica Sieb. & Zucc., and Hamamelis virginiana L. (which Sinnott & Bailey, 1914, regarded as the typical trilacunar dicotyledon); unilacunar, three-trace nodes in Parrotia persica C. Meyer, and unilacunar, four-trace nodes in Corylopsis spicata Sieb. & Zucc. In contrast, Harjal and co-workers (1984, table 2), considering only the petiole vasculature, reported only one strand entering the petiole in Corylopsis spicata, Distylium racemosum, Rhodoleia forrestii Chun ex Exell, and Trichocladus crinitis Pers., as opposed to three in Liquidambar styraciflua L. and Parrotiopsis Jacquemontiana Rehder. Thouvenin (1890), also considering only the petiole vasculature, reported three bundles departing from the stem (without reference to the number of lacunae) and entering the petiole in both Trichocladus crinitis and 7. e/liptica Ecklon & Zeyher. Thus, several taxa (Distylium racemosum, Hamamelis japonica, H. virgin- lana, Corylopsis spicata, Parrotia persica) have been reported or observed to have either trilacunar, three-trace nodes, or unilacunar nodes with one, three, or four traces. These and other inconsistencies suggest the need for a more thorough and detailed study of the stem-node-leaf continuum in the Hama- melidaceae, particularly with regard to intraspecific variation. Anatomical characters from the stem-node-leaf continuum have been used in some taxonomic studies regarding the relationships of genera and subfamilies in the Hamamelidaceae, including Mytilaria and Exbucklandia. Morvillez (1919) considered the similar petiolar structure in Disanthus (which now comprises subfam. Disanthoideae) and Exbucklandia (as Bucklandia, the two genera to- gether then constituting tribe Bucklandieae) to be intermediate forms, linking the tribe Balsamiflueae (Liquidambar and Altingia, now subfam. Liquidam- baroideae) to the remainder of the Hamamelidaceae. Skvortsova (1960), on the other hand, considered the data from petiole vascular structure to support the segregation of Liquidambar and Altingia as a family (Altingiaceae), to which she suggested that Mytilaria is related. The distinctive vegetative morphology shared by Exbucklandia, Chunia, and Mytilaria makes them easily identifiable as the only members of subfam. Ex- bucklandioideae. However, the three genera differ significantly in floral mor- phology, and in recent years they have been separated by some authors into 1990] BOGLE, MYTILARIA 117 two subfamilies. Chang (1973) segregated Mytilaria and Chunia in the new (and sixth) subfamily Mytilarioideae, leaving only Exbucklandia in the Ex- bucklandioideae. Huang and Lee (1982) supported this segregation on the basis of wood anatomy, suggesting a closer relationship of the Mytilarioideae to subfam. Disanthoideae than to any of the other subfamilies (i.e., Liquidam- baroideae, Rhodoleioideae, Hamamelidoideae). Takhtajan (1980), on the other hand, listed subfam. Chunioideae (rather than Mytilarioideae) as one of six subfamilies in the Hamamelidaceae but did not indicate which of the three genera of Exbucklandioideae he would assign to it or his reasons for removing them from the Exbucklandioideae. Mytilaria is clearly set apart from the other two genera of Exbucklandioideae by its single, conical, sheathing stipules; its gum ducts (secretory canals) in the ground tissue throughout the plant, including the flowers; its numerous, com- plete flowers with nearly inferior ovaries sunken in a long, fleshy spike; its petals and stamen filaments fused basally into a short tube; and its ten stamens with horned filaments surmounted by hooded anthers, connivent in a cycle over minute styles and stigmas. The presence of multilacunar nodes in Myti- laria, as opposed to trilacunar nodes and paired stipules in Exbucklandia, adds yet another distinctive character to this list, suggesting that separation of My- tilaria from Exbucklandia is more appropriate at the tribal level within subfam. Exbucklandioideae than as the separate subfam. Mytilarioideae. The position of Chunia in such a subdivision of the Exbucklandioideae is as yet uncertain. Its floral morphology is distinctive (naked, bisexual flowers sunken in a relatively-few-flowered, short spike) and is perhaps closer to that of Exbucklandia than that of Mytilaria. It also shares paired, connate stipules with Exbucklandia. However, my preliminary observations on the few nodes available for study indicate that it, too, has multilacunar nodes. This would tend to emphasize the unity of the three genera as a single subfamily. At the same time, it could be taken as support for the taxonomic separation of M)- tilariaand Chunia from Exbucklandia,as proposed by Chang (1973). However, the significant differences in floral morphology between Mytilaria and Chunia may militate against their being paired in this way. Investigation 1s continuing on Chunia. ACKNOWLEDGMENTS I wish to express thanks to the directors of the Harvard University Herbaria, the Museum of Natural History, Paris, and the New York Botanical Garden for access to or loans of specimens of Chunia, Exbucklandia, and Mytilaria. | am also deeply grateful to Shiu-ying Hu for translating the Chinese literature, and to Richard A. Howard for reviewing the original manuscript. Financial support from the University of New Hampshire Central University Research Fund, grant no. 980, is gratefully acknowledged. LITERATURE CITED Bisut, P. S., C. B. Maras, & G. S. PALIWAL. 1983. Nodal organization in Hamamel- ididae. Geophytology 13: 88-92. 118 JOURNAL OF THE ARNOLD ARBORETUM [VvoL. 71 CARLQUIST, S. 1961. Comparative plant anatomy. Holt, Rinehart and Winston, New ork. CHANG, H. T. 1948. Additions to the hamamelidaceous flora of China. Sunyatsenia 7: : 1973. A revision of the hamamelidaceous flora of China. Sunyatsen Univ. Bull. 1: 54-71. 1959. L’appareil libéro-ligneux foliaire des Hamamelidacées. Bull. Soc. Bot. N. France 12: 31-33. GRIFFITH, W. 1836. Description of two genera of the family of Hamamelideae, two species of Podostemon and one species of Kaulfussia. Asiat. Res, 19: 94-114. pls. = HarJAL, N., P. S. Bisut, & G. S. PALIWAL. 1984. Foliar vasculature as an aid to the classification of Hamamelididae. Pp. 282-300 in G. S. PALIWAL, ed., The vegeta- tonal wealth of the Himalayas. Puja Publishers, Delhi. ‘ . 79. The stem-node-leaf continuum of the Dicotyledoneae. Pp. 76- 87 in C. R. METCALFE & L. CHALK, Anatomy of the dicotyledons. ed. 2. Vol. 1. Huane, G., & C. Lee. 1982. Anatomical studies of Chunia wood. [In Chinese with English abstract.] Acta Bot. Sinica 24: 506-511. Lecomte, H. 1924. Une Hamamélidacée nouvelle d’Indochine. Bull. Mus. Hist. Nat. (Paris) 30: 503-507. Lussock, J. 1899. On buds and stipules. Kegan Paul, Trench, Truebner & Co., Ltd., Lond ondon. MorviLiez, M. F. 1919. L’appareil conducteur foliaire des Hamamélidacées et des formes voisines. Compt. Rend. Hebd. Séances Acad. Sci. 169: 542-545, RapForp, A. E., W. C. Dickison, J. R. MASSEY, & C. R. BELL. 1974. Vascular plant systematics. Harper an Sass, J. E. 1958. Botanical microtechnique. ed. 3. Iowa State University ane Ames. SmnNoTT, E. W. 1914. Investigations on the phylogeny of the angiosperms. I. The anatomy of the node as an aid in the classification of angiosperms. Amer. a Bot. 1: 303-322. pls. 30-35. —— & I. W. Bamey. 1914. Investigations on the phylogeny of 2 angiosperms. 3. Nodal anatomy and the morphology of stipules. Amer. J. Bot. 1: 441-453. p/. 44. Skvortsova, N. T. 1960. The anatomical structure of the ae system of leaf petioles of representatives of families Hamamelidaceae and Altingiaceae. Dokl. Akad. Nauk SSSR 133: 1231-1234. TAKHTAJAN, A. 1980. Outline of the classification of flowering plants (Magnoliophyta). Bot. Rev. (Lancaster) 46: 225-359. THOUVENIN, M. 189 Recherches sur la structure des Saxifragacées. Ann. Sci. Nat. Bot., ser. 7. 12: 1-174. Tonc, K. 1930. Studien ueber die Familie der Hamamelidaceae, mit besonderer Beriicksichtigung der eae und Entwicklungsgeschichte von Corylopsis. Bull. Dept. Biol. Sun Yatsen Univ 1990] BOUFFORD ET AL., FLORA OF CHINA 119 ADDITIONS TO THE FLORA OF CHINA DAvipD E. BOUFFORD,! ZHAN-HUO TsI,* AND PEISHAN WANG? taxa, Dryopteris kweichowicola (Dryopteridaceae), Gastrochilus nanus Ga Rubus fanjingshanensis reer te and Sabia swinhoei var. parvifolia (Sabiaceae), are described as new. Deinostema adenocaulon is re- ported for the first time from China and oe a Rae of ca. 1800 km from the closest populations in Korea. Joint fieldwork in northeastern Guizhou Province, China, in 1986 by Chinese and American botanists resulted in the collection of 2474 numbers and 28,633 sheets of flowering plants and ferns, including four previously undescribed taxa, Dryopteris kweichowicola Ching ex P. S. Wang ale desea Gastrochilus nanus Z. H. Tsi (Orchidaceae), Rubus fanjingshanensis L. T. Lu (Rosaceae), and Sabia swinhoei Hemsley ex Forbes & Hemsley var. parvifolia Y. H. Xiang & Q. H. Chen (Sabiaceae), and one, Deinostema ad ) Yamaz (Scrophulariaceae), previously unknown in China. The collections are all from the vicinity of the 38,000-hectare Fanjing Shan Nature Reserve, one of six Man and the Biosphere reserves in China (Chen, 1988). The reserve contains rich botanical resources in associations ranging from broad-leaved evergreen forests at the lowest elevations (ca. 550 m) through mixed deciduous—broad- leaved evergreen forests at middle elevations to subalpine scrub and meadows and dwarf bamboo thickets near the summits of the highest peaks at ca. 2550 m (Huang et a/., 1982). The area, an island of relatively undisturbed vegetation, is unusual within the vast, mostly deforested south-central part of China be- cause of the richness of its flora. Dryopteris kweichowicola Ching ex P. S. Wang, sp. nov. FIGURE 1. Affinis Dryopteris wallichianae (Sprengel) Hylander, sed minore, folliis tenui- ter chartaceis, soris prope marginem, venis utrinque facie conspicuis. Plants 50-70 cm tall. Rhizome short, erect. Fronds tufted, thin-chartaceous, the upper surface nearly glabrous, with few hairlike scales along costa, the lower surface with ovate-lanceolate scales; stipe 13-16 cm long, 3-5 mm in diameter, clothed in dense, brown scales, those at stipe base entire and those above ‘Harvard University Herbaria, 22 Divinity Avenue, Cambridge, Massachusetts 02138, U.S.A Institute of Botany, Chinese Academy of Sciences, Xiangshan, Beiing 100093, People’s Republic of China. 3Guizhou Institute of Biology, Academy of Sciences of Guizhou, Xiaohe, Guiyang, Guizhou, Peo- ple’s Republic of China. © President and Fellows of Harvard College, | Journal of the Arnold Arboretum 71: 119-127. ae 1990. JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 ( + eS = K f ne y te j : ROSS Se PD 35 .- we — : Nae” ay U WM psx \ g (oo AN as by A .\ Gir~ y Bolts ‘ AU “2 vay ee 9 uRE |. Dryopteris kweichowicola: a, frond; b, pinnule; c, scales from rhizome; d, scales from stipe. 1990] BOUFFORD ET AL., FLORA OF CHINA 121 fimbriate; leaf blade lanceolate, 35-50 cm long, 12-15 cm wide, gradually narrowed toward base, acuminate at apex, bipinnate; pinnae 25 to 30 pairs, patent, the lowest ones slightly reflexed and much shortened, the middle ones lanceolate, 7-8 cm long, 1—-1.3 cm wide, sessile, with base truncate, apex short- acuminate, pinnule segments ca. 20 pairs (these patent, oblong, ca. 5-6 mm long, 2-3 mm wide, the apex blunt, with 3 to 5 teeth, the veins 4- to 6-jugate, oblique, simple or forked). Sori 2 to 4 pairs per segment near margin; indusium brown, persistent. Type. China, Guizhou Province, Jiangkou Xian, vicinity of Jinding along the crest of the Fanjing Shan mountain range, elev. ca. 2200 m, growing at base of old wall in moist situation, 27 August 1986, Sino-Amer. Guizhou Bot. Exped. 45] (holotype, PE; isotypes, A, CAS, HGAS, TI). ADDITIONAL SPECIMENS EXAMINED. China. GUIZHOU PROVINCE: without further ra S. W. Teng 51367 (PE), 51368 (PE); eee Xian, by small stream, 1700 m alt., Z. S. Zhang et al. 402211 (PE This species is very similar to Dryopteris wallichiana in the shape of the fronds. The essential difference between the two is in the position of the sori: in the former they are near the margins of the segments, while in the latter they are near the costae. Gastrochilus nanus Z. H. Tsi, sp. nov. FIGURE 2. Species Gastrochilus toramani (Makino) Schltr. et G. raraensis Fukuyama affinis, ab illis floribus sine punctis purpureis; forma epichilii et petalorum; hypochilo cylindrico-subcylindrico differt. Epiphytic. Stems creeping, 3-4 cm long, with thick and slightly flexed roots. Leaves many, distichous, alternate, closely arranged, spreading, somewhat fleshy, elliptic-oblong, 8-10 mm long, 5-6 mm wide, apex acute, base jointed to sheaths, green with purple spots on both surfaces. Inflorescence leaf opposed, suberect-patent, 6-1 1 mm long, bearing 5 or 6 subumbellately arranged flowers; peduncle slightly fleshy, gradually thickened from base to apex, with tubular sheath near base; floral bracts with purple spots, somewhat fleshy, ovate-tri- angular, ca. |-1.3 mm long, apex acute; ovary greenish, 4-5 mm long, slightly angled; flowers spreading, greenish: sepals and petals somewhat fleshy, apex obtuse, |-veined; dorsal sepal elliptic, concave, 2.2 mm long, 1.2 mm wide, the lateral pair + oblique-oblong, as long as dorsal one; petals oblong, smaller than sepals; labellum adnate to lower half of column, obscurely 3-lobed; lateral lobes of hypochile small, suberect, triangular, villous internally; epichile mem- branous, patent, reniform, ca. 2 mm long, 2.5 mm wide at base, subtruncate, apex notched, edge and upper surface densely pubescent, with olivaceous, thickened patch in disc; spur straight, nearly parallel to pedicel, cylindrical, 2.8-3 mm long, ca. | mm broad, + contracted 1n middle, apex broadly rounded; column thick, ca. 0.5 mm long; anther hemispherical, abruptly contracted into triangular, acute apex; rostellum bifid; pollinia 2, globose, attached by linear caudicle to incrassate, oblong, apically bifid gland. 122 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Ly da Tee yf b 1.5mm FiGureE 2. Gastrochilus nanus: a, habit; b, flower, front view; c, flower, side view; d, operculum; e, stipe: f, leaf tip. Type. China, Guizhou Province, Jiankou Xian, Yuao, in the valley of the Heiwan River, elevation ca. 1000 m, 26 August 1986, Sino-Amer. Guizhou Bot. Exped. 407 (holotype, PE; isotypes, A, HGAS). Gastrochilus nanus is allied to G. toramanus and G. raraensis, but it differs from the former in having a reniform epichile and oblong petals and from the 1990] BOUFFORD ET AL., FLORA OF CHINA 123 latter in having smaller greenish flowers with a cylindrical to subcylindrical, vertical hypochile and differently shaped petals. Rubus fanjingshanensis L. T. Lu, sp. nov. FIGURE 3. Species ab R. treutleri J. D. Hooker, foliis 3-5-lobatis, lobis terminalibus apice acutis, lateralibus obtusis, margine irregulariter grosse-serratis; stipulis angustioribus, 10-15 mm longis, palmatis partitis, lobis linearibus vel lineari- lanceolatis, sepalis internis late-ovatis vel ovato-lanceolatis; externis foliolis, margine dissectis, lobulis lanceolatis, differt. Small, trailing shrub; branches terete, the older ones dark brown, the youn shoots brown or brownish, villous, with thin acicular prickles slightly dilated at base intermixed with stipitate glands. Stipules free, 1.4-1.8 cm long, 1-1.4 cm wide, villous and glandular-hairy on both surfaces, palmately parted, the lobes lanceolate or triangular-lanceolate, 3-7 mm long, 1.3-3 mm wide. Leaves with petiole 4-8 cm long, long-hairy and with acicular prickles and stipitate glands intermixed; blade orbicular or nearly orbicular, 7-11 cm long, 6.5—-11 cm wide, 5- (rarely 7-)lobed, the lobes obtuse to rounded with terminal one slightly longer than or equal to lateral ones, base deeply cordate, margins irregularly serrulate, apex obtuse, upper surface dark green, lower surface paler, both surfaces adpressed-villous but lower surface more densely so, veins gla- brescent and sparsely hairy at maturity, midrib and lateral veins on lower surface with acicular prickles and sometimes also sparse stipitate glands. In- florescences and flowers not seen; infructescences terminal short racemes, 4—5 cm long, or fascicles in axils of the leaves; rachises, fruiting pedicels, and calyces covered with long hairs and acicular prickles, sometimes with sparse stipitate glands intermixed; fruiting pedicels 8-11 mm long; bracts resembling stipules, but smaller; calyx cup shaped; sepals 5, broadly ovate-lanceolate to oblong- ovate, 8-14 mm long, 4-6 mm wide, the outer ones usually wider, dissected at apex or in upper part, with lobes linear to linear-lanceolate, 2-5 mm long, 1-1.5 mm wide, apex caudate; persistent stamens numerous, 3-4 mm long, filaments somewhat broadened at base, anthers globose or short-oblong; per- sistent pistils numerous, glabrous; torus elevated, long-hairy. Fruits aggregates consisting of many united drupelets, subglobose, 7-10 mm in diameter, red, glabrous, crowned with persistent calyx; stones subreniform, |.5—2.5 mm long, 1-1.8 mm wide, distinctly rugose. Fruiting in August and September. Type. China, Guizhou Province, Jiangkou Xian, vicinity of Jinding along the crest of the Fanjing Shan mountain range, elev. 2000-2300 m, 28-29 August 1986, Sino-Amer. Guizhou Bot. Exped. 634 (holotype, PE; isotypes, A, CAS, HGAS, TI). Rubus fanjingshanensis is apparently related to R. treutleri, but the latter differs markedly in having leaves 3- to S-lobed, with the terminal lobe acute at the apex, the lateral ones obtuse, and the margins irregularly and minutely serrate; narrower, fimbriate stipules 10-15 mm long, with linear to linear- lanceolate lobes; and sepals ovate to ovate-lanceolate, the outer ones foliose and dissected with lanceolate lobes. 124 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 71 SA ‘ Y pe Le <4 6 SN ~ y SS » SN AR a Z * Wuds i * Ay ay Lp SRO Z 4, See ann ah i HAA ack AVS yy Se PIR OS JH EHS RRS mae SSHN ai Mere SS ae mate 4c SAIS AY ae ee maa. rae < i C SAY NDR NAL) yee = GEN th BEANS NK \ ‘ * FIGURE 3. Rubus fanjingshanensis: a, upper portion of branch; b, lower leaf surface: c, fruit; d, seed. 1990] BOUFFORD ET AL., FLORA OF CHINA 125 Ficure 4. Deinostema adenocaulon: a, habit; b, portion of stem showing glandular hairs; c, immature capsule; d, flower opened to show I tofst d callosities on inner surface of corolla tube. 126 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Sabia swinhoei sagan ex a & Hemsley var. parvifolia Y. H. Xiang & Q. H. Chen, var. A typo pedunculis pedicellisque brevioribus, 2-3 mm longis, foliis minori- bus, 3-5 cm longis, 1-1.8 cm latis differt. Leaves ovate or elliptic to sublanceolate, 3-5 cm long, 1-1.8 cm wide, base acute to subrounded, apex acute to acuminate, upper surface glabrous except for pubescent midrib, lower surface + lax-pubescent, lateral nerves 5 to 7 pairs; petiole 2-3 mm long, densely pubescent. Cymes solitary, axillary; peduncles 3 mm long; pedicels 2 mm long. Drupelets obovoid-compressed, turning red at maturity. Type. China, Guizhou Province, Jiangkou Xian, Daiyenpeng along the Kaitu River on SW side of Fanjing Shan mountain range, elevation ca. 800 m, Sino- Amer. Guizhou Bot. Exped. 1082 (holotype, HGAS; isotypes, A, CAS, PE). The variety differs from the typical one in having much shorter cymes and smaller leaves. Deinostema adenocaulon (Maxim.) Yamaz. FIGURE 4. Deinostema Yamaz., separated from Gratiola L., in which it was once placed, by its ebracteolate calyx, valvate (vs. imbricate) calyx lobes in bud, and pilose (vs. glabrous) anthers (Ohwi, 1965), and from Dopatrium Hamilton ex Bentham by its bilocular (vs. unilocular) ovary (Yamazaki, 1953), is a genus of two species restricted to wet places in eastern Asia. The most widespread species, Deinostema violacea (Maxim.) Yamaz., ranges from Honshu, Shikoku, Kyushu, and the Ryukyu Islands in Japan, as well as Korea (Ohwi, 1965) and the northeastern provinces of China, to Jiangsu on the Asian mainland (Hong, 1979). Deinostema adenocaulon was previously known from only Honshu, Shikoku, and Kyushu in Japan and from Cheju-do at the southern tip of the Korean peninsula (Ohwi, 1965; Park, 1974). The specimens from northeastern Guizhou, Jiangkou Xian, SE side of the Fanjing Shan mountain range along the Heiwan River in the vicinity of the Ecology Station of the Guizhou Academy of Sciences, 30 August 1986, Simo-Amer. Guizhou Bot. Exped. 598 (A, CAS, GIB, PE, TI), were collected approximately 1800 km from the nearest previously known populations in Korea. In Guizhou the plants were abundant in a fallow paddy field with Microcarpaea minima (Konig) Merr., Eleocharis congesta D. Don, Eriocaulon sieboldianum Sieb. & Zucc., and Rotala mexicana Champ. & Schldl. ACKNOWLEDGMENTS We wish to thank Bruce Bartholomew and Stephen A. Spongberg, of the United States, and Qianhai Chen, Sizhao Fang, Jin-gen Qi, Yuling Tu, Yinghai Xiang, and Tsun-shen Ying, of China, all of whom participated in the fieldwork in northeastern Guizhou Province, China, in 1986, and Gustavo Romero and Rolla Tryon for reviewing the manuscript. We are particularly grateful to the 1990] BOUFFORD ET AL., FLORA OF CHINA 127 National Geographic Society, the Chinese Academy of Sciences, the Institute of Botany, Beijing, the Guizhou Academy of Sciences, and the Institute of Biology of Guizhou for their financial support of the fieldwork. LITERATURE CITED CHEN, S. C. 1988. Status of the conservation of rare and endangered plants in China. (Lecture presented at the International Symposium on Botanical Gardens, Nanjing, 25-28 September 1988.) Hone, D. Y. 1979. Deinostemma re Fl. Reipubl. Pop. Sin. 67(2): 9 Huanea, W. L., Y. L. Tu, L. YAna, S. Z. FANG, & J. L. Li. 1982. ae in the Fanjingshan Nature Preserve. (Group for the Scientific Survey of the Fanjingshan Mountain Preserve, eds.) Pp. 93-130 in Scientific survey of the Fanjingshan Moun- tain Preserve. (In Chinese, English abstract.) Foreign Language Press, Beijing. Oxnwt, J. 1965. Flora of Japan. F. G. Meyer & E. H. WaALKer, eds.) Smithsonian Institution, Washington, D. C. Park, M.K. 1974. Keys to the herbaceous plants in Korea (Dicotyledonae). (In Korean.) Chung Eum Sa, Seoul. YAMAZAKI, T. 1953. On the floral structure, seed development, and affinities of Dei- nostema, a new genus of Scrophulariaceae. J. Jap. Bot. 28: 129-133. JUDD, LYONIA 129 A NEW VARIETY OF LYONIA (ERICACEAE) FROM PUERTO RICO WALTER S. JUDD! Lyonia truncata var. proctorii is a member of Lyonia sect. Lyonia and as uch is characterized by an indumentum of unicellular hairs and multicellular, ferrugineous, biseriate- stalked, peltate scales. It is described and illustrated mera evolutionary line. It is considered conspecific with L. truncata, a species previously considered endemic to Hispaniola. Thus, two species of Lyonia now known from Puerto Rico: L. rubiginosa var. stahlii and L. truncata var. proctoril. Both taxa have their ane relatives in the mountains of central and southern Hispaniola. The genus Lyonia Nutt. (Ericaceae subfam. Vaccinioideae tribe Androme- deae; see Judd, 1979, 1981) comprises 35 species (51 taxa”), of which 27 belong to sect. Lyonia (Judd, 1981). This section is characterized by its indumentum of multicellular, ferrugineous, biseriate-stalked, peltate scales. The group is most diverse in the Greater Antilles, where 24 species occur: 12 species (16 taxa) in Cuba; ten species (11) in Hispaniola; two in Jamaica; two in Puerto Rico (including the one described herein); and one in St. Thomas. Most species occur in montane communities (usually cloud forests or pinelands) and are restricted to a single island or mountain range. Recent fieldwork in the Sierra Bermeja of Puerto Rico has resulted in the discovery of an undescribed variety of Lyonia truncata. This variety, described and illustrated here, brings the number of species and varieties recognized in the genus to 52. Terminology used in the description follows Judd (1981). Lyonia truncata Urban var. proctorii Judd, sp. nov. FIGURE. Varietas haec a Lyonia truncata Urban varictatibus truncata et montecristina (Urban et E. Ekman) Judd differt in pedicellis 2-5 mm longis (vs. 2.5-10(-13) mm longis), corollis urceolatis/cylindraceis, 3-4.5(-5) mm longis (vs. + cylin- draceis, 4-7 mm longis), antheris 0.7—1 mm longis (vs. 1-1.5 mm longis), et capsulis 3-4.5 mm longis (vs. (3—)4-6.5 mm longis). een of sonnet 220 Bartram Hall, University of Florida, Gainesville, Florida 32611. vonia lippoldii Berazain & Bisse (Berazain, 1987) is considered conspecific oa L. os (Judd, in ee addition, recent field studies have indicated that Lyonia obtusa Griseb. var. longipes (Urban) hudd eo Judd, 1981) is best considered to be specifically distinct from L. obtusa (Borhidi, 1983; Berazain, 1985) © President and Fellows of Harvard College, 1990. Journal of the Arnold Arboretum 71: 129-133. January, 1990. 130 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Et PETIT RG OLCLAAUIO RH WN CUNY b ie truncata var. proctoril (from Proctor & McKenzie 44010): a, habit, x 0.5; b- d, leaves, x 0.5; e, cross section of leaf blade, x 50; f, flower, x 6; g, stamen, x 12.5; h, Reine x 2551, ‘capsule, x 6. Evergreen erect shrub to ca. 1.5 m tall. Leaves with petiole 2-6.5 mm long; blade elliptic to ovate, 0.9-4.5 by 0.4-2.3 cm, + flat, coriaceous, 0.25—0.3 mm thick, the margin plane to slightly revolute toward petiole, distal portion clearly to obscurely and irregularly toothed to entire, proximal portion entire to + obscurely toothed, the abaxial surface sparsely to moderately lepidote, mod- erately to densely unicellular-pubescent. Inflorescences fasciculate, ca. 2- to 15- flowered; pedicels 2-5 mm long, sparsely pubescent; bracteoles opposite or subopposite, nearly basal to positioned on proximal 5 of pedicel, narrowly triangular, 0.9—2 mm long; bracts to ca. 2 mm long. Flowers 5- (or 6-)merous; calyx lobes triangular, 0.7-1.6 by 0.4-0.9 mm, the apex acuminate to acute, the abaxial surface lepidote, glabrous to densely pubescent (near base); corolla urceolate-cylindrical, 3—4.5(—5) by 2-3(-3.5) mm, white or sometimes slightly pink-tinged near mouth; filaments roughened, 2.5-3.5 mm long, unappend- aged, anthers 0.7-1 mm long. Capsules ovoid, 3—4.5 by 2.5-4 mm, very sparsely pubescent, especially near base: seeds ca. 2.5 mm long. 1990] JUDD, LYONIA 131 Type. Puerto Rico, municipio de Cabo Rojo, Sierra Bermeja, Barrio Llanos Costa, upper slopes and summit of Cerro Mariquita, elev. 250-301 m, 11 October 1987, G. R. Proctor & P. McKenzie 44010 (holotype, FLAS; isotypes, Herbarium, Dept. of Natural Resources, Puerta de Tierra, Puerto Rico (2 sheets)). EtyMo.ocy. The varietal epithet honors George R. Proctor (b. 1920), of the Department of Natural Resources, Puerta de Tierra, Puerto Rico, who collected the herbarium material and sent it to me for identification as possibly repre- senting an undescribed taxon. LEAF ANATOMY.’ Adaxial epidermis with inner periclinal walls lignified but not strongly thickened, height/breadth quotient of cells 0.5—1; adaxial hypodermis unilayered; lignified cells surrounding major veins transcurrent; mesophyll cells not lignified; stalk of peltate scales sunken into abaxial epidermis; petiole bundle bifacial, usually forming a cylinder. DISTRIBUTION AND ECOLOGY. Lyonia truncata var. proctoril has been collected only on the upper slopes and summit of Cerro Mariquita of the Sierra Bermeja in southwestern Puerto Rico (municipio de Cabo Rojo), where it is locally common on exposed rocky slopes. The region is occupied by dry rocky scrub to low forest on Upper Jurassic to Lower Cretaceous chert. According to Proctor (pers. comm.), immediately associated species include Acacia farnesiana (L.) Willd., Argythamnia candicans Sw., Aristida portoricensis Pilger, Bumelia obo- vata (Lam.) A. DC., Calliandra portoricensis (Jacq.) Bentham, Chamaesyce articulata (Aublet) Britton, Clusia rosea Jacq., Coccoloba microstachya Willd., Crossopetalum rhacoma Crantz, Digitaria eggersii (Hackel) Henrard (new to Puerto Rico), Eugenia sessiliflora Vahl, E. woodburyana Alain, Guettarda sca- bra (L.) Vent., Heteropteris purpurea (L.) Kunth, Lippia micromera var. helleri (Britton) Moldenke, Machaonia portoricensis Baillon, Melocactus intortus (Mil- ler) Urban, Opuntia repens Bello, Ouratea littoralis Urban, Plumeria alba L., Polvegala penaea L., Rheedia mariquitensis Proctor, ined., Sida jamaicensis L. new to Puerto Rico), Stvlosanthes viscosa Sw. (new to Puerto Rico), Tetramicra canaliculata (Aublet) Urban, Thrinax morrisii H. A. Wendl., Vernonia proctorii Urbatsch, ined., and Xylosma buxifolia A. Gray. Flowering specimens of var. proctorii have beer collected in October The Sierra Bermeja plants possess the synapomorphies (stamens lacking or with only minute spurs, and leaves at least sometimes strongly and irregularly toothed) delimiting a clade containing all previously recognized Hispaniolan and Puerto Rican species of Lyonia (Judd, 1981), as well as those (leaves densely unicellular-pubescent abaxially, with nonacuminate apices) characterizing a Hispaniolan clade called the ““Lyonia microcarpa—L. truncata—L. heptamera evolutionary line” (Judd, 1981). However, they lack the apomorphies of the more derived members of this “evolutionary line’ (see Judd, 1981, for char- acters involved). 3Techniques as in Judd, 1981. 132 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 The Sierra Bermeja plants are phenetically very similar to Lyonia truncata, a phenetically primitive, cladistically basal, metaphyletic species of the His- paniolan clade, as can be seen by comparing the description given above with that of L. truncata (Judd, 1981). The leaves of both are characterized by moderate to dense unicellular indumentum abaxially, secondary veins not much more prominent than the tertiary, higher-order veins not forming a raised-reticulate network on the abaxial surface, inner periclinal walls of the adaxial epidermis lignified but not thicker than other epidermal walls, and peltate scales with their stalks sunken into the abaxial leaf epidermis. The Sierra Bermeja plants differ from L. fruncata only in their shorter pedicels and their smaller flowers and capsules. The two are similar in habitat preference and in the low elevations they inhabit; they are nearly identical anatomically. Although individual features show overlap, plants of the two islands can be separated consistently by the characters in the diagnosis and the key. The Sierra Bermeja plants are believed distinctive enough to warrant taxonomic recog- nition and are considered to be a geographically isolated variety, L. truncata var. proctorii. Varietal rank has been preferred over specific recognition for this entity due to its strong phenetic similarity to L. truncata in both vegetative and reproductive features. Lyonia truncata occurs at lower elevations (100- 1600 m) than any other Hispaniolan species of Lyonia, and it is perhaps not surprising that this recently discovered low-elevation population of Lyonia in Puerto Rico is best placed in L. truncata. In 1981 I considered Lyonia truncata to be composed of two geographic varieties (equivalent to subspecies of many authors; see Borhidi, 1983) that differ in leaf size, orientation, and curvature (see Judd, 1981). Lyonia truncata, as now circumscribed, comprises three varieties, which can be distinguished as follows: _— . Corollas 3-4.5(—5) mm long, urceolate-cylindrical; anthers 0.7-1 mm long; pedicels 2-5 mm long; capsules 3-4.5 mm long; [Sierra Bermeja, Puerto Ole. exteare dence (ite Bg are eects he Eras a Pe tee eh anise seen Sees hae eee r. proctorii Judd. . Corollas 4-7 mm long, + cylindrical; anthers 1-1.5 mm long; eal 2.5-10(-13) mm long; capsules (3—)4-6.5 mm long; [Hispaniola]. 2. Leaves 0.8-3.5(-4) cm long, slightly concave to + flat, sometimes slightly recurved; shrub to 2.5(-3) m tall, often densely branched; [Sierra de Baoruco and Massif de la Selle) exe Bead ceed be panne sag akon 4 var. truncata. Leaves (2.5—)3—5(—6.8) cm long, + flat to recurved; shrub to small tree to 3-7 m tall, with branches more open; [Cordillera Central and Massif du Nord]. ...... bid feed slate ch ls Atego 2 a ene aie var. montecristina (Urban & E. Ekman) Judd. a" N A second species, Lyonia rubiginosa (Pers.) G. Don var. stahlii (Urban) Judd, occurs on Puerto Rico in cloud forests at high elevations in the Cordillera Central and a few outlying localities (Judd, 1981). This species is easily dis- tinguished from L. truncata var. proctoril in having leaves that completely lack unicellular hairs on the abaxial surface (Judd, 1981). It is noteworthy that L. rubiginosa and L. truncata, occurring on Hispaniola and Puerto Rico, are the only two members of sect. Lyonia that occur on more than one Antillean island 1990] JUDD, LYONIA je ACKNOWLEDGMENTS I thank George R. Proctor, who has collected this interesting Lyonia and has increased our knowledge of several other species of the genus through his numerous Antillean collections. I am also grateful to George Diggs and Peter Stevens for their helpful comments on a manuscript version of this paper. LITERATURE CITED BERAZAIN ITURRALDE, R. 1985. Revision critica del género Lyonia Nutt. (Ericaceae) en ee Feddes Repert. 96: 631-649. . Una nueva especie del género Lyonia (Ericaceae): L. lippoldii Berazain Bisse. Revista Jard. Bot. Nac. 8: 3-7. Boruip1, A. 1983. New names and new species in the flora of Cuba and Antilles, III. Acta Bot. Acad. Sci. Hung. 29: 181-215. Jupp, W. S. 1979. Generic relationships in the Andromedeae (Ericaceae). J. Arnold Arbor. 60: 477-503. 1981. A monograph of Lyonia. Ibid. 62: 63-209, 315-436. AL-SHEHBAZ & MARTICORENA, MENONVILLEA 135 MENONVILLEA ROLLINSI (BRASSICACEAE), A NEW SHRUBBY SPECIES FROM CHILE IHSAN A. AL-SHEHBAZ! AND CLODOMIRO MARTICORENA2 Menonvillea rollinsii, a new species described from northern Chile, is unique in the genus because of its shrubby habit and the development of cork in older stems. Menonvillea DC. is a genus of thirty species distributed primarily in the drier portions of northern and central Chile and the adjacent provinces of Argentina. The range of the genus extends southward into Patagonia, and of the eight species that grow there, only M. nordenskjoeldii (Dusén) Rollins reaches as far south as Prov. Santa Cruz, Argentina, and Region Magallanes y Antartica Chilena (Boelcke & Romanczuk, 1984). Seventeen species are restricted to Chile and eight to Argentina, while five grow in both countries (Rollins, 1955). Most species of Menonvillea are perennials with either an unbranched caudex or a deep, fleshy or woody, much-branched, rhizomelike caudex. About eight species are annuals, and two others, M. linearis DC. and M. pinnatifida Bar- néoud, are annuals that sometimes perennate under favorable conditions. From the underground caudex, all of the perennial species annually produce new herbaceous stems that bear flowers and fruits and then die at the end of the growing season. None of the previously recognized species has woody aerial stems, but M. rollinsii, described below, is unusual in the genus in being a shrublet with woody stems and well-developed cork. It gives us great pleasure to name the new species in honor of Professor Reed C. Rollins, former director of the Gray Herbarium of Harvard University, in recognition of his out g revision of Menonvillea (Rollins, 1955), in which he described nine new species and proposed new combinations for 12 others. Menonvillea rollinsii Al-Shehbaz & Marticorena, sp. nov. FIGURE 1. Fruticulus 15-20 cm altus, ad basin 4-6 mm in diametro; folia succulenta, lineari-subulata, scabra, 4-10 mm longa, 0.8-1.2 mm lata; sepala oblonga, scabra, 2—2.3 mm longa; petala oblonga, alba, 2-2.5 mm longa, marginibus crispis; pedicelli fructiferi recti, divaricati, 4-5 mm longi; fructus glaber, val- ‘Arnold Arboretum, Harvard University, 22 Divinity Avenue, Cambridge, Massachusetts 02138, U.S.A 2Depar nto de Botanica, Universidad de Concepcion, Casilla 2407, Concepcion, Chile. © President and Fellows of Harvard College, Journal of the Arnold Arboretum 71: 135-138. oe 1990. 136 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Thy Menonyillea FiGu : i! (holotype): a, plant; b, fruit, front view; c, fruit, lateral view. Scale bars = 1 cm (a), 1 mm (b, c). vibus late ovatibus et anguste alatibus, 2-3 mm longus; stylus 1—1.5 mm longus; semina ignota. Shrublet 15-20 cm high; sparsely and minutely scabrous almost throughout with trichomes 0.05-0.1 mm long. Old stems straw colored, 4-6 mm in di- ameter at base. Leaves sessile, fleshy, subulate-linear, scabrous, 4-10 mm long, 0.8-1.2 mm wide, entire, straight or incurved. Inflorescences terminal, few- flowered, ebracteate racemes. Sepals caducous, oblong, 2—2.3 mm long, ca. | mm wide, scabrous, scarious at margin. Petals white, oblong, 2—2.5 mm long, 1990] AL-SHEHBAZ & MARTICORENA, MENONVILLEA 137 E2. Scanning electron micrographs of Menonvillea rollinsii: a, flower bud; b, replum; c, leafy branch; d, portion of old stem; e, closeup of crack on old stem, showing trichomes, epidermis, and cork; f, cork. Scale Se = 1mm ear 0.1 mm (e ca. 0.8 mm wide, not clawed, uniform in width throughout, crisped at margin. Stamens 6, slightly tetradynamous; filaments erect, ca. 1.8 mm long; anthers oblong, 1-1.1 mm long. Median nectar glands solitary, toothlike; lateral glands ringlike. Fruiting pedicels straight, cnet 4-5 mm long. Fruits glabrous, narrowly 4-winged; valves broadly ovate, 2-3 mm long, 1.2-2.2 mm wide, inconspicuously veined on the back, pent in cross section, aa with narrow marginal wing 0.15-0.3 mm wide; style thick, 1-1.5 mm long; stigma capitate, 2-lobed. Seeds not seen. Type. Chile, Region IJ [Antofagasta], camino de Chuquicamata a Conchi, 2700 m alt., 5 April 1961, Ricardi, Marticorena, & Matthei 445 (holotype, CONC; fragment, GH). 138 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 Menonvillea rollinsii is unique in the genus because of its shrubby habit and its older stems with well-developed cork (see FiGurE 2e, f). It is also easily distinguished from the other species of Menonvillea in having short (4-10 mm), sessile, subulate-linear leaves (See FiGuRE 2c) and in being scabrous almost throughout (the petals, stamens, and fruits are glabrous) with minute, persistent trichomes only 0.05 to 0.1 mm long (see FiGuRrE 2e). The species is isolated in the genus and has no known close relatives. Menonvillea rollinsii grows in open, very dry habitats that have an average annual rainfall of only ca. 9 mm. It is associated with plants such as Adesmia atacamensis Philippi, Argvlia tomentosa Philippi, Cristaria Cav. (two species), Cryptantha linearis (Colla) Greene, C. parviflora (Philippi) Reiche, Philippiam- ra pachyphylla Philippi, P. fastigiata Philippi, Solanum sitiens 1. M. Johnston, Tetragonia trigona Philippi, and Trichocline caulescens Philippi. The shrubby habit in the Brassicaceae has evolved independently many times, as 1s evidenced by its occurrence in unrelated genera of various tribes (Al-Shehbaz, 1984). Apparently, woodiness also evolved independently in Menonvillea and in its nearest relative, Cremolobus DC., an Andean genus of seven species, of which four are woody and three are herbaceous annuals (Khanna & Rollins, 1965). ACKNOWLEDGMENTS We are most grateful to Donald H. Pfister for obtaining funds from the Harvard University Herbaria that supported the SEM portion of this research. We are thankful to Elizabeth A. Shaw for checking the Latin diagnosis, to Neil A. Harriman for his critical review of the manuscript, to Elizabeth B. Schmidt and Stephen A. Spongberg for their editorial advice, to Trisha Rice for the SEM work, to Barbara Nimblett for typing the manuscript, and to N. Moya B. for the illustration. LITERATURE CITED AL-SHEHBAZ, I. A. 1984. The tribes of Cruciferae (Brassicaceae) in the southeastern United States. J. Arnold Arbor. 65: 343-373. BoeLckE, O., & M. C. RoMANczuUK. 1984. Cruciferae. Fl. Patagonica 4A: 373-5 KHANNA, K. R., & R. C. ROLLINS. 1965. - taxonomic revision of Cremolobus Ge ciferae). Contr. Gray Herb. 195: 135-15 Rouuns, R.C. 1955. A revisionary study aie genus Menonvillea (Cruciferae). Contr. Gray Herb. 177: 3-57 1990] BOOK REVIEW 139 BOOK REVIEW Histoire du Concept d’Espéce dans les Sciences de la Vie, by Scott Atran ef al. Fondation Singer-Polignac, Paris, 1987. xi1 + 324 pp. ISBN 2-900927- 19-6. 150 francs softcover. This valuable book contains 15 chapters based on papers presented at a meeting in May, 1985, held in Paris under the auspices of the Fondation Singer- Polignac. This foundation is named after the Princesse Edmond de Polignac, née Winaretta-Eugénie Singer, who was born in New York in 1865. It is thus appropriate that some papers are in English and others are in French. The importance of this set of papers is in part, as M. J. S. Hodge notes in his article (““Darwin, Species and the Theory of Natural Selection’’), because the contributors have collectively engaged in a study of the /ongue durée— applied to the history of science, the long-term intellectual and institutional life of science. In the past we have tended to consider species concepts in the context of taxonomic species. In this volume many contributors take a refresh- ingly broad approach to their subject, stepping back from the issues of taxo- nomic species and species concepts and illuminating issues that are very much current; discussion among biologists is also becoming less introspective. Hodge himself emphasizes ambiguities between conceptualizations of species presup- posed by an explanatory program and the definition of species as a classificatory category and shows how Darwin’s species concept ultimately depends on the former. From late 1837 to late 1838, Darwin emphasized that cessation of interbreeding was necessary for the adaptation and formation of “‘good species.” However, as the analogy between artificial and natural selection assumed more prominence in his mind, he dropped this requirement, and with it the idea that species occupied a particular rank in the hierarchy of nature. Species as taxa existed, and Darwin recommended following the species limits of the masters of the art, but there were no criteria other than custom for ranking species Phillip Sloan (‘From Logical Universals to Historical Individuals: Buffon’s Idea of Biological Species’) takes a similar approach. He shows how Buffon integrated contemporary discussion on issues such as calculus and probability with William Harvey’s separation of two senses in which species were used in the work of Aristotle. Buffon distinguished between mathematical truth, which was arbitrary and abstract, and physical truth, which was grounded in fact and had a probability approaching certainty. Hence Buffon rejected species as uni- versals—they were abstract—and accepted species as wholes and individuals, rather than as collections of things. To Buffon, it was not the individual in the strict sense, but the chain of successive individuals that constituted the species, that was the great marvel. Genealogy, but not only of species, was all; classi- © President and Fellows of Harvard College, Journal of the Arnold Arboretum 71: 139-142. ae 1990. 140 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 fication was to be true ““Naturgeschichte,” not just “Naturbeschreibung.’’ How- ever, as Sloan has discussed elsewhere, the latter predominated. The reasons for this need detailed study. A contributing factor may be that the only tool by which genealogy could be directly established remained for almost two hundred years the ability to interbreed. This surely meant that Buffon’s ideas at the species level could be difficult to distinguish from more conventional species concepts. Yet in many of these, even the most “‘fixist,” genealogy and individuality of a sort were very important; it is only now that details of the relationship between morphology and genealogy are being teased apart. Richard Burkhardt, Jr. (“Lamarck and Species”) suggests that Lamarck did not alter the way in which he described taxonomic species, despite his changing ideas about what species represented in the natural world. Taxonomy even then had its own requirements (which it has been inclined to defend against the natural world ever since!). Toward the end of Lamarck’s life, it seems that his concept of species reverted from nominalist back to the position that he held when writing the Flore Francoise—that they were real. Unfortunately, Burkhardt’s discussion does not allow us to establish clearly enough the context of the change. Intergradation, for Lamarck, was a geographic as well as a temporal phenomenon, but at any one time and place species would be discrete. The issue of what such locally discrete species represent figures largely in Scott Atran’s paper (“The Early History of the Species Concept: an Anthro- pological Reading”). He observes that taxa in the fundamental rank of folk classification, the generic-specieme (in the strict sense, the representative of a genus growing in a local area), were recognized using morpho-ecological criteria; unlike some authors, Atran dismisses more culturally bound concerns, such as utility, as having little effect. Breeding criteria were introduced into species concepts by Andrea Cesalpino and John Ray as knowledge developed a broader geographic base: organisms that bred true showed only accidental variation and formed the abstract types that could be placed in a universal taxonomy. There are still major tensions between the folk species concept (or, more properly, the generic-specieme concept), the nondimensional species that Ernst Mayr (“The Species as Category, Taxon and Population”) defends vigorously, and the taxonomic species. Mayr emphasizes the interest of biologists in sym- patric situations and local populations, the potential status of an allopatric population being usually biologically rather uninteresting for Mayr. Some of these tensions have obvious causes, as in Mayr’s article: for instance, the term “population” is not clearly defined and the taxonomic-morphological species is equated with the typological (Donn Rosen’s species are branded as “‘typo- logically defined morphospecies’’). For Bernardino Fantini (“L’Entrée de la Biologie Moléculaire dans la Défini- tion de lEspéce’’) the definition of species as a taxonomic unity remained untouched by molecular biology, which has focused on mechanisms of spe- ciation. Untouched or passed by? The biological species concept was cham- pioned by Mayr in 1942 so that the origin of species could be discussed. But for this, species either have to be the units that evolved, or to stand in some static relationship to those that did (Mayr discusses stasis briefly). Birds (per- haps) aside, are taxonomic species in general units of this sort? Can they be, 1990] BOOK REVIEW 141 should they be, and what are the consequences for both taxonomy and biology of the various ways in which these questions are answered? Anne Diara introduces such questions in her paper, appropriately titled ““Les Espéces Sont-Elles Filles de la Nature ou du Naturaliste?”’ She discusses the apparent conflict between the species concepts of Alexis Jordan, a misanthropic royalist and fixist who circumscribed species very closely, and Charles Naudin, a much more central figure on the French biological scene who came to believe in evolution by 1852 (but never seems to have accepted natural selection). Diara suggests rightly that the argument was less over “facts” than how those “facts” should be interpreted (see also Jean-Louis Fischer, ““Espéce et Hybrides: a Propos des Léporides”’). As she notes, the taxonomic philosophy of Jordan (of whom Naudin contemptuously remarked that, with all his new species, he had not introduced a single new fact into science) was agreeable to Hugo de Vries. De Vries’s work, reinterpreted although it soon was, nevertheless focused attention on the nature of the Linnaean species. Diara briefly discusses a series of papers on species concepts in the American Naturalist of 1908, and she sees Linnaean classification as being maintained primarily for logical reasons. Fur- ther development of this point would have brought the arguments closer to home (and would probably have made the article far too long). Linnaean species, at least in the context of the polemics of that time, were broadly circumscribed, yet despite all the lip service paid to them, few of the botanists who extolled their virtues over the next fifty years—a list would be surprising in both length and content—were able to find biological reasons for maintaining them. One is left feeling that what one thinks might matter when it comes to the delimitation of species, does not, and what should not, does. The requirement that taxonomic species be readily recognizable is clearly responsible for much of this confusion. Mary P. Winsor (Louis Agassiz and the Species Question, Studies Hist. Biol. 3: 89-117. 1979) shows how hard it is to provide a reasonable evaluation of the link between species concepts and species taxa; overall, it was not Agassiz’s ideas that were unscientific but his dogmatism. It was not simply because Jordan did not believe in evolution that he described so many new species from France; splitters are not necessarily “typologists.” His fixist colleagues did not rise to his defense, while others found it hard to dismiss all his species as being absolutely worthless due to the very fact of evolution. Evolution has only gradually shaped species concepts and species taxa. In ornithology this change was greatly helped by the new “nomenclatural” code proposed by the American Ornithologists’ Union in 1885. Antonello La Vergata (“Au Nom de l’Espéce. Classification et Nomenclature au XIX* Siécle’’) shows that apparently arid nomenclatural debates over the name of a species may be, as he puts it, carried out in the name of the species, that is, a particular conception of nature. How authorities were cited when a species was transferred from one genus to another depended quite largely on whether species were conceived of as logically (and biologically) part of the genus in which they were placed (the Linnaean position), or the species (both as rank and taxon) was seen as more real than the genus. Priority does not necessarily have to do with the rights of the dead, but rather with the concepts of the living. Another 142 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 nomenclatural dispute that involves similar issues and that would repay study concerns species groups. These were ultimately accepted in zoology, in part because they were seen to be compatible with evolution, and rejected in botany. Returning to a theme of the book, discussions about species concepts must often range beyond taxonomy if they are to be properly understood: arguments about nomenclature and classification may be so bitter and protracted precisely because there are larger issues at stake. Furthermore, there are only partly overlapping groups of people using terms such as “‘species.”’ Pietro Corsi (““Ju- lien Joseph Virey, le Premier Critique de Lamarck”’) shows how Virey’s writ- ings, with their implicit and explicit species concepts, reached a wide segment of the francophone world (as is also true of Buffon’s work). What does today’s public think is being saved when a species is conserved, and where did they get their ideas? There are other useful essays: here I have focused on those most of interest to botanists. Although it may be a little hard to get hold of a copy of this book, it is most highly recommended for all biological libraries, and it is even not out of reach of less-than-well-paid taxonomists. The issues raised in it demand serious thought. The editing is on the whole good, although I think that three spellings for “‘armadillo” in eight lines must be something of a record; Qaddafi will have to look to his laurels in this department.—P. F. Stevens, Harvard University Herbaria, 22 Divinity Avenue, Cambridge, Massachusetts 02138. 1990] ANNOUNCEMENT 143 XV International Botanical Congress, Tokyo The Organizing Committee of the X V International Botanical Congress wish- es to announce that the XV IBC will be held in the Tokyo area during August and September 1993: the nomenclature session 23-27 August, and the general session 28 August—3 September. The first circular for the XV IBC will be prepared in 1990 and will be distributed to those who are interested in the Congress. Requests for information and other questions and comments may be sent to: The Secretariat XV Botanical Congress, Tokyo Department of Botany, Faculty of Science The University of Tokyo 7-3-1 Hongo Bunkyo-ku, Tokyo 113 Japan Postal Service STATEMENT OF OWNERSHIP, MANAGEMENT AND CIRCULATION Required by 39 U.S.C. 3685) 1A. Title of Publication ] 1B. PUBLICATION NO. | 2. Date of Filing JOURNAL OF ARNOLD ARBORETUM jo [0 [o 4 2{6|2|5| Sept 15, 388 | 3. Frequency of Issue 7 3A. No. of Issues Published 38. Annually Quarterly 4 $70.00 a City. County id ZIP +4 Code) (Not printers Hampshire St. Lawrence, Ks. 66044 Complete Mailing Address of the Hendquarters of General Business Office T as Full Names and Complete Mailing Address of Publisher, Editor Fubmhe (Name and Complete Mailing “Addrest) Arnold Arboretum of Harvard | University, 1041 New Hampshire St. Lawrence, Ks. 66044 | Editor (Name and Complete Mailing Address) S.A. Spongberg 22 Divinity Ave. Cambridge MA 02138 | Managing Editor (Name and Compleie Mailing Addrest) E.B. Schmidt 22 Divinit Ave. Cambridge MA_ 02138 7 Owner (if owned by a corporation, irs name and addres must he sated and also immediately thereunder the names and addresses of stockholders owning or holding I percent or more of total amount of stock If not owned by a Sareacdia. the names and addresses of the individual owners must be given. If owned by a partnership oF other unincorporated firm, irs namr and address, ax well ay that of each individual must he given If the publication ix published hy a nonprofit organization, its name and addres must he | 104] New Hampshire ST. Lawrence, KS 66044 | : : BR Known Bondholders, Mortaagres. Hold 1 Percent or More of Total Amount of Bonds, Mortgages or Other Securities (If there are none, co state) : Complete Malling Addr & nnd Serres - —_ = ae The purpose, function, heck yy) ~) Has Not Changed During cyte as Changed Ouring Oe sheer publisher must submit explanation of Preceding 12 Months Preceding 12 Months with this statement ) 10 Extent and Natura of Circulation lcsvaga No ] Actual No. Copies of Single Isaue (See instructions on reverse side) | Preceding 12 Months Published Nearest to Filing Date A. Total No. Copies (Net Press Run) = 750 750 ea; Paid and/or Requested souls tien 1. Sales and carriers, stree' I 2. Mail Subscription (Paid and/or requested) §27 558 c 7 (Sum or 10B! and /OR2) 527 558 _ Free Distribution by Mail Carrier or Other Means Samples, Complimentary, and Other Free Gepias 12 12 E_ Total Distribution (Sum of C and D) Vode eet 539 570 F. Copies Not Distributed 1. Office use, left over. unaccounted, spoiled after erinting 21 ] 180 2. Return from News Agents G. TOTAL (Sum of F, Fl and 2~s qual h in A) 750 750 ant Signature and Title of Editor, Publisher, Business Manager, or Owner | certit y 2) F me prae are correct and complete A : 7 PS Form 3526, Feb 1989 (See instrudtions on reverse) v O Journal of the Arnold Arboretum January, 1990 CONTENTS OF VOLUME 71, NUMBER 1 it ba ee ob a'p sc nace ane ac te een ee canoe aan teoameta il PVISGRION SIO, oc ko 8 ah ore aS ee wo eh eS ill The Genera of Betulaceae in the Southeastern United States. Jes 3. Pee 5 ve Gaeas eae 0k dees G5 dE ka 1-67 The Genera of Taxaceae in the Southeastern United States. ROSERT A. PRICE 66 4-asrecsdtascqews adh OG r bee ee ened ps 69-9 1 Generic Limits and Taxonomy of Brayopsis and Eudema (Brassi- caceae). ASAI IG PAP EMBAD 62 £5.655405 4 bobo d oe dep neha 2A Op aS ES 93-109 Multilacunar Nodal Anatomy in Myti/aria (Hamamelidaceae). ia AE OES f 0c 4-pa-5aRbaiks canecset-beveuoteaeesianeees 111-118 Additions to the Flora of China. AVID E. BOUFFORD, ZHAN-HUO TSI, AND PEISHAN WANG ..... 119-127 A New Variety of Lyonia (Ericaceae) from Puerto Rico. WEMETOR SI Yuccais ya che icnte datas eee Ter eLspaweleR ee 129-133 Menonvillea rollinsii (Brassicaceae), a New Shrubby Species from Chile IHSAN A. AL-SHEHBAZ AND CLODOMIRO MARTICORENA ........ 135-138 BOO OVO ooo we bo cba SASS AWS SE A eee es 139-142 Volume 70, Number 4, including pages 443-541, was issued 11 October 1989. JOURNAL oF tre ARNOLD ARBORETUM HARVARD UNIVERSITY VOLUME 71 NUMBER 2 ISSN 0004-2625 Journal of the Arnold Arboretum The Journal of the Arnold Arboretum (ISSN 0004-2625) is published quarterly in January, April, July, and October for $70.00 per year, plus $10.00 or $15.00 postage for addresses outside of the United States, by the Arnold Arboretum of Harvard University. It is printed and distributed by Allen Press, Inc., 1041 New Hampshire Street, Lawrence, Kansas 66044. Second-class postage paid at Lawrence, Kansas. POSTMASTER: send address changes to Journal of the Arnold Arboretum, % Allen Press, Inc., P.O. Box 368, Lawrence, Kansas 66044. Because publication of the Journal of the Arnold Arboretum is being suspended after Volume 71 (1990) has been completed, subscriptions can only be accepted for the present volume. Remittances for Volume 71 should be sent to Journal of the Arnold Arboretum, 1041 New Hampshire Street, Lawrence, Kansas 66044, U.S.A. Claims will not be accepted after six months from the date of issue. Volumes 1-51, reprinted, and some back numbers of volumes 52-56 are available from the Kraus Reprint Corporation, Route 100, Millwood, New York 10546, U.S.A. EDITORIAL COMMITTEE S. A. Spongberg, Editor E. B. Schmidt, Managing Editor P. F. Stevens, Book Review Editor P.S. Ashton K. S. Bawa C. E. Wood, Jr. Printed at Allen Press, Inc., Lawrence, Kansas COVER: The stylized design appearing on the Journal and the offprints was drawn by Karen Stoutsenberger. THIS PUBLICATION IS PRINTED ON ACID-FREE PAPER. JOURNAL OF THE ARNOLD ARBORETUM VOLUME 71 APRIL 1990 NUMBER 2 THE GENERA OF ARUNDINOIDEAE (GRAMINEAE) IN THE SOUTHEASTERN UNITED STATES!” GORDON C. TUCKER? Subfamily ARUNDINOIDEAE Tateoka, ois Jap. Bot. 32: 277. 1957, ““Arun- doidea Perennial or annual, small to very large herbaceous plants of wetlands, wood- lands, and lowland and montane grasslands [semideserts]. Rhizomes often present. Stems erect or spreading (stolons sometimes present); nodes solid, 'Prep G ic Fl f the Sout! United States, a long-term project made possible by grants ee the National Science Foundation and at this writing supported by BSR-8716834 (Norton G. eae sags investigator), under which this account d, and BSR-8717333 (Carroll E. Wood, Jr., ipal ae carte This treatment, 132nd in the s series, follows the format established in the es paper (Jour. Arnold Arb. 39: 296-346. 1958) and continued to the present. The area covered by the Generic en includes North and South Carolina, Georgia, Florida, Ten- plants of this area, with information about Sarena members of a family or genus in brackets. Those references I did not verify are marked with asteris I have continued to enjoy working with Norton Miller os Carroll Wood on the Generic Flora project and I thank them for their ee and advice. William J. Crins, Ihsan A. Al-Shehbaz, and J. Darbyshire (Agriculture Canada, Ottawa) reviewed the manuscript an numerous valuable suggestions. Elizabeth B. Schmidt and Stephen ong - improved the final manuscript with their editorial skills. Thanks are extended to the staffs of the New York State Library (especially Alta Beach, Senior Librarian) and the Botany Libraries of a University (especially Geraldine C. Kaye, Librarian) for providing many penne . Mink as curators of the following herbaria who have sent specimens or provided access spitality during my visits: A, ALU, CCNL, NN, DUKE, ECON, GH, NEBC, NYS ntribution number 625 of the a ey ae Science Servic The auacan (by Karen Stoutsenberger from dissections by ai Wood) are rearrangements of parts of Ficures 3, 5, and 7 in C. S. Campbell’s account of the family Gramineae, its subfamilies, and tribes in the southeastern United States (Jour. Arnold Arb. 66: 123-199. 1985). 3Biological Survey, New York State Museum, Albany, New York 12230 © President and Fellows of Harvard College, Journal of the Arnold Arboretum 71: 145-177. ae 1990. 146 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 glabrous, pubescent, or puberulent. Leaves several, basal and cauline, or cauline only; sheath margin open; blade linear to narrowly lanceolate, bundle sheaths double; chlorenchyma not radiate (except in Aristida and Neyraudia); silica bodies of various shapes, longitudinally oriented; stomata not dominating in- tercostal zones, subsidiary cells domed. Inflorescences terminal or terminal and axillary (axillary cleistogamous ones wholly enclosed within sheaths in Dan- thonia and Nassella). Spikelets few to very numerous, laterally flattened to subcylindrical, 1- to several-flowered; articulation above the glumes. Glumes 2, subequal to decidedly unequal, lanceolate, equaling or shorter than the lemmas; lemmas conduplicate, involute, or convolute, membranaceous (or indurate). Paleas shorter and narrower than the lemmas. Lodicules [3] 2 (or lacking), oblong or obtruncate, entire or shallowly emarginate, the margin sometimes ciliate. Stamens | or 3; filaments slender; anthers ellipsoid to linear. Ovaries ellipsoid to cylindrical. Fruit a caryopsis [achene], cylindrical or flat- tened, more or less clasped by persistent lemma and palea. Pericarp adnate to (free from] the seed. Hilum ovate-oblong to linear. Base chromosome numbers 9,10, 11, 12. (Including Aristideae C. E. Hubb., Centothecoideae Soderstrom, Danthonieae Zotov, Stipeae Dumort.) Type GENUS: Arundo L. A subfamily of about 60 genera and some 1400 species, in five tribes, dis- tributed worldwide but with the greatest diversity, both in species and genera, in the Temperate Zone of the Southern Hemisphere. The Arundinoideae are represented in the Southeast by seven genera in four tribes (Aristideae C. E. Hubb., Arundineae Dumort., Centotheceae Ridley, Stipeae Dumort.) and a total of 43 species, half of them in Aristida The circumscription of the subfamily has and is accepted here in the sense of Campbell and Davidse and fie to include the Centotheceae, and in the sense of Barkworth & Everett to include the Stipeae, which were treated as an unplaced tribe by Campbell. Narrower circumscriptions have been proposed by Clayton & Renvoize, Conert (1987), and Renvoize (1982). In an extensive cluster analysis using 71 characters, Renvoize (1982) dem- onstrated the importance of leaf-blade anatomy in defining a core group of genera. Several genera, such as Neyraudia Hooker f., combine features of the Arundinoideae with those of other traditional subfamilies. In a phylogenetic analysis of the family, Kellogg & Campbell concluded that the subfamily is polyphyletic and probably consists of several groups that are basal to other major clades (i.e., subfamilies) of the grasses. Danthonia DC. and its allies were included in the Aveneae Dumort. (Pooi- deae) by Bentham (1883) and Hitchcock (1951) because of their glumes about as long as the spikelets, their several florets per spikelet, and their twisted geniculate lemma awns. Various studies (summarized by Campbell) have point- ed out differences between the avenoids and the danthonioids in cytology, leaf anatomy, and embryology. The danthonioids were first accorded tribal status by Zotov, and their similarity to the arundinoids has since been further dem- onstrated by Hilu & Wright (1982) and Renvoize (1982). The reed grasses, Gynerium Beauv., Phragmites Adanson, and Thysanolaena Ast 1990] TUCKER, ARUNDINOIDEAE 147 Nees, have chlorenchyma cells with invaginated walls, a bambusoid feature. Neyraudia has radiate chlorenchyma (possibly C,) and is similar and transi- tional to the Chloridoideae. However, it lacks other chloridoid features such as egg-shaped microhairs, cruciform silica bodies, and triangular subsidiary cells. It has slender arundinoid microhairs, but its embryo is eragrostoid. Its resemblance to Arundo and Phragmites seems to be due to convergent evolution of habit, not taxonomic affinity (Clayton & Renvoize). The presence of kranz anatomy suggests that species of Neyraudia are C,, but biochemical and isotopic studies have not yet been conducted to confirm this. Cortaderia Stapf is a danthonioid genus of about 24 species, with greatest diversity in the Andean and Pampan regions of South America. There are also four species in New Zealand and one in New Guinea. Cortaderia Selloana (Schultes) Ascherson & Graebner, pampas grass, is cultivated as an ornamental in warm regions of the United States. Evidence that it spreads from cultivation is lacking, although plants persisting after cultivation have been reported from North Carolina (Radford et a/.) and Texas (Gould). In California, it 1s a non- spreading and popular cultivated species. Another cultivated species, C. jubata (Lem.) Stapf (‘“C. atacamensis’’), a South American native, is an invasive weed along the California coast (Cowan). In California, only carpellate plants are known and abundant seed is produced apomictically (Costas Lippmann, 1977). It is also a bad weed in New Zealand (Knowles & Ecroyd). Another danthonioid genus, Schismus Beauv., is represented in the south- western United States and adjacent Mexico by two Eurasian adventives, S. barbatus (L.) Thell. (Gould & Moran) and S. arabicus Nees. The Aristideae, the only largely C, tribe of the Arundinoideae, include just three genera: Aristida, Sartidia De Winter, and Stipagrostis Nees. It is distinc- tive both in spikelet morphology and leaf-blade anatomy (Renvoize, 1986). Anatomically, the inner layer of cells in the two-layered bundle sheaths is as large as or larger than the outer layer. The three genera can also be distinguished anatomically. In Sartidia the chlorenchyma is not radiate and the photosyn- thetic pathway is C;. Stipagrostis (sometimes included in Aristida) and Aristida are both C, genera. In Stipagrostis there are no chloroplasts in the inner layer, while in Aristida they are present. The unigeneric tribe Micraireae Pilger is endemic to northern Australia (Clay- ton & Renvoize; Lazarides). There are 13 species of Micraira F. Mueller, all of which grow in very thin soil and show an adaptation to this habitat condition in their ability to recover from extreme desiccation. The genus is unique in the family in having spiral phyllotaxy. The base chromosome number is 10. The monotypic tribe Thysanolaeneae C. E. Hubb. includes only Thysano- laena maxima (Roxb.) O. Kuntze, 2” = 24, of tropical southern and south- eastern Asia. The genus is related to Phragmites and may not warrant tribal status (Clayton & Renvoize). The Centotheceae are a small group of 11 genera and 26 species (ca. 30, fide Clayton & Renvoize; Tenorio). Except for Chasmanthium Link, which is en- demic to southeastern North America, and Gouldochloa Valdés, Morden, & Hatch, endemic to northeastern Mexico, they are pantropic. Centothecoids are similar to the herbaceous bambusoids in habit but are decidedly arundinoid 148 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 in leaf anatomy. The distinctive embryos have a mesocotyl, lack an epiblast (present in other arundinoids), and have a scutellum cleft and a rolled first leaf. Centothecoid grasses were first accorded taxonomic notice as a subtribe of the Festuceae (= Poeae) by Bentham (1883). They were subsequently treated as subfam. Centothecoideae (Clayton & Renvoize; Soderstrom; Tenorio) but are here considered to be a tribe of the Arundinoideae following Campbell and Davidse and colleagues. A strongly papillate abaxial surface (absent in other grasses) characterizes all genera (Davidse ef al.). The spikelets are strongly laterally compressed, the caryopsis is laterally flattened, and the hilum is basal and punctiform to shortly linear. Many genera have pulvinate petiolate leaves (Chasmanthium and Gouldochloa are exceptions), double bundle sheaths, and leaf tissue differentiated into palisade and spongy layers (in grasses otherwise known only in Laciasis (Griseb.) Hitche., Panicoideae A. Br.). Bicellular, often filiform microhairs are commonly present (Davidse et a/.). A base chromosome number of 12 characterizes nearly all cytologically known genera. The only large genus, Zeugites R. Br., 1s confined to the New World tropics, as are Calderonella Soderstrom & Decker and Pohlidium Davidse, Soderstrom, & Ellis. In the Old World tropics are Bromuniola Stapf & Hubb., Centotheca Desv., Chevalierella A. Camus, Lophatherum Brongn., and the most primitive genus, Megastachya Beauv. Only Orthoclada Beauv. has representatives in both the Old World and the New. Tenorio proposed (but did not validly publish) a monotypic genus for the African species of Orthoclada. The Stipeae are here included in the Arundinoideae, as has been suggested by Barkworth & Everett and Watson and colleagues. However, Kellogg & Campbell hypothesized that the Stipeae form a clade that is the sister group of the Pooideae. Their amino-acid profiles are intermediate between those of the pooids and the bambusoids. Two-celled microhairs are absent from the Stipeae, as they are from the Pooideae. However, single-celled microhairs, called “‘pylidial hairs” by Renvoize (1985), are present in some species of Stipa and Oryzopsis sensu lato. These hairs may have evolved from the two-celled, fingerlike microhairs of Bambusoideae. In the Southeast the Stipeae are rep- resented by two genera, Nassella (Trin.) Desv. and aaa J. Presl, each of which has been included in Stipa by some workers The stipoid genus Oryzopsis Michx. (including Piptatherum Beauv.) is rep- resented by ten or more species in North America, mostly in the West. Three species, O. asperifolia Michx., O. canadensis (Poiret) Torrey, and O. racemosa (J. E. Sm.) Ricker, reach their southern range limits in the mountains of West Virginia and may yet be found in the mountains of North Carolina and Ten- nessee. The western species O. hymenoides (Roemer & Schultes) Ricker, 2” = 48, comes eastward to Manitoba, Kansas, and western Texas and might be expected as an adventive or waif in Arkansas or Louisiana. Although the genus Milium L. was included in the Stipeae by Clayton & Renvoize, who made the amazing statement that it is merely an awnless version of Oryzopsis, it was excluded from the tribe by Barkworth & Everett. According to anatomical data provided by Renvoize (1985), Milium is anomalous in the Stipeae in having tapering long cells and in lacking short cells in the leaf epidermis. It 1s typically pooid in its leaf anatomy, amino-acid profile, suscep- tibility to rusts, and cytology (see Barkworth & Everett for references). 1990] TUCKER, ARUNDINOIDEAE 149 Haustorial synergids, an unusual micromorphological feature not otherwise known in the grasses, have been found in Chionochloa Zotov, Cortaderia, Danthonia (including Rytidosperma Steudel), Erythranthera Zotov, Lampro- thyrsus Pilger, Pyrrhanthera Zotov, and Sieglingia Bernh. (Philipson & Con- nor). These massive synergids extend through the micropyle to lie between the ovule and the ovary wall. Normal pyriform synergids, the typical graminoid condition, characterize Arundo and Phragmites. The ovary in A. Donax L. has an unusually long stalk (Bhanwra). The outer epidermal cells of the ovary accumulate an unidentified darkly staining material in A. Donax and P. Karka (Retz.) Trin. ex Steudel (Bhanwra). Intercalary growth of the chalaza after fertilization characterizes all arundinoid grasses investigated. Information on the reproductive biology of the arundinoid grasses 1s incom- plete. Most genera have perfect flowers. Species of Cortaderia and Lamprothyr- sus are gynodioecious. The Neotropical Gynerium sagittatum (Aublet) Beauv., the only species of the genus, is dioecious. Self-compatibility is known in some species of Aristida, Chasmanthium, Chionochloa, Cortaderia, Danthonia (in- cluding some species assigned to the segregate Rytidosperma), Nassella, and Stipa. Phragmites australis is apparently self-incompatible. Apomixis is re- ported from Cortaderia jubata (Philipson, 1978). Four genera in our area (4/isti- da, Chasmanthium, Danthonia, and Nassella) produce cleistogamous as well as chasmogamous spikelets in at least one species. In most genera (all those in the Southeast) the pericarp is fused to the seed coat (ordinary grass caryopsis). However, in several (none in the Southeast) the pericarp is free or separable and the fruit is thus an achene. The genera with achenes are Amphipogon R. Br. (Australia), Anisopogon R. Br. (Australia), Dregeochloa Conert (southern Africa), Elytrophorus Beauv. (tropical Africa, India, Australia), Hakonechloa Honda (Japan), Pentameris Beauv. (southern Africa), Pyrrhanthera (New Zea- land), and Urochlaena Nees (southern Africa). With “reluctantly free pericarp” (Clayton & Renvoize, p. 182) are Molinia Schrank (temperate Eurasia) and Tribolium Desv. (southern Africa). Except for Hakonechiloa, all the genera with a free pericarp are confined to the Southern Hemisphere (or to continents derived from Gondwanaland). The dispersal mechanisms of most arundinoid grasses depend on hairs or awns. The mature spikelets of Neyraudia and Phragmites break at the joints of the fragile rachilla. The hairs of the rachilla allow the florets to be wind dispersed. The silky, clasping lemmas of Arundo probably serve the same function. The long, sharp awns and short, stiff callus hairs of the lemmas of Aristida species cause the mature spikelets to stick in fur or skin of animals and thus to be dispersed. Peart, however, examined carcasses of some 100 marsupials and found no awned diaspores in their fur, a surprising result in Australia, where both Aristida and Danthonia are diverse. African species of Stipagrostis have plumose awns, enabling dispersal by wind. The awns of the lemmas of Danthonia species are hygroscopic, twisting and coiling and moving the floret into soil crevices, where chances for its germination are improved (Peart). (The hygroscopic awn of Piptochaetium may function in the same way.) The long hairs of the lemmas of the Australian D. tenuior (Steudel) Conert collapse on drying and rehydrate on wetting, moving the diaspores across soil and into crevices (Peart). 150 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 The dinoid grasses are little known chemically. The leaves of Phragmites australis contain flavonoid aglycones (chrysoeriol, isorhamnetin, luteolin, quer- cetin, tricin), flavonol and flavone O-glycosides (e.g., glucosyl-3 quercetin and glucosyl-7 tricin), and flavone C-glycosides (7,3'-dimethoxy-isoorientin, iso- orientin, isoscoparin, swertiajaponin, swertisin). Tricin and C- glycoflavones are known from about 90 percent of the grasses that have been studied (Jay & Viricel), The major flavonoid components of the ‘‘flower tissue” (= spikelets?) of P. australis are isoswertiajaponin and swertiajaponin (Nawwar et al.); also noted were 3’O-gentiobioside and the 3'-O-glucoside of swertiajaponin. In addition, there are two flavonol glycosides, rhamnetin 3-O-rutinoside and rhamnetin 3-O-glucoside, a class of compounds rare in grasses. Phragmites australis has a minute (0.27 percent) hydrocarbon fraction and no appreciable polyisoprenes (Buchanan et a/.). The epicuticular layer of Chionochloa species contains long-chain carbon compounds (alcohols, aldehydes, alkanes, esters, and fatty acids) up to 52 carbons in length (Savill et a/.). Numerous fungi parasitize arundinoid grasses. Balansia hypoxylon (Peck) Atk. (anamorph: Ephelis borealis Ellis & Everhart) parasitizes Danthonia com- pressa Austin, D. sericea Nutt., and D. spicata (L.) Beauv., as well as Aristida glauca (Nees) Walp. and three species of Stipa. It forms segmented sclerotia (Sprague). Uromyces Danthoniae McA\lp. occurs on six Australian species of Danthonia, however, no rusts are reported from any New World species of Danthonia. Puccinia invenusta H. Sydow & P. Sydow grows on all three species of Phragmites in the Old World but has no other hosts. Puccinia Magnusiana Korn. (worldwide) and P. Trabutii Roum. & Sacc. (northern Africa, southern Asia) both occur on Arundo and Phragmites. Puccinia Neyraudiae H. Sydow & P. Sydow is reported only from Neyraudia madagascariensis (Kunth) Hooker f. in southern India (Cummins). The earliest evidence of arundinoid grasses dates from the Miocene. In a fine study Thomasson (1984) demonstrated anatomical features in silicified remains from Nebraska. Fossilized florets of the extinct stipoid genus Berrio- chloa M. K. Elias were found in the abdominal cavity of a fossilized rhinoceros, Teleoceras major Hatcher, and provide direct evidence of grass diet in Miocene times (Voorhies & Thomasson). Quaternary remains of Phragmites australis from Egypt could be confidently identified because the internal and external structures of rhizomes and attached stem bases were preserved (El-Saadawi et Plants of the subfamily have some economic importance. None are used for grain. Stems of Arundo and Phragmites are harvested for paper pulp in Europe and Asia, and they are sometimes important for forage. Danthonia species are important for forage in Australia (and to a lesser extent in western North America), as are immature plants of Aristida in the American Southwest, but neither is important in these respects in the Southeast. Nassella leucotricha is an important cool-weather forage grass in Texas. Stipa tenacissima L. is gath- ered for paper pulp in Algeria and Spain (Hitchcock, 1951) REFERENCES: ACEVEDO DE VARGAS, R. Las especies de Gramineae del género Cortaderia en Chile. Bol. Mus. Nac. Hist. Nat. Chile 27: 205-246. 1959.* 1990] TUCKER, ARUNDINOIDEAE 151 ar ene V., & L. Watson. Comparative ultrastructure of microhairs in grasses. t. Jour. Linn, Soc. 98: 303-319. 1988. [Amphipogon strictus R. Br. has ““Ennea- os on-type” microhairs; cf. RENvoIzE (1986), who found no microhairs in A. car- cinus F. Mueller. ArBER, A. Monocotyledons. A morphological study. 258 pp. Cambridge, England. 1925. (Reprint, J. Cramer, Weinheim. 1961.) e Gramineae, a study of cereal, bamboo, and grass. 480 pp. New York. 1934. BAMey, L. H., E. Z. Bamtey, & STAFF OF THE BAILEY HortTorium. Hortus third. xiv + 1290 pp. ‘New York. 1976. Barkwortu, M. E. Embryological characters and the taxonomy of the Stipeae (Gra- mineae). Taxon 31: 233-243. 1982. —. Prtilagrostis in North America and its relationship to other Stipeae (Gramineae). Syst. Bot. 8: 395-419. 1983. [Includes numerical analysis and many informative illustrations. ] . Everett. Evolution in the Stipeae: identification and relationships of its monophyletic taxa. Pp. 251-264 in T. R. SoDERSTRoM, K. Hinu, C. S. CAMPBELL, & M. E. BARKwortTH, eds., Grass systematics and evolution. Wash neion, D.C. 1987. , J. MCNEILL, & J. Maze. A taxonomic study of Stipa Nelsonii (Poaceae) with a key Shinai it from related taxa in western North America. Canad. Jour. Bot. 57: 2539-2553. ane BARTOLOME, J. W. Sti ulchra, a survivor from the pristine prairie. Fremontia 9(1): 1-6. 1981. [Seed ecal exceeds 200 kg/ha in dense stands in California.] BENTHAM, G. Notes on Gramineae. Jour. Linn. Soc. Bot. 19: 14-134. 1882. . Gramineae. Jn: G. BENTHAM & J. D. Hooker, Gen. Pl. 3: 1074-1215. 1883. HANWRA, mbryology in relation to eeotes of Gramineae. Ann. Bot. II. 62: 215- 233. 1988. [Illustrations s.] Brown, W. V. rer nae in grass systematics. Bot. Gaz. 119: 170-178. 1958. BUCHANAN, R. A., C. L. Swanson, D. WEISLEDER, & I. M. CuLL. Gutta-producing grasses. Phytochemistry 18: 1069-1071. 1979. [Phragmites australis; contains mi- nute hydrocarbon fraction.] CAMPBELL, C. S. The subfamilies and tribes of Gramineae (Poaceae) in the southeast United States. Jour. Arnold Arb. 66: 123-199. 1985. Caro, J. A., & E. SANCHEZ. Las especies de Stipa (Gramineae) del subgénero Jarava. Kurtziana 7: 61-116. 1973. CLAYTON, W. D., & S. A. RENvoizeE. Genera graminum: grasses of the world. Kew Bull. Add. Ser. 13. [viii +] 389 pp. 1986. CLEWELL, A. F. Guide to the vascular plants of the Florida Panhandle. viii + 605 pp. Tallahassee. 1985. [Gramineae, 106-167. ConerT, H. J. Die Systematik und Anatomie der Arundineae. 208 pp. Weinheim, West oe 1961 e Chionochloa-Arten von ee und Neuguinea (Poaceae: Arundino- ras eacieabee Biol. 56: 153-164. 1975. Current concepts in systematics of nd Arundinoideae. Pp. 239-250 in T. R. SODERSTROM, K. Hu, C. S. CAMPBELL, & M. E. BARK WORTH, eds., Grass systematics and evolution. Washington, = C. 1987 & A. UrPE. Revision der Gattung ieee A aa Arundinoideae: Danthonieae). Abh. oe. Naturf. Ges. 532: 974. Connor, H. E. Reproductive pense in grasses. Pp. ven iG in T. R. SODERSTROM, K. Hin, C. S. CAMPBELL, & M. E. BARKworTH, eds., Grass systematics and evolution. Washington, D. C. 1987 . Breeding systems in New Zealand grasses. X. Species at risk for conservation. New Zealand Jour. Bot. 26: 163-167. 1988. Costas LippMANN, M. A. Ecology and reproductive biology of the genus Cortaderia in California. 365 pp. Unpubl. Ph.D. thesis, Univ. Calif., Berkeley. 1976.* [52 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 e on the weedy “‘pampas grass” in California. Fremontia 4(4): 25-27. 1977. [Good eee of Cortaderia jubata, C. Selloana.] Cowan, B. D. The menace of pampas grass. Fremontia 4(2): 14-16. 1976. Cummins, G. B. The rust fungi of cereals, grasses, and bamboos. xv + 570 pp. New York. 1971 DAHLGREN, R. M. T., H. T. CLIFFORD, & P. F. YEo. The families of monocotyledons: structure, oe and taxonomy. x11 + 520 pp. Berlin, Heidelberg, New York, and Toky >; DAVIDSE, G., T. R. SopersTROM, & R. P. Exuis. Pohlidium petiolatum (Poaceae: Cen- totheceae), a new genus and species from Panama. Syst. Bot ae 131-144. 1986. [Includes much information on circumscription of Centothece Depecca, D. M. Studies on the Californian species of Stipa once: Madrono 12: 129-160. 1954. DeWert, J. M. J. The genus Danthonia in grass phylogeny. Am. Jour. Bot. 41: 204-211. 1954 _ Leaf anatomy and phylogeny in the tribe Danthonieae. [bid. 43: 175-182. 1956. DE WINTER. B. The South African Stipeae and Aristideae. Bothalia 8: 201-404. 1962. Dore, W. G., & J. MCNEILL. Grasses of Ontario. Agr. Canada Monogr. 26. 566 pp. 1980 Exias, M. K. Tertiary prairie grasses and other herbs from the High Plains. Spec. Pap. Geol. Soc. Am. 14: 1-176. 1942. [Includes Stipeae.] Exuts, R. P. Leaf anatomy of the South African Danthonieae (Poaceae). I. The genus Dregeochloa. Bothalia 12: 209-213. 1977; XVI. The genus Urochlaena. Ibid. 18: 101-104. 1988 EL-SAADAwI, W.E., A. A. BADAWI, & A. A. EL-AWAmri. On silicified rhizome fragments of Phragmites a Trin. from the Pleistocene of El-Fayum, Egypt. Palaeon- tographica, Abt. B. 154: 172-178. pls. 1-5. 1975. Fow er, N. L. ae requirements for germination and establishment of three grass species. Am. Midl. Nat. 115: 131-145. 1986. [Aristida longiseta, Stipa leucotricha (=Nassella leucotricha) in Texas. — & D. W. Dun ap. Grassland vegetation of the eastern Edwards Plateau. Am. Midl. Nat. 115: 146-155. 1986. [Ecological information on Aristida longiseta, Stipa leucotricha (=Nassella leucotricha) in Texas.] FreitTaGc, H. The genus a anaes in Southwest Asia. Notes Roy. Bot. Gard. Edinburgh 33: 342-408. . The genus Stipa ae oet in ees and South Asia. [bid. 42: 355-489, 1985. Guosat, S., R. K. CHAuDuRI, & S. Dutta. Alkaloids of flowers of Arundo Donax. Phytochemistry 10: 2852, 2853. 1971. S. K. BHATTACHARYA. Occurrence of curarimetic indoles in flowers of Arundo donax. Pl. Med. 21: 22-28. 1972. [Nine simple indolic bases isolated. ] GouLp, F. W. The grasses of Texas. vil + 653 pp. College Station, Texas. 1975. — &R. Moran. The grasses of Baja California, Mexico. San Diego Nat. Hist. Soc. Mem. 12. 140 pp. 1981. HaAtL, D. W. The grasses of Florida. 498 pp. Unpubl. Ph.D. thesis, Univ. Florida, Gainesville. 1980. Hatcu, S. L., J. VALDEs R., & C. W. Morden. Stipa hirticulmis (Poaceae: Stipeae), a new species from Nuevo Léon, Mexico. Syst. Bot. 11: 186-188. 1986. Hivu, K., & K. WriGut. Systematics of Gramineae: a cluster analysis study. Taxon 31: 9-36. 1982. Hircucock, A.S. The North American species of Stipa. Synopsis of the South American species of Stipa. Contr. U.S. Natl. Herb. 24: 215-289. 1925. 1990] TUCKER, ARUNDINOIDEAE 153 . Manual of the grasses of the United States. ed. 2 (revised by A. CuAse). U. S. Dept. Agr. Misc. Publ. 200. 1054 pp. 1951. Hom, L., J. V. PANCHO, J. P. HERBERGER, & D. L. PLuckNetTr. A geographical atlas of world weeds. (Arabic, Chinese, English, French, German, Hindi, Indonesian, Jap- anese, Russian, and Spanish introductions.) xliv + 391 pp. New York and other cities. 1979. Husparp, C. E. Gramineae. Pp. 871-903 in J. Hutcuinson, The families of flowering plants. Oxford. 1973. JACQUES-FELIx, H. Les Graminées (Poaceae) d’Afrique tropicale, I: généralités, classi- fication, description des genres. Inst. Recherches Agron. Cult. Vivéres Bull. Sci. Paris 8. xi + 345 pp. 1962.* Jay, M., & M.-R. Viricev. Les flavonoides de feuilles du Phragmites australis: essai de definition du profil polyphénolique de lespéce. Phytochemistry 19: 2627, 2628. ene 'B. L. Cytotaxonomic studies in Oryzopsis. Bot. Gaz. 107: 1-32. 1945. Natural hybrids between Oryzopsis and Stipa. I. Oryzopsis hymenoides x Stipa speciosa. Am. Jour. Bot. 47: 736-7 QO. Kam, Y. K., & J. Studies on the relationships and evolution of supraspecific taxa using med data. IJ. Relationships and evolution of Oryzopsis hymenoides, O. virescens, O. Kingii, O. micrantha, and O. asperifolia. Bot. Gaz. 135: 227-247. 1974. KELLOGG, E. A., & C. S. CAMPBELL. Phylogenetic analyses of the Gramineae. Pp. 310- 322 in T. R. SoDERSTROM, K. Hivu, C. S. CAMPBELL, & M. E. BARKWORTH, eds., Grass systematics and evolution. Washington, D. C. 1987. — B., & C. Ecroyp. Species of Cortaderia (pampas grass and toetoe) in New Zealand. ‘Rotorua Forest Res. Inst. Bull. 105: ii + 24 pp. 1985. [Adventive C. jubata a forest weed. ] LAzARIDES, M. Micraira. Brunonia 2: 67-84. 1979. [Australian endemic genus forms tribe Micraireae.] Lona, R. W., & O. Lakes. A flora of tropical Florida. xvii + 962 pp. Coral Gables, Florida. 1971. [Gramineae, 132-202.] MATTHEI, O. R. Estudio critico de las gramineas del género Stipa en Chile. Gayana Bot. 13: 1-137. 1965. [Thirty-nine species; Nassella excluded; illustrations. ] Maze, J. ree on the awn anatomy of Stipa and Oryzopsis (Gamincce: Syesis 5: 169- 171. 1972 = .R. Boum. Comparative embryology of Stipa E/meri (Gramineae). Canad. Jour. Bot. 51: 235-247. 1973. . Lin. A study of the mature megametophyte of Stipa E/meri. Canad. Jour. Bot. 53: 2958-2977. 1975. Munz, P. A. A California flora. 1681 pp. a & Los Angeles. 1959. Nawwar, M. A. M., H. I. Ex Sissi, & H. H. BarAcaAtT. The flavonoids of Phragmites australis flowers. a egg oe 19: ee 1980. PALMER, P. G., & A. E. TUCKER. A scanning electron microscope survey of the epidermis of East African grasses, L ‘Smi thson. Contr. Bot. 49: 1-84. 1981. Ldsthenatherum glaucum, Bromuniola Gossweileri, Crinipes abyssinicus, Elytrophorus globularis, Ha- brochloa Bullockii, Megastachya mucronata, Neyraudia arundinacea, Orthoclada africana, Pentaschistis borussica, Phragmites mauritianus, Triraphis Schinzii.] Paropl, L. R. Las especies de Stipa del oo Pappostipa de la Argentina y Chile. Revista Argent. Agron. 27: 65-106. 1960. Peart, M. H. Experiments on the ees significance of the morphology of seed- dispersal units in grasses. Jour. Ecol. 67: 843-863. 197 — M.N. Haustorial synergids in Cortaderia (Gramineae). New Zealand Jour. ot. 15: 777, 178. 1977 —. Apomixis in Cortaderia jubata (Gramineae). Ibid. 16: 45-59. 1978. 154 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 —. The haustorial syneneids of Cortaderia (Gramineae) at maturity. Acta Soc. Bot. Polon. 50: 151-160. 1981. & H. E. Connor. Haustorial synergids in danthonioid grasses. Bot. Gaz. 145: 78-82. 1984. Raprorp, A. E., H. E. AHLES, & C. R. BELL. Manual of the vascular flora of the Carolinas. Ixi + 1183 pp. a Hill. 1968. ReEDER, J. R. The embryo in grass systematics. Am. Jour. Bot. 44: 756-768. 1957. Renvoize, S. A. The subfamily Arundinoideae and its eats in relation to a general classification of the Gramineae. Kew Bull. 36: 85-102. 1982. A survey of the leaf blade anatomy in grasses. VI. ca Ibid. 40: 731-736. 1985; VIII. Arundoideae. /bid. 41: 323-342. 1986. Rosinson, E. R. Naturalized species of Cortaderia (Poaceae) in southern Africa. S. Afr. Jour. Bot. 3: 343-346. 1984. [Adventives C. Selloana, C. jubata are invasive weeds.] RosENGuRTT, B., & B. R. ARRILLAGA DE MArFEI. Nuevas especies y sinopsis de Stipa en el a Bol. Fac. Agron. Montevideo 72: 1-34. 1964. SAVILL, M. G., R. BickesTAFF, & H. E. ConNor. Interspecific variation in epicuticular waxes of ‘Chionochioa. Phytochemistry 27: 3499-3507. 1988. [Each species has distinct eae of ar chain carbon compounds. ] SCHECHTER, Y., L. Jounson. A new species of Oryzopsis (Gramineae) from Wy- oming. oe 18: 342-347. 1966. & The probable origin of Oryzopsis contracta. Am. Jour. Bot. 55: 61 1- 618. 1968. SODERSTROM, T. R. The grass subfamily Sn eee Taxon 30: 614-616. 1981. [Validation of subfamily: synopsis of included gene 0KER. Calderonella, a new genus a fae and its relationships to the centostecoid genera. Ann. eneaa Sa Gard. 60: 427-441. 1973. SPELLENBERG, R. W., & L. E. MEHLENBACH Anatomical and cytological studies of an as hybrid, Oryzopsis rier x Stipa Lemmonil (Gramineae). Canad. Jou t. 49: 1565-1574. 1971. SPRAGUE, = ee of cereals and grasses in North America (fungi, except rusts and smuts). xvi + 538 pp. New York. 1950. STAPF, O. The pampas grasses. FI. Sylva 3: 171-176. 1905. [Cortaderi ia.] Tarra, H. Studies on amino acid contents in plant seeds IV. Amino acid contained in the seed of Gramineae (part 3). Bot. Mag. Tokyo 79: 36-48. 1966. Tenorio, E. C. The subfamily Centostecoideae (Gramineae). xi1 + 420 pp. Unpubl. Ph.D. dissertation, Univ. Maryland, College Park. 1978. THomasson, J. R. Late Cenozoic grasses and other angiosperms from Kansas, Nebraska, and Colorado: biostratigraphy and relationships to living taxa. Kansas Geol. Surv Bull. 218: 1-68. 1979. . Paleoeriocoma (Gramineae, Stipeae) from the Miocene of Nebraska: taxonomic and phylogenetic ef mani Syst. Bot. 5: 233-240. 1980 [1981]. iocene grass (Gramineae: Arundinoideae) showing eran micromorphol- og ical and internal see features. Bot. Gaz. 145: 204-209. 1984. Tom.inson, K. L. Comparative anatomical studies in Danthonia sensu lato (Danthon- ieae: ea Aliso 11: 97-114. 1985. [Leaf and lodicule anatomy.] Townrow, J. E.S. The genus Stipa in Tasmania. Part 3—revised taxonomy. Pap. Proc. Roy. Soc. Tcmane 112: 227-287. 1978. [Twelve species described and illustrated; all rdienone] TsveLev, N. N. Zlaki SSSR. 2 parts. Leningrad. 1976. (Grasses of the Soviet Union. English uaa by B. R. Sharma. 2 parts (1196 pp.). Rotterdam. 1984.) [See pp. 848-936 in transl.] TuTin, T. G. Dea ae Cortaderieae, Danthonieae, Molinieae, and Aristideae (Gra- mineae). Pp. 252-255 in T. G. Tutin et al, eds., Flora Europaea. Vol. 5. Cambridge, England. 1981. 1990] TUCKER, ARUNDINOIDEAE 153 VALDES R., J., C. W. Morpen, & S. L. Hatcu. Gouldochioa, a new genus of centothecoid grasses from Tamaulipas, Mexico. Syst. Bot. 11: 112-119. 1986. [Ilustrations.] VickerRY, J. W., S. W. L. Jacons, & J. Everett. Taxonomic studies in Stipa in Australia. Telopea 3: 1-133. 1986. Vooruies, M.R., & J. R. THOMASSON. Fossil grass anthoecia ee Miocene rhinoceros skeletons: diet in an extinct species. Science 206: 331-333. 1979. Wa ters, S. M., et ai, eds. The European garden flora. Vol. ro pp. Cambridge, England, and other cities. 1984. [Arundinoideae, 34-36; six genera.] Watson, L., H. T. CLirForp, & M. J. DaLttwirz. The classification of the Poaceae: subfamilies and supertribes. Austral. Jour. Bot. 33: 433-484. WUNDERLIN, R. P. Guide to the vascular plants of Central Florida. viii = 472 pp. Tampa and other cities. 1982. [Poaceae, 51—96.] Yates, H. O. Morphology and cytology of Uniola (Gramineae). Southwest. Nat. 11: 145- 189. 1966. —_—— vision of grasses traditionally referred to Uniola, Il. Chasmanthium. Ibid. 415-455. 1966 [1967]. Yeou, H. H., & L. WAtson. Systematic iat in amino acid compositions of grass caryopses. Phytochemistry 20: 1041-1051. l. Zotov, V. D. Synopsis of the grass subfamily te asides in New Zealand. New Zealand Jour. Bot. 1: 78-136. 1963. KEY TO THE GENERA OF ARUNDINOIDEAE IN THE SOUTHEASTERN UNITED STATES General characters: air or annual herbaceous plants of dry to wet places. Leaf blades linear to lanceolate, flat or involute. Inflorescences paniculate (rarely Ores): terminal an lee Gils laterally compressed or subterete. Glumes 2, conspic uous; flowers i to several, disarticulating above glumes; lemmas often awned. A. Flowers imperfect or perfect (plants gynodioecious), forming dense tussocks. ..... f selav kanes tel esd oat toes eee desparate acca an eae encanto eS Cortaderia. | A. Flowers perfect o florets sometimes imperfect); plants rhizomatous, or loosely to densely caespito B. Plants large; ee cauline only. C. Rachillas glabrous; lemmas pilose throughout. ................ 1. Arundo. C. a pilose; lemmas glabrous or pilose on margins only. ternal ligules a line of hairs: lateral inflorescences often present; lemmas oe EXCUIVER AWNY s..3in cee Ae es a a panes . Neyraudia. D. External ligules absent; inflorescences strictly terminal; _ awnless ra eet eee thee th ohare at etier goa eae aes 2. Phragmites. B. Plants small to medium; leaves basal and cauline. E. Spikelets subterete, |-flowered; lemmas indurate, involute. Lemma awns |-parted. G. Paleas rudimentary. .... 0000.00. ccc cece ee eee ee 6. Nassella. G. Paleas slightly longer than and protruding from apex of lemmas. Be aetna dele eee gad cee pe des eee aa ea eae ees 7. Piptochaetium mma awns 3-parted. 0.0.00... ee 5. Aristida. E. Spikelets laterally flattened, 2- to 12-flowered; lemmas membranaceous, con- H. Glumes and spikelets about the same length; lemmas awned; oo s subcylindrical; leaves mostly basal. ................... 4. nthonia. H. Glumes much shorter than spikelets; lemmas awnless; ene ae flattened; leaves mostly cauline. ................. 8. hasmanthium. 156 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Tribe ARUNDINEAE Dumortier, Obs. Gram. Belg. 82. 1824. 1. Arundo Linnaeus, Sp. Pl. 1: 81. 1753; Gen. Pl. ed. 5. 35. 1754. Very large perennials of sunny, damp soils and shallow, fresh waters. Roots fibrous; rhizomes stout. Stems stout, solid, glabrous; nodes glabrous. Leaves several, cauline; sheath glabrous; ligule membranaceous, margin minutely cil- iate; blade lanceolate, auriculate, the margin and apex scabridulous; cross veins numerous. Inflorescences single, terminal, much branched, plumose. Spikelets lanceolate; glumes 2, lanceolate, as long as the spikelets; rachilla slender, flat- tened, glabrous; florets [1 or] 2; lemmas lanceolate, clasping, membranaceous to firm, 5- to 9-nerved, the lower '2 pilose; paleas lanceolate, '2—* as long as the lemmas, hyaline, the margin ciliate. Stamens 3; anthers linear, the apex and base of thecae divergent. Ovaries slenderly cylindrical; styles 2, free, very slender; stigmas slightly longer than styles, plumose, laterally exserted. Cary- opses oblong, smooth; hilum short; embryo large. Base chromosome number 12. Lecrorype species: 4. Donax L.* (Name from Latin word for cane.)— GIANT REED. A genus of three species native to the Old World. Arundo Donax L., 2n = 110, the ig wide-ranging species, occurs from Spain to India. The other species are A. Plinii Turra, 2n = 72, of the Mediterranean region, and 4. formosana raed endemic to Taiwan Arundo Donax is naturalized from seed of planted specimens in the south- eastern and southwestern United States. I have examined specimens from Virginia, North Carolina, Florida, Louisiana, Texas, and California. Seed set is reported to be poor in 4Arundo Donax in India because of the failure of meiosis in a majority of the ovules (Bhanwra). The reeds for woodwind instruments are cut from the stems of Arundo Donax (Clayton & Renvoize). The stems are used for thatch in southern Europe, Asia, and Africa, cellulose for paper pulp, and fodder. Arundo Donax is also cultivated around pools and on streambanks as an ornamental. REFERENCES: Under subfamily references see BAILEY ef a/., BHANWRA; CLAYTON & RENVOIZE; CONERT (1961); Munz; RENvoIzeE (1986); TuTIn; and WALTERS ef @ 2. Phragmites Adanson, Fam. Pl. 2: 34, 559. 1763. Tall, rhizomatous perennials of swamps, marshes, ditches, and roadsides. Roots fibrous; rhizomes stout, scaly, horizontal to oblique; long stolons often produced by terrestrial plants. Stems glabrous, often glaucous. Leaves cauline, 10-15; sheaths overlapping, with margin free for entire length, glabrous; ligule ‘Linnaeus included six species in his diverse genus Arundo. The genus was typified by removal of five of these: 4. Bambos L. to Bambusa (““Bambus’’) by Gmelin (Syst. Nat. ed. 13 2(1): 579. 1791); A. Phragmites L. to Phragmites by Adanson (Fam. Pl. 34, 559. 1763); A. epigejos L. and A. Cala- magrostis L. to Calamagrostis by Adanson (/bid. 31, 530); and A. arenaria L. to oT by Host (Gram. Austr. 4: 24. p/. 41. 1809), leaving only 4. Donax. Linnaeus (Gen. Pl. 35. 1754) described the lemmas as pilose (valvulae ... basi lanugo’’) in Arundo, a feature fitting only , roa among the species he included in Arundo. 1990] TUCKER, ARUNDINOIDEAE E57 FiGurE 1. Tribe Arundineae, spikelets or their parts. a, b, Arundo Donax: a, spikelet, x 3; b, floret (palea and lemma enclosing ee from adaxial side (note ere rachilla segment and hairs along edge of lemma), x 3. c, d, Phragmites australis: c, spikelet (parts spread out and hairs Reciie for clarity), x me “ floret from adaxial side ee abundant hairs from rachilla segment), x 4. e, f, Danthonia spicata: e, spikelet, x 3; f, floret from adaxial side (note geniculate awn borne from middle of notch between bilobed lemma apex), x 5. a short membrane bearing dense fringe of multicellular hairs; blades distichous below, becoming secund above, linear-lanceolate, constricted basally, widest at about '4 the length, tapered to a long, attenuate tip, flat, the midvein much wider and thicker than the lateral veins, cross veinlets visible on the abaxial surface. Inflorescences terminal, solitary, plumose, each node bearing several primary branches (1 or 2 of these larger than the others); nodes of the primary branches each with | or 2 secondary branches; secondary branches bearing single tertiary branch at each node; inflorescence axis, secondary, and tertiary branches terminated by solitary spikelets; nodes of inflorescence with ring of long, multicellular trichomes (these giving a woolly appearance to inflores- cence). Spikelets numerous, lanceolate; glumes 2, unequal, the second shorter than the florets; rachilla pilose, giving the inflorescence a silky or fluffy ap- pearance; flowers 2 to several, the lowest staminate or empty, the succeeding perfect; lemmas lanceolate, long-acuminate, 3-nerved; paleas less than '2 as long as the lemmas. Stamens 3 (lacking, 1, or 2 in lowest, staminate flower); anthers ellipsoid. Ovaries oblong; styles separate, short; stigmas plumose. Cary- opses subterete, oblong; styles persistent, indurate; basal furrow broad: hilum ’s as long as grain. Base chromosome number 12. TyPE species: Arundo Phrag- mites L. = P. australis (Cav.) Trin. ex Steudel. (Name from Greek phragmites, 158 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 hedge, probably applied to this genus because of its dense, hedgelike growth along ditches and waterways.)— COMMON REED. A genus of three (or more) species, nearly cosmopolitan in distribution. The only species in the New World, Phragmites australis (P. communis Trin.), is widespread, but not especially abundant or common, in the Southeast. Occurring on all continents except Antarctica, Phragmites australis may have the widest distribution of any angiosperm. It was present in the American Southwest at least 1000 years before European contact (Kane & Gross), and in southern New England at least 3000 years ago (Niering & Warren). Reported in 1843 as occasional throughout New York State (Torrey), 1t was not reported from Georgia or South Carolina by Elliot (1821-1824) and remained unknown in those states until the 1970’s (Stalter). It is now abundant in New York, where it may be in part adventive (Mitchell). It spread rapidly on Long Island during the early twentieth century, which may be explained by the introduction of a more aggressive biotype from Europe or elsewhere. Distinctive morpho- types, presumably genetically based, have been reported from southern Lou- isiana (Schott & White). The complex and confused taxonomy of Old World populations makes it difficult to match the North American populations with any of the numerous European varieties and cultivated forms. Phragmites japonicus Steudel, 2n = 48, was recognized by Tsvelev, who distinguished it from P. australis by its shorter glumes and spikelets and its zigzag (not straight), elongate rhizomes. The species occurs in Japan, China, and the extreme eastern Soviet Union. Phragmites Karka (Retz.) Trin., 21 = 18, 36, 38, 48, grows in tropical Africa, southeastern Asia, and northern Aus- tralia (where its distribution complements that of the more temperate P. aus- tralis). Phragmites mauritianus Kunth is found in tropical Africa and the islands of the Indian Ocean (Fanshawe). The species are barely distinguishable (Clayton & Renvoize). There is considerable variation in the chromosome complement of Phrag- mites australis. Haslam (1972) reported diploid numbers of 36, 48, 84, and 96. Aneuploids are also reported: 2” = 42, 44, 46, 49, 50, 51, 52, 54, 56 (Gorenflot et a/., 1972). In Europe tetraploids and octaploids account for most populations; hexaploids are rare, being reported from Sweden (Bjérk), the Mediterranean region, and Iran and Afghanistan (Bahrman & Gorenflot). In the New World hexaploids have been reported from Costa Rica (Pohl). Tet- raploid and octaploid clones grow together in the Danube Delta (Bahrman & Gorenflot). No correlation between ploidy, habitat, and geography was found in a survey of 40 European populations (Raicu ef al.). Apparently the only report from North America is 2n = 48 from a Canadian population (Hunter). Bahrman & Gorenflot reported variation in soluble proteins of leaves in an extensive survey of European, North African, and Central Asian populations. Principal-components analysis revealed four major groups on the basis of iso- zymes present. The diversity of isozymes could not be correlated with ploidal level. The greatest diversity in isozymes was found in populations in Iran and Afghanistan, while the greatest variation in chromosome number was in south- ern Europe. Evidently, macromolecular diversification in this species has pro- ceeded independently from chromosomal changes. 1990] TUCKER, ARUNDINOIDEAE 159 There are numerous reports worldwide of lack of seed set in Phragmites australis. A complete explanation has yet to be provided, and the question offers possibilities for further research. Meiosis is regular in certain populations from France (Cartier & Lenoir). However, chromosome fragmentation during microsporogenesis was detected in populations from Ireland that failed to set seed (Curran). In populations from India, megasporogenesis was normal, but microsporogenesis was arrested at the tetrad stage, and there was no mitotic division to produce the vegetative and generative nuclei (Satyamurty & Ses- havatharam). Abundant seed set has been reported in populations from Af- ghanistan and Iran (Bahrman & Gorenflot), South Africa (Curran), Minnesota (Harris & Marshall), and the Mackenzie District, Northwest Territories, Canada (Cody). In Swedish populations, however, reproduction by seed was limited: only 0.3 -8.0 percent of the florets produced viable seed (Gustafsson & Simak). No viable fruits have been reported from Ontario populations, and reproduc- tion there has apparently been entirely vegetative (Dore & McNeill). Self- incompatibility may account for low seed set because clonal growth has pro- duced genetically homogeneous populations (genets). Germination is influenced by temperature, with the rate increasing linearly from 16 to 26°C, while the number of days needed for germination decreases from 25 at 16°C to only ten at 26°C. The seeds cannot germinate under more than 5 cm of water. Seed germination 1s little affected by salt concentrations below one percent, reaching a limit of tolerance at two percent (Kim et a/.) in Korean populations. Plants of Phragmites australis are able to grow in brackish as well as fresh water. Evidently tolerance of salinity varies widely (Hocking et al.): a maximum tolerance for mature plants was reported as 1.2 percent in Britain, 2.9 percent in New York, and 4.0 percent on the Red Sea coast. For detailed reviews of the autecology and physiological ecology of Phrag- mites australis, see Haslam (1972, 1973) and Hocking and colleagues. For a popular nontechnical account of general ecology and economic importance, see Brown. Plants of Phragmites are used by humans in several ways. Stems are used for basketry and thatch in several European countries, and as a source of pulp for paper production, especially in eastern Europe. Immature stems provide forage for cattle in Australia (Hocking et a/.) and southern Africa. Phragmites australis is an invasive weed in the northeastern United States. The rhizomes are able to grow under pavement and cause damage by cracking and piercing it (Amano & Maki; Hocking ef a/.). Unwanted populations can be effectively removed by cutting them several times annually, particularly in midsummer when rhizome starch reserves are lowest. Dense stands provide wildlife habitat in Europe (Bibby & Lunn). Reeds are planted on recently drained lands in the Netherlands and Japan (Kamio) to remove excess water and prepare the soil for agriculture. REFERENCES: Under subfamily ea see CLAYTON & RENVOIZE; Dore & MCNEILL; HOLM et al., Munz; and Tsv AMANO, K., & T. MAKI. Studies on the method of random paving for the slope protection of embankments. (Part 3) The mechanism of the destruction in asphalt concrete 160 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 paving by reeds (Phragmites communis Trinius). (In Japanese; English summary.) Jour. Agr. Sci. Tokyo 29: 89-100. 1984. Apinis, A. E., C. G. C. Cuesters, & H. K. TALIGOoLA. Colonisation of Phragmites communis leaves by ime Nova Hedwigia 23: 1 13- ee 1972. Aicrofungi nternodes of aerial standing culms of Ff RTES coca Trin. Ibid. 26: 495- 507. 1975. BAHRMAN, N., & R. GORENFLOT. Apport des protéines foliaires solubles dans l’inter- prétation ‘du complexe polyploide du Phragmites australis (Cav.) Trin. ex Steudel. Rév. Gén. Bot. 90: 177-184. 1983. Bay Ly, I. L., & T. A. O'NEILL. Seasonal tonic Ae ee in a Phragmites communis community. Canad. Jour. Bot. 50: 2103-2109. Best, E. P. H., M. Zrpprn, & J. H. A. DASSEN. nike i and production of Phragmites australis in Lake Vechten (The Netherlands). Hydrobiol. Bull. 15: 165-173. 1981. Brissy, C. J., & J. Lunn. Conservation of reed beds and their avifauna in England and Wales. Biol. Conserv. 23: 167-186. 1982. [Phragmites australis. ] ByOrRK, S. Ecologic investigations of Phragmites communis. Studies in theoretic and applied limnology. Folia Limnol. Scand. 14: 3-248. 1967. BorRNKAMM, R., & F. RAGHI-Atri. Uber die Wirkung unterschiedlicher Gaben von Stickstoff und Phosphor auf die Entwicklung von Phragmites australis (Cav.) Trin. ex Steudel. Arch. Hydrobiol. 105: 423-441. 1986. [Effects of nitrogen and phos- phorus on growth. Brix, H. Light dependent variations in the composition of the internal atmosphere of Phragmites australis (Cav.) Trin. ex Steudel. Aquatic Bot. 30: 319-329. 1988. Brown, L. Reed discovery. Horticulture 59(2): 32-37. 1981. Cartier, D., & A. LeNorr. Contribution a l’étude du développement de Povule du Phragmites australis (Cav.) Trin. ex Steud. Rev. Gén. Bot. 87: 289-295. 1980. CHASHCHUKHIN, V. A. Ecological a ha gas regime in the rhizome of the common reed. Soviet Jour. Bot. 10: 68, 69. CLayton, W. D. Studies in the eae aN Kew Bull. 21: 111-117. 1967. [World- wide synopsis of Phragmites; three species.] . The correct name of the common reed. Taxon 17: 168, 169. 1968. Copy, W.J. A contribution to the knowledge of the flora of the southwestern Mackenzie District, N. W. T. Canad. Field-Nat. 77: 108-120. 1963. Curran, P. L. Fertility of Phragmites communis. Irish Nat. Jour. 16: 242. 1969. Davis, A. N., & T. L. Briccs. Dispersion pattern of aerial shoots ra common marsh reed Phragmites australis (Poaceae). Rhodora 88: 325-330. De LA Cruz, A. A. The production of pulp from marsh grass. ae Bot. 32: 46-50. 1978 Denny, P. Permanent swamp vegetation of the Upper Nile. Hydrobiologia 110: 79-90. 1984. 'Phraemites Karka.] Dinka, M. The effect of mineral nutrient enrichment of Lake Balaton . the common reed (Phragmites communis). Folia Geobot. Phytotax. 21: 65-84. 1986. Drum, R. W. Electron microscopy of opaline ae in cs Rie and other Gra- mineae. (Abstract.) Am. Jour. Bot. 55: 713. Durska, B. Changes in the reed (Phragmites pecee Trin.) condition caused by diseases of fungal and animal origin. Pol. Arch. Hydrobiol. 17: 373-396. 1970.* DyxysovA, D. Ecomorphoses and ecotypes : Phragmites communis Trin. (In Czech; English summary.) Preslia 43: 120-138. & D. Hr KA. Production ecology a Phragmites communis. 1. Relations of two ecotypes to the microclimate and nutrient conditions of habitat. Folia Geobot. Phytotax. 1: 23-61. 1 ,K. Vesper, & K. PripAn. Productivity and root/shoot ratio of reedswamp species wing in Suidoor hydroponic cultures. Folia Geobot. Phytotax. 6: 233-254. 1971. Euuiot, S. Sketch of the botany of South Carolina and Georgia. 2 vols. 606 + 743 pp. Charleston, South Carolina. 1821-1824. 1990] TUCKER, ARUNDINOIDEAE 161 FANSHAWE, D. B. The biology of the reed—Phragmites mauritianus Kunth. Kirkia 8: 147-150. 1972. FERNALD, M. L. The generic name Phragmites. Rhodora 24: 55, 56. 1922. —. Phramites communis Trin. var. Berlandieri. Ibid. 34: 211, a 1932. hragmites communis vs. P. maximus. Ibid. 43: 286, 287. GALLAGHER, J. L., & M. J. MILs. Seasonal patterns in storage es in Phrag- mites australis. (Abstract.) Am. Jour. Bot. 73: 667. 198 GoOrRENFLOT, R. Le complexe polyploide du Phragmites australis ree ) Trin. ex Steud. = P. communis Trin.). Bull. Soc as pls 123: 261-271. , J.-M. Hupac, M. Jay, & P. on NDE. Geographic a polyploidy, and patterns of flavonoids in Phragmites australis (Cav.) Trin. ex Steud. Proc. NATO Adv. Stud. Inst. G. 1: 474-478. 1982. —., P. Raicu, D. Cartier, I. CIoBANU, V. STOIAN, & S. Staicu. Polyploid complex en communis Trin. Compt. Rend. Acad. Sci. Paris, D. 274: 1501-1504. ——. a M. SANEI CHARIAT-PANAHI. Le complexe polyploide du Phragmites australis (Cav.) Trin. ex Steud. (= P. communis Trin.) en Iran. Rév. Cytol. Biol. Vég. Bot. 2: 67-81. 1979. GorHAM, E., & W. H. PEARSALL. Production ecology III. Shoot production in Phragmites in relation to habitat. Oikos 7: 206-214. 1956. A. GustTaAFsson, A., & M. SIMAK. X-ray photography and seed sterility in Phragmites commuants Trin. Hereditas 49: 442-450. 1963. HALDEMANN, C., & R. BRANDLE. Seasonal variations of reserves and of fermentation processes in 1 wetland plant rhizomes at the natural site. Flora 178: 307-313. 1986. Hansson, L.-A., & W. GRANELI. Effects of winter harvest on water and sediment chem- istry in a stand of reed See See australis). Hydrobiologia 112: 131-136. 1984. Harris, S. W., & W.H. MARSHALL. Experimental germination of seed and establishment of seedlings of Phragmites communis. Ecology 41: 395. 19 HAsLaM, S. M. Stem types of Phragmites communis Trin. Ann. ‘Bot. II. 33: 127-131. 1969a. . The development and emergence of buds in Phragmites communis Trin. Ibid. 289-301. 1969b. —. The development of shoots in Phragmites communis Trin. [bid. 695-709. 1969c. . The development of the annual population in Phragmites communis Trin. Ibid. 34: 571-591. 1 —. Biological flora of the British Isles. Phragmites communis Trin. Jour. Ecol. 60: 585-610. 1972. . Some aspects of the life history and autecology of Phragmites communis Trin., a review. Polsk. Arch. Hydrobiol. 20: 79-100. 1973 Ho, Y. B. Shoot development and production studies of Phragmites australis (Cav.) Trin. ex Steudel in Scottish lochs. Hydrobiologia 64: 215-222. 1979, HockinG, P. J., C. M. Fintayson, & A. J. CHick. The biology of Australian weeds. 12. Phragmites australis (Cav.) Trin. ex Steud. Jour. Austral. Inst. Agr. Sci. 1983: 123- 132. 1983. [Detailed review emphasizing ecology. Hu, Y. H., & C. L. Lt. Comparative anatomy of the culm and fibers of Phragmites Pes d Miscanthus sinensis. In Chinese; English summary.) Acta Bot. Sinica 1: 252-260. 1963. tes A. W.S. A karyosystematic investigation in the Gramineae. Canad. Jour. Res. 11: 213-241. 1934. Hurauisiu, I. Soluble proteins in Phragmites australis (Cav.) Trin. ex Steud. growing in different ecological conditions. Hydrobiologia 15: 275-278. 1977. INGRAM, H. A. P., A. M. BARCLAY, A. M. Coupar, J. G. GLover, B. M. Lyncn, & J. I. SpreNT. Phragmites performance in reed beds in the Tay Estuary. oe, Ror, Soc. Edinburgh 78B: 89-107. 1980. Kamio, A. Studies on the drying of marshy and heavy clay soil ground by means of 162 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 vegetations—on the process of polder land drainage and structural changes of Phrag- mites communis community in the Hachirogata central polder. Jap. Jour. Ecol. 32: 357-364. 1982. Kang, A. E., & G. T. Gross. Anasazi communities at Dolores: early Anasazi sites in the Sagehen Flats area. 985 pp. Denver. 1986. [Phragmites plant parts associated with several ruins in southwestern Colorado dated a.p. 600-800. Kaneta, M., & N. SucrvAMA. The constituents of Arthraxon hispidus Makino, Mis- canthus tinctorius Hackel, Miscanthus sinensis oo and Phragmites communis Trinius. Bull. Chem. Soc. Japan 45: 528-531. Kauprl, P., J. SELKAINAHO, & P. PUTTONEN. A ona for estimating above-ground biomass in Phragmites stands. Ann. Bot. Fenn. 20: 51-55. 1983. Kin, K.S., Y.S. Moon, & C. K. Lm. Effect of NaCl on germination of Atriplex Gmelinii and Phragmites communis. (In Korean: English abstract.) Korean Jour. Bot. 28: 253- 259. 1985. Lau, E., M. Gotporras, V. D. BALDwin, P. DAYANANDAN, J. SRINIVASAN, & P. B. KAUFMAN. Structure and ene of silica in the leaf and internodal epidermal system of the marsh grass Phragmites australis. Canad. Jour. Bot. 56: 1696-1701. 1978. LAWRENCE, D. B. Arboretum’s reed marsh in esa perspective: a plea for conser- vation oh natural resource. Minn. Agr. Exper. Sta. Misc. Rep. 111: 24-27. 1972. Leror, M. The common reed, ne ae “astra (Cav.) Trin. 7 Steud. (Gra- eae a buck overview. Newsletter Con t. Soc. 15(1): 7-10. 1987. Lenoir, A., V. STOIAN, D. CARTIER, R. sarees a P. RaIcu. ace et méiose nollinique du Phragmites australis (Cav.) Trin. ex Steud. Compt. Rend. Acad. Sci Paris, D. 280: 621-624. 1975. LuTHeER, H. nae era iiber ae fruktifikative Vermehrung von Phragmites com- munis Trin. Acta Bot. Fenn 1950. ANTLE, P. G. Studies on Cae purpurea (Fr.) Tul. parasitic on Phragmites com- munis Trin. Ann. Appl. Biol. 63: 425-434. 1969 ETZLER, K., & R. Rozsa. Additional notes on ey tidal wetlands of the Connecticut River. Newsletter Conn. Bot. Soc. 15(1): 1-6. MITCHELL, R. 7 A checklist of New York State aes ie York State Mus. Bull. 458. x + 272 p Mook, J. H., « J. VAN DER ToorRN. The influence of environmental factors and man- agement on stands of aes oo, II. Effects on yield and its relationships with shoot density. Jour. Appl. E 19: 501-517. 1982. [See also TooRN & Mook. NIeRING, W. A., & R. S. WARREN. Ou ae tidal marshes: vegetation changes as revealed by peat analysis. Conn. Arb. Bull. 12. 22 pp. 1977. Onpbok, J. P. The horizontal structure of reed stands (Phragmites communis) and its relation 1 to productivity. Preslia 42: 256-261. 1970. Patuts, M. The structure and history of Plav: the floating fen of the delta of the Danube. Jour. Linn. Soc. Bot. 43: 233-290. 1916. PazoureEK, J. Density of stomata in leaves of different ecotypes of Phragmites communis. Folia er Phytotax. Praha 8: 15-21. 1973. Pearcy, R. W., J. A. Berry, & B. BARTHOLOMEW. Field photosynthetic performance and Relea of Phragmites communis under summer conditions in Death Valley, California. Photosynthetica 8: 104-108. 1974. a P.,S. Staicu, V. STOIAN, & T. ROMAN. Phragmites communis Trin. chromosome mplement in the Danube Delta. Hydrobiologia 39: 83-89. 1972. ee. K. P., & Y. I. MoLotKovskil. Water ca in Phragmites australis in southern Tadzhikistan. ae Jour. Ecol. 10: 384-393. Ramirez G., C., & N. ANAzco R. Variaciones eae en el desarrollo de Scirpus californicus, Typha angustifolia, y Phragmites communis en Pantanos Valdivianos, Chile. (English summary.) Agro Sur 10: 111-123. 1982. 1990] TUCKER, ARUNDINOIDEAE 163 SANGSTER, A. G. Electron-probe microassays for silicon in the roots of Sorghastrum nutans and Phragmites communis. Canad. Jour. Bot. 56: 1074-1080. 1978. . Anatomical features and silica depositional patterns in the rhizomes of the grasses Sorghastrum nutans and Phragmites australis. Ibid. 61: 752-761. 1983. SATYAMURTY, T. V. C., & V. SESHAVATHARAM. Sterility in Phragmites communis (Retz.) Trin. Curr. Sci. Bangalore 53: 1158, 1159. 1984. inlustration| SCHIERUP, H.-H., & V.J. LARSEN. M copper, lead in the littoral zone of a polluted and < a non- -polluted lake. ei TG bie. uptake, and translocation of heavy metals in Phragmites australis (Cav.) Trin. Aquatic Bot. 987 A qd ScuoTt, M. J., Jn., & D. A. WHitE. Morphometric analysis of two adjacent populations of Phragmites australis in the Mississippi River Delta. ASB Bull. 35: 54. 1988. STALTER, R. Phragmites communis in South Carolina. Rhodora 77: 159. 1975. [First record for a ] SToIAN, V., A. LENorr, P. Raicu, & R. GORENFLOT. La méiose et la taille ranean duindividus octoploides du Phraainiles ean Compt. Rend. Acad. Sci. Pari 278: 457-459. 1974, SZAJNOWSKI, F. Relationship bet leaf area i d st ducti f Phragmites communis Trin. Pol. Arch. Hydrobiol. 20: 257-268. 1973. TALIGOOLA, T. K., A. E. Apinis, & C. G. C. CHEsSTERS. Microfungi colonizing collapsed aerial parts of Phragmites australis Trin. in water. Nova Hedwigia 23: 465-472. 1972. Tuompson, D. J., & J. M. SHAy. The effects of fire on Phragmites australis in the Delta Marsh, Manitoba. Canad. Jour. Bot. 63: 1864-1869. 1985. Toorn, J. VAN DER, & J. H. Moox. The influence of environmental factors and man- Torrey, J. A flora of the state of New York. 2 vols. 1056 pp. /6/ pls. Albany. 1843, TSCHARNTKE, T. Variability of the grass Phragmites australis in relation to the behaviour and mortality of the gall-inducing midge Giraudiella inclusa (Diptera, Cecidomyi- idae). Oecologia 76: 504-512. 1988. [Eggs laid on the fourth internode, which is high in nutrients but low in silica.] VesBeER, K. Evaluation of introduced forms in experimental plantations oe reed (Phragmites communis Trin.). Pol. Arch. Hydrobiol. 22(4): 33-42. WAISEL, Y., & Y. RECHAW. Ecotypic differentiation in Phragmites communis Trin. Hydrobiologia 12: 259-266. 1971. WALKER, J. M., & E.R. Waycoop. Ecology of Phragmites communis. I. phos wee ofa single shoot in situ. Canad. Jour. Bot. 46: 549. 1968. [Eight percent of phot synthesis carried out by leaf sheaths. ] WEISNER, S. E. B. Factors affecting the internal oxygen supply of Phragmites australis (Cav.) Trin. ex Steudel in situ. Aquatic Bot. 31: 329-335. 1988. [Oxygen transport probably es depth tolerance.] WEISSER, P. J., & R. J. PARSONS. Monitoring Phragmites australis increases from 1937 to 1976 in se Siyai Lagoon (Natal, South Africa) by means of air photo interpre- tation. Bothalia 13: 553-556. 1981. [Area covered increased fourfold i in 20 years. ] Wu, G.-L., H.-C. Ye, & G.-F. Li. Emt of Phragmites communis. (In Chinese: English abstract.) Acta Bot. Sinica 29: 361- 366. 1987. YAMASAKI, 8S. Role of plant aeration in zonation of Zizania latifolia and Phragmites australis. Ea ae Bot. 18: 287-297, 1984. & I Growth responses of Zizania latifolia, Phragmites australis, and Fe sacchariflorus to varying inundation. Aquatic Bot. 10: 229-239. 1981. 164 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 7] 3. Neyraudia Hooker f., Fl. Brit. India 7: 305. 1896. Caespitose perennials of damp, sunny places. Roots fibrous; rhizomes short, solid. Stems erect; nodes glabrous. Leaves several, cauline; sheath glabrous; ligule membranaceous, pilose (abaxial ligule a conspicuous cartilaginous ridge, pilose, becoming glabrous with age); blade linear-lanceolate, very slightly au- riculate, the margin scabridulous, the surfaces glabrous. Inflorescences terminal (and sometimes axillary also), paniculate, much branched, plumose. Spikelets oblong, 3- to 6-flowered; glumes 2, broadly lanceolate, hyaline; rachilla inter- nodes pilose; calluses pilose; lemmas lanceolate, 3-nerved, pilose abaxially toward the edges, the apex bifid, aristate from the notch, the awn scabridulous, stifly excurved or recurved, sometimes slightly spiraled; paleas lanceolate, hyaline, 2-nerved. Stamens 3; anthers ellipsoid. Ovaries oblong, glabrous; styles 2, free; stigmas short, feathery. Caryopses linear, subterete, slightly dorsiven- trally flattened; hilum short; embryo large. Base chromosome number 10. Type species: N. madagascariensis (Kunth) Hooker f. = N. arundinacea (L.) Henr. (Name an anagram of Reynaudia, a monotypic genus of Cuban grasses.)— SILK-REED. Two species, both endemic to the Old World tropics. One species, Neyraudia Reynaudiana (Kunth) Keng, 2” = 40, silk-reed, Burma-reed, is adventive in southern Florida (Hall; Long & Lakela; several specimens, the earliest from 1930, examined from Dade County). Nevraudia arundinacea (L.) Henr., 2n = 40, a native of southern Asia, has been cultivated in southern Florida (Hall) and might be anticipated as an adventive. In N. arundinacea the first lemma subtends a perfect flower, while in NV. Reynaudiana it subtends a sterile one. Leaves of Neyvraudia have an abaxial (external) ligule, which has been over- looked by some workers. It consists of a cartilaginous ridge convergent at its ends with the adaxial ligule. The abaxial ligule is pilose at first but usually becomes glabrous as the leaf matures. REFERENCES: Under subfamily references see ee (1883); CLAYTON & RENVOIZE; HALL; LONG & LAKELA; and PALMER & TUCK 4. Danthonia A. P. de Candolle in Lamarck & A. P. de Candolle, Fl. Frang. ed. 3. 3: 32. 1815, nom. cons. Caespitose perennials. Roots fibrous or wiry. Rhizomes lacking. Stems sev- eral, unbranched, terete, more or less scabridulous, glabrous or pilose; nodes with a medial constriction. Leaves several to many, basal and cauline; sheath open, shorter than the blade, glabrous or pilose; ligule a dense fringe of short hairs; blade flat (sometimes becoming involute in age), glabrous or pilose, the margin and midvein scabridulous. Inflorescences terminal and axillary (ter- minal cleistogamous or chasmogamous; axillary cleistogamous, enclosed in sheath); unbranched except for 2-5 primary branches, these appressed or spreading, scabridulous. Spikelets |-3 per branch, narrowly ellipsoid, 4- to 8-flowered; glumes 2, equal, lanceolate, (3- to) 5- to 7-nerved, acute, mucron- ulate, persistent; rachilla straight, glabrous; calluses pilose; lemmas elliptic, 1990] TUCKER, ARUNDINOIDEAE 165 5-nerved, pilose abaxially especially near the margin, bidentate, bearing an awn from the notch between the teeth, awn with spiraled base and straight apex (axillary florets awnless); paleas broadly elliptic. Lodicules 2 or lacking, oblong to ovate, entire or truncate. Stamens 3; anthers narrowly ellipsoid, the thecae divergent apically and basally. Ovaries ovoid; styles very short; stigmas 2, feathery. Caryopses oblong to ellipsoid, dorsiventrally flattened (the abaxial face convex, the adaxial concave), the base stipitate, the apex obtuse; embryo about 3 as long as endosperm. Base chromosome number 12. (Sieglingia Bernh., nom. rejic.; including Rytidosperma Steudel, Notodanthonia Zotov.) TYPE SPECIES: Avena spicata L. = D. spicata (L.) Roemer & Schultes, typ. cons. (Named in honor of Etienne Danthoine, early nineteenth-century French bot- anist.)— OAT GRASS. A genus of about 80 species (Conert, 1987), here accepted in a broad sense to include the segregate Rytidosperma. Rytidosperma was centered in Australia and New Zealand and was treated as a genus, on shaky grounds, by Clayton & Renvoize (but not by Conert, 1987). Reviewing all species attributed to Danthonia, Conert found no constant feature to distinguish it from Rytidosper- ma. The patterns of lemma pubescence, often used to distinguish between the two genera, show greatest diversity in Australia. Tomlinson, in an anatomical survey of the tribe, noted that Danthonia and Rytidosperma are alike in leaf- blade anatomy but differ in lodicule morphology. In Danthonia the lodicules lack both macrohairs and microhairs, while in Rytidosperma both are present. There are 32 species in Australia (Conert, 1987; Vickery) and 16 in New Zealand. There are two endemics in the Himalayas, two species in Northeast Africa, and 18 species in South America. All the North American represen- tatives of Danthonia are 2n = 36 (Gray et al.). Only three species occur in the Southeast. Danthonia spicata (L.) Beauv. (leaf blades involute; panicle branches erect; glumes 7-14 mm long; lemma teeth (not central awn) less than 2 mm long) is widespread, occurring from Labrador to southeastern Alaska, south to western Florida and eastern Texas, and in the mountains of New Mexico and northern California. A disjunct population occurs in Hidalgo, Mexico (Conert, 1987). Canadian populations were recognized by Fernald as var. pinetorum Piper, said to differ in having mostly straight (not curled) basal leaves and broader, weakly nerved glumes. ore & McNeill, however, thought most Ontario populations belonged to the typical variety. The second southeastern species, Danthonia compressa Austin (leaf blades flat; panicle branches spreading; glumes 7-14 mm long; lemma teeth about 3 mm long) occurs from southern Quebec to Ohio and south to the mountains of North Carolina and Tennessee. The third, and probably the most abundant, southeastern species is Dan- thonia sericea Nutt. (leaves involute, generally silky-pubescent; panicle branches short, erect; glumes 12-18 mm long; longer lemmas 7-10 mm long), which occurs on the Coastal Plain and Piedmont from southern New Jersey and southern Kentucky to northern Florida and Louisiana. It consists of three races, which have been accorded varietal or specific status by some workers. The 166 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 major taxonomic problem in the southeastern taxa concerns this species and its segregate D. sericea subsp. epilis (Scribner) Bloma. (D. epilis Scribner; leaves glabrous), reported to be restricted to the upper Piedmont of North Carolina and Georgia. Quinn & Fairbrothers made cytological preparations from 21 populations and found no karyological differences (all were 2n = 36). Popu- lations of D. sericea subsp. epilis in New Jersey begin growth and flowering in response to increasing soil temperature and are day-length neutral. However, opulations on granitic outcrops in Georgia have their flowering time fixed genetically (Rotsettis et a/.) and start to flower early in the year, which may be advantageous in allowing seed set before the thin soil in which the plants typically grow dries out in the summer. Gray and colleagues also studied these two regional groups of subsp. epi/is: populations from bogs in New Jersey lacked stomata and macrohairs on the abaxial leaf surfaces but had them on the adaxial ones, while plants from well-drained sites of the Piedmont and Coastal Plain had hairs and stomata on both surfaces. Some intermediate populations from sites that were intermittently wet had abaxial hairs and stomata, but they were more plentiful on the adaxial surface. All populations had the same chromo- some number, 2” = 36. When grown in the greenhouse, both races responded to decreasing soil moisture by reduction in the width of the stomatal aperture. There is no information on compatibility or sterility between these differen- tiated populations, so it seems premature to recognize them taxonomically. The potential for further research 1s enticing. Five other species of Danthonia occur in North America. Danthonia inter- media Vasey is boreal and cordilleran in distribution, occurring from New- foundland to Alaska and south to the mountains of New Mexico and northern California (Cayouette & Darbyshire). It also grows on the Kamchatka Peninsula of the eastern Soviet Union (Conert, 1987; Tsvelev). Danthonia Parryi Scribner of the Rocky Mountains (Alberta to Colorado) is morphologically similar and probably closely related. The second pair of western species is characterized by few-flowered inflo- rescences. Danthonia californica Bol. has an amphitropical distribution, oc- curring in the West from Montana to southwestern British Columbia and south to New Mexico and California, and is also reported from Chile (Munz). Pilose plants have been segregated as var. americana (Scribner) Hitche. Danthonia unispicata (Munro ex Thurber) Munro ex Macoun has inflorescences consisting of only one (to three) spikelet(s). Its range is similar to that of D. californica, and Munz considered the two species as doubtfully distinct. The Australian Danthonia pilosa R. Br., 2n = 48, is adventive in California from seed planted for forage (Hitchcock, 1951; Munz). Baum & Findlay emphasized lodicule morphology in their revision of the Canadian species of Danthonia. They detected four lodicule patterns, two of which are represented in the southeastern species (lodicules were absent from both D. sericea and D. spicata; club-shaped lodicules— presumably two per floret, although they did not say so—with truncate apices characterized D. compressa). Thus they recognized the North American species chiefly by the lodicules, a taxonomic scheme not receiving much if any subsequent accep- tance. In a second paper, Findlay & Baum described a new species, D. cana- 1990] TUCKER, ARUNDINOIDEAE 167 densis Findlay & Baum, which occurs across Canada and the northern United States, and which differs from D. Parryi in the shape of the lodicules and other quantitative characters usable only by means of principal-components analysis. Such “‘one-character” taxonomy is problematic in a genus beset with taxonomic difficulties (see Rotsettis et a/.; Vicke notable feature of Danthonia is the production of cleistogamous spikelets by some species, including the three southeastern representatives. Cleistogamy was apparently first noted in the genus by Austin, who described “‘spikes’’ on short branches wholly enclosed by upper leaf sheaths. Clay (1982) investigated the reproductive biology and population genetics of D. spicata in North Car- olina. All plants produced both axillary (cleistogamous) and aerial (““chasmog- amous”’—actually cleistogamous and chasmogamous) flowers,’ but the pro- portion of each kind was variable. Using vegetatively produced individuals and growing these under different conditions, Clay (1982) found that plants of the same genotype consistently produced more aerial florets in the greenhouse (40 percent) than in the field (28 percent). Larger plants produced a higher percentage of axillary florets. From genetic analysis of related individuals, he determined that 50 percent of the observed variation in production of axillary florets was genetically based and 50 percent environmentally based. Genetic differences were possibly the result of natural selection. In other natural pop- ulations the percentage of axillary florets was lower: eight or nine percent in Wyoming and zero in Michigan (Scheiner & Teeri). Clay & Antonovics (1985b) also compared genetic variation in certain mor- phological characters for greenhouse- and field-grown plants of Danthonia spi- cata. They found significant genetic variation present for every character ex- amined, both vegetative (height of flowering stem; length and width of uppermost leaf) and reproductive (number of cleistogamous flowers; length of second glume). The degree of variability among closely related individuals (plants from seeds of the same carpellate parent) suggested that the aerial florets are in part cleistogamous or self-pollinated. The three southeastern species vary in the relative proportions of aerial and axillary florets (Clay, 1983b). Danthonia compressa averaged 50 percent aerial florets (minimum, 35 percent); D. spicata ranged from zero to 70 percent, with an average of 25 percent; and D. sericea, including plants assignable to D. epilis, averaged only five percent (maximum, ten percent). Grazing or mowing appears to favor the production of axillary florets over aerial, since such populations had the highest percentage of axillary florets. Clay (1983b) compared the weight of diaspores (caryopsis plus clasping lem- ma and palea) from aerial and axillary florets in the southeastern species. In Danthonia compressa the average weight for both kinds was 0.80 mg, while the axillary ones were 25 percent heavier in D. sericea and 35 percent heavier In sense from Durham County, North Carolina, studied by Clay, all aerial florets were pel chasmo ogam s. Ho owever, Clay sent seeds from his study population to Philipson, who grew them atte ealand: About half of the aerial spikelets in the resulting an were cleistogamous aes. 1986). In the ele of genetic studies by Clay, I have used the cate- gories aerial and axillary florets where he used “‘chasmogamous”’ and “‘cleistogamous.’ is discussion below and Philipson for further details, a see Darbyshire & Cayouette for commen 168 JOURNAL OF THE ARNOLD ARBORETUM [voL. 7] in D. spicata. In the two latter species about two thirds of the difference was due to heavier lemmas and paleas in the axillary florets. A comparison of the fitness of plants derived from aerial and axillary cary- opses in Danthonia spicata showed only a slight difference (Clay & Antonovics, 1985a). The advantage in survival was mostly during the seed and seedling stages (Clay, 1983a). Both kinds of diaspores germinate in mid- to late spring in North Carolina. Maximum germination was about 30 percent for seeds stored dry at room temperature for several months. Axillary cleistogenes ger- minate over a longer period (30 percent germination in ten days) than aerial (10 to 15 percent in 30 to 40 days). Subsequent studies by Philipson show the diversity of reproductive features in Danthonia spicata. She grew 49 plants (122 panicles) from a single North Carolina population and found that 52 percent of the florets from aerial panicles were in fact cleistogamous (rather than chasmogamous, as was assumed by Clay). There was wide variation in the percentage of cleistogamy, and 58 panicles (nearly 50 percent) were wholly cleistogamous. In the aerial cleistog- amous florets the anthers were tiny and indehiscent; the pollen grains germi- nated in situ, and the tubes grew through the anther wall into the adjacent stigmas. Often, both chasmogamous and cleistogamous florets were produced in the same spikelet (the cleistogamous generally proximal to and maturing sooner than the chasmogamous). Thus, many plants produced three kinds of diaspores: chasmogamous and cleistogamous ones from aerial inflorescences, both dispersed some distance by the hygroscopic lemma awns, and atelechoric awnless axillary cleistogenes generally germinating quite close to the parent plant. Danthonia spicata is able to grow in diverse environments, from open sandy soil in full sun to closed-canopy oak-pine forest with only ten percent of full sunlight in Michigan (Scheiner & Teeri). Individual plants have the ability to recover from severe droughts. Both genetic differentiation (microevolution or genetic drift) and phenological flexibility appear to contribute to this ecological amplitude. Population variability based on genetics has also been demonstrated in the Australian Danthonia caespitosa Gaudich. (Quinn & Hodgkinson). The re- sponse to density and to temperature varied along a latitudinal gradient. The infraspecific diversity of species from different regions tends to counter the suggestion of Clayton & Renvoize that the Arundinoideae are not successful and are outside the mainstream of grass evolution. Seed dormancy in the Australian species Danthonia carphoides F. Mueller ex Bentham and D. caespitosa was investigated by Hagon. Seeds sown within two weeks of ripening had 30 percent germination. After storage at room temperature (dry) for six weeks, all dormancy was broken. Germination re- sponses were not appreciably affected by temperatures in the 15—35°C range. REFERENCES: Under subfamily references see BENTHAM (1882, 1883); CLAYTON & RENVOIZE; CLE- WELL; CONERT (1961; 1975; 1987); ConNor (1987); DEWetT (1954, 1956); Dore & McNEILL; Hitcucock (1951); MUNz; TOMLINSON; TSVELEV; and ZoTov. 1990] TUCKER, ARUNDINOIDEAE 169 ABELE, K. Cytological studies in the genus Danthonia. Trans. Proc. Roy. Soc. S. Austral. 82: 163-173. 1959.* Austin, C. F. Danthonia DC. Bull. Torrey Bot. Club 3: 21, 22. 1872. [Four species.] Baum, B. R., & J. N. FINDLAy. Preliminary studies in the taxonomy of Danthonia in Canada. Canad. Jour. Bot. 51: 437-450. 1973. Biake, 8S. T. Plinthanthesis and Danthonia and a review of the Australian species of Leptochloa. Contr. Queensland Herb. 14: 1-19. 1972. Brock, R. D., & J. A. M. Brown. Cytotaxonomy of Australian Danthonia. Austral. Jour. Bot. 9: 62- 91. 1961. CAYOUETTE, J., & S. J. DARBYSHIRE. La répartition de Danthonia intermedia dans lest du Canada. Nat. Canad. I4: 217— 220. 1988. [Distribution map; specimen citations. ] Cray, K. in a natural population of the grass Danthonia spicata. Evolution 36: 734-741, The differential establishment of seedlings from chasmogamous and cleistog- mous flowers in natural populations of the grass Danthonia spicata. Oecologia 57: 183- 188. 1983a. . Variation in the degree of cleistogamy within and among species of the grass Danthonia. Am. Jour. Bot. 70: 835-843. 1983b. . The effect of the fungus Atkinsonella hypoxylon (Clavicipitaceae) on the repro- ductive system and demography of the grass Danthonia spicata. New Phytol. 98: 165-175. 1984. & J. ANTONovics. Demographic genetics of the grass Danthonia spicata: success of progeny from chasmogamous and cleistogamous flowers. Evolution 39: 205-210 1985a. Quantitative variation of progeny from chasmogamous and cleistog- amous flowers ; in the grass Danthonia spicata. Ibid. 335-348. 1985b. & ONES. Transmission of Atkinsonella hypoxylon (Clavicipitaceae) by cleistogamous seed of Danthonia spicata (Gramineae). Canad. Jour. Bot. 62: 2893- 2895, & : . A. Quinn. Density of stomata and their responses to a moisture gradient in Danthonia sericea populations from dry and wet habitats. Bull. Torrey Bot. Club 105: 45-49. 1978 ConerRT, H. J. The genus Danthonia in Africa. Mitt. Bot. Staatssamm. Miinchen 10: 299-308. 1971. r Danthonia domingensis Hackel & Pilger (Poaceae: Arundinoideae: Dan- thonieae). Senckenberg. Biol. 56: 293-313. 1975. DARBYSHIRE, S. J. The oldest Canadian specimen in a Canadian herbarium? PI. Press (Mississauga) 4: 107. 1986. [Danthonia spicata from Newfoundlan —— &]J. OUETTE. Biology of Canadian weeds. 92. Danthonia spicata (L.) Beauv. in Roem. & Schult. Canad. Jour. Pl. Sci. 69: 1217-1233. 1989. DeWET, J. M. J. Leaf anatomy and morphology in South African species of Danthonia. Bothalia 7: 303-310. 1960. FERNALD, M. L. Notes on Danthonia. Rhodora 45: 239-246. 1943. FInDiay, J. N., & B. R. BAUM. The nomenclatural implications of the taxonomy of Danthonia in Canada. Canad. Jour. Bot. 52: 1573-1582. 1974. [See also BAuM & FINDLAY. ] GRAY, J. R., J. A. Quinn, & D. E. FArRBROTHERS. Leaf epidermis morphology in popula ons of the Danthonia sericea complex. Bull. Torrey Bot. Club 96: 525-530. 1969 Hacon, M. W. Germination and dormancy of Themeda australis, Danthonia spp., Stipa bigeniculata, and Bothriochloa macra. Austral. Jour. Bot. 24: 319-327. 1976. Hopckinson, K. C., & J. A. Quinn. Adaptive variability in the growth of Danthonia caespitosa Gaud. populations at different temperatures. Austral. Jour. Bot. 24: 381- 396. 170 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 & Environmental and genetic control of reproduction in Danthonia caespitosa populations. Ibid. 26: 351-364. 1978. PHILIPSON, M. A re-assessment of the form of reproduction in Danthonia spicata (L.) Beauv. New Phytol. 103: 231-243. 1986. —— & M. C. Curistey. An epiphytic/endophytic fungal associate of Danthonia a transmitted through the embryo sac. Bot. Gaz. 146: 70-81. 1985. [Illustra- tions. Quinn, J. A., & D. E. FAIRBROTHERS. Habitat, ecology and chromosome numbers of natural populations of the Danthonia sericea complex. Am. Midl. Nat. 85: 531-536. 1971. & K. C. HopGkinson. Population variability in Danthonia caespitosa Sia ae) In response to increasing density under three different temperature regimes. Jour. Bot. 70: 1425-1431. 1983. Rotsettis, J., J. A. Quinn, & D. E. ere vier Growth and flowering of Danthonia sericea ae Ecology 53: 227-234. 1972. SCHEINER, S. & C. J. Goopnicut. The comparison of phenotypic plasticity and genetic ee in populations of the grass Danthonia spicata. Evolution 38: 845- 855. 1984. , J. Gurevitcu, & J. A. TEERI. A genetic analysis of the photosynthetic properties of populations of Danthonia spicata that have different growth responses to light level. Oecologia 64: 74-77. 1984. TeERI. Phenotypic flexibility and genetic adaptation along a gradient of secondary forest succession in the grass Danthonia spicata. Canad. Jour. Bot. 64: 739-747. 1986. Simpson, M. Value of the awn in establishing seed of Danthonia penicillata (Labill.) Palisot. New Zealand Jour. Sci. Tech. A. 1952: 360-364. 1953. Toots, V. K. Germination of the seed of poverty grass, Danthonia spicata. Jour. Am. Soc. Agron. 31: 954-965. 1939.* VickeRY, J. W. A revision of the Australian species of Danthonia DC. Contr. New South Wales Natl. Herb. 2: 249-325. 1956. Tribe ARISTIDEAE C. E. Hubbard in Bor, Grasses Burma Ceylon India Pakistan, 5. Aristida Linnaeus, Sp. Pl. 1: 82. 1753; Gen. Pl. ed. 5. 35. 1754. Caespitose annuals or perennials of dry soils. Roots fibrous; rhizomes short or lacking. Stems several to many, erect or slightly oblique, branched from the axils, especially in the lower portion, Leaves several, cauline and basal; sheath glabrous or pilose; ligule a short, pilose membrane; external ligule (sometimes present) a cartilaginous ridge interrupted by the midvein, bearing a row of trichomes; blade linear, about as long as the sheath [absent], involute or flat, the margin and surface scabridulous. Inflorescences paniculate, more or less open. Spikelets pedicellate, 1-flowered; glumes 2, strongly unequal to subequal, sometimes awned; lemmas lanceolate, shorter or longer than the glumes, in- volute [convolute], 3-nerved, calluses prominent, shortly hispid, the awn con- spicuous, scabridulous [plumose], with 2 lateral teeth whose bases are some- times fused to the lower portion of the awn forming the awn column; paleas much shorter than lemmas, 2-nerved. Lodicules 2. Stamens 3 or 1; anthers linear. Ovaries shortly cylindrical; styles 2, short; stigmas plumose, laterally exserted. Caryopses compressed or terete, tightly enclosed by lemma and palea, TUCKER, ARUNDINOIDEAE 171 Ficu Spikelets or their parts. a-d, Tribe Aristideae, Aristida longispica (sect. Aristida): a, spikelet with glumes spread apart, x 6; b, floret (emma and palea) containing mature caryopsis, the 3-awned lemma completely enclosing palea and caryopsis, x 6; c, caryopsis, x 6; d, palea, x 12. e-h, Tribe Stipeae, Piptochaetium avenaceum (Stipa avenacea): e, spikelet with single floret (only basal portion of lemma awn shown), x 2; f, entire floret, showing relative length of hygroscopic lemma awn, x 4; g, floret (note hairy base of lemma and rachilla forming bearded, sharp-pointed callus), lemma clasping pointed palea (only base of lemma awn shown), x 5; h, palea, x 5. i-k, Tribe Centotheceae, Chasmanthium latifolium (Uniola latifolia): i, spikelet, x 1%; j, floret (lemma and palea enclosing flower), x 3; k, caryopsis, x 5. T2 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 sometimes longitudinally furrowed. Base chromosome number 11. TYPE SPE- cies: A. adscensionis L., the only species included in the genus by Linnaeus. (Name from Latin arista, a beard or awn.)— THREE-AWN GRASS. A genus of about 330 species distributed nearly throughout the tropical and warm-temperate regions of the world. Twenty species occur in the Southeast. The genus is reasonably well known, thanks to the worldwide studies of Henrard (1926; 1927; 1928; 1929; 1932; 1933a, b). Asynopsis of the southeastern species (including those in Delaware, Maryland, Virginia, West Virginia, and Ken- tucky) has been presented by Allred (1986). His keys allow easy identification of the species in eastern North America. The subgeneric classification of Aristida is confused and difficult. There are probably parallels in the evolution of spikelet characters noted by Henrard (1926; 1927; 1928; 1929; 1932; 1933a, b). Section ARISTIDA (sect. Chaetaria (Beauv.) Trin.) (awns of the lemma persistent) accounts for about 200 species worldwide and 18 of the 20 species in the Southeast. The section was divided by Hitchcock (1924, 1951) into informal groups, a scheme also followed by Allred (1986) and seemingly prudent in view of the lack of understanding of the phylogeny of the genus. About one third of the southeastern species are annuals (group Dichotoma of Hitchcock, 1951), the remainder perennials. A common situation in the section is pairs of intergrading species; for example, A. basiramea Engelm. ex Vasey and A. dichotoma Michx. Polyploidy charac- terizes several species of the southwestern United States (DeLisle, 1973) of the A. purpurea Nutt. complex (group Purpureae of Hitchcock, 1951). Aristida Roemeriana Scheele, 2n = 22, includes only diploids; 4. /ongiseta Steudel, A. Fendleriana Steudel, and A. glauca (Nees) Walp., diploids and tetraploids (2n = 22, 44); A. Wrightii Nash, diploids, tetraploids, and hexaploids; and A. purpurea Nutt. (2n = 88), octaploids as well. In none of these species were morphological criteria found to distinguish the various autopolyploids (De- Lisle, 1969). The polymorphic and pantropic A. adscencionis L., 2n = 22, has numerous infraspecific categories. In the New World it has been reported from Texas westward to California and southward through Central and South Amer- ica. Section ARTHRATHERUM (Beauv.) Reichenb. (awn column deciduous at its base) is most diverse in Africa. There are only two species in the Southeast. Aristida desmantha Trin. & Rupr. (longer glume less than 2 cm long; awn column 2—5(—7) mm long) occurs in our area only in Arkansas and Louisiana. Aristida tuberculosa Nutt. (longer glume more than 2 cm long; awn column 8-— 15 mm long) is widespread in the eastern United States and in all of the southeastern states except Tennessee and Arkansas. Chromosome counts (e.g., A. brevisubulata Maire, 2n = 22, and A. pallida Steudel, 2n = 44) suggest the occurrence of polyploidy in African representatives of this section. The chiefly African sect. PSEUDARTHRATHERUM Chiov. 1s not represented in the New World. This section differs from the preceding in having the awn column deciduous at its summit (1.e., just below the level at which the lateral awns branch from the central one). Also not represented in the New World is sect. STREPTACHNE (R. Br.) Domin 1990] TUCKER, ARUNDINOIDEAE 173 (awn column present, not articulated). This section, recognized by Lazarides and by Bourreil & Reyre, was included without comment in sect. ARISTIDA by Clayton & Renvoize. One of its species, Aristida Humbertii Bourreil & Reyre, known from Angola, is unique in the genus in that all three awns are reduced to short mucros. REFERENCES: Under subfamily references see BENTHAM (1883); CLAYTON & RENVOIZE; CLEWELL; Connor (1987); DE WINTER; FOWLER; FOWLER & DUNLAP; HITCHCOCK (1951); and WUNDERLIN. ALLRED, K. W. Studies in the genus Aristida (Gramineae) of the southeastern United States. I. Spikelet variation in A. purpurescens, A. tenuispica, and A. virgata. Rhodora 86: 73-77. 1984a; II. Morphometric analysis of A. intermedia and A. longespica. Ibid. 87: 137-145. 1985a; III. Nomenclature and a taxonomic comparison of A lanosa and A. palustris. Ibid. 147-155. 1985b; IV. Key and conspectus. [bid. 88: 367-388. 1986. [Twenty species; synoptic treatment with key, distributions, illus- — of spikelets. ] rphologic variation and classification of the North American Aristida pur- purea eee (Gramineae). Brittonia 36: 382-395. 1984b. Boacpan, A., & A. SroRRAR. Control of Aristida and other annuals in Kenya Rift Valley pastures. Empire Jour. Exper. Agr. 22: a 1-223. 1954.* BourreIL, P. Réflexi ur Pécologie, la rphogénése et l’évolution, fondées sur la culture d’Aristida rhiniochloa, graminée ocr africaine. Adansonia II. 10: 409- 427. 1970. & A. Gestotr. Contribution a l’étude caryologique de diverses graminées afri- caines des genres Aristida L. et Stipagrostis nes mossou ns 11: 125-134. 1971. ,& H. GIL_eT. Contribution a Aristida rhiniochloa (Graminée d’aprés des spécimens d’ Noe boréale. cre II. 11: 685-690. en GILLET. Synthése des connaissances et des recherches nouvelles sur Aristida rhiniochloa, graminée africaine amphitropicale. Mitt. Bot. Staatssamm. Miinchen 309-340. 1971. & Y. Reyre. Un nouvel aristide de l’Angola de la section Streptachne. Adansonia II. 9: 421-427. [Includes illustrations of pollen.] OUIN. Contribution a l’étude caryologique de quelques aristides (Gra- eae Afrique boréale. Conséquences taxonomiques. Nat. Monspel. Bot. 21: 29- 6. 1970 DELISLE, D. G. Chromosome number and pollen size in the genus Aristida. Proc. lowa Acad. Sci. 86: 74-81. 1969. . Chromosome numbers in the Aristida purpurea complex. Southwest. Nat. 18: 79-83. 1973. HENRARD, J. T. A critical revision of the genus Aristida. I. Meded. Rijks-Herb. Leiden 54: I- VIII, 1-220. 1926; II. Ibid. 54A: 221-464. 1927; II. Ibid. 54B: 465-701. 1928; aaa Ibid. 54C: 703-747. 1933a. nograph of the genus Aristida. I. Ibid. 58: 1-153, io 1-60. 1929; II. Ibid. 58 A: 57. 325, pls. 61-169. 1932; Index. Ibid. 58B: I-XII. Hitcucock, A. §. The North American species of Aristida. One U. S. Natl. Herb. 22: 517-586. oe LAZARIDES, M. tida L. (Poaceae, Aristideae) in Australia. Brunonia 3: 271- 1980. Siar seven species; discussion of problems in the genus; keys, descriptions, representative illustrations. } LONGHI- WAGNER, H. M. Uma nova éspecie de Aristida L. (Gramineae) do Brasil. Bradea 5: 59-62. 1988. [A. pendula; illustrated.] 174 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 MATTHEI J., O. The species of Aristida L. (Poaceae) in Chile. Gayana Bot. 44: 17-23. 1987. [Three species; illustrated. ScHoz, H., & P. KOnic. Eine neue Aristida (Gramineae) aus Arabien. Willdenowia 17: 111-113. 1988. pe coae ] Warnock, B. H. A new three-awn grass from Trans-Pecos, Texas. Sida 9: 358, 359. 1982. ‘Aristida Brownii; jie d. Tribe STIPEAE Dumortier, Obs. Gram. Belg. 83. 1824. 6. Nassella (Trinius) Desvaux in C. Gay, Hist. Chile Bot. 6: 263. 1854. Caespitose perennials. Stems hispid or glabrous; nodes appressed-hispid, becoming glabrous with age (lower nodes usually geniculate). Leaves several to many, mostly basal; sheath pubescent or glabrous; ligule an unfringed mem- brane (abaxial ligule absent); blade setaceous to linear, more or less involute, hispid abaxially, less so adaxially. Inflorescences solitary, terminal, paniculate; branches flexuous, scabrid, sometimes hairy. Spikelets 1-flowered, slightly lat- erally compressed, disarticulation above glumes. Glumes 2, equal, broadly lanceolate, acute [awned], 3-(to 5-)ynerved. Lemmas elliptic, convolute, becom- ing indurate, hispid basally, the surface densely covered with short, stout prick- les (seemingly tuberculate), contracted into a smooth, basally fringed collar at the base of the awn, the awn 5-10 times longer than lemma, geniculate, scabrid, hairless. Paleas elliptic, much shorter than lemmas (or absent), nerveless, hya- line. Lodicules 2, glabrous, fleshy. Stamens 3; anthers oblong. Ovaries oblong, glabrous; styles free, stigmas 2. Caryopses oblong to pyriform, laterally com- pressed, smooth; hilum linear; embryo large, with epiblast. Base chromosome number 14. (Urachne sect. Nassella Trin. Mém. Acad. Sci. St. Pétersb. Sci. Nat. VI. 1: 73. 1830.) Lecrorype species: N. pungens Desv., designated by Parodi (1947). (Origin of name unknown.) A genus of about 50 species (Barkworth & Everett), in an area from Patagonia north through the Andes into the Caribbean region, Mexico, and the United States, and western Canada. The only southeastern representative, Nassella leucotricha (Trin. & Rupr.) Pohl (Stipa leucotricha Trin. & Rupr.), Texas win- tergrass, 2n = 28 (Gould), occurs sporadically in western Louisiana and south- western Arkansas and ranges southward through Texas into South America. It is abundant in Texas, where it is important as a cool-season forage grass. The caryopses of Nassella leucotricha have prolonged, erratic germination. They probably germinate more or less continuously throughout the year, except during lengthy dry spells (Fowler). On the basis of controlled-environment experiments, however, Call & Spoonts characterized the optimal parameters for germination and hypothesized that in central Texas most germination would occur from late September through mid-November and, during mild winters, also from December through February. REFERENCES: Under subfamily references see BARKWORTH & EVERETT; FOWLER; FOWLER & DUNLAP; and GouLp. Brown, W. V. A cytological study of cleistogamous Stipa leucotricha. Madrono 10: 97- 107. 1949, 1990] TUCKER, ARUNDINOIDEAE 175 — ——. The relation of soil moisture to cleistogamy in Stipa leucotricha. Bot. Gaz. 113: 1952. CALL, C. A., & B. O. Spoonts. Characterization and germination of chasmogamous and basal axillary See aoe florets of Texas wintergrass. Jour. Range Managem. 42: 51-55. 1989. DyYKSTERHUIS, E. J. Axillary cleistogenes in Stipa leucotricha and their réle in nature. Ecology 26: 195-199. 1945. Paropl, L. R. Las especies de gramineas del género Nassel/a de la Argentina y Chile. Darwiniana 7: 369-394. 1947. [Nine species. ] 7. Piptochaetium J. Presl in K. Presl, Reliq. Haenk. 1: 222. p/. 37. 1830. Caespitose perennials of dry, open places. Stems branched only near the base, glabrous; nodes glabrous, constricted medially. Leaves several per stem; sheath ribbed, glabrous; ligule longer than wide, membranaceous (abaxial ligule ab- sent); blade linear, flat to involute, midvein prominent, surface and margin scabridulous. Inflorescences solitary, terminal, open, paniculate; branches sca- bridulous. Glumes 2, equal, lanceolate-acuminate, conspicuously 3- to 5-veined, margin hyaline. Spikelets solitary, 1-flowered, terete. Calluses subulate [cu- neate], covered with long, stiff, extrorse hairs. Lemmas narrowly oblong [el- liptic], 3-nerved (nerves visible from adaxial surface), coriaceous, often tuber- culate above, margin involute, fitting into the sulcus of the palea, apex thickened, scabridulous [spiny]; awn several times longer than lemma body, spiraled (hy- groscopic), scabridulous. Paleas lanceolate, slightly longer than the lemmas, sulcate medially, glabrous, the apex protruding between the lemma margins. Lodicules 3. Stamens 3; anthers narrowly ellipsoid. Ovaries cylindrical; stigmas 2, plumose. Caryopses slenderly cylindrical, firmly enclosed by palea and lem- ma. Base chromosome number 11. Type species: P. setifolium J. Presl, the only species. (Name from Greek piptein, to fall, and chaete, bristle, referring to the deciduous lemma awns of the type species.) — NEEDLEGRASS. A genus of about 30 species. The circumscription follows Barkworth & Ev- erett, and Parodi. Piptochaetium is distinguished from other genera of the Stipeae by its grooved palea into which the margins of the involute lemma fit (Barkworth & Everett; Parodi). It is characterized by phloem fibers in the leaf blades, colorless cells between the abaxial epidermis and the bulliform cells, and circular to paradermally rounded vascular bundles (Parodi & Freier). Only two species are found in the Southeast. Piptochaetium avenaceum (L.) Parodi (Stipa avenacea L.), 2n = 22, 28, occurs from Massachusetts to south- western Ontario (Dore & McNeill) and Michigan, south to northern Florida and eastern Texas. It is known from all the Southeastern States, where it grows in oak or pine woods, along roadsides, and in fields Piptochaetium avenacioides (Nash) Valencia & Costas (Stipa avenacioides Nash) is endemic to central Florida (Hall; Hitchcock; Wunderlin). It has longer awns (6-10 cm, vs. 4-6) and longer lemmas (12-18 mm, vs. 8-10) than P. avenaceum (Hall). The two species are otherwise similar and perhaps closely related, but because of considerable parallelism in the Stipeae (Barkworth & Everett), this is perhaps a premature statement. The range of P. avenacioides, although restricted, does not appear to overlap that of P. avenaceum. 176 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 REFERENCES: Under subfamily references see BARKWORTH & EVERETT; DorE & MCNEILL; HALL; Hircucock (1951); PARopI; and WUNDERLIN Paropt, L. R., & F. Frerer. Observaciones taxondmicas sobre las gramineas estipeas. Ciencia Invest. 1: 144-146. 1945. VALENCIA, J. I, & M. Costas. Estudios citotaxonémicos sobre Piptochaetium (Gra- mineae). Bol. Soc. Argent. Bot. 12: 167-179. 1968. Tribe CENTOTHECEAE Ridley, Mater. Fl. Malay Penin. 3: 122. 1907. 8. Chasmanthium Link, Hort. Berol. 1: 159. 1827. Single-stemmed or loosely clustered perennials of open woods. Roots fibrous; rhizomes short, approximately horizontal. Stems unbranched or sparingly branched. Leaves cauline, several; sheath glabrous or pubescent; ligule a short hyaline membrane or fringe of hairs; blade linear to linear-lanceolate, scabridu- lous on veins, otherwise glabrous or sparsely pilose. Inflorescences terminal and/or axillary, open or contracted. Spikelets oblong, laterally flattened, sessile or pedicellate, 2- to 12-flowered; glumes 2, equal or subequal, deltoid to lan- ceolate, acute to acuminate, conduplicate, 3- to 7-nerved, the midvein often scabridulous; lemmas narrowly oblong, acuminate, acute, or bifid, 5- to 15- nerved, the midvein scabridulous or ciliate; paleas about as long as the lemmas, bicarinate, bulged out near the base, the keel scabridulous. Lodicules 2, lobed- truncate, 2- to 4-nerved. Stamens solitary; anther broadly oblong to linear. Ovaries oblong; styles short; stigmas 2, plumose. Caryopses ovoid to ellipsoid, 12. (Excluding Gouldochloa Valdés, Morden, & Hatch.) Type species: C. gracile Link = C. laxum (L.) Yates, the only species included by Link. (Name from Greek chasma, open or gaping, and anthos, flower.) A genus of five species, all endemics of eastern North America and all occurring in the Southeast. Chasmanthium was long included in Uniola (Chlori- doideac). Detailed studies by Yates established the affinities of Chasmanthium with the Centotheceae and the heterogeneity of Unio/a as circumscribed by Hitchcock and other agrostologists. All species of Chasmanthium are self- compatible (Yates Chasmanthium latifolium (Michx.) Yates (large spikelets pedunculate, pen- dent), 2 = 48, is the most distinctive species and presumably the least closely related to the others. It is found from New Jersey to Kansas south to the Florida Panhandle and south-central Texas, with outlying populations in central Ne- braska and the mountains of northeastern Mexico.*® It occurs in all of the ‘Reported from Manitoba by Yates as follows: Canada, Manitoba, Cow Creek Bridge, sales Gruver, 13 July 1936 (Ncu). I have examined the specimen, an it is C. latifolium as annotated b Yates. However, it is unlikely this specimen came from Manitoba. The printed heading af ihe abel reads ‘Herb. of Margaret G. Dudley/Winnipeg, Canada.” pie is no locality called Stillwater in Manitoba. There is a Cow Creek Bridge in Stillwater, Payne County, Oklahoma, well within the range of C. latifolium. Darwin Gruver is unfamiliar as a collector of Canadian plants to the Canadian botanists with whom I have corresponded. Thus I believe the oe in question came from Oklahoma, and C. oe should be excluded from Canada’s flor. 1990] TUCKER, ARUNDINOIDEAE we) Southeastern States. Plants produce both cleistogamous and chasmogamous florets. Chasmanthium latifolium is tetraploid, while the other species are dip- loid. e four remaining species have more or less sessile spikelets and narrow, spikelike inflorescences. The closely related Chasmanthium sessiliflorum (Poir- et) Yates, 2n = 24 (southeastern Virginia to Missouri and Oklahoma, south to central Florida and eastern Texas), and C. /axum (L.) Yates, 2n = 24 (south- eastern New York to southeastern Virginia, Kentucky, and southeastern Mis- souri to central Florida and eastern Texas), are distinguished by their tiny (S5— 10 mm long) spikelets and purple anthers. In both species the lemma and palea spread, exposing the caryopsis at maturity; this distinguishes this species-pair from the other species of the genus, in which the caryopsis remains covered. Chasmanthium laxum differs from C. sessiliflorum only in having glabrous rather than pubescent leaf sheaths and collar Large sessile spikelets characterize the two remaining species, both of which have more restricted ranges than the preceding species: Chasmanthium orni- thorhynchum (Steudel) Yates, 2n = 24, occurs on the Gulf Coastal Plain of western Florida, southern Alabama, and southern Mississipp1, to southeastern Louisiana. Chasmanthium nitidum (Baldwin ex Ell.) Yates, 2” = 24, grows on the Atlantic and Gulf Costal Plain of South Carolina, southern Georgia, and central and western Florida. In both the spikelets are 7-18 mm long, and the aryopses are covered completely by the palea and lemma. Chasmanthium nitidum has glabrous inflorescence axils and only one sterile lemma per spikelet, while C. ornithorhynchum has densely pilose axils and two to four sterile lemmas. The hybrid Chasmanthium laxum x C. ornithorhynchum is known from the Gulf Coast region of southern Mississippi (Yates). Hybrid populations are associated with both parental species and have enlarged sterile spikelets. REFERENCES: Under subfamily coe see CLAYTON & RENVOIZE; CLEWELL; HALL; HITCHCOCK (1951); WUNDERLIN; and YATES. STEVENS & CULLEN, LINNAEUS 179 LINNAEUS, THE CORTEX-MEDULLA THEORY, AND THE KEY TO HIS UNDERSTANDING OF PLANT FORM AND NATURAL RELATIONSHIPS P. F. STEVENS! AND S. P. CULLEN? Linnaeus’s ideas on the composition of the various parts of the plant and the generation of plant form by the interactions of these tissues both with each other and with the sap of the plant are outlined. Linnaeus used the analogy he drew between the vegetative plant and the larva of an insect to justify his emphasis on parts of the fructification in the classification of genera; as with insects, he used the adult, rather than the larva, in classification. His thoughts on the relationship between the various appendicular parts of the plant in the context of his theories of prolepsis and metamorphosis are summarized. Lin- naeus’s ideas of plant form support his theories on the generation of plant ranged form that also incorporated findings from the microscopic world. Here, as elsewhere, the medulla can be shown to be the seat of life, but only its interaction with cortex allowed stable form to become manifest. Goethe’s and A. Candolle’s notions of the “‘metamorphosis”’ of plants are briefly situated with regard to Linnaeus’s ideas, as is the work of some earlier botanists. In the discussion the interrelationship is explored between Linnaeus’s different approaches to looking at plants and their significance for an understanding both of his taxonomic work and of his more general thinking about the diversity of life. The ideas of Linnaeus, and especially those of Goethe, on the meta- morphosis of plant form are situated at one end of a spectrum of responses to wn taxonomic system represents the opposite end. The archaic cast to Linnaeus’s cortex- medulla theory is confirmed, although its coherence and explanatory powers are abundantly evident. Most commentaries on Linnaeus’s biological work have focused on his idea that the taxonomically important characters of an organism reflect, or are, that organism’s essence, and especially on his arrangement of organisms in a clas- sification using these characters. Any ideas that Linnaeus may have had as to how the form of an organism becomes manifest, and how the diversity of form in the living work is generated and organized, are less frequently discussed. It is usually suggested (see, for example, Stearn, 1957, and Gustafsson, 1985) that Linnaeus was perhaps more interested in simply describing the world than in understanding laws or principles underlying the diversity of form present in ‘Harvard University Herbaria, 22 Divinity Avenue, Cambridge, Massachusetts 02138. 21350 Old Trail, Maumee, Ohio 43537. © President and Fellows of Harvard College, 1990. Journal of the Arnold Arboretum 71: 179-220. April, 1990. 180 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 it. However, it is becoming abundantly evident that this is not an accurate perception of his activities, as perusal of the numerous theses written by Lin- naeus and defended by his students shows. Largely using these theses, Guédés (1969) outlined the development of Linnaeus’s ideas on plant morphology and development; Larson (1971) emphasized the Aristotelian dimensions of Lin- naeus’s thought and discussed his ideas on plant hybridization in some detail; and Broberg (1985), ina remarkable article, succeeded in pulling together much of the underlying body of theory in these theses and showed clearly the im- portant synthetic component in Linnaeus’s writings. One goal of this paper is to bring together these different parts of the story, emphasizing its botanical dimension in particular. We also hope to deepen the appreciation of Linnaeus’s important efforts to develop an understanding of organic form and diversity, at the same time stressing the relationship of his thought to that of his im- mediate predecessors and successors. Larson (1971) clearly outlined what is, as he noted, largely implicit in the work that embodies Linnaeus’s taxonomic principles, the Philosophia Botanica (Linnaeus, 1751a): all genera are natural—the work of nature—and form the foundation of theoretical botany. Such genera were to be recognized by dis- tinctive and constant features in the various parts of the fructification, these being the most essential parts of the plant. To quote Larson (1971, p. 93): “The ‘naturalness’ of Linnaean genera rests, then, upon assumptions about the prin- ciples of activity for the performance of which plants have come into being. The nature of the genus has a narrower sense than ‘reality’: it is the formative factor in reality.” Cain (1958, pp. 154, 155, quoting Maritain, 1983) suggested that “nature” or “essence,” although originally in Aristotelian thought referring to those principles of the activity for which an organism came into being—the final cause—might also be restricted simply to ‘‘what a thing is,” its visible char- acters. To quote Larson (1971, p. 93) again, ““The simple elements of the fructification, when isolated, and given explicit form, teach the naturalist the characters of the genera spelled out by the hand of God.” But between God (a cosmic final cause) and explicit form (formal cause), there are other levels of causality. There is the organism with its role in the community (see, for example, Linnaeus, 1749, 1760b),> a local final cause. There are also the material and efficient causes of explicit form, which cause any particular form to be what it is, and it is with these levels of understanding of form in Linnaeus’s botanical thought that we are mainly concerned here. Although the focus of this paper is not Aristotelian causality in Linnaean thought, it should be remembered that Linnaeus’s biological thought as a whole has a decidedly Aristotelian back- ground (Larson, 1971). This leads to an underemphasized aspect of Linnaeus’s work that he devel- oped most actively after about 1749. He suggested that the “formative factor” >Following authors such as Stearn (1966), authorship of the Linnaean theses is to be ascribed to Linnaeus himself, not to the defendant whose name is on the title page. One of the very few cases where the actual defender of the thesis had some hand in its writing is the important Gemmae Arborum (see also below). This thesis alone is cited under its defender, Léfling. 1990] STEVENS & CULLEN, LINNAEUS 181 ofa plant is the medulla: without the medulla, or when the medulla is depleted, the plant dies; through the medulla, the unchanging essence of the plant is transmitted through successive generations. Interaction of the medulla with the cortex leads to the development of all the parts of the plant from cotyledons to pistils; in the flower, in particular, the medulla interacts successively with the different tissues derived from the cortex, each interaction producing a different part of the flower. Although all these parts other than the pistil come entirely from the cortex, the pistil being largely medulla, they cannot develop without the stimulation of the medulla. The parts of the fructification, which include the flower, are of course the basis of most Linnaean genera. Ideas such as these are discussed at length in a series of theses, especially Gemmae Arborum (L6fling, 1749), Metamorphoses Plantarum (Linnaeus, 1755a), Prolepsis Plantarum (1760c, 1763), Fundamentum Fructificationis (1762b), and Mundum Invisibilem (1767a). These theses are often extended justifications for positions Linnaeus only outlined in better-known works such as the Philosophia Botanica (1751a) and the tenth and subsequent editions of the Systema Naturae (e.g., 1759b, 1767b). They owe much to classic notions of the plant body, 1n particular those of Cesalpino, whom Linnaeus frequently acknowledged, and through Cesalpino, Theophrastus and Aristotle. They are also intimately linked with complex analyses of the relationships between plant parts, which are little connected with Linnaeus’s use of plant structure in his systematic studies. In these more purely morphological analyses, Linnaeus suggested that there was a fundamental similarity (and often interchangeability) between cotyle- dons, leaves, sepals, stamens, and the like —in fact, between all the appendicular parts of the plant (this term is used without any implication as to what these structures “really” are) except the carpels, and that exception, as we shall see, holds only in some respects. The observations that Linnaeus emphasized in establishing these similarities were very different from those he used 1n grouping species to form genera, and genera to form families or ordines naturales, as in the fragments of his natural system. He focused on nonessential variation, adopted what might be called a physiological-balance theory of plant repro- duction and morphogenesis, and emphasized a parallel between vegetative and floral buds, and, more generally, a whole variety of “budlike”’ structures in the plant, including the embryo in the seed. Both his systematic and his morpho- logical studies led to an emphasis on continuity of form, but the forms em- ohasized were different.4 Focusing on these aspects of Linnaeus’s thought allows us to understand his later work more clearly and perhaps puts his so-called failure or inability to develop more than the bare outlines of a natural system, the development of which he clearly considered to be very desirable, in a somewhat different light. In addition to classifying all the new material that was being sent to him, ‘Johann Wolfgang Goethe and Augustin-Pyramus de Candolle are usually considered two of the most important early proponents of ideas ee the fundamental similarity of all plant parts, although the former had little immediate influence in systematic botany (see Guédés, 1969; Cusset, 1982). C. F. Wolff outlined similar os in 1766 (Mueller, 1952). 182 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 Linnaeus was developing ideas very different from those we commonly attrib- ute to him. Gustafsson (1985, p. 126) noted, “‘For Linnaeus the crucial aspect of his work (and the only way in which he could, under the assumptions and axioms of his time, proceed) consisted 1n observing differences in nature at the cost of the similarities. That - he stressed the hierarchic aspect and was forced to ignore continuity.”’ He went on to discuss the attempt to depict the chain of being as a specially aie analytical language. Even if, following Gus- tafsson, such an attempt is deemed in principle impossible, we see Linnaeus effectively dealing with continuity 1n his morphological work, and he of course adumbrated the problem of continuity in the context of his natural system (e.g., Linnaeus, |751a; see especially Broberg, 1985). Between Linnaeus’s more or less analytical system and his belief that God’s handiwork was visible in the natural world lies a complex and incomplete set of theory and observation that explains at an intermediate level organic form in the botanical world. It is centered on the ancient idea that the plant body is made up of cortex and medulla, and it emphasizes continuity. Here Linnaeus’s “passion for synthesis” (Broberg, 1985, p. 179) finds its full expression Below we outline the different aspects of this “passion for synthesis,” treating each more or less separately. These aspects are cross-referenced to indicate the interrelationship of the several strands of Linnaeus’s thought that together form a skein of considerable intricacy. THE CORTEX-MEDULLA THEORY OF PLANT CONSTRUCTION INTRODUCTION The cortex-medulla theory of plant construction, in which all parts of the plant are equated with tissues derived from the cortex, or with medullary tissue, first figured prominently in Linnaeus’s work in the thesis Gemmae Arborum, defended by Pehr L6fling in 1749. Although mentioned in earlier works, 1t was there always in a much less central position. L6fling’s thesis coupled the classical conception of the flower in which the outer, vegetative cortex splits to reveal the inner tissues of the plant with an extensive, but sketchy, elaboration of the cortex-medulla theory. L6fling drew an analogy between the vegetative parts of the plant and the larva of an insect; he looked at budlike structures in general and considered them to be directly comparable; and he examined in some detail the growth of floral and vegetative parts of the plant, entertaining the notion that at least some flowers were precociously developed shoots. This leads to the idea that the parts of the plant other than the stem and root (i.e., all the major organs borne on the shoot) are fundamentally equivalent. It is interesting to note that this thesis is one of the very few in which the student defending it is supposed to have had an appreciable part in its writing.° ‘This student, Pehr L6fling (1729-1756), also helped to write the Philosophia Botanica (Linnaeus, 175la) at Linnaeus’s dictation, Linnaeus being unwell at the time. There are further developments in the ideas mentioned above in this book. During dictation, stig questioned things he could not understand, so becoming throughly grounded in Linnaean botany (Blunt, 1971; see Léfling, 1758, p. [3] of the Férera/ written by Linnaeus). One of Linnaeus’s best and favorite students, Léfling unfor- tunately died young while collecting in South America 1990] STEVENS & CULLEN, LINNAEUS 183 PROBLEMA BOTANICUM in gratiam D:n1 ae a ee proees tum PR ASIDE, . Orpus VEGETABILE conftat Medulla, veltita Ligne, fatto ex I. CO Libero, Daclebe feceflit ab interior’ fubftantia Corticés , qui iple Epsdermide induitur 2 4 cum tegumentis crefcit fefe in (onginuenen przpri- mis extendendo, fibrasque in Jatitudinem verf{us a protrudendo, 3. Fibre medullaris extremitas per ¢ coniean, prot, folvitur fe~ pius in Gemmam ex foliolis imbricatam 4. Folssum expanditur attrahendo faceum hutritiuim, quem mo- dicum Sone concedit, antequam cadit, nunquam renafciturum, ~ 5. a (3) compendium, future herba extenditur in ramum hic in ete inque eee » donec fructificatio imponat uiitatns terminum antique ee oni. 6. €a4yx fit in Demat ex foliis ‘non fecedentibus 5 an te vi eadenr neers expandenteque3 probant hoc Nr Rofa, apenfia, Persanthtum commune, Inwvolucrnm, & ¢x §: Ale ee intra, rola ~ 7. Hec facto m umpitur | intra calycem ramuli apex, fecundum leges cuigue fabltgutia (1) prepriass inque Florem expanditur anauo {patio Eee S. 8. Cerollam teneram, magis mollem & caducam oriti ex corticis, nunc Calycis (6), propria fub(tantia, pulpofo libero, confrmant Féres ruernales, prxprimis Daphnes, co tempore, quo Liberi fubftantia a eortice non recellit. tamina fieri ex fubftantia lignea (), olim libero , probat confiftentia fetus, plenstedo florum e vegetatione Nis aise abi, ex — fo) ertum eft, cum alia hoc in loco fee nulla, Pas In. Fradles ex pititlo medullari nequit vitam ‘nove plante in- choare, “nif prius ftantinum eflentia lignea abforpta fuerit ab humore eet piftilli, 2, Quz see: aS connexionis foliorum in Calycem ? xcocius rumpatur ift Hfouem,t Que vis mraviiie hujus effectus? Nodum extremum vegetationis videtur ’ i ‘fol viffe, gui hunc explicer Gordium. Ficure 1. Linnaeus’s questions in the thesis Gemmae Arborum (Lofling, 1749, facing p. [i)). 184 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 There is, however, a further complication that makes the authorship of the ideas in this thesis uncertain. Facing the first page is a series of 11 aphoristic statements concerning plant construction and development and a final series of questions addressing the forces involved in the phenomenon of flowering. This page, headed “Problems in botany proposed to the glory of God by the president [Linnaeus] of the author [LGfling],” is placed at the end of the re- printed versions of the thesis (e.g., Linnaeus, 1751b, 1786) and has largely been overlooked (but cf. Stafleu, 1971). A reproduction of this crucial page is given here (see FiGureE 1), and a free translation follows. 1. The VEGETABLE body is composed of the Medulla, covered by Lignum ees made from Liber [phloem and cambium], after it cas s from the inner substanc of the Cortex, which itself is overlain by the Epidermi 2. The Medulla with coverings he by itself, Sie mainly in length and pushing out fibers to the side toward the leave 3. The extremity of the medullary Shee protrudes — the cortex, often being broken up in the imbricate Bud [made up] of sma . The expanded Leaf, attracting nutritive sap, She oe it falls, allows a mod- icum to the bud, never being borne again 5. The Bud (3), the compendium oP the future vegetative plant being extended into a branch, first as a twig, then to infinity, up to the time when the fructification imposes the final termination of the growth of the old plant. The Calyx being within the bud, [made up] of leaves that do not separate, lacking the ne to separate and expand; Nigella, Rosa, els common Perianth, Involucre, and Salicis amento insectifera rosa demonstra the apex of the branch is Hea ae within the calyx, the substances (1) [the eG cof the plant] following laws that are proper, I say the Flower expanding a year eva sly. 8. n corolla, very soft and caducous arising from the cortex, now [part] of the Calyx (6), a true substance, fleshy liber, spring Flowers, especially eee confirm at that time when the substance of the Liber does not separate from the . The Stamens being derived from the woody substance (1), once liber, a axisienee position, the completeness of the flowers from more vigorous plant show, where the hardened lignum did not allow stronger propulsion from the soft liber 10. The Pistil of the center of the flower, [made] from its special substance of the medulla, with nothing else remaining in that place. The Fruit ea the pistil [made up] of the medulla is unable to lay the foundations for the life of the new plant, unless first the woody essence of the stamen is absorbed by the medullary humor of the pistil. 12. What then [is] the cause of the joining ae ak in the Calyx? by what is the precocious apex of the plant torn apart in the flow What miraculous force does this? It appears to me that the ultimate node of the plant is dissolved, as was loosened the Gordian [knot]. — The contents of this introductory page outline many of the points that Lin- naeus developed over the next fifteen years. It is interesting, and perhaps relevant in attempts to understand authorship of the various parts of this thesis, to note that the emphasis on those forces in the plant that might account for the morphological phenomena observed is not so evident in the body of Léf- ling’s thesis. This, however, is clearly seen to be the central point at issue, as the emphasis on the final statements, possibly written by Linnaeus himself, suggests. 1990] STEVENS & CULLEN, LINNAEUS 185 THE Basic CONSTRUCTION AND DEVELOPMENT OF THE PLANT Bopy Linnaeus (e.g., 1751a) consistently divided the plant body into three basic parts: the radix, or root; the herba, or vegetative region of the plant, sometimes simply called “‘planta” (e.g., Linnaeus, 1741, p. 12); and the fructificatio, every- thing from the calyx to the seed.*° The root, which took up food from the soil, produced the vegetative part of the plant as well as the fructification. The plant body as a whole was made up of a central medulla, roughly corresponding to the pith, covered by the lignum (wood). This latter arose from the liber (inner bark: phloem and cambium), which was in turn derived from the cortex, itself overlain by the epidermis (see TABLE 1). These several tissue types could be reduced to two, the cortex and the medulla (Linnaeus, 1759a, p. 4; the ‘““corporea externa”’ and the “medulla interna” in Linnaeus, 1767b, p. 7), by emphasizing that the lignum and the liber were both derived from the cortex. In Linnaeus’s taxonomic analyses the fructification was divided into two parts, the flower and the fruit, the former being made up of calyx, corolla, stamens, and pistil, and the latter of pericarp, seed, and receptacle (e.g., Lin- naeus, |175la). Linnaeus had very early integrated some of the organs of the fructification with the cortex-medulla division of plant tissue (see TABLE 1). He thought that the thicker outer cortex formed the calyx, and the thinner inner cortex, the petals; the stamens came from the wood, and the pistil from the medulla (Linnaeus, 1738; see FiGurE 2A). Thus the inner parts of the plant were displayed by the splitting of the cortex in the flower (see, for example, SInitially the ‘planta’? was thought of as being composed of the trunk, the leaves, the “fulcra” (stipules, prickles, and the like), and the fructification (Linnaeus, 1736a, p. 7). TABLE 1. The main organ categories of the plant and their origin. TISSUE TYPES MAIN PLANT PARTS Normal Main categories Organs arrangement Linnaeus, 1746+ Radix Root Root Cortex + medulla — Herba (planta) Vegetative Stem Cortex + medulla Cortex + medulla plant Leaf Cortex rtex Thorns, etc Presumably cortex Presumably cortex (“fulcra’”) Fructificatio Flower ‘aly Cortex Outer cortex Corolla* Liber Inner cortex Stamens* Lignum Alburnum Pistil* Medulla — Fruit Pericarp _ ign ed* Medullat Medullat Receptacle* *The seven parts of the fructification. tThis is closer to Cesalpinian ideas as to the equivalence of organ categories with tissue types, see below The cortex became involved in the production of the seed via the process of fertilization. 186 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 3 4 Figure 2. Linnaeus’s ideas on the construction of the plant body: A, flower vegetative bud (stipple = cortex, black = liber, white = wood, ascending diagonal l lines = membranes surrounding medulla, descending diagonal lines = medulla. | = first year’s growth, 2 = second year’s growth, and so on; c = leaf on current year’s growth). In flower all growth occurs in a single year (anticipation), vegetative bud it occurs over several years; for diagrammatic purposes, a single leaf represents each year’s growth. Lofling, 1749, p. 13, footnote s; Linnaeus, e.g., 1755b, p. 5). However, the exact details of this oft-repeated comparison differed. ee equating the various organs of the flower with those of the vertebrate reproductive system, Linnaeus (1746) briefly mentioned that the calyx was made up of outer cortex, the corolla from inner cortex, the stamens from nutritive alburnum (sapwood, correspond- ing to the liber), the pericarp from lhgneous substance, and the seed from medulla. Here both parts of the fructification, flower and fruit, derive from the several parts of the basic tissue system of the plant, but usually 1t was the flower alone that was compared with these tissues (see FiGURE 2A; LOfling, 1749, cf. facing p. [1] and p. 13; Linnaeus, 1751a, 1751c (the medulla produces the seed), 1755a, 1759b). If bracts were included in this comparison, they, too, were considered to be formed from the cortex (Linnaeus, 1763). In the vegetative plant the ends of the medullary fibers protruded through 1990] STEVENS & CULLEN, LINNAEUS 187 the cortex, liber, and wood and stimulated the formation of leaves from the cortex (see FIGURE 2B). The young leaves formed imbricate buds with medulla in the center; the buds themselves did not develop from leaves. Thus the vegetative region of the plant was effectively an assemblage of buds, a com- posite, or—in modern terminology—a modular organism, that could repeat itself ad infinitum unless terminated by the fructification (L6fling, 1749) or the death of the medulla. The medulla was both the form-generating part of the plant and the part that ensured the plant’s continued growth; all parts of the plant with the potential for growth had medulla. Linnaeus (1763) expanded this point in the thesis Prolepsi Plantarum. Each of the numerous buds on a tree contained medulla, which, growing out and dividing incessantly, continued the life of the plant, the ultimate fibers of the medulla being in the next generation of buds. Thus, only a few years after a willow twig 1s put in the ground, the subdivision of the medulla in the buds of that original twig results in a large, copiously branched tree from which twigs can be plucked and planted; the whole process can then repeat itself. In an annual plant, on the other hand, Linnaeus thought that there were no buds, only flowers in the axils of the leaves. This meant that the medulla in each axil was used up in the production of the flowers; the plant then died since it had no medulla left (Linnaeus, 1760c). Hence the only difference between annuals and perennials was that in the former the medulla was quickly used up in the fructification, while in the latter it was for the most part retained in the trunk or root and could produce more buds. Linnaeus’s concept of plant growth contained two antagonistic tendencies or movements involved in growth and development. A movement from the outside of the plant that was manifest in the direction of development of the cortex-derived tissues was nutritive and descending, while that from the inside to the outside was ascending and generative: “the life of the medulla, spongy, divisible by multiplication and with an infinity of endings, growing upwards, incarcerated in the Cortical Body, which is nutritive, descending, joined to the ground ...” (Linnaeus, 1759b, pp. 826, 827; there are extended comparisons with animals in several places—e.g., Linnaeus, 1751a, 1759a). The cortex could be compared to the vascular system; the mouth, or ventricle, of the plant was in the ground, from which the roots drew up nourishment that was purified in the leaves. As elsewhere in this theoretical edifice, there are problems in the interpretation: the force in the cortex is descending, but food is taken (ingested) from the ground. In any case these forces do not exactly correspond to other visualizations of the plant: the three main morphological subdivisions of the plant body are root, vegetative plant, and fructification (see TABLE 1). In the seed, which was the rudiment of the new plant, the plumule was the scaly, ascending part of the corculum (the plumule-radicle junction and, to many classical authors, the “heart” of the plant). When bathed in sap (humore), the bud could grow ad infinitum. The rostellum was the simple, descending part (Linnaeus, 175la). The above-ground plant was produced from the root. Nothing new was formed; instead, there was simple continuity through the medulla (“Nova creatio nulla, sed continuata generatio, cum Corculum seminis constat parte radicis medullari’’: ‘bid., p. 38). As will be seen, the balance of 188 JOURNAL OF THE ARNOLD ARBORETUM [vot 71 the upward and downward forces in the plant affected the kind of organs it produced, especially whether they were floral or vegetative. The particular form of each organ was a manifestation of the essence of the species, although taxonomically unimportant details were due to accidental causes such as a change in the environment or the growth of galls. Linnaeus returned to the interaction of these tissues with each other and the growth forces of the plant in numerous works (e.g., Linnaeus, 1759a; 1759b, p. 827, translated below; 1767c): the plant with the rootlets sucking the humid envelopment (tincturam aquo- sam) of the ground, which by the heat [that] is daily added, is pushed through substance, holding up the ascending stem, within which the multiplying me- dulla, the base dissolving, the apex infinite, although I conceive the bundle of fibers as a growing isosceles [triangle], in which stronger force tears apart the outer fibers, diverging outward and penetrating the cortex, terminating in a bud that multiplies. similarly; from the obstacle of the cortex appear the ex- the medullary fibers converge, and in protruding they lay bare the substance of the cortex in the ca/yx, of the liber in the corolla, of the wood in the stamen, of the medulla in the pisti/, the plant ending in new life, the collected threads in the seeds are the last of the medulla. Thus the medulla was indeed involved in the generation of all the appen- dicular organs. The leaves could not reproduce themselves or give rise to buds; when a bud, containing medulla, was removed, leaves would never be produced again (see above, also Linnaeus, 1760b). Leaf, petal, and stamen alike were dependent on medulla for their existence, although their substance was cortical in nature. In a very brief summary of the growth of the plant, Linnaeus (175 1a, p. 301) even suggested that the whole vegetative part of the plant above the ground (herba) was the product of the medullary substance of the root. THE ANALOGY BETWEEN LARVAE OF INSECTS AND THE VEGETATIVE ORGANS OF PLANTS L6fling (1749) described the bud as being the larva of the herb—t.e., of the vegetative part of the plant. Linnaeus quickly developed this analogy. There was a metamorphosis in the plant, like that recorded by Jan Swammerdam in cabbage white butterflies: the caterpillar pupated; later the pupa or chrysalis broke through the cortex and metamorphosed, and the butterfly, the perfect form of the insect, emerged (Linnaeus, 1749). Note that this and other analogies Linnaeus used go far beyond mere comparability; they betoken more what we would think of as homologies (Stevens, 1984a, 1984b; Broberg, 1985; see especially Beer, 1983, chapter 3, and Atran, 1990, including references). And so it is and will be true, that whoev d d pl tl to understand them from their aul eirachife: to understand them in the same bed as insects, should expect their metamorphosis. For example, whoever lly ex mine sialon oleracea, and afterwards should see Crambe maritima, would be q 1 that Brassica and Crambe by relationship 1990] STEVENS & CULLEN, LINNAEUS 189 (“‘cognatione”’) should border near one another, and never think other than two so very dissimilar’ plants should produce similar fructification, would, however, suppose the opposite when the fructification appears in Crambe, which differs so very much from that in Brassica that it comes close to Rap- istrum maximum rotundi folium monospermum Cornuit, to such a degree that these two, although dissimilar, more properly are cognate and are themselves associated by qualities, more than the aforesaid, and this the internal structure of the plants, placed in front of our eyes in the fructification, shows. — Linnaeus (1755a, p. 15) Thus larvae, effectively equivalent to the vegetative, external part of the plant, could be very similar, masking taxonomically important differences in the internal structure of the plant as manifest in the flowers. Other examples were Compositae with rayed flowers and those lacking rays, but which were very similar vegetatively, including Conyza bifrons and C. raguzina, Bidens cernua and Coreopsis bidens, etc. (Linnaeus, 1755a). Linnaeus observed that such plants were called “Bifrontes’”’ by botanists; the Conyza bifrons radiata of the first edition of the Species Plantarum (Linnaeus, 1753) became the Jnula bifrons of the second and subsequent editions. Since the petals, stamens, and stigma were at the same time the internal structure of the plant and the adult plant that became visible after its metamorphosis, this disposed of potential objections that the parts of the fructification, on which classification was based, were transient structures. They were not, because they were manifest during the whole life of the adult plant, or at least a good part of that life (see Ray, 1696; Sloan, 1972).8 The significance of this set of analogies is clear. This way of visualizing the plant meshed with Linnacus’s statements as to the relative importance for classification of the different parts of the plant. The external form of the plant, its habit, was largely the impression on our senses made by its most obvious parts, the stems and leaves, which were cortical in nature. The habit was comparable to the larva and was a poor guide to taxonomic relationships at the generic level. This is because what is classified is not the larva, crust, or skin—not the vegetative parts of the plant derived from the cortex—but the imago, or fructification, generated by the medulla or the interaction of the medulla with the cortex (see, e.g., Linnaeus, 1762b). It was the inner parts of the plant that were displayed in the flower. Following Linnaeus’s basically Cesalpinian appreciation of the principles of life, it is in these parts that the vital principles of the organisms resided (see below), and so it is on them that classifications should be based. The fructification became evident only after the metamorphosis of the plant, the calyx, or cortex, being torn. Of course, although all the parts of the fructification were essential taxonomically, when they fell from the plant the essence of the plant had not been lost, since the 7Later (Linnaeus, 1759c) corrected to “‘similar.” ®To strengthen the analogy, Linnaeus frequently described the fructification as the ‘‘animalcule” of the plant, with the calyx corresponding to the elytra, or wing cases, the petals to the wings, and the stamens and pistils to the genitalia (see also Linnaeus, 1759b, 1760a, 1762b, 1767c, 1776). 190 JOURNAL OF THE ARNOLD ARBORETUM [VvoL. 71 medulla, which produced them or stimulated their production, persisted in the seed. Hardly surprisingly, then, the seed, with its corculum nestling between the upwardly growing plumule and the downwardly growing radicle, and en- closed by cortical layers that were ‘“‘sloughed off’ or “‘molted,” was the real seat of life in the plant. The bud, and indeed vegetative variation in general, was of little taxonomic importance at the generic level (Linnaeus, 1762b). However, as Linnaeus later (1767b, p. 10) remarked, the differentiae of the larvae—that is, vegetative differences—could be used for the names of species since such differences were used in distinguishing between species. PROLEPSIS, OR ANTICIPATION, AND THE COMPARISON OF FLORAL AND VEGETATIVE SHOOTS Despite the variety of form manifested by the appendicular organs, both within and among plants, there were nevertheless important similarities among these organs. One of the ways of looking at such similarities was the theory developed by Linnaeus in which the growth of the flower, which occurred in a single year, could be equated with several years’ growth of the vegetative plant. The two theses in which this subject was most extensively treated are Prolepsis Plantarum and Prolepsi Plantarum (Linnaeus, 1760c, 1763)—liter- ally, the anticipation of plants. These theses, although with similar titles, differ in the approaches that Linnaeus adopted. In the first (Linnaeus, 1760c), de- fended by Ullmark, emphasis was placed on the fact that each type of floral organ could be considered as equivalent to a year’s growth of the vegetative shoot and were themselves the “shoots” (soboles)’ of that year’s growth. In the second (Linnaeus, 1763), defended by Ferber, more emphasis was placed on what is basically a physiological explanation for the relationship among the different parts of the plant in general, the parts of the branch and bud, and the parts of the flower, as well as the tissues out of which all the parts of a branch, bud, and flower are constructed. Of course, instances that apparently did not fit this general explanation received special attentio However, the notion of the opening of the flower Siiineene several years’ growth of the vegetative shoot was itself anticipated in earlier writings. In L6fling’s (1749, pp. [1]. 4) pivotal thesis a similarity among bud scales, coty- ledons, and calyx was noted; all enclosed younger parts of the plant and were forced open, or fell off, as they expanded. The plant growing by vegetative buds was like a polyp, and that growing from seeds (““generatio’’) more like an animal with eggs (see also Linnaeus, 1746). Linnaeus (1759a) later described three types of reproduction in plants; the third was vivipary, itselfa kind of precocious growth (see also Linnaeus, |760a). The bud was simply a continuation of plant, rather than something new; the seed, however, was entirely new when grown °Linnaeus used the term “soboles” in connection with the development of leaves, sepals, petals, and the like, but these are not branches or twigs, for which he used the words “ramus” or “‘ramulus.” Here the word “‘sobol” is translated as “shoot,” and the word is always in quotes; without quotes, the word is used to mean a young branch 1990] STEVENS & CULLEN, LINNAEUS 191 (Léfling, 1749, p. 4, “‘continuata” and “propagata”’; see also Broberg, 1985). However, as 1s apparent above, both flowers and vegetative shoots were en- closed in buds, the scales of which fell off, and were actually formed in the preceding year (L6fling, 1749). Flowers in the axils of the scales of a catkin (ament) could thus be compared to small buds in the axils of leaves of a leafy shoot, the latter, however, developing (‘‘germinandis”’) only in the following year (see also Guédés, 1969). Buds were not found in leaf axils where there are fructifications, hence Linnaeus thought about how the fructification, just like the vegetative plant reduced to rudiments (“in compendium redacta’’), could develop from a bud that had undergone metamorphosis (L6éfling, 1749, pp. 12, 13) Thus the fructification could develop a year before comparable structures on a leafy branch. Annual plants, too, were precocious (ibid., p. 12): “If pre- cocious flowers, which may be considered to be unexpanded buds, mature seeds in that year [in which they flower], and the same year are dispersed onto the ground, and may germinate in the following year, you may see nature hastening to open the outer bud scales of the fructification to such a degree that germination from the seed corresponds completely to the germination of leaves [which give rise to the developing leaves of the vegetative bud]; herbs [“herbae’’] therefore avoid undergoing both vegetative and reproductive growth at the same time.” The seed is in some way the “bud” of the annual plant; perhaps this comparison was in part to circumvent the apparent absence of buds in annuals (see Ray, 1686). Here we see the idea of anticipation combined with that of the metamorphosis or change of one part into another; this latter concept we shall deal with shortly. Both in turn are combined with intimations of the physiological kind of ex- planation that Linnaeus would later adopt for them. As Léfling (1749) noted elsewhere in the same thesis, buds were nothing but the vegetative part of the plant contracted because of a deficiency in the vegetative force. There is a very terse summary of the mature version of the theory of antic- ipation in the important tenth edition of the Systema Naturae (Linnaeus, 1759b, p. 826; see also Guédés, 1969): ‘The ‘shoots’ of the present year are the leaves; of the next are bracts; of the third the perianth (calyx); of the fourth the petals; of the fifth year the stamens, and the stamens being produced (exhaustis), the pistil. These things are clear: of themselves; because of Ornithogalum; luxuriant [growth]; proliferous flowers; doubled flowers and Carduus.’ A slightly more expanded account appeared soon after in the edition of the thesis Metamorphosis Plantarum prepared by Linnaeus for the collected edition of these theses defended by his students, the Amoenitates academiae (Linnaeus, 1759c). It should be noted that this account (see FiGuRE 3) is an addition and is not found in the original version (Linnaeus, 1755a; 1759c, cf. p. 372). Much of this addition was translated (into French) by Guédés (1969), who did not realize that it was not part of the thesis as originally published. Linnaeus inserted this new section after discussing ‘‘budding” and individ- uality in Taenia, polyps, and the like (see also Linnaeus, 1748). After making an analogy between the cerebral and vascular systems of an animal and the medulla and cortex of a plant, he compared the various parts of the flower es JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 372 METAMORPHOSIS quadrupeda pull Se utero tuniculo um licali & placenta ‘alligant > que in illis marceflit, fed in vermibus , & quidem tetris Salen thi mus matrem p eidem que adherere, quo fit, ut pro vis generatione FS novus ¢xoriacur a articulus, qui dint elt vita, feu ne individaum, ut in radice Tritics repentis. His intellectis ree a een ori infpieere de- pennies qui codem modo ac Tx artus fuos, aw cet plures planta fimiles event quam ide dak im matres ru cll » &t um velramum eee tunes eS ut Pa iviventes i in Ane mitatibus tantummodo ofint, confimiliterac in vege- tabilibus, in guibus quvis gemma ae am contti- Pal vitam, licer fed facto illt obnoxia elt, quod, poll ee peractam florefcen+ tiam vel exuitionem lar cae lam ex cadem planta aliqua nuda, vel fl ov Uti Animalium machina conftat Syftemate ce- rvebvofe & va/culofo nutriente , fic etiam vegetabi um; in his Medulla loco cerebri f Medalla fpinalis & t o yaforum pe em fuccus aliment defertur, Ex cortice deponitur quotannis Liber ex Libro fir Lignum rigidum loco oflium. —Adeco- que be ave partes plemincs conftitaunt Larvam piangane & hw partes dein mutantur in llorem, amquam in Infectumvolt ans co coptrato calyce cor- dal, fixis, mafculis Stas minibus c ligno & femineo Piftillo Medullari. Quomodo haec me oe. fiat vel non fiat apeiidiine ae ntu cillam aut Or- ithogalum cane. er a, Bulbus conttans eee tunicati - prioris annires liétis bafibus toliorum b. Vo PLANTARUM. 373 b. Folia bafi intra bulbum vaginantia pra fentisanat foboles & herba funt. c. Gemma minima refidet intra bafin finguli foliiin bulbo. d, Hxc gemma (c) d imo inf folia, ni ae be ce. Sivero hrc Sail (d) Hie ae mediante enato pene n folia, ft, Bractex in hac pica (e) tum ort fuere ex foe liorum pracocioribus, ad- g. Dum Braéteas has (f) concipio uti gemm x (c) Equa in "ol Eanes d) tesla r etiam ered oncipiam in corum alis herere alias Bemrass minors, fi lbs 5 permantier Sed ultim he ate evadun h. Corol coque uti gemma tertii anni _{B» qu ideoqe dusbos annis eet, exiftic. i, ) tamquam in bulbi gem ma concipio coke» minores nla se hos muratos cffe in ftamina; adeo orf ue Sta mina Q quarti anni foboles fint ; ti anni, ui naturam inies orem fcrutati funt, invene- runt, quod fic dicta metamorphofis in infectis, pro- ri¢ non fit Frenfubteanrsatic > qu al cm OVIDIUS - ireprefentavit, fed at SWAMMERDAM clare obfervavit, fe in larva Papilionis Brafica,cum etiamnum in terram reptat, ala: & totus Papilio, fub ipfo cortic abfco nditus effet; adeo ut hac metamorphofis, denudt Oo pa tium tantum effet, codem modo ac in abn ts. Si jam Hores infpiciamus , Pak bimus cos nih i] ali vss — 3. Addition to thesis Metamorphosis Plantarum (in Linnaeus, 1759c, pp. 31, 332: paragraphs between arrows added in this version). with parts of the adult insect and the particular tissues from which they arose. He then went on to discuss this metamorphosis in the context of the growth of Scilla and Ornithogalum. He suggested that a bud in the axil ofa leaf might produce additional leaves the following year, or it might develop into a spike in the same year that the leaf was produced. The leaves on the spike—the bracts—were tender because they were precocious, representing the leaves of the next year’s growth; like other leaves, they bore buds in their axils, but these developed immediately into flowers. The corolla (perianth) was thus the bud of the third year and developed two years in advance, the stamens were the “shoots” of the fourth year, and the pistil that of the fifth year. The details of the equivalence of the several parts of the flower with a particular year’s growth might differ, as with the version in the Systema Naturae just mentioned, but the principle was the same. Linnaeus equated each mor- phologically different part of the flower (or inflorescence) with one year’s growth. Since both Scilla and Ornithogalum lack bracteoles and have a perianth rather than a readily distinguishable calyx and corolla (see below), the differences in the two versions of prolepsis outlined relate to differences between these mono- cotyledons and the more common arrangement in dicotyledons. A similar variation of this basic theme is found in Linnaeus’s explanation of double 1990] STEVENS & CULLEN, LINNAEUS 193 flowers in the Compositae: the involucre represented the “shoots” of year two; the receptacular bracts, those of year three; the pappus, year four; the corolla, year five; and the petaloid pistil, year six (Linnaeus, 1760c). These two theses on prolepsis thus represent the extended justification and discussion of the mature form of the theory of anticipation. There was a close parallel between the growth of the vegetative branch and that of the flower ‘‘wherever leaves are found, there between the substance of the cortex is the fiber of the medulla”; “*. . . the cortex produced leaves, thus leaves are nothing other than ‘shoots’? advanced from the cortex and can produce new life” (Linnaeus, 1760c, pp. 4, 11). In the flower, of course, the medulla produced bracts and calyx with cortex, stamens with lignum, and so on; all these structures were basically cortical in origin. Within “the substance of the cortex” (e.g., Linnaeus, 1763, p. 7), which made the floral organs, was to be found the pure medulla in the pistil; in buds, imbricate leaves enclosed the medulla. In both vegetative and floral buds the leaves were associated with buds (but see below). “The tree (sic) then produced the flower, nature having anticipated the progeny of five years then producing together [and] forming from budded leaves (‘‘foliis gemmaceis”’), bracts, calyx, corolla, stamens, and pistil, and the seed filled up with granular medulla terminating the life of the plant” (Linnaeus, 1767b, p. 9: see also 1763). Note, however, that as with the earlier simpler equation of “tissue” types with different parts of the flower, the parts of the flower involved in prolepsis do not correspond exactly to the seven parts of the fructification that provide the natural characters of genera—calyx, corolla, stamen, pistil, seed, pericarp, and receptacle (see above and TABLE 1). Also, the cortex proper is involved in two years’ growth, the other tissues in one. Linnaeus noted that prolepsis occurred in some animals; in Volvox globator, which he included in the Zoophyta (the sixth and last order of the Vermes, in the animal kingdom: see below), he described a comparable series of generations as being visible inside the body of the minute adult (Linnaeus, 1758) and a similar phenomenon was known in species of Aphis. Anton van Leeuwenhoek was the first to note what he thought were “‘seeds” in Volvox in 1700 (see Dobell, 1932). Further, the figure five in the five years of the mature theory of anticipation was in line with Linnaeus’s favored quinarian numerology (e.g., Jonsell, 1979; Lindroth, 1983: Broberg, 1985). But there is a tension in Linnaeus’s reasoning here, at least to a twentieth- century reader (Guédés, 1969). Linnaeus did not distinguish clearly between axillary and terminal buds but considered that all buds occurred in the axils of leaves of flowering plants; he had no notion of a terminal bud continuing the growth of the stem (see Léfling, 1749). Thus if a flower bud really repre- sented six years’ growth, the parts of the flower would come from a series of buds successively borne in the axils of the leaf or leaves of the preceding years’ growth, and this is how Linnaeus (e.g., 1 760c) discussed that relationship. All the parts of the flower would then be opposite one another. However, Linnaeus (e.g., 175la, pp. 57, 61) knew that the sepals and petals, at least, alternated in their positions. There are additional complications caused by attempts to equate each year’s growth in vegetative branches with a particular tissue (see FiGuRE 2B). Linnaeus 194 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 (1763) discussed the detailed structure of the vegetative bud and compared it to that of the flower in the thesis Prolepsi Plantarum. The leaves of the first (current), second, and third years came from the cortex, those of the fourth year from the liber, and those of the fifth from the wood. For the leaves of the sixth year, he invoked membranes almost touching the medulla, which he compared to the meninges surrounding the central nervous system of verte- brates. Of course, the sixth year’s growth, or the “shoots” of the sixth year, corresponding to the pistil in the flower, could not be the medulla itself, because the medulla was needed to continue the growth of the shoot, whereas in the flower the medulla could be used up in the formation of the pistil, since with flowering the shoot that bore the flower died. As Linnaeus (ibid., p. 11) went on to say, ‘““Whenever the budded leaves of the first year, which are outside, develop into a branch with its leaves and buds, then as much medulla [as remains] in the axils of the leaves of the sixth year, which are concealed in the intimate shades of the plant, protrudes as the new budlet rudiment (nova protrudit rudimenta gemmulacea) for the seventh year, and thus growth 1s prolonged” (cf. Guédés, 1969, p. 338; translation of ‘‘pro- trudit” as the ambiguous “produit’’). The protrusion of the medulla beyond the cortex agrees with the ideas Linnaeus expressed elsewhere (see above) of the medulla pushing upward and outward and forcing the more rigid cortex apart; continuation of growth is by the medulla in the ultimate (sixth) year’s budlet itself budding. Since the mature leaves are always cortical in origin, although this is not always clear in Linnaeus’s writings (Guédés, 1969), the development of the vegetative bud described above is at some variance with the normal pattern of tissue development. The inner leaves effectively have to change their nature from wood to liber and then to cortex before they expand, and this direction is the reverse of that in which the tissues develop in the stem THE RELATIONSHIP BETWEEN PLANT PARTS With Linnaeus, as with later proponents of the foliar theory of the flower, the analogy between floral and vegetative shoots is connected with that between all appendicular organs. It is not clear which analogy came first to Linnaeus, although the evidence suggests 1t was the second (see below). Evidence for the equivalence of all lateral organs came largely from the nontaxonomic findings of teratology (note that Sachs (1890) somewhat underrated this aspect of meta- morphosis). A number of interesting abnormal phenomena were early discussed by Lin- naeus, although not initially from the point of view of interchangeability of plant parts. Thus in a double flower (‘‘flos luxurians’’) some parts were mul- tiplied and others destroyed; there might be many petals but few stamens, yet no interconversion between the two was noted (Linnaeus, 1736a). The calyx and corolla sometimes could not be sharply distinguished since in a number of taxa, including Daphne and Ornithogalum, they formed a single body. This was green and tough on the outside, so showing its calycine quality, and thin and colored on the inside, due to its corolline nature (Linnaeus, 1738, in a discussion of Cesalpino’s work; see below). This arrangement is particularly 1990] STEVENS & CULLEN, LINNAEUS 195 common in spring-flowering plants, and it was later suggested (L6fling, 1749; see also above) that in such cases the substance of the liber did not draw back or separate from the cortex. There was abundant evidence for the basic similarity of appendicular organs in the changes and intergradations manifest in the various terata, or growth monstrosities, known to Linnaeus and his contemporaries, information on which was summarized in the thesis Metamorphosis Plantarum (Linnaeus, 1755a). Guédés (1969) gave a particularly thorough analysis of this aspect of the ‘“‘“metamorphosis”’ of plants. These terata included doubled flowers, in which the stamens, and sometimes also the pistils, were changed into petals, resulting in sterility (Linnaeus, 175la, 1755a). Extreme changes also occurred when whole flowers, series of flowers, or even a leafy branch bearing flowers, were produced from the pistil of a simple flower or the receptacle of an aggregate flower like a scabious or a thistle; this is the phenomenon Linnaeus (175 1a) called proliferation. Galls and other kinds of insect infestation provided similar evidence of metamorphoses— Pistacia produced long, purple follicles, Ceras- tium, imbricate capitula, and so on (Linnaeus, 1755a). In such terata organs changed their form, and such changes, appropriately interpreted, gave evidence at the same time of the potentiality, nature of, and relationships between organs. Hence, despite the fact that petals were not produced directly from plain cortex, they were interchangeable with—and in some way fundamentally the same as—leaves, which were. Since the liber and the wood were derived from the cortex, it 1s perhaps hardly surprising to see that under the appropriate “physiological” condition — excess of sap—they could become plain cortex again (see Guédés, 1969). This leads to the other major line of argumentation bearing on the relationship between plant parts, which can loosely be described as a physiological-balance theory of development. Much nutriment led to the production of leaves, or more leaflike structures, while less nutriment caused flowering, or flowers with less leafiness. This argument was also in evidence early; the opening questions of the thesis Gemmae Arborum (L6fling, 1749; see also above) seem to expect a physiological answer, and in the important addendum to the Philosophia Botanica entitled ““Metamorphosis Vegetabilis’” (Linnaeus, 175la, p. 301, translated in part below; see also Celakovsky, 1885, and Guédés, 1969), the explanation for the rudimentary theory of anticipation advanced two years before by L6fling is expanded in terms of such physiological ideas (it should be remembered that L6fling was acting as amanuensis for Linnaeus during the writing of the book): Buds or budlets or flowers or both are fertile. The plumule of the seed is often terminated by a flower or bud. The principle of flowers and leaves is the same. The principle of buds and /eaves is the same. The bud contains rudiments of leaves. Stipules are appendices of leaves. The perianth is made up of connate rudiments of leaves. By (‘‘derivato’’) diverted nutriment to the scales of an ament, the destroyed florets are changed to Leaves. By nutriment diverted to the florets ofan ament, the leaves are made Calyces. Moderate growth produces flowers from the terminal leaves. 196 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 This style of explanation was elaborated over the next 15 years (see especially Linnaeus, 1763). It was a well-established horticultural observation, and several times commented upon by Linnaeus (e.g., 1755a, 1760c, 1762a), that plants in fertile ground were wont to branch profusely and produce lots of leaves but few flowers. When the cortex was well nourished, the medulla could not emerge (Linnaeus, 1759a). In poorer and especially drier conditions this relationship was reversed, with less vegetative growth and more flowering. Similarly, plants (such as species of Dianthus, Papaver, and Anemone) that produced flowers with numerous petals when grown on good soil bore simpler flowers with fewer petals when grown in poorer ground (e.g., Linnaeus, 1755a, 1760c). With an excess of the descending, vegetative force, either more leaves or more petals resulted (petals, of course, although leaflike, were derived from the liber, not the cortex itself; leaves were directly derived from the cortex). As the vegetative force became weaker, the activity of the ascending, generative force in the medulla became more evident; there was freer movement (‘‘propulsioni’”’) in the medulla, the cortex was weaker, and so the medulla emerged further from the cortex and produced flowers. The general balance theory extended to taxa like Lilium bulbiferum and Dentaria, where plants with bulbils did not set seed (e.g., Linnaeus, 1763); vegetative forces were again in excess. The different sexes of flowers could in part be at least similarly explained; in male flowers the medulla found its way through the cortical substances but did not have the force to expand into the pistil, dying off or drying out (Linnaeus, 1763). There is thus abundant evidence that Linnaeus envisioned the fairly ready transition between the different parts of the flower. (This differed from the situation in animals, where in one ambiguous passage Linnaeus (1760c, p. 19; see also Guédés, 1969) suggested that a liver could not change into a heart, or a heart into a stomach; each had its own nature (“sed singula suum retinent principium’’).) Nevertheless, although transitions occurred, intermediate struc- tures on normal plants were uncommon; they were varieties, variations from the taxonomic or essential norm that depended on fixed and discrete gaps between both the organs of plants and the taxonomic groups. The situation was not so clear when the pistil was considered; it was normally medulline, yet in doubled flowers it could become petaloid-cortical. Linnaeus suggested that the pistil was indeed covered by a very thin layer of cortex, and it was this that developed and made the pistil—the “shoots” of the sixth year— on occasion foliaceous. However, when this was the case, there was nothing— , no medulla—present in those “leaves” (“Quod si ulterius pistilli mutati- onem in folia ostendere foret animus. . .”,—Linnaeus, 1760c, p. 18). Medulla and cortex were not interchangeable, yet ane largely medullary pistil could still be the leaf ofthe last year in the proleptic series. This version of the development of the flower is similar to that of the vegetative bud discussed above, with “special” tissue of cortical origin, yet not part of the normal series of tissues of cortical derivation, giving rise to the leaves of the ultimate year. 1990] STEVENS & CULLEN, LINNAEUS 197 THE CORTEX-MEDULLA THEORY AND HyBRIDIZATION The cortex-medulla theory also played an important role in Linnaeus’s ideas on hybridization. In normal reproduction the medulla, the internal structure of the plant, its essence, was effectively continuous through time and successive generations since creation because it remains unchanged in the seed (Linnaeus, 1759a). Or, more accurately, the pistil, derived from the medulla, was unable to lay the foundations for the life of the new plant until the woody essence of the stamen had been absorbed by the medullary humor of the pistil (Léfling, 1749; see also above). Later Linnaeus (1767b; see also 1759a) talked about the copulation of the “cortex externa” with the medulla, the medulla in this case being compared to nerve fibers, producing the new life of the plant. Thus the woody essence from the stamen, representing the cortex, and the medulla, from the pistil, were both represented in the seed. After 1747 in particular, Linnaeus became much interested in the phenom- enon of hybridization, although it took him about ten years to develop his views on this subject (Larson, 1971). Excellent summaries have been provided by Bremekamp (1953), Hagberg (1953), Hofsten (1958), Larson (1968, 1971), Stafleu (1971), and Broberg (1985), while Zirkle (1935) included extensive translations of Linnaeus’s writings on hybridization. In hybridization the con- stancy of species form required for Linnaeus’s taxonomic system would appear to break down. There is, however, less conflict when such simple hybridization is explained in terms of the cortex-medulla theory. As Linnaeus saw it, hy- bridization was like any other fertilization event, with involvement of the stamen, basically cortex, and the pistil, pure medulla. The results were perhaps even predictable. Although the hybrid might resemble the father in overall appearance, “with regard to the inner medullar substance and fructification it is the image of the mother” (Linnaeus, 1771, p. 107; see also 1751c, 1759d, 1760a, 1762b) and was thus more likely to be classified in the same genus as the mother (see also Hull, 1985). In an addition to the thesis Plantae Hybridae (Linnaeus, |756a), the analogy between larvae and the vegetative part of the plant was drawn (see above), implicitly suggesting a comparison between the male contribution and the nonessential covering of the metamorphosing insect. As Hofsten (1958, p. 80) aptly remarked, “‘the new species were the old ones in new array.” They were the old ones because they had the same medulla and were in the same genus; they were in new array because they had a different cortex. There were, however, possible taxonomic problems for Linnaeus when infrageneric hybridization was considered; varieties, and variation in general, might be the result (Linnaeus, | 762b). Toward the end of his life, Linnaeus developed larger ideas as to how much plant diversity could be explained by hybridization (see particularly Hofsten, 1958; and Larson, 1971). The details of these complex theories need not concern us here, but the basic results of hybridization at any level would be expected to be the apparent physical dominance of the father but the taxonomic dom- inance of the mother. God created organisms with medulla covered by the 198 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 71 principles of the various kinds of cortex; these were the ordines naturales. Successive hybridizations produced genera, species, and varieties (Linnaeus, 1764; see Bremekamp, 1953, and Larson, 1971, for translations). Bremekamp (1953) calculated that up to 216,000 species could be produced by this whole process—rather more than the 10,000 or so Linnaeus believed existed. During these hybridizations, what was originally undifferentiated medulla is gradually modified by the different cortices that God had created. Bremekamp (1953) saw this hybridization theory (as in Linnaeus, 1762b, 1764) as justifying the levels of the Linnaean hierarchy. He thus suggested that it not only explained the generation of taxonomic diversity in plants, but also the fact that there was a taxonomic hierarchy. However, it should not be forgotten that there are five levels of the Linnaean hierarchy—the four just mentioned, as well as the classes into which the ordines naturales were grouped. Also, there were suggestions that hybridization provided Linnaeus with a reason why his families (ordines naturales) could not be defined. Problems surrounding the definition of such families were not simply caused by the tension between the recognition of such families by their habit and their definition in terms of fructification characters. Hybridization might even lead to the recombination of these fructification characters, which would mean that they would not be restricted to any particular family (Malmestr6m & Uggla, 1957; cf. Larson, 1971; see also above). But hybridization between species or even genera would not necessarily disturb the economy of nature (Atran, 1990; see also below); the properties of congeneric plants are largely the same and are unaffected by hybridization (Linnaeus, 1762b). Finally, if all possible hybrid combinations are produced, the “form” that taxonomic diversity takes is likely to be very regular, and depending on the nature of the interactions of cortex with medulla, without any particular gaps (Eriksson, 1983), a subject to which we now turn. CONTINUITY AND THE CORTEX-MEDULLA THEORY Linnaeus, although believing both species and genera to be discrete, consid- ered that at higher taxonomic levels there were no particular gaps.!° Linnaeus’s great interest in the pe/oria variant of Antirrhinum (see the thesis De Peloria, defended by Rudberg; Linnaeus, 1744) shows his concern over change that transgressed established boundaries. Pe/oria was an example: how could a genus arise de novo, since genera were immutable and discrete? But Pe/oria could be fitted into his general taxonomic scheme of discrete, generic-level entities. Additionally, in De Peloria the intermediate nature of corals, Abraham Trem- bley’s work on polyps, and the production of wingless aphids by winged aphids also received attention. All these phenomena were later integrated with his ideas of systematic continuity at higher taxonomic levels, although it was perhaps not so much the change from winged to wingless in aphids that was emphasized by the later Linnaeus, but the way in which they reproduced. 'Broberg (1985) rightly noted that although Linnaeus crossed out “natura non fecit saltus”’ in his own copy of the Philosophia Botanica, this was simply an editorial correction; the phrase was repeated twice in the same section, and Linnacus allowed the second mention to stand. 1990] STEVENS & CULLEN, LINNAEUS [39 The comparison of the vegetative bud with a polyp, and of seeds with an egg-bearing animal (e.g., Linnaeus, 1749), was important in heralding the in- tegration of the cortex-medulla theory with ideas of continuity (for a good treatment of this, see Broberg, 1985). Practically all the different aspects of the theory were involved. The thesis Animalia Composita (Linnaeus, 1759a) pre- sented the justification for Linnaeus’s remodeling of the arrangement adopted in the last part of the animal kingdom in the tenth edition of the Systema Naturae (Linnaeus, 1758), a rearrangement that can be understood only in the context of the cortex-medulla theory. As noted above, Linnaeus frequently compared the medulla to part or all of the nervous system of animals, and in the thesis Animalia Composita, for example, animals were said to have the same general cortex-medulla construction as plants (see also Linnaeus, 1759d; Lindroth, 1983; Broberg, 1985). The medulla of plants was equivalent to the spinal medulla of animals, that is, to the spinal cord itself. Thus all organisms possessed the cortex-medulla type of construction. However, the constraint exercised on the medulla spelled one of the differences between most animals and plants. The latter were composite organisms because the medulla was able to push through the cortex and form buds; in animals, the medulla remained strictly enclosed by cortex, so buds could not form and the organism remained simple. Plants were branched, and this branching was the manifestation of the inherent multiplicative property of the medulla (see below). But in worms and some other animals, there was also no constraint to the medulla, no hard vertebral column, and so the medulla could escape in the same way as it did in plants (Linnaeus, 1759a). Thus a single articulation of 7aenia had life, just as a single articulation of the root (rhizome) of Triticum repens could give rise to a complete new plant (Linnaeus, 1767a). Composite animals were placed next to flowering plants in Linnaeus’s scheme of things (see TABLE 2). In TABLE 2, the distribution of this feature along with others that come from plant and animal construction are superposed on Linnaeus’s grouping of the Vermes (see Linnaeus, 1758). The class Vermes ended with three orders. The first, the Testacea, were bivalve and gastropod mollusks in which a single animal was covered by a hard shell. The next group, the Lithophyta, were ““composite molluscan animals, sprouting out from strong underlying coral in which they are grafted and which they build” (/bid., p. 789). A shell is there, but in this group the animals are “composite,” being organically connected. Linnaeus described the animals involved as nereids, which he would probably have put in the order Intestina if they had not been enclosed in stony matrix, or as hydroids, which he would likely have placed in his last order, the newly cir- cumscribed Zoophyta, if they had been free living (see Linnaeus, 1745, for earlier arguments as to whether corals were plants, animals, or stones, and if animals, whether they were to be classified by the coverings or the organisms contained). The Zoophyta consisted of ‘““composite flowering animals with an animated body (stirps vegetans)” (ibid., p. 799), one of its members, Hydra, being described as a sensitive flower (note that the Zoophyta were originally placed with plants—e.g., Linnaeus, 1737). These three orders, particularly the Zoophyta, were all more or less anom- alous in the context of typically animalian features. The Zoophyta in particular 200 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 TABLE 2. The Linnaean scala at the juncture of animals and plants.' ANIMALIA - VERMES PLANTAE (flowering plants) Intestina Mollusca Testacea Lithophyta® Zoophyta’ S + —_—_________——_cortex-medulla construction > Le) " . sete C+ sensitive —————_|] animate | ||argely insensitive > e owers ~< no prolepsis {| prolepsis ———-» Dd < no roots {| ——_______» S + _______ {f roots » #—\——— no _ metamorphosis {I metamorphosis ———® ww D . ‘ ‘ A S <——-simple organism ——_1}—_————— composite organism ——————_> — © Se ston ston often bark and & + naked—_] | ony —_s rs) ering® cora | stipe, etc. | wood aD = 6 = Serpula — Tubipoera- Volvox— —»# prolepsis common arr tubular > subcylin- — prolepsis a2 univalve drical oS 3 & Echinus-«Chiton— BL - many many shells a plates 1S) 'This table depicts the distribution of different features ‘‘separating” animals and plants. Most of them also occur in some, but not all, members of the Animalia-Vermes and sometimes isolated give some indication of the degree of separation between the groups afforded by each feature. Also shown are some bridging genera that blur the lines of separation between families, making smaller- scale linkages. *The animals inhabiting the shells of the Testacea are usually referred to molluscan genera, but Teredo belongs to the Intestina (see also Linnaeus, 1771b, for problems in classifying shells; and Lindroth, 1983, for references). 3The animals inhabiting the Lithophyta belong either to the Mollusca (Nereis) or the Zoophyta (Hydra “The “flowers” of the Zoophyta are usually referred to Hydra, itself a zoophyte, rarely to Medusa, nollusk. ene of a various stony and woody coverings differ, but the fact that these groups all have coverings of one sort or another is important. 6Serpula, He last genus in the Testacea, 1s the only one in which the animal is described as being a Teredo; this may be in part because of the similarity ofa Serpula “shell” to the calcareous li ining living, would probably have been placed by Linnaeus in his Intestina. Serpula penis, however, is a lamellibranch mollusk; paired shells can be seen at the end of the tube (K. Boss, R. Turner, pers. comm.). The genus adjacent to Serpula, Dentalium, has a tubular shell and is included in its entirety in the Mollusca today (see also Dodge, 1952, for the taxa included in the Testacea; Winsor, 1976, for a description of serial sequences in Linnaeus’s waaada Classiicavon: 7Linnaeus did not mention the distinctive, rather ] here, but these are very evident. The general arrangement of the Testacea is: first, genera in which the shells have many valves, then genera with two valves, then genera with one spiral valve, and finally genera with one valve that is without a regular spiral (see also Ga 1983). hi ee 1990] STEVENS & CULLEN, LINNAEUS 201 were constructed like plants, yet they were more obviously like animals in their behavior. They often had roots; they were generally caulescent, with life mul- tiplying in branches; they had buds that could be removed; and they meta- morphosed into an animate flower that had the power of voluntary movement and that itself changed into seed-bearing capsules (Linnaeus, 1758, p. 643). Although the Zoophytes lacked leaves, Linnaeus (e.g., 1751b) suggested that the leaves of a plant might indeed be organs of motion that were passively moved by the wind. The articulations of the Zoophytes were the ‘“‘exuviae”’ of the animal, solid cortical tissue, the result of a process similar to that by which the cortex of the bark changed into solid wood (Linnaeus, 1759a). Linnaeus noted that the Zoophyte Hydra behaved like Salix in that its separated buds could grow into independent organisms;''! the ability of Sa/ix cuttings to grow so readily provided him with a good example of the role of the medulla in vegetative growth, just as its catkins provided an early example of prolepsis. The last genus in the Zoophyta, Vo/vox, included V. globator, which as we have already noted showed yet another plantlike feature, prolepsis. To summarize, the species placed in the Zoophyta and Lithophyta were complex organisms in which the notion of individuality developed for higher animals was inap- propriate, but which were more plantlike in this respect as well. The Zoophyta even included 7aenia, the tapeworm, which had earlier been classified in the Vermes-Reptilia (Linnaeus, 1756; of the Vermes-Intestina of Linnaeus, 1758), but which Linnaeus came to believe was also a composite organism, the ar- ticulations of which had internal flowers! TABLE 2 shows clearly the nature of Linnaean continuity with an overlapping, catenalike distribution of characters, all of which come from Linnaeus’s un- derstanding of the cortex-medulla theory; we can find no other relevant char- acters. It is interesting that both Linnaeus (e.g., 1759a, p. 4) and Fabricius (in Giseke, 1792, pp. [2], 4) used the word ‘‘catena”’ in the same context of con- tinuity. Plantlike features are found in many animals, so making the distinction groups—a part that Linnaeus placed immediately adjacent to groups in which these features were constant. In a world in which continuity rules, “groups” are circumscribed by more or less arbitrarily selected characters and cannot be defined by covarying characters (see also Lamarck, Note that in the Testacea and Lithophyta, Linnaeus sane the general identity of the organism inhabiting the hard covering the particular nature of which defined these groups—e.g., ““Animalia Teredo.” In the Zoophyta there are comparable references to these organisms, but as flowers: for example, “Flores "Abraham Trembley’s then recent, but already celebrated, work on Hydra paid great attention to its capacity to grow and regenerate (e.g., Trembley, 1744). Dawson (1987) discussed Trembley’s work in some detail and showed that some of his experiments were performed because of his beliefs that Hydra was a plant and that plants were inside-out (not upside-down) animals. Dawson also detailed the impact of Trembley’s work on Charles Bonnet, who, unlike Trembley, interpreted this and other evidence as support for the existence of a Ladder of Natural Beings. Bonnet’s (1745) arrangement of continuity along this ladder differs in detail from that of Linnaeus. 202 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Medusae” and ‘Flores Hydrae.”’ It is interesting to see that the actual meta- morphosis represented by the appearance of those flowers was interpreted as being more like that of plants than that of insects, since the zoophyte flower did not separate from its stalk, unlike the imago of the insect (Linnaeus, 1762b). From the Zoophyta with their “flowers” and largely vegetable (flowering-plant) type of construction, it was of course a short step to the flowering plants, the next group in the scheme of things. It should be remembered, however, that Linnaeus (e.g., 1766a) considered the Zoophyta to be the link among all three kingdoms, Ae mineral kingdom included. Unfortunately, it is impossible to follow the linkages within flowering plants by studying the arrangement of plants in the Genera Plantarum or the Species Plantarum, since both followed the sexual system that did not represent natural relationships as Linnaeus saw them. The fragments of a natural arrangement that Linnaeus (e.g., 1738, 175la) produced lack the characterization of the groups recognized that might allow the kind of analysis presented here to be taken further. However, the partially characterized groups in Giseke (1792) may be susceptible to such an analysis, although the form that the continuity is likely to take will probably not be linear, as the illustration of the Siar any of these groups (rbid., figure facing p. [623]) suggests. Of course, any “‘combi- natorial” hybridization of cortex and medulla (see above) is also nee to generate linear continuity. Although Linnaeus grouped genera into natural families, from Giseke’s figures it is clear that some genera bridged gaps between groups; there were also plants not yet discovered that might fill the gaps. In addition, a property of “natural” groups in a world of continuity needs em- phasizing here. Such groups contain a portion of the real (continuous) order yet are artificial in that their boundaries are not evident in nature. That our groups have more or less “real” boundaries should not blind us to the distinctive nature of Linnaean higher taxa (as well as those of Lamarck and De Jussieu). VOLVOX CHAOS, CHAOS PROTHEUS, AND UNRESTRAINED MEDULLA The last species mentioned in the first volume of the tenth edition of the Systema Naturae (Linnaeus, 1758) 1s Volvox chaos, which Linnaeus found to lack definite form, in this respect being “‘more inconstant than Prometheus (sic)” (tbid., p. 821). In the twelfth edition of the Systema Naturae, Linnaeus (1767b) again remodeled the last part of the Zoophyta, and he added the genera Furia and Chaos, the latter including Chaos protheus (sic), the erstwhile Volvox chaos, and other minute organisms. (The word “‘chaos” has several connota- tions: Chaos was a primordial world of disorder, formlessness, and confusion; Linnaeus’s genus Chaos was certainly all three.) Linnaeus (1767a) included details of recent discoveries of submicroscopic life in the thesis Mundum Invisibilem. He discussed Otto von Miinchhausen’s work on fungi both extensively and with approval. Von Miinchhausen had claimed to have found animalcules developing from the seeds of such fungi as Ustilago, Lycoperdum, and Agaricus, so perhaps making fungi animal, rather than plant, in nature.'* Linnaeus (1767a) speculated that fungi might better be put in a new kingdom, perhaps along with polyps and infusoria (‘‘Moleculae vivae’’). He (1767b) placed the animalcules coming from fungi in the genus 1990] STEVENS & CULLEN, LINNAEUS 203 Chaos (as C. fungorum and C. ustilago), Chaos being the genus that contained infusoria and animalcules of all sorts. Linnaeus even thought some diseases were perhaps caused by organisms that should be placed in this genus. To Ramsbottom (1941, p. 297) this recognition of the genus Chaos represented a “lamentable lapse” on Linnaeus’s part: ‘‘So acute in his sense of affinities, so sure footed in so many different fields, Linnaeus here came sadly to grief.”” But Miinchhausen’s findings must first be interpreted from Linnaeus’s point of view (for other literature on Miinchhausen, see Ramsbottom, 1941; Ainsworth, 1976; Broberg, 1985). Chaos itself fitted readily into the general superstructure of the cortex-medulla theory. This new kingdom to which Linnaeus alluded was ser ute between the C lacking precise form. Thus the last scholium in the thesis addressed the issue of whether these smallest animals were pure medulla, lacking an organic body. Pure medulla might be the seat of life, yet it was formless, or at least without constant form; only when constrained by and interacting with the cortex was definite form generated. In addition, that Miinchhausen’s fungi should develop into microscopic worms would surely find ready resonance in Linnaeus since he had earlier (175 1a) noted that the fungi were in classificatory chaos, with specific and varietal limits being indistinct because of the lack of constant form in these organisms. Perhaps the problem with fungi was that they lacked much hard cortex; to Giseke (1792), probably reflecting Linnaeus’s later thoughts on the subject, this lack certainly explained how fast fungi grew (here Giseke is apparently reporting on his studies with Linnaeus in 1771). The variability of fungi, the greatest of any plant, again occasioned comment. It was almost to be expected that classification, which depended on constant morphology (this in turn depending on the interactions of rigid cortex and medulla), would be so difficult in the fungi. Animation was the major characteristic of life, a characteristic that resided in the medulla. Linnaeus (1767a, p. 19; see also above) even considered the possibility that the medulla of the inanimate plant itself might be animated, although being constrained by the cortex it could not show this property. Small wonder that when organisms were made up of pure medulla they showed active movement and sensitivity, but not constant form Linnaeus toyed with the idea that the felioneiiio between the plant and animal kingdoms was similar to the metamorphosis that occurred in the de- velopment of plants and insects, that of the plant in particular occurring during prolepsis and the shedding of its covering (“‘ut viderentur ipsa naturae adyta penetrari detegendo pro/epsin transformatum per antipraegnationem”’ — Lin- naeus, 1767a, p. 20). Zoophytes also showed this metamorphosis, but becoming '2In the eighteenth century, fungi growing out of insects or other animals were discovered. Linnaeus (1753) initially included these fungi in Chana, as chai like C. Hitec were ae ide atin 10 Cordyceps. Although there was whether or not such “ table flie ““vegetable wasps and plant worms” (Cooke, 1892) ae evidence of ee they - seem to have played an important role in that element of Linnaeus’s thinking under ee i (see Liitjeharms, 1936; Ramsbottom, 1941). 204 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 more like animals in the process (“per metamorphosin abire in animalia, que- madmodum plantae in flores” (‘bid.)); hence, plants could possibly metamor- phose into animals—witness Miinchhausen’s findings. So prolepsis, metamor- phosis, and the cortex-medulla theory were adequate to the challenge presented by this unexpected microscopic world, although clearly an even larger issue is raised—that of the nature of the generation of life itself (Broberg, 1985). Pre- formation was, however, clearly not involved (see also Farley, 1982, but cf. Goethe, 1891, p. 322) GOETHE, DE CANDOLLE, AND METAMORPHOSIS Goethe developed his ideas on the relationship between the appendicular organs of plants in his Versuch die Metamorphose der Pflanzen zu Erklaren (Goethe, 1790; additional references are to the translation by Arber, 1946) during his travels in Italy, and especially in Sicily in early 1786 On the surface there is general congruence between Linnaeus’s and Goethe’s views on plant form, both involving a fundamental similarity between organs of the plant that appeared at first sight distinct. Goethe emphasized annual plants (Arber, 1946), while Linnaeus stressed perennials and trees, although as we have seen, the latter did not entirely ignore annuals (to Goethe, Linnaeus’s ideas seemed to mean that annuals were plants originally destined by nature to live for six years). In his almost exclusive consideration of the above-ground, appendicular parts of the plant, Goethe effectively deemphasized Linnaeus’s downward, vegetative movement, being little concerned with the root, but so was Linnaeus in his morphological writings. Goethe considered that the gen- erative tissue of the plant was the liber, not the medulla, and that the various parts of the flower represented a series of expansions and contractions of tissue under the influence of an ever-more-refined sap. He saw all seven appendicular parts of the plant Geaves) calyx, corolla, stamens, ee fruit, and seed) as representing Six steps, three successi stem leaves (expansion) and calyx (contraction), petals (expansion) and sexual organs (con- traction), and fruit (great expansion) and seed (great contraction). This idea of the development of the plant is more complete than that in most versions of Linnacus’s theory of prolepsis (but cf. Linnaeus, 1746), all parts of the plant being involved; Goethe also did not discuss particular tissues not being re- sponsible for producing particular organs. Goethe emphasized the interchange- ability of floral and vegetative organs, citing evidence very similar to that given by Linnaeus and placing great weight on teratologies like doubled flowers and on transitional organs. For Goethe, the group of structures that Linnaeus called nectaries played an important role both in the progressive refinement of the sap of the flower, refined sap being needed for the production of the sexual organs, and in demonstrating intermediacy in form between the petals and the stamens.'3 ‘Interestingly, others found Linnaeus’s term “‘nectary,”’ which he used in a taxonomic context but with an almost physiological definition (e.g., Linnaeus, 175 1a), to be unsatisfactory for their largely taxonomic concerns. male us (1762a) listed 18 different kinds of nectaries in addition to other glandular the flowers; all secreted nectar, or apparently did. However, his contem- poraries and their immediate successors thought that there should not be a single term, “‘nectary,” since these structures were very different in nature (see, for example, Turpin, 1815; Stevens, 1984a). 1990] STEVENS & CULLEN, LINNAEUS 205 The main difference between Goethe and Linnaeus is less in the detail of their explanations of how tissue and form were generated, although there are substantial differences, and more in the role these ideas played in their thought. The two belonged to entirely different intellectual generations. At the risk of oversimplification (see below), it can be said that Linnaeus ultimately needed constant and discrete characters to be able to classify, and his theories on metamorphosis, prolepsis, and hybridization involve a fairly precise and cir- cumscribed causality of form. Goethe, on the other hand, adopted a more Neoplatonic approach, seeing unity in nature, indeed in life as a whole, and was looking for ideas behind (or in front of) manifest form, more real than the form itself. That form might intergrade was not worrying but the reverse. Goethe (1891, 1901) himself considered Linnaeus’s approach limited. There was a further connection between the two men. When Goethe devel- oped his ideas in 1786, he had with him an old edition of the Genera Plantarum (Goethe, 1890; probably ed. 4, published in 1752). He had also apparently read the Philosophia Botanica (Mueller, 1952), in which, as we have seen, Linnaeus’s ideas on metamorphosis were clearly, if concisely, expressed. Goethe of course made repeated reference to Linnaeus when he later wrote the Versuch die Metamorphose der Pflanzen zu Erkldren. Ideas of metamorphosis are gen- erally evident in Linnaeus’s work from the 1750’s onward. Goethe’s actual discovery of the idea of metamorphosis was, according to his own accounts (e.g., Mueller, 1952), independent. De Candolle (e.g., 1827, Vol. 1; see also Guédés, 1972), who also developed the notion of the fundamental similarity of all the appendicular parts of the plant, did so largely independently of both Goethe and Linnaeus; he could not even read German. His ideas are best expressed in his Théorie Elémentaire (1813) and especially in the Organographie Végétale (1827). In the former the evidence for this similarity came from the intermediacy of form of different organs and the relationship between parts in “normal” flowers, but in the latter he made more of evidence from teratology and also noted briefly the effects of cultivation and the amount of sap on whether or not a plant would flower. He was early (e.g., 1807) interested in the doubling of flowers. He incorporated some of Goethe’s terminology into his work but apparently largely ignored Linnaeus. This was perhaps because it had long been evident to De Candolle that Linnaeus’s notion that the pistil was made up of pith could not be true of the monocotyledons (De Candolle’s ““Endogens”’). De Candolle (in Lamarck & De Candolle, 1805) and others considered that the monocotyledons had no pith. (Compare Lamarck, 1778—he considered pith to be essential for life, its death with age causing the death of the individual.) Although he largely used the same kind of evidence for establishing the similarity of appendicular parts of the plant as did Linnaeus, he did not cite Linnaeus; the approach of Grew and Malpighi was clearly more congenial to him. De Candolle was perhaps primarily a taxonomist; his morphological work helped his taxonomic studies in that it made the often variable and deceptively simple structure of the flower more comprehensible and regular to him and helped in the development of his ideas of floral ‘“‘symmetry” (not the same as “type” —Stevens, 1984a). De Candolle’s emphasis on the similarity of sepals, petals, leaves, and the like was immediately seen by his contemporaries as 206 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 being of great importance, and it was singled out for attention in several of his obituaries (e.g., Dunal, 1842; Brongniart, 1846). De Candolle’s work initially stimulated a rather typological approach to floral organization in people like Michel Dunal, Christian Moquin-Tandon, and Auguste de Sainte-Hilaire. Since, however, the limits of plant groups remained vague, the typological approach did not flourish in systematics (Stevens, 1984b). LINNAEUS AND ANTECEDENTS TO THE CORTEX-MEDULLA DOCTRINE Some of Linnaeus’s predecessors had ideas that are—or might be—supposed to have had some influence on the cortex-medulla theory (see also Guédés, 1969). We discuss briefly some aspects of the work of Cesalpino, Mariotte, Grew, Malpighi, and Vaillant, so as to understand more clearly the background to Linnaeus’s work. Andrea Cesalpino, the noted Italian systematist, was considered by Linnaeus (1751a) to be the first, and one of the greatest, “‘true botanists” and was freely acknowledged (Linnaeus, 1738, 1749) to be the immediate source of the cortex- medulla theory. There is no detailed treatment of his work, although Greene (1983, and references therein) provided an entry into the literature. Cesalpino was a noted Aristotelian scholar, his book Quaestionum Peripateticarum Libri V (which we have not seen) being of importance in this context. Characters of the fructification, and particularly those of the seed, were most important in classification for Cesalpino because the multitude of parts they provided al- lowed distinctions between groups to be made (Cesalpino, 1583; see also Mor- ton, 1981). But they were, of course, also very important functionally, being involved in that vital aspect of the plant’s life, reproduction; they were vege- tative substance allowing the plant to reproduce (Cesalpino, 1583). The cortex-medulla distinction and the role of those two parts in the life of the plant pervade many of the introductory chapters of book | of Cesalpino’s work, but as Sachs (1890) correctly observed, the relation between plant organs and these tissues is different from that usually described by Linnaeus (but cf. Linnaeus, 1747). Leaves in general were indeed produced from the cortex. Thus ggested that deciduous leaves were produced from the outer cortex, evergreen leaves from the inner cortex (liber: Cesalpino, 1583). The calyx and corolla, both on occasion called leaves (folium), were also of cortical origin, but the stamens (flocci) were not mentioned in this context. They were small, and their importance for reproduction was then unknown. It is in the discussion of the origins of the different parts of the fruit that the cortex-medulla distinction is most focused. The fruit arose from the inside of the plant and was made up of three tissues, medulla, lignum, and cortex, which in turn formed the seed and the woody (‘“‘cortex’’) and fleshy (“‘pericarp’’) parts of the fruit (7bid., p. 18). The fruit developed after the flower, the stigma plus style (“stamen”’) of the former being the young fruit (/bid., pp. 14, 19; see also Sachs, 1890). The position of the medulla provided further evidence of its nature and significance, there being the overwhelming power of the analogy that Cesalpino (1583, p. 3; translated in Sachs, 1890, p. 46) drew at some length between the vital parts of animals and plants. Nature always concealed the principles of 1990] STEVENS & CULLEN, LINNAEUS 207 life (“vitalia principia’’) in the innermost parts, such as the viscera in animals. Hence in plants the principle (“‘principium’’) was to be found not within the cortex, but more internally—that is, in the internal medulla—of which much was in the stem but not in the root. The heart or soul of the plant, the ‘‘cor,” was the junction of the root and the stem but extended throughout the plant. As Cesalpino (1583) noted, the leaves and fruits followed the nature of the cortex, the internal seeds that of the medulla. Cesalpino’s ideas on the parts of plants, their origins, development, and relationships need a more extended treatment than can be given here. However, it may be noted that catkins provided him with a challenge; he thought that in such structures flowers had changed into a different substance— more spe- cifically, that the amentum was produced from the seat of the flower, with the “stamina” (pistils) forming the ament and the petals and sepals (‘‘folia’’”) and stamens degenerating into scales (Cesalpino, 1583; Sachs, 1890). Other flowers were also distinctive in the context of this discussion. Cesalpino surmised that in Ornithogalum and Helleborus the calyx and corolla were joined, since the leafy organs surrounding the flower were green on the outside and colored on the inside. The flowers of Cucurbita and Punica presented a different problem: here the calyx was continuous with the outer cortex of the fruit, the flower originating from the fruit (‘‘flos in radice fructus exoritus” —Cesalpino, 1583, p. 16; Greene, 1983, p. 820). Cesalpino also noted the rarity of plants in which flowers were borne directly on branches with thick bark and the progressive purification of the sap of the plant in the flower. In none of the other works mentioned below (or in Guédés, 1969) does the cortex-medulla distinction assume such prominence. Nehemiah Grew (col- lected in his Anatomy of Plants, 1682; see also Arber, 1941, and Morton, 1981) indeed noted that there were only two parts of the plant that were fundamentally (“essentially’’) distinct: the pithy part and the ligneous part, “‘or such others as are analogous to either of these” (Grew, 1682, “Philosophical History,” p. 19). However, in his discussion of the anatomy of different plant organs, including seed and fruit, Grew emphasized not “pith” and “‘wood” but the particular nature of the tissues making up these parts. Although the several tissues of the plant were compounds of these two parts (Grew, 1682, ““Anatomy”’) there were other interesting levels of analysis. However, the pith was very important because sap moved through it in large part, and the energy and nutritive quality of the sap determined the particular part of the plant that developed; in general, there was progressive purification of the sap as it moved through the plant (Grew, 1682, ““Anatomy’’). The flower promoted the ascent of the sap, so if there was no flower, the fruit would die. If the flowers were large, much sap would be present, but it would be used up by the flower “like a greedy Nurse, that prepares the Meat for her Child, and then eats it up herself’ (Grew, 1682, “Anatomy,” p. 37). The intrinsic rate of ascent of the sap ofa species was also important; grapes, with rapidly ascending sap, had almost no flower since no increase in sap that the flower could provide was needed. This is not so different from either the physiological-balance theory of de- velopment that Li led, or the ideas put forward by Edme Mariotte Fr r 208 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 7] (1679), although Sachs (1890) attributed to Mariotte in particular Linnaeus’s idea that the medulla grows by extending itself and its envelopes in a form of intussusception; the pressure of the sap makes the parts of the plant expand. Mariotte (1679, pp. 54, 55) noted that in cuttings the pith (“mouelle’’) imbibes water like a sponge ‘“‘and transmits it in the little fibers between the bark and the wood, from where it is pushed in part towards the end of the base to produce roots at the extremity of the little point, ... and in part to the nodes which are in the air, to make the buds there swell up, and to make them extend in branches and leaves.” In his extended analogy between plant and animal nutrition, Mariotte conceived of the former as progressive purification of the sap after initial ““digestion”’ by the root. Guédés (1969) discussed at length the several indications Grew gave that he recognized an equivalence among cotyledons, leaves, and even calyx and co- rolla. Such equivalence was widely noted. Marcello Malpighi (1686), who with Grew is considered to be a founder of plant anatomy, discussed such transitions extensively (see also Mobius, 1901; Arber, 1941; Guédés, 1969). Sebastian Vaillant, in his important Sermo de Structura Florum (1718) (see Stearn, 1957; Larson, 1971; Stafleu, 1971) dealt largely with sexual reproduction in plants. However, when discussing the transition of stamens and carpels into petals in doubled flowers, he adopted a style of explanation very similar to Grew’s (see also Guédés, 1969). Enough has been said to discern the relationship between Linnaeus and earlier authors in the context of the cortex-medulla theory. The general outline of this theory as found in Linnaeus is evident in Cesalpino (1583), although some details differ substantially. Cesalpino’s emphasis was on the fruit and seed, the functions of the flower being poorly understood in his time (see Greene, 1983). Linnacus, stimulated especially by the discovery of the sexual functions of the different parts of the flower, focused more on the flower. Cesalpino emphasized the importance of the internal structure of the organism as the place where the principles of life resided, hence Linnaeus’s (1738) early suggestion that the flower represented the insides of the plant exposed by the tearing of the cortex. By and large, similar flowers presented problems to both Cesalpino and Lin- naeus— Ornithogalum and catkin-bearing plants figure prominently in the work of both. Embryonic ideas of the transformation, metamorphosis, or general equivalence of plant parts were widespread, being evident in the works of Grew, Malpighi, Vaillant, and other authors (see Arber, 1946; Guédés, 1969). The expansive power of the sap is alluded to in Mariotte’s work, while the constrictive effect of thick outer cortex on the development of flowers and fruits was suggested by Cesalpino. The general idea of plant nutrition in the seven- teenth and early eighteenth centuries, exemplified in the work of Mariotte and Grew in particular, is that of progressive purification of the sap coming from the roots, but there was known to be movement of fluid in the other direction as well. The development of particular plant structures was generally considered to be dependent on the presence of the right kind or amount of sap. Cesalpino himself developed the cortex-medulla theory from a rather ten- tative analyses of the plant advanced by Theophrastus (for details, see also Greene, 1983). Theophrastus distinguished between core and bark, with wood 1990] STEVENS & CULLEN, LINNAEUS 209 occupying a somewhat ambivalent position—perhaps a part of the bark or a tissue coordinate with it. This is the third level of organization; these tissues are in turn made up of varying amounts of flesh, fibers, and sap, while moisture and warmth are the fundamental properties of the plant body (the formal structure of Grew’s idea of the plant is similar to this). The composition of plant organs is discussed (Theophrastus, 1916) in terms of the tissues making up the second level of organization. In line with arguments later advanced by Linnaeus and others, Theophrastus (1976) noted that the distribution of food in the tree affected fruiting, excessive food to the vegetative growth being prejudicial to the fruit and leading to a failure to bear. DISCUSSION Linnaeus’s classically inspired idea that in the flower the inner structure of the plant, its rea/ structure, becomes evident to the human eye is central to an understanding of his botanical thought in particular, but to much else besides. The medulla, the form-generating part of the plant, entered most completel into the formation of the pistil and seed and was surrounded by structures representing the different parts of the cortex. The cortex, especially in its veg- etative aspect, was largely disposable from the point of view of the adult plant, and this, when joined with his views on plant sexuality, justified the emphasis on the characters of the fructification in the formation of natural genera (e.g., Linnaeus, |751a). The cortex-medulla theory is also intimately involved in the notion of both prolepsis and metamorphosis, and to a considerable extent in the broad outlines of his arrangement and understanding of living beings, especially of plants and “‘plantlike” animals. Clearly, Linnaeus’s development of this idea in plants was also influenced by discoveries about Hydra, Volvox, the “individuality” of each segment of the tapeworm, and so on, as extensively detailed by Broberg (1985), but Bro- berg’s thesis (p. 167) that ‘‘zoology rather than botany steered Linné’s mental activity” is perhaps a little overstated. To him (p. 162), ‘“‘The reproduction observed for the polyp had to affect the conception of reproduction in the main, and it is necessary to consider Linné’s very personal marrow-bark [medulla- cortex] doctrine in light of what the polyp-tapeworm had enlightened him about; less so as a result of his botany.” In a somewhat similar vein, Larson (1971, p. 106) noted that “Linné had taken note of this theory [the cortex- medulla theory] as early as 1738, but he only found a use for it fifteen years later” in his ideas about hybridization. In the thesis Taenia (Linnaeus, 1748) the animal body is indeed discussed in much the same style, although with a terminology rather different from that Linnaeus adopted very soon afterward for plants. However, the cortex-medulla theory is evident in his earliest work. The classical flower became integrated into a new, all-encompassing vision, but the general background to this was accessible to even the young Linnaeus. The assertion that there was a fundamental similarity among all the appen- dicular parts of the plant, whether via metamorphosis or anticipation, is the culmination of this vision. In particular, the theory of prolepsis, or anticipation, seems itself not to have 210 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 been anticipated in any detail in earlier literature. Linnaeus considered his ideas on prolepsis to be very important (he wrote concerning prolepsis in one of his autobiographical fragments, “Nobody has penetrated further into the secrets of the creation” (Pulteney, 1805, p. 558)); the mysteries of Nature were revealing themselves to him. Linnaeus (1760c, pp. 1, 2; see also Linnaeus, 1763) strongly asserted that his ideas were new: Nobody will readily have doubted that the nature of the plant i is much simpler than that of animals to such a degree that it is not surprising if in the latter it others by physiology. In truth I have advanced by a third way proposed by me, the lead to be followed in section 10, p. 826, in the Systema Naturae [ed. 10, 1759b, in part translated above], so that others who are unaccustomed to it can follow it without a mistake This is by way of introduction to the thesis Prolepsis Plantarum, in which he developed the parallel between the growth of the flower and that of the vegetative shoot. Of course, Linnaeus was in general not one to hide his light under a bushel: he also thought that his fragments of a natural method were a masterpiece, and that much else he had done was of similar importance (Mal- mestrém & Uggla, 1957; Larson, 1971; Lindroth, 1983). Here, perhaps, is a clarification of his apparently hasty dismissal of anato- mists as mere likers of botany—botanophiles—-whose work did not pertain to the science of botany (Linnaeus, 1736b, 175 1a). Stephen Hales, Johann Gesner, and Christian Theophilus Ludwig, all interested in the “laws” of botany, the anatomists Grew and Malpighi, and the physiologist Bernhard Feldman were all included in this sorry group. Their discoveries, whether in anatomy or physiology, did not pertain to an understanding of that aspect of plants by which their essence was revealed and their form in general made manifest. Those physiologists who were included among the “philosophers” of botany — Thomas Millington, Joachim Camerarius, Vaillant, and Johann Gustav Wahl- bom—were mentioned because their work contributed to an understanding of plant sexuality. This was of much more central importance to Linnaeus’s thought than findings on cell structure, and he credited Millington and colleagues with having revealed the laws of nature and the mystery of sex. Earlier, Linnaeus (e.g., 1741) had characterized the work of the botanophiles by its lack of interest in the fundamentals of botany; that is, it was not involved in the disposition of plants (into species, genera, orders, classes), or in their naming It may also be noted that Linnaeus (1751a) observed, albeit with some reluctance, that the use of a lens was not essential to an understanding of his sexual system, despite some people’s protestations. Microscopes were, however, essential in anatomical work, even ifan anatomist like Grew deliberately started out by recording what the unaided eye could see, only after that using a lens. Interestingly, Linnaeus seems to have been dissuaded by Albinus (probably Bernard Sigefred Albinus; see Mirbel, 1810) from following up on an initial wish to study anatomy. Anatomy and taxonomy did not begin to be integrated in botany until the late nineteenth centu The ideas of prolepsis and metamorphosis in particular allowed Linnaeus to account for much variation that was not relevant to his classification but 1990] STEVENS & CULLEN, LINNAEUS Pa that nevertheless existed and so could not be entirely dismissed. This point is a very important one. As Guédés (1969) correctly observed, in this new way of looking at a plant, Linnaeus paid close attention to teratological phenomena and similar variation that he necessarily ignored when working on the limits of genera in the sexual system. Examples of the term “‘varietas” included double and proliferous flowers and other trivial variation of no taxonomic importance (Linnaeus, 1744). Luxuriant flowers, including both the double and proliferous flowers mentioned above, were always monstrous, never natural (Linnaeus, 1751a). All double flowers were at most varieties, and variety never could be the basis of real species (Linnaeus, 1 762b). This was reasonable from Linnaeus’s taxonomic viewpoint. The essences of species never changed, so how could these variants, the expression of which often depended on the fertility of the ground in which they were grown, represent that essence? (However, see Rams- bottom, 1939.) In Linnaeus’s morphological work they could, however, be evidence of the potentiality of plant form. In this alternative way of under- standing plants, what is accidental variation in one situation becomes important evidence of the real nature of structure in another As with Goethe, there is no clear idea of a type'‘ or ideal plant in Linnaeus’s morphological writings; there is simply a continuum of form. Although at first sight all appendicular organs are best interpreted as modified leaves (this is a reading of some passages in Linnaeus—e.g., 1760c, p. 19), this would seem to run counter to Linnaeus’s statements that the anthers, stigmas, and seeds were the essential parts of the flower: how could structures essential at high levels of classification be modified from those essential only at low levels? However, since Linnaeus noted that the medulla was involved in the production of all appendicular organs, the leaf or calyx could be considered simply as two parts of the “shoot” from the cortex modified by the medulla; the taxonomically most important structures were those closest to the maternal forces in the medulla. Thus the form-making potentiality of the plant largely or entirely resided in the cortex, although for its expression in visible form involvement of the cortex with the medulla was essential. That the pistil (largely pure medulla) changed its form in some terata is not to be interpreted as a demonstration of the interchangeability of medulla and cortex (cf. Guédés, 1969) but simply as the medulla failing to penetrate the cortex or penetrating it more strongly than was usual in a flower (a discussion on staminate flowers is along the same lines— Linnaeus, 1763). The medulla stimulated the production of organs appropriate to the balance of physiological forces in the plant, only in the pistil perhaps in part determining form. Otherwise the medulla itself was largely formless; even in the flower it could be argued that carpels took different forms, especially in terata, because they had a thin cortical covering, although Linnaeus does not seem to have gone that far.'* However, his explanation of the malleability of '4“As Brady (1987) and some others (e.g., Lenoir, 1987) have a emphasized, there is also no notion in Goethe’s early botanical work of a definite form that is the “‘type” of all appendicular structures and to which all others must ultimately be reducible; ‘al appendicular organs are not ified leaves, or modified anything else for that matter (see also Arber, 1946). édés (1969) thought that Linnaeus’s flower polyaxial, but this seems too ‘ interpretation. As we saw, Linnaeus did not aeeaeeeh clearly between axillary and terminal buds. pale JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 form in species of Chaos and his theory of hybridization, in which God created a single medulla that became differentiated only by contact with a series of different cortices during successive hybridizations, confirm that pure medulla lacked particular form. Hence Linnaeus had achieved a fairly comprehensive explanation of the material cause of form. The substance of the organism (cortex) was dissociated from the essence (medulla) that generated the form that substance assumed, and essence became manifest in substance. The efficient cause of plant form is the physiological-balance theory described above. But hybridization and continuity compromise any attempt to make simplistic explanations of the later Linnaeus as being a rigid typologist One strand of thought in pure sores? is an emphasis on the continuity of form, as illustrated by the comparison of different organs showing inter- mediate structures more or less out of any context supplied by the taxonomic relationships of the organisms bearing those organs (see Stevens, 1984c). It 1s this kind of continuity that is evident in Linnaeus’s morphological writings. The “‘plenitudo” that he emphasized (Linnaeus, |755a, p. 7) 1s both a simple fullness or doubling of a flower and also a phenomenon that leads to the confirmation that there is continuity or plenitude in the world of pure form. Morphological plenitude in particular interrelated a number of diverse phe- nomena in the world of plant form; unruly form was at least reduced to order, even if not completely understood. Indeed, the ““Linnaean”’ cortex-medulla theory of plant tissues and the physiological-balance theory of the vegetative- floral distinction provided him with a rather full and complex understanding of form. His work on what we would call pure morphology concentrated more on nature’s similarities than on its differences and led him to an apparently satisfying and truly systematic comprehension of the plant world. But Linnaeus was also the quintessential classifier and namer, and here he roceeded in a largely analytical fashion. Yet despite these extensive and time- consuming classificatory studies on which he dealt with the flood of novelties pouring into Uppsala from correspondents around the globe, he attempted the partial synthesis of a world of systematically grouped form as a continuum, at least at the suprageneric level (e.g., Linnaeus, 1737, 1751a; Linnaeus’s corre- spondence with Albrecht von Haller—-see Smith, 1821, Vol. 2; Daudin, 1926).'® As already noted, hybridization—in which the cortex-medulla theory played an important, if poorly understood, role—could generate this continuity. It explained the arrays of similar species being discovered 1n genera like Geranium and Erica and the lack of distinction between genera assembled into natural groups (Linnaeus, 1744 (see Ramsbottom, 1939), 1762b). Ata yet higher taxo- nomic level, the cortex-medulla theory and prolepsis together enabled Linnaeus to demonstrate that a number of animals were like plants in important respects, and so continuity was evident there as well (see TABLE 2). New discoveries in '6Since many plants remained to be discovered, Linnaeus had a ready explanation for apparent gaps. Such gaps are clear in the diagram of relationships of Linnaeus’s ordines naturales (families), believed to be based on one drawn by Linnaeus canny 7 ee 1792). Some families touch, but most are separated by gaps of varying sizes 1990] STEVENS & CULLEN, LINNAEUS 213 the microscopic world hardly disturbed the link between Hydra, with its sen- sitive flowers, and the other Zoophyta, with flowering plants. Volvox chaos was well positioned. The dissolution of form it seemed to suggest in its terminal position in the animal kingdom was not the change from animals to plants, which was a gradual change rather than dissolution anyway, but the loss of form that heralded the discovery ofa new kingdom. In its formlessness, capacity for increase, and activity was evidence of the very principles of life, life in its purest and certainly simplest form—pure medulla (see also Broberg, 1985). The cortex-medulla theory or, more generally, Linnaeus’s ideas on the con- struction of organisms also helped his understanding of how and why organisms could live together in the world (1.e., order between organisms in the living world; it was used to explain a yet higher level of causality) (see, for example, Lindroth, 1983). Many of the properties of plants, including their palatability to insects and, perhaps not surprisingly, their medicinal attributes (Linnaeus, e.g., 1747, 1752; Giseke, 1792),'’ are constant in genera and higher groups in a natural classification. Such properties reside in the medulla, which is at the same time at the heart of natural classifications in general. The close association between plants and insects was evident in that some insects tended to eat plants of different species of the same genus (Linnaeus, 1752; 1760a); perhaps insect larvae could teach us about the medicinal properties of plants (Linnaeus, 1749; see also Linnaeus, 1751b). This close, although rather one-sided, association was strengthened for Linnaeus because the larvae of both groups metamor- phosed to produce the adult, hence the listing of both insects and plants in the thesis Pandora et Flora Rybyensis (Linnaeus, 1771). Such an interlocking of ideas would allow Linnaeus at least partially to sidestep the issue of whether God had created a definite number of species; even if he had not, hybridization might not increase the number of different ecological or functional units in nature. Commenting on the later Linnaeus, Broberg (1985, p. 180) remarked: ““What he needed was a principle which gave causality to creation, not blind faith.” That principle was largely based on the cortex-medulla theo Linnaeus’s world of systematic continuity was different in nature from that of morphological continuity. Although both are to an extent based on the cortex-medulla theory, the latter more directly than the former, they use dif- ferent observations to support the different constructions of continuity. Lin- naeus would surely have approved of the comment that P. J. F. Turpin (1815, p. 429) made when discussing possible interrelationship of inflorescence types: “But why have we need to make abstractions, since nature never fails to show herself all the intergradations which can illuminate us?” (Early in his life in particular, Turpin was strongly influenced by Goethe.) There is little evidence of discord between the two kinds of continuity in Linnaeus’s own work; both stemmed from his observations of nature. As a further complication, genera "The opposition between the secur external, vital cortex and the feminine, internal, animal medulla is the key to Linnaeus’s concept of medicine. Since it also involves an opposition between taste and smell and a hefty dose a aces (the number 5 again), Linnaean medicine is less than easy to understand (Linnaeus, 1766b; see also Hjelt, 1907; Lindroth, 1983) 214 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 and higher taxa that in the taxonomic scheme of things needed to be separated by distinctive and constant features were associated in an arrangement of a kind that allowed the reader to develop ideas of continuity and even change— as also in Charles Bonnet and J. B. A. P. M. de Lamarck (see Lovejoy, 1936— he barely mentioned Linnaeus). Further, prolepsis, metamorphosis, and hy- bridization are more directly subversive ofa world of discrete, separate entities. The tension is however, evident to us, the tension between nature dynamic and flexible and nature ordered rationally, preferably with gaps. Historically, however, there has all too often been overt or covert conflict between systematists and evolutionists (the work of the latter is dependent for a considerable part on the systematic patterns produced by the former) and proponents of pure morphology (see, for example, Cusset, 1982; Wetzels, 1985; Appel, 1987; Brady, 1987; Portmann, 1987). Wetzels (1985, p. 145) wrote eloquently about Goethe’s approach: “It is this delicate simultaneity of the concrete immediacy of individual observation and the equally concrete presence ofa picture ofa whole, that Goethe had described earlier as ‘tender empiricism,’ an empiricism which was capable of making the individual phenomenon trans- parent so that the whole of which it was not so much a part, but a manifestation, became visible.” But of course words such as ““empiricism”’ and the like have different meanings for proponents of these two approaches; there are tensions between the phenomenological Goethean and the somewhat-more-circum- scribed Linnaean approaches to morphology (see above) and more conventional systematics and science (see Amrine & Zucker, 1987; Sattler, 1986). The form of the organism may not have the same meaning or significance for the pro- ponents of the two approaches. Lindroth’s (1983) important essay almost cas- ually captures this dilemma. He emphasized that Linnaeus was preeminently an empiricist, an acute and enthusiastic observer, a voluptuary of nature, and an empirical genius, yet somebody who made no major contributions to science. In his systematic work empiricism was subservient to an overwhelming desire for order. Yet in the cortex-medulla theory an almost “tender empiricism” combines with order because the theory explains why and how the world 1s as it is. As a theory, it proved to be of little interest even to his contemporaries. Linnaeus’s way of establishing order, as Lindroth so clearly demonstrated, was hasty, superficial, and to a high degree inductive, as well as being based on a questionable philosophy of life. The conflict is less evident in Goethe’s work. As is well known, he had earlier found the Linnaean system, and the terms associated with it, intellectually unsatisfying. Goethe disliked counting and analyzing, which he thought that Nature abhorred, but these were needed if he was to use the Linnaean system. Goethe also observed how the same organ varied in shape on a single plant, and to him this suggested problems with terminological categorization (see, e.g., Arber, 1946; Guédés, 1969; Wetzels, 1985). Of course, the Linnaean sexual system was not the most natural arrangement when it came to the delimitation of larger groupings of plants (cf. Stafleu, 1971), although a few groups, such as the Gynandria-Diandra, the Didelphia-Decandra, and the Tetradynamia-Silic- ulosa were more or less composed of related genera (Orchidaceae, Leguminosae, and Cruciferae, respectively: see Linnaeus, 1753). Thus, neither the terms 1990] STEVENS & CULLEN, LINNAEUS 21S Linnaeus used for plant parts nor Linnaeus’s sexual system itself led Goethe to a satisfactory understanding of plant form or diversity. However, his ideas on metamorphosis did just this, albeit in a largely asystematic context. Goethe and Linnaeus drew the data they synthesized into their respective visions of continuity— metamorphosis, prolepsis, a modified scala naturae— from their appreciation of the world of external form. Belief in the continuity of organic form persisted in the systematic community well into the nineteenth century; it was believed, echoing Linnaeus, that if there were gaps in the system of nature, they would eventually be filled. But the similarities go further. Lin- naeus’s valuation of anatomy resonates with the work of A.-L. de Jussieu, a founder of the new “natural method” in systematics. For De Jussieu (1778) the functions of the relatively few plant structures that were used in classification were known; those structures were almost all external, and there was no need to study their anatomy. Of course, his remarks were made in a systematic context, and those of Linnaeus some 20 years before in a largely morphological context, yet both men emphasized the external appearance of the plant as being a suitable object for study, and the two produced natural arrangements that are similar in their basic principles. The world of external form may turn out to be an unreliable guide to both systematics and anatomy, and continuity in one guise or another, or at least reticulating relationships, are the likely results of an analysis of external form. De Jussieu’s natural method is as explicitly based on assumptions of continuity as Linnaeus’s conception of life, and his Genera Plantarum (De Jussieu, 1789) can be analyzed in the same way as Linnaeus’s Vermes, and with similar results. Linnaeus, with a view of life notably archaic or anachronistic even for its time (see, for example, Cain, 1958; Stafleu, 1971; Lindroth, 1983; Hull, 1985), embraced the Aristotelian notion of plenitude and worked it out in the context of a natural arrangement of groups, the characters of which were all to be discerned by the naked eye. The hylomorphic cortex-medulla theory aids in our understanding of this natural arrangement. It provides features that serve to bind the larger units of the arrangement into an indivisible, albeit branching, continuum rather than to separate them, in the process showing clearly that the units are not discrete. It also justified the selection of characters used in classification. In a similar fashion, Linnaeus developed an approach to the analysis of plant form that embraced another manifestation of plenitude, the fundamental similarity and interconvertibility of the appendicular organs of the plant. The rather archaic basis of this thought should not blind us to its evident power in synthesizing a very disparate body of observations. It is noteworthy that the cortex-medulla theory is best developed in plants and fungi. For plants, Linnaeus based his ideas on his own observations of exterior form and on Cesalpino’s seventeenth-century theory, largely ignoring the work of Grew and Malpighi; for fungi, his thoughts were sparked by Miinch- hausen’s disputed findings. In animals, the theory is most evident in Linnaeus’s discussion of the least-understood and smallest organisms, albeit those on which some of the most exciting discoveries of the day were being made. It was most successful where knowledge was least well established, serving to guide him through these areas of uncertainty; the theory colored both his 216 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 observation and his interpretation of other people’s data (see Le Guyader, 1988). As with the “nemesis divina,” ‘reality must—once again—adjust itself to the scheme” (Lindroth, 1983, p. 53). Linnaeus’s theory was epitomized in the name changes of that almost animal, almost plant, almost a member ofa new kingdom: Volvox chaos became Chaos proteus. Chaos itself was a genus of half-misinterpreted observation at the limits of one particular man’s late-eighteenth-century vision. The fate of Linnaeus’s theory was like that of Furia, a new genus described by Linnaeus and placed adjacent to Chaos in the twelfth edition of the Systema Naturae (see Lindroth, 1983). The sole species in the genus, Furia infernalis, was Linnaeus’s contri- bution to the book of imaginary beings; 1t never existed, despite the fact that a vicar, no less, reported that one had fallen onto his plate. The vicar and Linnaeus alike were mistaken, and as we saw above, De Candolle rejected the cortex-medulla theory because monocotyledonous plants simply did not have pith. Similarly, J. G. Kélreuter’s experiments on hybridization (see Roberts, 1929: Mayr, 1986) showed that crosses in which the same species was first the female parent, and then the male parent, tended to produce the same kind of hybrid: there was no maternal dominance; there was not even any hybridiza- tion. Even so, ideas of metamorphosis in particular and change in general are so pervasive in Linnaeus’s later works that their influence should not be entirely discounted. ACKNOWLEDGMENTS We are very grateful to K. Boss, G. Broberg, A. J. Cain, A. H. Dupree, D. Hendler, A. Kabat, J. L. Larson, E. Lord, R. J. O’Hara, D. H. Pfister, S. A. Roe, and R. Turner for comments on the manuscript and/or helpful discussion during its preparation. Our thanks are also due to Ihsan Al-Shehbaz for pre- paring the diagram, to John Lupo for taking the photographs, to the staff of the libraries in the Harvard University Herbaria for their unfailing help, and most especially to Rose Balan and Oanh Tran for typing the manuscript. Finally, we also owe a great deal to the tolerant and helpful audiences that have listened to versions of these complex ideas. LITERATURE CITED ArinsworTH, G.C. 1976. Introduction to the history of mycology. Cambridge University Press, ee England. . ZUCKER. 1987. Postscript. Goethe’s science: an alternative to modern science or within it—or no science at all? Pp. 377-388 in F. AMRINE, F. J. ZUCKER, . WHEELER, eds., Goethe and the sciences: a reappraisal. Riedel, Boston. APPEL, T. 1987. The Cuvier-Geoffroy debate. Oxford University Press, Oxford. ARBER, A. 1941. Nehemiah Grew and Marcello Malpighi. Proc. Linn. Soc. London 153: 218-238. 1946. Goethe’s botany. Chron. Bot. 10: 67- Ames. S. 1990. Foundations of natural history. eae University Press, Cam- bridge, England. Beer, G. 1983. Darwin’s plots: evolutionary narrative in Darwin, George Eliot, and nineteenth century fiction. Routledge, Chapman and Hall, London. 1990] STEVENS & CULLEN, LINNAEUS 217 Biunt, W. 1971. The compleat naturalist: a life of Linnaeus. Viking, New York. BonneT, C. 1745. Traité d’insectologie. Durand, Paris. Brapy, R. H. 1987. Form and cause in Goethe’s morphology. Pp. 257-300 in F. AMRINE, F. J. ZUCKER, & H. WHEELER, eds., Goethe and the sciences: a reappraisal. BREMEKAMP, C. E. B. 1953. Linné’s views on the hierarchy of the taxonomic system. Acta Bot. Neerl. 2: 242-253 BRoBERG, G. 1985. Linné’s systematics and the new natural history discoveries. Pp. one 181 in J. WEINSTOCK, ed., Contemporary perspectives on Linnaeus. University ess of America, Lanham, Maryland. Seeee A. T. 1846. Notice sur Aug. Pyr. de Candolle. Mém. Agric. Soc. Roy. Centr. Agric. [Reprint.] Cain, A. J. 1958. Logic and memory in Linnaeus’s system of taxonomy. Proc. Linn. Soc. London 169: 144-163. 983. paar ere of shells. [Abstract.] Amer. Malac. Bull. 2: 82. CANDOLLE, A. P. bE. 1807. Considérations générales sur les fleurs doubles, et en par- ticular sur ie de la famille des Rénonculacées. Mém. Phys. Chim. Soc. Arceuil 3: 385-404. —. 1813. Théorie élémentaire de la botanique. Déterville, Paris. ; 27. Organographie végétale. 2 vols. Déterville, Paris sae eit L. 1885. Linné’s Anteil an den Lehre von der Metamorphose der Pflanze. t. Jahrb. 6: 146-186. aa A. 1583. De plantis libris XVI. Marescottum, Florenc C C. 1892. Vegetable wasps and plant worms. Society for . pause eas Rena ew ee etropila. E. Sepals soon caducous; fruits glabrous F. Basal leaves numerous; petioles persistent, straw colored, 3-3.5 cm long, about as long as blades, overlapping and forming dense a hs wae Seee gg ate Hal aus aceateeiaia ntti eet hese anne coda W. densifolia. F. Basal leaves few to many; petioles soon caducous or if persistent, then not straw Se usually less than 3 cm long, shorter than blades, not forming dense G. wn. see usually branched above; plants with some dendritic trichomes. _ . parvifolia. G. ae nga! unbranched above; plants without dendritic trichomes. — and cauline leaves ciliate with simple trichomes. “Basal leaves pinnatifid, with 3 to 5 pairs of lateral lobes; in- florescences modified, 1- to 3-flowered, cymelike racemes; pet- als less than 2 mm long; caudices unbranched; stems less than 2 CM IONS: . eke oi) dae soalnen ene anes W. cymosa. Basal leaves entire, rarely dentate or sinuately lobed: inflores- cences usually sonra racemes; petals 2-5.5 mm long; sitet Sh plipatana rade a Gieedee. ba eh Ro anerece ees W. suffruticosa. Basal leaves flat, thin, oblong to oblanceolate or spatulate, rarely linear-lanceolate. K. Styles obsolete or rarely up to 0.6 mm long in fruit. Fruits torulose; infructescences lax racemes; fruit- ing pedicels slender, divaricate, 4—-8(—12) mm long; basal leaves entire, to 1.5 mm wide. ........... Seagisard once inser gue areas eae 13. W. lagunae. L. Fruits smooth; infructescences rae dense sub- umbellate sed, |.5- 4.5(-7) mm m long: basal leaves sual ae 2- 4.5(-6) mm wide. ....... 10. W. colchaguensis. Styles 1-3 mm sone in fruits, if shorter then fruits con- spicuously flatten . Lower a of basal leaves usually with tri- chomes shorter than a e€ on margins or upper surfaces; fruits torulos lly leafless; pet- ioles of basal leaves on swollen. ............ 15. W. imbricatifolia. — rs 1990] AL-SHEHBAZ, WEBERBAUERA 227 M. Lower surfaces of basal leaves glabrous; fruits smooth; stems with few leaves; petioles of basal leaves slender, not swollen. N. Basal leaves entire; petals (3.5—)4-5 mm long; styles (0.8-)1.5-2 mm long in fruit. ........ sate geeegee see oe 12. W. stenophylla. N. Basal leaves dentate; petals 2.5-3.5 mm long; styles 0.5-0.9(—1.1) mm long in fruit. ...... aed She a eee Ll. W. chillanensis. H. Basal and cauline leaves not ciliate, sometimes nie or pu- bescent with at least some branched tri O. Leaves ne saee ates petals 6.5— 3 mm long: caudices thick, 1.5-2 cm in diameter. ................... 4. W. smithii. Leaves a a. nee pubescent or rarely glabrous; petals 2-3.5(-4) mm long; caudices slender, almost always less than 1 cm in diameter. Trichomes minute, 0,03—0.1(-0.15) mm long; basal leaves © P. filiform to narrowly linear; cauline leaves coarsely dentate- SEMALCs vanced xa ara ieee ed 3. W. minutipila. P. Trichomes coarser, (0.2-)0.4—0.7(-1.1) mm long; basal linear; cauline leaves entire to repand or dentate. ...... 1. W. spathulaefolia. 1. Weberbauera spathulaefolia (A. Gray) O. E. Schulz, Pflanzenr. IV. 105(Heft 86): 193. 1924 Sisymbrium spathulaefolium A. Gray, ee Expl. nage Phan. 15(1): 60. 1854. Hes- peris spathulaefolia (A. Gray) Kuntze, Revis. Gen. Pl. 2: 935. 1891. Type: Peru, [Junin,] eas Wilkes Expedition, Soe Ate s name (holotype, us 5526!; isotype, Arabis Se Walp. in Meyen, Observ. Bot. 248. 1843, non A. Ce DC. Syst. Nat. 2: 227. 1821, non Nutt. ex Torrey & A. Gray, Fl. N. A : 81. 1838. TYPE: Peru, Altos de Toledo, Meyven s.n., April 1831 (aot seen). Sisvmbrium orophilum Wedd. Ann. Sci. Nat. Bot. V. 1: 288. 1864. Hesperis orophila Kuntze, Revis. Gen. Pl. 2: 935. 1891. Type: Bolivia, Prov. Larecaja, vicinity ie = pea Anilaya, Juriguana, 4500 m alt., Mandon 914 bis (holotype, P!; D: ,G! Sisymbriume ae Wedd. Ann. Sci. Nat. Bot. V. 1: 289. 1864. Type: Bolivia, rochers de La Laucha, Cordillera de La Paz, aie 5.n., 1851 (holotype, P!). Sisymbrium septaceum Wedd. Ann. Sci. Nat. Bot. V. 1: 289. 1864. Type: Bolivia, Potosi, D’Orbigny 1447 (holotype, P! (photo, rl)). Braya densiflora Muschler, Bot. Jahrb. Syst. 40: 275. ser Weberbauera densiflora (Muschler) Gilg & Muschler, Bot. Jahrb. Syst. 42: 481. 1909. Type: Peru, Hacienda Arapa, Yauli, Lima—Oroya road, aes W ste 304 (lectotype (designated by Macbride, 1938), B!; isolectotype, G Weberbauera spathulaefolia (A. Gray) O. E. Schulz var. integrifolia O. E. Schulz, ies Field Mus. Nat. Hist. Bot. Ser. 8: 80. 1930. Type: Peru, [Depto. Lima,] Rio Blanc in rocks, uplands, 8-19 May 1922, 15,000 ft [4572 m] alt., Machride & nny &I1 (holotype, F! (photos, F!, G!, GH!); isotype, GH!). Caudices simple or sometimes branched, slender, less than 1 cm in diameter, usually covered with petiolar remains of previous years. Stems usually decum- bent, unbranched, (2—)4—23(-—43) cm long. Trichomes (0.2-)0.4—0.7(-1.1) mm 228 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 long, short stalked and submalpighiaceous, or long stalked and furcate, some- times simple, rarely absent. Basal leaves petiolate, blades oblong or lanceolate to spatulate, rarely linear to ovate, (1-)2—6(—7) cm long, (2-)3—9(-11) mm wide, obtuse or rarely acute at apex, entire or repand to dentate, rarely lyrately pinnatifid, thin and nonfleshy, pubescent or glabrous; petioles (0.5-)1—4.5 cm long. Cauline leaves subsessile to short petiolate; blades oblong to obovate or lanceolate, 5—13(-—18) mm long, 2—8(—1 5) mm wide, entire to repand or dentate. Inflorescences bracteate to ebracteate; infructescences lax to compact, (0.5-) 1-9(-17) cm long. Sepals caducous, oblong to ovate, (1.5)2—2.5(-3) mm long, 1-1.4(-1.7) mm wide, obtuse, scarious at margin, sparsely pubescent to glabrous. Petals white, spatulate, attenuate to clawlike base, 2-3.5(—4) mm long, 1.5-1.8(-2) mm wide. Filaments white, erect, 1.5-2.5 mm long, somewhat dilated at base; anthers oblong to ovate, 0.4-0.7 mm long. Fruiting pedicels ascending, subappressed at base, straight to curved, (2-)3—6(-8) mm long. Fruits divaricate to erect and subappressed to rachis, linear to oblong, terete, non- torulose, usually abruptly ending in style, straight or rarely curved, (6—-)8-15(- 20) mm long, 1.5-1.8(-2) mm wide; valves smooth, glabrous, conspicuously to obscurely veined; septa hyaline; styles 0.2-0.8(-1.5) mm long; stigmas entire. Seeds subbiseriately arranged, oblong to ovate, (0.8-)1-1.4 mm long, (0.5-)0.6— 0.8 mm wide, light to dark brown; funicles usually strongly differentiated into persistent broad base and filiform distal portion. Flowering late December to mid-February, fruiting late February to mid- May. Growing on sand or clay in puna-grassland, boggy areas, valley bottoms, calcareous cliffs, dry steep slopes, and rocky places at altitudes of 3600- 4800 m. REPRESENTATIVE SPECIMENS EXAMINED. Argentina. Prov. CATAMARCA: Depto. Andalgala, Cerro Yutuyaco, S/eumer 2721 (LiL); Rio Potrero, Sleumer 1905 (B). PRov. Jusuy: Depto. Humahuaca, Mina Aguilar, S/ewmer 3402 (LIL). Prov. La Rioja: Sierra Famatina, Cueva de Pérez, Hieronymus & Niederlein 376 (8). Bolivia. Depro. LA Paz: Prov. Aroma, raco Pendiente, Fisel U- 188 (Gu), U-472 (cps); Prov. Larecaja, vicinity of Combaya, Mandon 914 (gM, G, P); Prov. Murillo, La Cumbre, on road to Unduavi, Solomon 5029 (mo); ca. 15 km NNE of La Paz, Beck 9130 (GH); 17 km SE of Collana on La Paz- Calacoto road, Beck 4290 (Gu); Prov. Pacajes, Charafia, Asplund 2664 (s, ups), 6201 (us); Corocoro, Asplund 2418 (s, ups), Panacachi, pos 2582 (s). DEPTO. Potosi: Prov. Frias, Cerro Potosi, Petersen & Hyerting 1030 (c, “drdenas 398 (us). Peru. DEPTO. ANCASH: Prov. Bolognesi, between Tallenga and eae aque, Cerrate 749 (GH); Prov. Carhuas, Huascaran National Park [HNP], Quebrada Ishincea, Smith, Valencia, & Gon- zales 9440 (mo); Prov. Huaras, HNP, Quebrada Shallap, Smith, Valencia, & Gonzales 9670 (Mo): Sah fee HNP, Quebrada Los Cedros, Smith, Valencia, & Minaya 9924 (GH, MO); Prov. Recuay, HNP, pass between Nevado Pasto Ruri and Nevado Raria, Smith & oa 10182 (mo): Quebrada Quena Ragra, Smith, Valencia, & Torres 11730 (Mo); Quebrada Queshque, Smith, Valencia, & Torres 11845 (GH, Mo); Rio see drainage, 15 mi from highway, Smith, Stein, & Todzia 9373 (GH, MO). DEPTO. HUANC. Lica: Prov. Huancavelica, Bunbunya, Tovar 2/9 (GH). Depro. JUNiN: between Cerro ¢ - Pasco and La Quinua, Asplund 11871 (s); Morococha, Haapala s.n., 15 Feb. 1949 (H); Ondores, Pattersson 293 (s); vicinity of Oroya, Kalenborn 132 (GH, Ny), [32a (us); between Tarma and La Oroya, Weberbauer 2550 (B). Depto. Lima: Rio Blanco, Macbride 2991] (F, NY); Saltacuna, Soukup 1940 (us); Ticlio Bajo, Diers 979 (GH); Visco, Macbride & Featherstone 590 (F, G, Ny); Prov. Huarochiri, Casapalca, Asplund 11425 (s), Prov. co 1990] AL-SHEHBAZ, WEBERBAUERA 229 Yauyos, Huacrococha, 17 km to Tupe, Cerrate 1226 (GH). Depro. MoqueGua: Prov. Moquegua, cordillera above Torata, W ee 7471 (BM, F, G, US). DEPTO. PAsco: Prov. Cerro, Cerro de Pasco, Asplund 11779 (s), Macbride 3065 (cas, F, Us). DEPTO. Puno: Prov. Carabaya, Antapampa, eae 6837 (F Weberbauera spathulaefolia is one of the most variable South American crucifers. The variation is most noticeable in length of infructescences, occur- rence of bracts, type and density of trichomes, and shape and margin of leaves. The infructescences are usually racemes a few to several centimeters in length, but in Asplund 11779 they are subumbellate and only 0.5 cm long, and in Cerrate 749 they are lax and to 17 cm. Perhaps the most significant variation is in the length of the bracteate portion of the inflorescence. In many populations the inflorescences are either bracteate throughout or ebracteate, while in others only the lowermost portion of the racemes are bracteate. In some collections (e.g., Asplund 2418, Machbride 2991) both ebracteate and bracteate plants are found within the same population. As indicated above, the holotype of W. spathulaefolia has two plants; the raceme of one plant is ebracteate, while that of the other is bracteate only on the lower half. Evidently, the occurrence of bracts 1s not a good character for the generic delimitation of Weberbauera. Stem and leaf pubescence of Weberbauera spathulaefolia is also highly vari- able. Glabrescent forms, as well as forms with simple or asymmetrically furcate trichomes, are widespread in specimens from throughout the range of the species. In contrast, plants with submalpighiaceous trichomes are apparently more common in the northern portions of the species range than elsewhere. However, in the majority of collections more than one trichome type is found, and in some (e.g., Mandon 9/4) both glabrous and densely pubescent plants are present. Variation in the length of infructescences, the occurrence of bracts, and the pubescence of stems and leaves does not correlate with the geography of We- berbauera spathulaefolia. The species cannot be subdivided morphologically into practical subordinate taxa. 2. Weberbauera densifolia Al-Shehbaz, sp. nov. FIGURE 2. Herba perennis caespitosa, radicibus longibus crassibus. Folia basales nu- 5 iqua anguste oblonga, glabra, subtorulosa, paucisperma, 5-6 mm longa, 1|.2- 1.4 mm lata. Semina oblonga, uniseriata, ca. 1.3 x 0.7 mm longa; cotyledones incumbentes. Deep-rooted, cespitose perennial. Main roots thick, more than 15 cm long, 3-5 mm wide. Caudices short; stems decumbent, 4-6 cm long, glabrous. Basal leaves numerous, rosulate, petiolate; blades lanceolate, 2-3 cm long, 4-6 mm wide, obtuse to subacute at apex, attenuate at base, pinnatifid to sinuate or repand-dentate, pubescent with straight, simple or asymmetrically furcate tri- chomes 0.2-0.4 mm long; petioles persistent, erect to spreading, densely over- 230 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 URE 2. Weberbauera densifolia (holotype): a, plant; b, basal leaf; c, portion of infructescence with fruit; d, seed. Scale bars = 1 cm (a, b), | mm (c, d). lapping, straight, 3-3.5 cm long, 2-3 mm wide, straw colored and becoming papery with age, glabrous. Cauline leaves few, petiolate; blades 8-11 mm long, 4-6 mm wide, repand to entire. Flowers not seen. Infructescences ebracteate, glabrous; fruiting pedicels erect to only slightly ascending, subappressed, straight, 3-5 mm long. Fruits narrowly oblong, subtorulose, 5-6 mm long, 1.2-1.4 mm wide, glabrous, few seeded; septa complete, membranaceous; styles 0.5—0.7 1990] AL-SHEHBAZ, WEBERBAUERA 231 mm long. Seeds uniseriately arranged, oblong, ca. 1.3 mm long, 0.7 mm wide, obscurely reticulate; cotyledons incumbent. Type. Argentina, Prov. Catamarca, Depto. Santa Maria, Sierra del Aconquija, 4600 m alt., 20 Feb. 1925, S. Venturi 66/0 (holotype, Us (photocopy, GH)). Weberbauera densifolia is easily distinguished from the other species of the genus by its numerous, densely overlapping leaves with persistent, straight, straw-colored petioles 3—-3.5 cm long and about as long as the leaf blade. Other species of Weberbauera have few petioles that are much shorter than the blades and that do not form dense crowns. Furthermore, W. densifolia is also unique in the genus in having deeply penetrating, thick, few-branched main roots that are more than 15 cm long. Weberbauera densifolia superficially resembles W. spathulaefolia in several aspects of leaf blade and fruit morphology. The two are apparently related, but such a relationship cannot be fully assessed without adequate material of the former 3. Weberbauera minutipila Al-Shehbaz, sp. nov. FIGURE 3. Herba perennis tenella, caudicibus tenuibus, caulibus decumbentibus 2-7 cm longis, sparse vel dense pubescentibus, pilis minute furcatis vel simplicibus 0.03-0.1(-—0.13) mm longis. Folia basales petiolata filiformes vel anguste linea- res, Integra vel sparse dentata, longitudinaliter plicata vel plana (1—)2—3(-4.5) cm longa; petiolis complanatis ad basim expansis persistentibus; folia caulina ovata vel lanceolata dentato-serrata subsessilia (3—)4—8(—10) mm longa. Racemi ebracteati. Sepala erecta oblonga |.5—2 mm longa; petala spathulata alba 2- 2.7 mm longa. Pedicelli fructiferi recti subappressi 2.5—3.5(—5) mm longi. Sili- qua anguste oblonga glabra subtorulosa 6-9 mm longa; stylus 0.2-0.3 mm longus. Semina uniseriata ovata 3 vel 4 in loculo. Delicate perennial herbs. Caudices simple, slender, covered with petiolar bases of previous years. Stems decumbent to rarely ascending, simple, 2—7 cm long, sparsely to densely covered with minute, furcate or simple trichomes 0.03-0.1(-—0.13) mm long, sometimes glabrescent. Basal leaves petiolate; blades filiform to narrowly linear, rarely linear-lanceolate, (1—)2—3(—4.5) cm long, 0.5- 1.5(—2.5) mm wide, acute at apex, entire or rarely dentate, longitudinally plicate to flat, pubescent to glabrescent; petioles persistent, (3—)5—8(—10) mm long, flattened, conspicuously expanded at base, straw colored, glabrous. Cauline leaves short petiolate to subsessile; blades ovate to lanceolate, (3—)4—8(—10) mm long, 1.5—-3.5(-5) mm wide, dentate-serrate or rarely sublaciniate. Inflo- rescences ebracteate racemes, slightly elongated in fruit. Sepals caducous, erect, oblong, |.5-2 mm long, 0.7—1 mm wide, rounded at apex, nonsaccate, narrowly scarious at margin, drying lavender, glabrous or only sparsely pubescent below apex. Petals white, spatulate, 2—-2.7 mm long, 1-1.2 mm wide, rounded at apex, attenuate to clawlike base. Nectar glands obscure, ringlike, subtending bases of filaments. Filaments white, erect, 1.1-1.5 mm long; anthers oblong, 0.4—0.5 mm long. Fruiting pedicels somewhat ascending, subappressed to rachis, straight, 2.5—3.5(-5) mm long, glabrous. Fruits narrowly oblong, subtorulose, 6-9 mm pe) JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Figure 3. Weberbauera minutipila (holotype): a, plant; b, flower with | sepal and 1 petal removed; c, sepal; d, petal. Scale bars = 1 cm (a), 1 mm (b-d). long, ca. | mm wide, glabrous; styles 0.2-0.3 mm long. Seeds 3 or 4 per locule, uniseriately arranged, ovate, ca. | mm long, 0.7 mm wide, brown; cotyledons incumbent; funicles slightly expanded at placental end. Type. Peru, Depto. Puno, Prov. San Roman, “‘shale barren” with rock-chip covered bare ground, on bare hills on Puno—Arequipa road at km 112.8, ca. 4 km (air) E of Tinocopalca (km 119.5), ca. 10 km W of road-turnoff to Sta. Lucia (at km 101), ca. 4000 m alt., 12 Jan. 1963, H. H. & C. M. IItis and D. & V. Ugent 1455 (holotype, GH; isotype, wIs). ADDITIONAL SPECIMENS EXAMINED. Peru. DepTo. AREQUIPA: Prov. Arequipa, S of Sta. Lucia on road from Puno to Arequipa, ca. 4—5 km E of Sta. Lucia, //tis & Ugent 1415 (wis). Bolivia. Depto. La Paz: Prov. Murillo, 14 km N of La Paz, Mina Milluni, 15 km hacia Tuni Condoriri, Beck 3832 (Gu), Valle Chuquiaguillo, Asplund 1888 (s). Weberbauera minutipila is distinguished from the other species of the genus in having minute trichomes rarely to 0.13 mm long (see FiGure Ic), filiform to linear basal leaves 0.5—1.5(—2.5) mm wide, usually coarsely dentate-serrate, ovate to lanceolate cauline leaves, and ebracteate inflorescences. Both Weberbauera minutipila and W. spathulaefolia grow in Prov. Murillo, Bolivia. It is not known, however, if the two species grow sympatrically. 1990] AL-SHEHBAZ, WEBERBAUERA 233 a e me) SR) Steet ert bag qees, o> ie xa Ss Tee Ficure 4. Weberbauera smithti (holotype): a, plant; b, flower; c, petal; d, stamen. Scale bars = | cm (a), 1 mm (b-d). 4. Weberbauera smithii Al-Shehbaz, sp. nov. FIGURE 4. Herba glabra perennis caespitosa, caudicibus lignosis crassis 1.5—2 cm latis, caulibus decumbentibus 4—7 cm longis. Folia basales rosulata, late spathulata, succulenta, glabra, integra, apice obtusa, base attenuata, 2-4.5 cm longa, 0.6- 1.3 cm lata. Racemi ebracteati. Sepala erecta, oblonga, glabra, decidua, 4.2- 5.5 mm longa, 2.2—2.6 mm lata; petala alba, late spathulata, 6.5-8 mm longa, 2.8-3 mm lata. Pedicelli fructiferi, recti, 6-10 mm longi, divaricati-adscen- dentes. Siliqua oblonga vel linearis, obtusa 0.5-1.5 cm longa, 1.5—2 mm lata, valvis crassibus, glabris; stylus 1-1.5 mm longus. Semina matura ignota. Perennial, glabrous, cespitose herbs with unbranched, thick, woody caudices 1.5—2 cm in diameter. Stems few, decumbent, 4-7 cm long. Basal leaves rosu- late, fleshy; blades broadly spatulate, 2-4.5 cm long, 6-13 mm wide, obtuse, attenuate at base, entire, nonciliate, glabrous; petioles thick, 1-2.5 cm long. Cauline leaves subsessile; blades oblanceolate, 1-1.8 cm long, 3-5 mm wide, subacute, attenuate to broad base, entire, glabrous or with few, simple, subapical trichomes. Inflorescences ebracteate racemes, elongated in fruit. Sepals ca- ducous, erect, oblong, 4.2—5.5 mm long, 2.2—2.6 mm wide, obtuse, nonsaccate, narrowly scarious at margin, glabrous. Petals creamy white, broadly spatulate, 6.5-8 mm long, 2.8-3 mm wide, rounded at apex, attenuate to broad, clawlike base. Nectar glands confluent, low, ringlike, subtending bases of filaments. Sta- mens slightly tetradynamous; filaments white, erect, slender, 3.5-4 mm long; 234 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 anthers oblong, 1-2 mm long. Fruiting pedicels divaricate-ascending, straight, 6-10 mm long. Fruits oblong to linear, 0.5—1.5 cm long, 1.5-2 mm wide, obtuse at apex, abruptly ending in style; valves thick, glabrous, with obscure midvein; styles 1-1.5 mm long. Mature seeds not seen. Growing in cracks of exposed rocks or among rock outcrops in grasslands and margins of bare needle-ice zones at altitudes of 4770-4870 m. Type. Peru, Depto. Ancash, Prov. Yungay, Huascaran National Park, Llan- ganuco Sector, Quebrada Ancosh at Portachuelo (77°35'W, 9°03’S), 31 Dec. 1984, D. N. Smith & K. Goodwin 8894 (holotype, Mo). ADDITIONAL SPECIMEN EXAMINED. Peru: Same as type locality, Smith 11298A (Mo). Weberbauera smithii, named after one of it collectors, is easily distinguished from the remainder of the genus by its fleshy, glabrous leaves, thick, unbranched caudices, considerably larger flowers with petals 6.5-8 mm long, and rounded fruit apices abruptly ending in a distinct style. Nothing can be said about the variability of this handsome species because only two plants were available for study. Field notes of the two collections above indicate that the species has purple petioles, sepals, and young fruits. 5. Weberbauera retropila Al-Shehbaz, sp. nov. FIGURE 5. Herba perennis, caudicibus tenuibus ca. 2 mm latis, caulibus simplicibus adscendentes vel subdecumbentes, 2.5—5.5 cm longi. Folia basales haud rosu- lata, petiolata, late spathulata vel oblanceolata, dentata, 1.3—2.5 cm longa, 4— 7 mm lata, sparse pubescentibus, pilis furcatis. Racemi ebracteati. Sepala erecta, oblonga, persistens, purpurea, 2—2.5 mm longa, I-1.2 mm lata. Petala alba, spathulata, attenuata, 3-3.5 mm longa, I-1.5 mm lata. Pedicelli fructiferi 2- 3.5(-6) mm longi. Siliqua oblonga, subacuta, 5—8(—10) mm longa, 1.5-1.8 mm lata, valvis sparse pubescentibus pilis retroris simplicibus 0.1-—0.25 mm longis; stylus 0.5-0.8 mm longus. Semina uniseriata ovata 1.4-1.5 mm longa. Perennial herbs with slender rootstocks ca. 2 mm wide. Stems ascending to subdecumbent, simple, 2.5—5.5 cm long, glabrous. Basal leaves not rosulate, petiolate, caducous, often spatulate to oblanceolate, 1.3-2.5 cm long, 4-7 mm wide, dentate, sparsely pubescent with stalked, Y- or T-shaped furcate tri- chomes; petioles to | cm long. Cauline leaves similar to basal ones but smaller. Inflorescences ebracteate racemes, slightly elongated in fruit. Sepals persistent, erect, oblong, 2—2.5 mm long, I—1.2 mm wide, obtuse at apex, membranaceous at margin, purplish, glabrous or rarely with few subapical trichomes. Petals white, spatulate, 3-3.5 mm long, 1-1.5 mm wide, attenuate to short, clawlike base. Nectar glands ringlike, poorly developed outside bases of filaments. Fil- aments white, erect, slender, |.5-1.8 mm long; anthers ovate, 0.4—-0.5 mm long, apiculate. Fruiting pedicels ascending, straight, 2-3.5(-6) mm long, glabrous. Fruits narrowly oblong, slightly torulose, 5-8(-10) mm long, |.5—1.8 mm wide, 0.8 mm long; stigmas entire. Seeds usually 3 or 4 per locule, uniseriately 1990] AL-SHEHBAZ, WEBERBAUERA 235 d ; oA Schley GURE 5. Weberbauera retropila (holotype): a, plant; b, sepal: c, petal; d, stamen; e, fruit and fruiting pedicel; f, seed. Scale bars = 1 cm (a), 1 mm (b-f). arranged, ovate, 1.4—-1.5 mm long, 0.9-1 mm wide, dark brown; funicles slightly expanded proximally, slender distally; cotyledons incumbent. Type. Bolivia, Depto. La Paz, Prov. Pacajes, Charafia, 4000 m alt., Asplund s.n., 2 March 1921 (holotype, us). ADDITIONAL SPECIMEN EXAMINED. Bolivia: same as type locality, ca. 4050 m alt., Asplund 2700 (s) Despite the altitudinal differences between the two collections cited above, it is likely that both were made from the same population. Comparisons of Asplund’s collections of several other species show that labels of his duplicates at us usually differ from those of the original set at s in only a few minor details. Therefore, the specimen above at s might well be an isotype. Weberbauera retropila is easily distinguished in having pubescent fruits with retrorse simple trichomes, persistent calyces, slender rootstocks, fewer (three or four) seeds per locule, and sparsely pubescent leaves with Y- or T-shaped, furcate trichomes (see FiGurE Id). Because the rootstocks are very slender and the basal leaves are caducous, the species could easily be mistaken for an annual. This aspect of habit and the presence of persistent sepals make W. retropila 236 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 somewhat anomalous in the genus. However, all other features of the species are characteristic of Weberbauera. 6. Weberbauera herzogii (O. E. Schulz) Al-Shehbaz, comb. nov. Sarcodraba herzogii O. E. Schulz, Notizbl. Bot. Gart. Berlin-Dahlem 10: 563. 1929. Type: Bolivia, auf erde zwischen Felsplatten der ‘ica iiber dem Titicacasee bei Guaqui, ca. 3900 m, 7. Herzog 2510 (holotype, B! (photos, F!, GH!, Mo!, Ny!); isotypes, G! (photos, F!, Mo!), z!). Caudices slender, simple or rarely few branched. Stems decumbent to erect, unbranched, (1.5—)3—12 cm long, hirsute with spreading to reflexed simple tri- chomes 0.4—-1 mm long, rarely glabrous or glabrescent. Basal leaves petiolate; blades broadly obovate to spatulate, (1-)1.5-3.5 cm long, (4-)5—10 mm wide, rounded at apex, entire to repand or rarely dentate, usually ciliate; petioles ciliate, (0.5-)0.7—2 cm long. Cauline leaves 0.3-1.1(-—2) cm long, 2—-5(-10) mm wide, dentate to repand or entire. Inflorescences bracteate racemes, elongated or rarely unexpanded in fruit; bracts 3-toothed to entire. Sepals caducous, oblong to ovate, (2.2-)3-4 mm long, |.5—2 mm wide, obtuse, scarious at margin, glabrous to sparsely hirsute. Petals white to lavender, spatulate to broadly obovate, 3.5—5(-7) mm long, 1.5—2.5 mm wide. Filaments white, erect, slender, 2.2-4 mm long; anthers oblong, 0.6-0.9 mm long. Fruiting pedicels ascending to divaricate, usually straight, (3—)4-10 mm long. Fruits linear to oblong, terete, nontorulose, (S—)8—18 mm long, |.6—2.1 mm wide, usually attenuate to slender style; valves glabrous, with prominent to obscure midvein; septa hyaline; styles slender, (0.8—)1.2—2.5 mm long; receptacles 4-angled, usually expanded. Seeds uniseriately arranged, oblong, 1.5—1.8 mm long, 0.8-1 mm wide, light brown; funicles strongly differentiated into broad, persistent base and filiform distal portion. Flowering early December through February; growing at altitudes of 3125- 4700 m REPRESENTATIVE SPECIMENS EXAMINED. Argentina. Prov. Jusuy: Depto. Humahuaca Esquinas Blancas, between Tres Cruces and Humahuaca, Ruthsatz 13/18 (Gu). Bolivia. Depto. La Paz: 5 km E of Villa Santa Fe (ca. 100 km N of Oruro), Conrad 2692 (mo); 2 km W of Villa Santa Fe, Conrad 2709 (mo); Prov. Ingavi, Guaqui, Asplund 2226 (s). oTost: Orocoro, near Ventilla, Ceballos, Charpin, Fernandez Casas, & Valdés- Bermejo 252 (Gc); Potosi, Cardenas 173 (GH), 399 (us); Miraflos, D’Orbigny 1347 (Pp). Peru. Depto. Puno, Puno, lakes te a us), Soukup 106 (F); Santa Lucia, Sharpe 143 (F); without rales locality, Gay 2256 (Pp). Weberbauera herzogii, which is often confused with W. spathulaefolia, is easily distinguished by its ciliate basal leaves, exclusively simple trichomes, inflorescences bracteate along their entire length, large (1.5-1.8 mm) seeds, and attenuate fruits that terminate in slender styles (0.8—-)1.2—2.5 mm long. The closely related W. spathulaefolia has nonciliate leaves, furcate trichomes some- times mixed with simple ones (or the plant is glabrous), ebracteate or partially bracteate (rarely completely bracteate) inflorescences, smaller ((0.8—)1-1.4 mm long) seeds, and usually blunt fruits with styles almost always less than | mm long. 1990] AL-SHEHBAZ, WEBERBAUERA Zo) Both Weberbauera herzogii and W. spathulaefolia grow in Depto. Potosi, Bolivia. Cardenas 399, cited under the former species, and Cardenas 398, cited under the latter, were probably collected from the same general area. It is not known, however, if the two species are actually sympatric. In a few collections (e.g., Cardenas 173, 399) the rachises of the central infructescences do not elongate, and the fruits appear as if borne on short scapes. However, this abnormality is insignificant because it is found within certain collections (e.g., Ceballos et al. 252) that otherwise have normal in- fructescences. The placement of Weberbauera herzogii in Sarcodraba by Schulz (1929) was not based on a careful evaluation of the boundaries of these genera. Schulz (1924, 1936) stated that Sarcodraba differs from Weberbauera mainly in having silicles instead of siliques. Apparently, he did not examine any fruiting material of W. herzogii. The species is so closely related to W. spathulaefolia that the two have probably evolved from a common ancestor. 7. Weberbauera bracteata (O. E. Schulz) J. F. Macbr. Candollea 5: 356. 1934. Pelagatia bracteata O. E. Schulz, Pflanzenr. IV. 105(Heft 86): 192. 1924. Type: Peru, Depto. Ancash, Prov. Pallasca, Cordillera of Pelagatos, 4600 m alt., 23 Jan. 1920, A. Weberbauer 7234 (holotype, B! (photos, F!, GH!, Mo!, Ny!): isotypes, F(2 sheets)!, G!). Small, cespitose perennials. Caudices simple, thick, 6-12 mm wide. Stems decumbent, 3-6 cm long, glabrous. Basal leaves rosulate, petiolate; blades oblanceolate to spatulate, 1.5—4 cm long, 4-9 mm wide, lyrate-pinnatifid, with 2 to 4 lateral lobes, rarely repand or entire, ciliate with simple, straight tri- chomes 0.3-0.9 mm long, glabrous on both surfaces or sometimes pubescent on upper. Cauline leaves petiolate; blades obovate to spatulate, 7-13 mm long, 2-4.5 mm wide, gradually reduced in size upward, rounded at apex, entire or rarely repand, ciliate. Inflorescences corymbose racemes, bracteate throughout, elongated in fruit. Sepals caducous, erect, oblong to ovate, 1.5-2.2 mm long, 1-1.1 mm wide, rounded at apex, nonsaccate at base, membranaceous at mar- gin, glabrous. Petals white, spatulate, 2.5-3 mm long, to | mm wide, rounded at apex, attenuate to clawlike base. Filaments white, erect, 1.5-2 mm long; anthers oblong, 0.6-0.7 mm long. Fruiting pedicels 3-7 mm long. Fruits as- cending, oblong, terete, straight, nontorulose, 6-9 mm long, |.5-2 mm wide, obtuse at apex; valves glabrous, with conspicuous midvein; styles to 0.2 mm long. Seeds 3 to 5 per locule, uniseriately arranged, oblong, 1.4—1.6 mm long, 0.6-0.7 mm wide, brown; funicles broad at proximal end. Weberbauera bracteata is known only from the type collection. It is most closely related to W. herzogii, from which 11 is distinguished in having glabrous stems, lyrately pinnatifid basal leaves, small petals to 3 mm long, obtuse fruits, and minute styles to 0.2 mm long. Weberbauera herzogii usually has retrorsely pubescent stems, entire to dentate basal leaves, larger (3.5-5(-7) mm long) petals, attenuate fruits, and conspicuous styles (0.8—)1.2—2.5 mm long. The differences between W”. bracteata and the closely related W. cymosa are given under the latter. 238 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 f weer ee aS Je FiGure 6, Weberbauera cymosa: a, plant; b, basal leaf, c, cauline leaf; d, fruit. Scale bars = 1 cm (a), | mm (b-4). As indicated above, Schulz’s (1924, 1936) placement of Weberbauera brac- teata in the monotypic Pelagatia is not supported by any morphological evi- dence. His allegation (1924; see key, p. 181) that Pe/agatia differs from We- berbauera in having flattened instead of slender funicles 1s erroneous because 1990] AL-SHEHBAZ, WEBERBAUERA 239 almost all plants that I have examined of W. spathulaefolia, the generic type, have flattened funicles. 8. Weberbauera cymosa Al-Shehbaz, sp. nov. FIGURE 6. Herba caespitosa perennis tenella, caudicibus crassis, caulibus decumben- tibus, glabris, tenuibus, 1-1.7 cm longis. Folia basales petiolata, pinnatifida, rosulata, 5-15 mm longa, 2-3 mm lata, lobis lateralibus 3- vel 5-jugis, oblongis vel ovatis, ciliatis, 0.8-1.2 mm longis, 0.4-0.6 mm latis; folia caulina anguste oblonga, sessiles, ciliata, integra, 2.5—4 mm longa. Inflorescentia ebracteata, 2- vel 3-floribus, dichasialium plus minusve faciens. Sepala oblonga, erecta, non- saccata, glabra, |.4-1.6 mm longa. Petala alba, spathulata, ca. 1.4 mm longa. Pedicelli fructiferi recti 2.5-S mm longi. Siliqua oblonga, obtusa, glabra, 4-5.5 mm longa; stylus obsoletus usque ad 0.3 mm longus. Semina immatura uniseri- ata, ovata, ca. 1.2 x 0.7 mm, funiculis filiformibus. Small, cespitose perennials. Caudices simple, thick, 4-6 mm wide. Stems decumbent, slender, |-1.7 cm long, glabrous. Basal leaves rosulate, petiolate; blades ovate to lanceolate, S—15 mm long, 2-3 mm wide, pinnatifid, the lateral lobes in 3 to 5 pairs, oblong to ovate, 0.8-1.2 mm long, 0.4-0.6 mm wide, obtuse, ciliate with simple trichomes 0.3-0.5 mm long; petioles 3-8 mm long, glabrous. Cauline leaves few, sessile; blades narrowly oblong, 2.5-4 mm long, Q.8-1.3 mm wide, obtuse to rounded, entire, ciliate. Inflorescences ebracteate, highly modified, dichasiumlike or 2- or 1-flowered racemes. Sepals erect, ob- long, 1.4-1.6 mm long, ca. 0.7 mm wide, rounded at apex, glabrous. Petals white, spatulate, ca. 1.4 mm long. Filaments erect, ca. 1.3 mm long; anthers oblong, 0.3-0.4 mm long. Fruiting pedicels straight, 2.5-5 mm long. Fruits oblong, terete, 4-5.5 mm long, ca. 1.5 mm wide, obtuse at apex, glabrous; styles obsolete to 0.3 mm long; stigmas entire. Seeds (immature) 4 or 5 per locule, uniseriately arranged, ovate, ca. 1.2 mm long, 0.7 mm wide, brown; funicles filiform. Type: Bolivia, Depto. La Paz, Murillo, near Palca, base of Illimani, 4800-5000 malt., 25 Feb. 1979, 4. Ceballos, A. Charpin, J. Ferndndez Casas, & E. Valdés- Bermejo 543 (holotype, c). Weberbauera cymosa is known only from the type collection. It is most closely related to W. bracteata, which it resembles in having ciliate, pinnatifid, rosulate basal leaves, ciliate, entire cauline leaves, oblong fruits with minute styles, and few, uniseriately arranged seeds. However, it is easily distinguished by its ebracteate, two- or three-flowered, cymelike, modified racemes, sessile cauline leaves, and smaller basal leaves. Weberbauera bracteata has bracteate racemes, petiolate cauline leaves, and larger basal leaves. Weberbauera cymosa resembles W. trichocarpa in having highly modified, cymelike racemes, but the two appear to have little else in common. In fact, W. cymosa is probably a sister species of W. bracteata, whereas W. trichocarpa is closer to W. spathulaefolia than to any other species of the genus. It is most likely, therefore, that cymelike racemes evolved independently in those two species of Weberbauera. 240 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 9. Weberbauera trichocarpa (Muschler) J. F. Macbr. Candollea 5: 356. 1934. Eudema trichocarpum Muschler, Bot. Jahrb. Syst. 40: 276. 1908. Brayopsis trichocarpa Muschler) Gilg & Muschler, Bot. Jahrb. Syst. 42: 484. 1909. Alpaminia trichocarpa (Muschler) O. E. Schulz, Pflanzenr. IV. 105(Heft 86): 191. 1924. Type: Peru, above Lima, near Alpamina (as Alpamia in the original publication), 4500 m alt., 2 March 1904, A. Weberbauer 5119 (holotype, B!). Small, cespitose perennials, densely covered throughout with appressed, ses- sile, malpighiaceous trichomes 0.5—1.1 mm long, these oriented parallel to long axes of organs they cover. Caudices simple or branched, slender, usually covered with petiolar remains of previous years. Stems decumbent, rarely ascending or procumbent, unbranched, terete, (1—)2—4.5(-6) cm long. Basal leaves petiolate; blades linear to linear-lanceolate, (1-)2—4(-4.5) cm long, (1.5-)2-3.5 mm wide, acute, attenuate at base, entire, conduplicate or tarely flat, densely pubescent on both surfaces; petioles persistent, flattened, 5-10(-13) mm long, to 4 mm wide at base, somewhat membranaceous at margin. Cauline leaves short pet- iolate: blades ovate to lanceolate, 5—-9(-11) mm long, 1.5-3(—3.5) mm wide, entire. Inflorescences cymelike, umbellate, usually 3-flowered racemes, often subtended at base by leafy bracts. Sepals caducous to persistent, erect, oblong, 3.5-4 mm long, 1.5-2 mm wide, obtuse at apex, nonsaccate, narrowly scarious at margin, densely pubescent on outside. Petals pink to pinkish-yellow, spat- ulate, 5-6.5 mm long, |.5-2.2 mm wide, rounded at apex, attenuate to pu- bescent claw. Filaments white, erect, 3-4 mm long; anthers narrowly oblong, 0.8-1.1 mm long, minutely apiculate. Fruiting pedicels straight, (6-)7—15(-17) mm long, densely pubescent. Fruits oblong to linear, terete, straight, 7-1 3(- 15) mm long, 1.7-2.4(-2.6) mm wide, obtuse at both ends; valves densely pubescent, with obscure midvein; septa hyaline; styles (0.4-)0.6-0.8(-1) mm long, glabrous. Seeds numerous, biseriately arranged in each locule, oblong to ovate, 1-1.2 mm long, 0.6-0.7 mm wide, light to dark brown; funicles flattened proximally, filiform distally. Endemic to Peru, where it grows on limestone slopes and cliffs at altitudes of 4200-4800 m. REPRESENTATIVE SPECIMENS EXAMINED. Peru. — ANCASH: Prov. Pallasca, Conchucos, Weberbauer 7229 (B, F, GH), 7229a (F, GH). Depro. LIMA: Rio Blanco, Macbride 2990 (F, GH, MO, NY, US). ees Pasco: Cerro de a Asplund 11778 (s), 11835 (s), Macbride 3073 (CAS, F, GH, MO, us). Weberbauera trichocarpa is readily distinguished from the other species of the genus by its dense covering of appressed, sessile, malpighiaceous trichomes on almost all parts of the plant (including sepals, outer portion of claws, and fruit valves) (see FiGure la) and by its cymelike inflorescences. Submalpigh- iaceous trichomes are found on stems of W. spathulaefolia, cymelike racemes also characterize W. cymosa, and pubescent fruits occur in W. retropila. The presence of pubescent claws is restricted to W. trichocarpa. However, this character alone does not justify the separation of the species to a distinct genus, and some genera (e.g., Sisvmbrium) of the Brassicaceae include species with glabrous petal claws and others with pubescent ones. 1990] AL-SHEHBAZ, WEBERBAUERA 241 Perhaps the most remarkable feature of Weberbauera trichocarpa and W. cymosa 1s the cymelike, modified raceme. To my knowledge, no other species of the Brassicaceae has this inflorescence type. The flowers and fruits in W. trichocarpa are typically arranged in threes, with one generally terminating the stem. Of the 216 inflorescences studied, this ““dichasiumlike”’ raceme was ob- served in 112 (ca. 52 percent). Two-flowered inflorescences were found in 88 (ca. 41 percent) of the total, and most of these have a third abortive flower. Therefore, the basic inflorescence type for the species is a three-flowered raceme. One-, four-, and five-flowered inflorescences constitute the remaining seven percent. Variation in the number of flowers could be observed on various stems of a given plant. An example is found in Macbride 2990 (mo). 10. Weberbauera colchaguensis (Barnéoud) Al-Shehbaz, comb. nov. Cardamine? colchaguensis Barnéoud in C. Gay, Fl. Chile 1: 115. 1846. Sisymbrium colchaguensis (Barnéoud) Wedd. ex Fourn. Rech. Anat. Tax. Fam. Crucif. (Thesis, Paris), 134. 1865. Hesperis colchaguensis (Barnéoud) Kuntze, Revis. Gen. Pl. 2: 934 1891. Stenodraba colchaguensis (Barnéoud) O. E. Schulz, Notizbl. Bot. Gart. Berlin- Dahlem 11: 644. 1932. Type: Chile, [Region VII, Del Libertador General Bernardo O’Higgins,] Colchagua, cordillera del Cajon del Azufre, cerca del volcan de Talcaré- gue, ee al ft [ca. 2438-2743 m] alt., Gay 771] (holotype, P, not seen; isotypes?, G(2 sheet aan ae ee Gillies ex sane rary Bot. Misc. 3: 140. 1833, non E. pusillum Bory & Chaub. in Bory, Exp. Sci. Morée, Bot. 3(2): 190. 1832. Braya pusilla A. Gray, U. S. Expl. Exped. Phan, 15(1): 57. 1854. aH pusillum Wedd. ex Fourn. Rech. Anat. Tax. Fam. . egy ee 131. 1865, non S. pusillum Villars, Fl. Delph. in Gilib. Syst. Pl. 1: 139; ieper pusilla ie Revis. Gen. Pl. 2: 935. 1891. ae ne O. E. Schulz, Pflanzenr. IV. 105(Heft 86): 194. 1924. Stenodraba pusilla Boelcke, ae Bot. Ark. m 143. 1968. Type: Chile, el Cerro de la Porcura and la Cumbre de los Andes, 12,000 ft [ca. 3656 m] alt., Gillies 8 (holotype, E! (photo, A!)). Draba andina Philippi, Linnaea 28: 669. 1856. Stenodraba andina (Philippi) O. E. Schulz, Pflanzenr. IV. 105(Heft 86): 187. 1924. Type: Chile, [Region VIII, Maule,] “in andibus prope oppidum Linares legit Germain” (holotype, sGo, not seen). Draba patagonica Philippi, Linnaea 28: 669. 1856. Stenodraba andina (Philippi) O. E. Schulz var. patagonica (Philippi) O. E. Schulz, Pflanzenr. IV. 105(Heft 86): 188. 1924. Stenodraba 1 (Philipp1) Ravenna, Nordic J. Bot. 1: 141. 1981. Steno- draba pusilla Boelcke var. patagonica (Philippi) Boelcke, Fl. Patagonica 4a: 530. 1984. Type: Chile, [Region es Los Lagos,] Volcani de Osorno, Philippi s.n., March 1852 (holotype, sGo, not see Arabis drabaeformis Schldl. Flora 39: 410. 1856. Type: Chile, ae X, Araucania,] Cordillera de Ranco, Lechler 2958 (holotype, HAL; isotypes Stenodraba andina (Philippi) O. E. Schulz var. Airticaulis O. E Se hd Te Pflanzenr. IV. 105(Heft 86): 188. 1924. Type: Chile, ““Gipfel des Berges Pichiguan,” 1852, Philippi 67 (holotype, B! Stenodraba andina (Philippi) O. E. Schulz var. stylosa O. E. Schulz, Pflanzenr. IV. 105(Heft 86): 188. 1924. Type: Chile, [Region X, Araucania,] Volcan Lanin, 1800 m alt., Neger s.n., April 1897 (holotype, B; isotypes, M(2 sheets)!). Cespitose perennials. Caudices much branched, slender, the branches ter- minated in rosettes and covered at base with petioles of previous years. Stems subdecumbent to ascending or erect, simple or rarely branched, (1.5—)3—13(- 242 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 22) cm long, glabrous or pubescent with simple and/or rarely furcate trichomes. Basal leaves rosulate, short to long petiolate; blades oblanceolate to spatulate or obovate, rarely oblong, (5—)7—22(—25) mm long, 2—4.5(-6) mm wide, rounded at apex, cuneate to attenuate at base; entire to repand or dentate, rarely sinuately lobed, conspicuously ciliate with simple trichomes 0.6—1 mm long; petioles persistent, to 10 mm long, ciliate. Cauline leaves few, sessile; blades oblong to elliptic or obovate, (2-)4-8 mm long, 1—2.5 mm wide, ciliate and often densely pubescent with simple trichomes. Inflorescences ebracteate, corymbose ra- cemes, rarely lowermost flowers bracteate; infructescences subumbellate or rarely elongated and to 4 cm long; rachises glabrous or pubescent. Sepals caducous, erect, oblong, 1.7—2.5 mm long, scarious at margin, glabrous to hispid. Petals white, sometimes with purplish patch on back, oblong-oblan- ceolate, 2—3.5 mm long, rounded at apex. Filaments |.4-1.5 mm long; anthers 0.3-0.8 mm long. Fruiting pedicels erect and subappressed to rachis, very rarely divaricate-ascending, stout, straight, 1.5—4.5(-7) mm long, glabrous to pubes- cent. Fruits subsessile, oblong to linear, (4-)5-10(-13) mm long, 1.4—2(-2.2) mm wide; valves smooth, glabrous, with prominent midvein; septa hyaline; styles obsolete to 0.2(-0.6) mm long. Seeds 3 to 6 per locule, uniseriately to somewhat biseriately arranged, oblong to ovate, |.1—1.4(—1.6) mm long, 0.6- 0.8(-1) mm wide, dark brown; funicles slender throughout or somewhat ex- panded at placental end. Growing on stony hillsides at altitudes of 900-3350 m. REPRESENTATIVE SPECIMENS EXAMINED. Argentina. Prov. MENDoza, Atuel Valley: near road to Volcan Orero, Bécher, Hjerting, & Rahn 1872 (c); Laguna Atuel, Bocher, Hjerting, & Rahn 1976 (c, Mo); El Angulo, Hyerting, Petersen, & Rahn 452 (c), Bocher, Hjerting, & Rahn 1911 (c); Piedra del Burrero, Val Tordilla, Wilczek 425 (G, 2 sheets); Cajon del Burro, Wilczek 439 (G, US). PRov. NEUQUEN: Pino Hachado, Hauman s.n., Feb. 1920 (BA); Los Lagos, Filo Machete al Co. Rothleugal, ane 929 (LIL). PRov. Rio NEGRO: ua del Rio Colorado, ory Gorra, Moreau s.n., 19 Feb. 1940 (BA, 2 sheets); Cerro milon, Moreau s.n., 2 Feb. 1940 (Ba); an Nac. Nahuel Huapi [PNNH], Cerro =e | Pedehsen 1599 (c), ae g.n., 23 Jan. 1940, Feb. 1945 (Ba, 3 sheets); PNNH, Co. Catedral, Moreau s.n., 4 March 1943 (Ba, 2 sheets); PNNH, Hito Mirador, Moreau s.n., 19 Feb. 1943 (BA); PNNH, Hito Millaqueo, Moreau s.n., 20 Feb. 1943 (BA). Chile. [REGION ITV, Coquimmso:] Prov. Choapa, La Vega Redonda, E of La Vega Escondida, Morrison 16995 (ps, uc). [REGION VII, DEL LIBERTADOR GENERAL BERNARDO O’HIG- GINS:] Prov. Colchaqua, Las Damas, Philippi 91b (sco), Valle Hermoso, Philippi 91a (sco). [REGION VHT, MAULE:] Prov. Curic6, El Valle de los Ciegos, near the volcano of Peteroa, Bridges 1120 (Bm, 2 sheets; E); Prov. Talca, Cordillera de Talca, Philippi 91c (sco); Prov. Linares, Cordillera de Maule, Germain s.n., 1856 & 1857 (BM; G, 2 sheets); Linares, Philippi s.n., 1876 (B, F, G). [REGION X, ARAUCANiA:] Prov. Cautin, Villarrica, Neger s.n., 1897 (mM). [REGION XI, Los peas Prov. Osorno, Paso Puyehue, Sparre & Constance 10812 (uc). REGION XII: Cerro Agudo, Arroyo & Squeo 870116 (CONC). Schulz (1924) misinterpreted the limits of Weberbauera colchaguensis (as pusilla), because except for the type, all his other citations belong to W. spath- ulaefolia. The type collection, however, 1s conspecific with the plants that Schulz recognized first (1924) as Stenodraba andina and later (1932) as S. colcha- 1990] AL-SHEHBAZ, WEBERBAUERA 243 guensis. Boelcke (1968) was the first to reduce the last two names to the syn- onymy of S. pusilla, but as is shown below, he maintained Stenodraba as a distinct genus. When published, Erysimum pusillum Gillies ex Hooker & Arn. was a later homonym of £. pusillum Bory & Chaub. According to Stafleu & Cowan (1976, 1979), the first and second homonyms were published in March 1833 and September 1832, respectively. Therefore, all combinations based on the former, including Weberbauera pusilla (Gillies ex Hooker & Arn.) O. E. Schulz, are illegitimate and should be treated as new names based on the type of Erysimum pusillum Gillies ex Hooker & Arn., rather than as new combinations (see ICBN, Article 72). The earliest legitimate name for the species is Cardamine colcha- guensis Barnéoud, and therefore a new combination based on that name is needed in Weberbauera. Unaware of the fact that Gillies’ name was a later homonym, Boelcke (1968) proposed the combination Stenodraba pusilla (Gil- lies ex Hooker & Arn.) Boelcke, but this name is illegitimate, and the species should be known in Stenodraba as S. colchaguensis (Barnéoud) O. E. Schulz. Weberbauera colchaguensis is variable in shape and margin of basal leaves, length of infructescences, pubescence of stems and leaves, and length and width of fruits. Plants with pubescent stems were recognized by Schulz (1924) as var. hirticaulis, and those with dentate leaves were treated as var. patagonica (as Stenodraba andina). However, plants with glabrescent or densely pubescent stems are found within the same population, while dentate and entire leaves are often found on the same plant. In my opinion, none of the numerous variants of the species deserves formal recognition because the variation usually occurs within populations and does not correlate with geography. The species has been reported to grow as far south into Patagonia as Prov. Santa Cruz. Although I have not seen any material from that area, I have used Boelcke’s (1984) work as the basis for mapping the five southernmost localities of the species (see MAP 2). Ravenna (1981, p. 141) argued that Stenodraba andina var. patagonica should be recognized as a distinct species, S. patagonica (Philippi) Ravenna, because it differs from S. andina in the “larger size of all parts of the plant,” as well as in having “less flattened and more obtuse pods.” These alleged differences are unrealistic, and Weberbaurea colchaguensis shows continuous variation in most parts, particularly the fruits. Boelcke (1984) reduced S. patagonica to a variety of S. pusilla, but his new combination was illegitimate for two reasons. First, he chose the illegitimate name S. pusilla instead of S. colchaguensis for the species. Second, his new combination was invalidly published because the basionym (Draba patagonica Philippi) and a full and direct reference to its author and place of valid publication were not given (see ICBN, Article 33.2). Weberbauera colchaguensis is distinguished by its slender, much-branched caudex, ciliate leaves, inconspicuous flowers with petals less than 3 mm long, subumbellate infructescences, stout fruiting pedicels usually less than 5 mm long, and short styles rarely to 0.6 mm. It is most closely related to W. steno- phylla and W. chillanensis, from which it is distinguished by features listed under the latter species. 244 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 11. Weberbauera chillanensis (Philippi) Al-Shehbaz, comb. nov. Draba chillanensis Philippi, Anales Univ. Chile 2: 377. 1862. Stenodraba neers Le aut E. Schulz, Pflanzenr. IV. 105(Heft 86): 188. 1924. Type: Chile, [Re Biobio; Prov. Nuble,] Termas de Chillan, Philippi s.n. (SGo 49187 (photo, ved). Stenodraba chillanensis (Philippi) O. E. Schulz f. Jaxa O. E. Schulz, Pflanzenr. IV. 105(Heft 86): 189. 1924. Type: Chile, [Region IX, Biobio,] valley of Nieblas, Termas de Chillan, Philippi g.n., Jan. 1877 (as 1887) (B!, sGo!). Cespitose perennials. Caudices multibranched, the branches slender, ter- minated in rosettes, covered with petiolar remains of previous years. Stems subdecumbent to ascending, (3-)6-14(-20) cm long, glabrous. Basal leaves rosulate, petiolate; blades spatulate to oblanceolate, sometimes linear-lanceo- late, (5-)14—30(-40) mm long, (1-)2-3.5(-6) mm wide, obtuse to subacute at apex, attenuate at base, dentate, ciliate with simple and/or long-stalked furcate trichomes to | mm long. Cauline leaves few, linear-lanceolate, 7—15(-30) mm long, 1-1.5 mm wide, entire, ciliate. Inflorescences ebracteate, corymbose ra- cemes, elongated in fruit. Sepals spreading, oblong to ovate, 2—-2.6(-3) mm long, 1.2-1.5 mm wide, scarious at margin, glabrous or sparsely pubescent. Petals white, drying purplish in center, oblanceolate to spatulate, 2.5-3.5 mm long, 1.2-1.5 mm wide, rounded at apex, attenuate to clawlike base. Filaments somewhat spreading, 2.2-3 mm long; anthers oblong, 0.7-0.8 mm long. Fruit- ing pedicels divaricate-ascending, straight, (4-)6—-11(-18) mm long, glabrous. Fruits divaricate-ascending, linear, flattened parallel to septum, nontorulose, straight, 9-19(-22) mm long, 1.5-1.8(-2.2) mm wide; valves glabrous, with prominent midvein; styles slender, 0.5-0.9(-1.1) mm long. Seeds uniseriately arranged in each locule, oblong, 1.4-1.6 mm long, 0.7-0.9(-1) mm wide, dark reddish-brown; funicles thickened at placental end. REPRESENTATIVE SPECIMENS EXAMINED. Argentina. PRov. MENDOzA: Depto. Malargiie, Bafios del Azufre, Castellanos s.n., 19 Jan. 1941 (BA). Chile. [REGION VIII, MAULE:] Prov. Curic6, Cajon del Azufre, Albert s.n., Feb. 1891 (sco); Cordillera Curico, Reiche $.n. (B, CONC); cae (nee (sco); Volcan Peteroa, Werdermann 604 (B, BM, E, F, G, MO, NY c, z); Prov. Talca, eee Talca, Laguna de aguas calientes, Philippi 16150 (SGO); is Linares, Thermae Longavi, Schénemann 2277 (sco). [REGION Biosio:] Prov. Nuble, Cordillera de Chillan, Termas de Chillan, Jaffuel 3722 (Gx), Philippi s.n., Feb. 1892 (sco). Weberbauera chillanensis is closely related to W.. colchaguensis and W. steno- phylla. All three have narrowly spatulate to oblanceolate, ciliate basal leaves, slender caudex branches each terminated by a rosette, persistent petioles, and ebracteate inflorescences. Weberbauera stenophylla is easily distinguished by its entire basal leaves, elongate infructescences, large petals (3.5—)4—5 mm long, and styles 1.5-2 mm long. Both W. chillanensis and W. colchaguensis usually have dentate leaves, shorter petals (to 3.5 mm), and stout or slender styles rarely reaching | mm in length. Weberbauera chillanensis is separated from W. colchaguensis in having conspicuously flattened fruits 9-19(-22) mm long, slender, divaricate-ascending fruiting pedicels usually more than 6 mm long, elongate infructescences, and slender styles 0.5—0.9(—1.1) mm long. In contrast, W. colchaguensis has subterete or slightly flattened fruits (4-)5—10(-13) mm 1990] AL-SHEHBAZ, WEBERBAUERA 245 long, stout, erect fruiting pedicels almost always less than 6 mm, usually sub- umbellate infructescences, and obsolete or stout styles rarely to 0.6 mm. Fur- thermore, the occurrence of furcate trichomes on the leaves of W. chillanensis and their absence in W. colchaguensis and W. stenophylla is a useful diagnostic feature. 12. Weberbauera stenophylla (Leyb.) Al-Shehbaz, comb. nov. Draba stenophylla Leyb. Anales Univ. Chile 16: 679. 1859. Fiala stenophylla (Leyb.) O. E. Schulz, Pflanzenr. IV. 105(Heft 86): 189. 1924. E: Chile, Cord. Santiago, Cerro Colorado, Mapocho Valley, 6000-7000 ft [ca. ne 2134 m] alt., Leybold s.n? Draba leyboldii Philippi, Linnaea 33: 10. 1864, non Dalla-Torre & Sarnth. Stenodraba stenophylla (Leyb.) O. E. Schulz var. /eyboldii (Philippi) O. E. Schulz, Pflanzenr. IV. 105(Heft 86): 190. 1924. Type: Chile, [Regién IV, Coquimbo,] Cordillera Donia Rosa, Vo/kmann s.n. (fragment, sGo!). Draba cauquenensis Philippi, Anales Univ. Chile 81: 330. 1893. Type: Chile. [Regién VII, Maule,] Hacienda de Cauquenes, Cajon del Arriero, Dessauer s.n. (lectotype (here designated), sGo!; isolectotypes, B!, M!). Cespitose perennials. Caudices slender, much branched, the branches ter- minated in rosettes, covered with petiolar remains of previous years. Stems erect to subdecumbent, (3-)5—13 cm long, glabrous; basal leaves rosulate, pet- 1olate; blades narrowly oblanceolate to rarely spatulate, (1—-)1.5-—4.5 cm long, (1-)1.5—2(—3) mm wide, subacute to rounded at apex, usually attenuate at base to conspicuous petiole, entire, ciliate with simple trichomes to | mm long. Cauline leaves few, subsessile; blades oblong-lanceolate, 4-10(-14) mm long, 1-1.5 mm wide, entire, usually ciliate. Inflorescences ebracteate, corymbose racemes, elongated in fruit. Sepals green to lavender, oblong, 2.5-3 mm long, 1-1.5 mm wide, scarious at margin, glabrous. Petals white, broadly obovate, (3.5-)4—5 mm long, 2-2.5 mm wide, clawed. Filaments white, linear, 2.5—4 mm long; anthers oblong, ca. 0.8 mm long. Fruiting pedicels slender, divaricate- ascending, 5—8(-10) mm long. Fruits oblong to linear, 7-12 mm long, ca. 1.5 mm wide, smooth, glabrous; styles slender, (0.8—)1.5—2 mm long; stigmas entire. Mature seeds not seen REPRESENTATIVE SPECIMENS EXAMINED: Chile. [REGION VIII, MAULE:] Prov. Curic6, An- des de Curico, Cag n., 1892 (sGo); El Valle de los Ciegos, near the volcano of Peteroa Bridges 112] (BM, E, GH); Prov. Talca, Cordillera de Talca, Reiche s.n. (s); Turrieta, in Cordillera de Tales. Philippi 1612 (sGo); Prov. Linares, Hacienda de Cauquenes, La Chapa, Dessauer s.n., 1875 (M). Weberbauera stenophylla is apparently restricted to the three provinces of Region Maule, Chile. I have not seen any recent collections among the holdings of major herbaria. It is not known if the species has become extinct or is very rare and endangered. Weberbauera stenophylla 1s easily distinguished from its closest relatives by characters listed under the preceding species. ?According to Stafleu & Cowan (1979), the types and bestsenua - Fnegneh Peybols, wo selee in Chile in 1855, are unknown. However, if the type cannot be loca of the species is accompanied by a well-illustrated plate that ee serve as the type. 246 JOURNAL OF THE ARNOLD ARBORETUM [vov. 71 13. Weberbauera lagunae (O. E. Schulz) Al-Shehbaz, comb. et stat. nov. gee suffruticosa (Barnéoud) O. E. Schulz var. lagunae O. E. Schulz, Notizbl. Bot t. Berlin-Dahlem 10: 469. 1928. Type: Chile, [Region II, Atacama, Prov. aco ale Cordillera Laguna Chica, ca. 4000 m alt., Jan. 1924, Werdermann 262 (holotype, B!; isotypes, BM!, CONC!, GH!, UC’). Cespitose, scapose perennial. Caudices multibranched, the branches slender, terminated in rosettes, covered with persistent petiolar remains of previous years. Scapes leafless or rarely with small cauline leaf, slender, 2-5(-10) cm long, purplish, glabrous. Basal leaves rosulate, petiolate; blades narrowly ob- lanceolate to oblong-oblanceolate, 4—-10(-15) mm long, (0.5-)1-1.5 mm wide, rounded at apex, somewhat attenuate at base, entire, flat, glabrous beneath, densely ciliate and hispid above with unbranched trichomes to | mm long, midrib prominent; petioles persistent, thick, subterete, 1-5 mm long. Inflo- rescences ebracteate, few-flowered, corymbose racemes, elongated considerably in fruit. Sepals oblong, 1.5-2 mm long, 0.8-1 mm wide, nonsaccate, pubescent with straight trichomes. Petals yellowish white, with purplish area on back, oblong, 1.5-1.7 mm long, 0.4-0.8 mm wide, undifferentiated into blades and claws. Filaments |.2-1.5 mm long; anthers oblong, 0.5-0.6 mm long. Fruiting pedicels divaricate to occasionally divaricate-ascending, slender, 4-8(—12) mm long, glabrous. Fruits linear, terete, torulose, 7-1 1 mm long, 0.9-1.2 mm wide, tapering at both ends, glabrous; septa hyaline; styles 0.3-0.4 mm long; stigmas entire. Seeds uniseriately arranged in each locule, oblong, 1.2-1.3 mm long, ca. 0.7 mm wide, brown; funicles thickened at placental end. Growing on dry rocky slopes and gravelly, sod-covered banks at altitudes of 3800-4000 m. REPRESENTATIVE SPECIMENS EXAMINED. Chile. [REGION HI, ATACAMA:] Prov. Huasco, vicinity of Laguna Chica, Johnston 5950 (GH), vicinity of Laguna Valeriano, Johnston 6056 (CONC, GH, US). Although Weberbauera lagunae was originally described by Schulz (1928) as a variety of Stenodraba suffruticosa (now W. suffruticosa), it is so different in leaves, flowers, and fruits from both this and the closely related W. imbri- catifolia that it should be recognized as a distinct species. Weberbauera lagunae has narrowly oblong, nonclawed petals less than 2 mm long and styles 0.3-0.4 mm long in fruit. In contrast, both W”. imbricatifolia and W. suffruticosa have clawed, broadly spatulate petals 3.5-5.5 mm long and styles 1-3 mm long in fruit. Furthermore, WW’. /agunae differs from W. imbricatifolia in having the lower leaf surface glabrous instead of pubescent, and from W. suffruticosa in having the leaves flat and oblanceolate instead of semiterete and linear. Weberbauera lagunae differs from the related W. colchaguensis in having lax, racemose infructescences, slender, divaricate fruiting pedicels 4—-8(-12) mm long, torulose fruits, and entire basal leaves to 1.5 mm wide. Weberbauera colchaguensis has dense, subumbellate infructescences, stout, subappressed fruiting pedicels 1.5—4.5(-6) mm long, smooth fruits, and usually dentate basal leaves 2—4.5(-6) mm wide. 1990] AL-SHEHBAZ, WEBERBAUERA 247 14. Weberbauera suffruticosa (Barnéoud) Al-Shehbaz, comb. nov. Draba suffruticosa Barnéoud in C. Gay, Fl. Chile 1: 157. 1846. Sisymbrium suffruti- cosum (Barnéoud) Fourn. Rech. Anat. Tax. Fam. Crucif. (Thesis, Paris), 132. 1865. Hesperis suffruticosa (Barnéoud) Kuntze, Revis. Gen. Pl. 2: 935. 1891. Draba im- bricatifolia Bernéoud var. suffruticosa (Barnéoud) Reiche, Fl. Chile 1: 116. 1896. Stenodraba suffruticosa (Barnéoud) O. E. Schulz, Pflanzenr. IV. 105(Heft 86): 190. 1924. Type: Chile, [Region IV, oon Cordillera Ovalle, 12,000 ft [ca. 3658 m] alt., Gay s.n. (holotype, P; isotype, B! (photos, GH!, Mo! ar, — Cespitose, scapose perennial. Caudices branched, the branches slender, ter- minated in rosettes, covered with petiolar bases of previous years. Scapes leafless or rarely with | or 2 leaves, 3-5 cm long, glabrous. Basal leaves rosulate, petiolate; blades linear, thick, usually semiterete, (3-)4-8 mm long, 0.5-0.8 mm wide, the margin somewhat incurved, entire, ciliate, the lower surface glabrous, the upper surface usually densely hispid with simple trichomes to 1 mm long; midribs and petioles persistent. Inflorescences ebracteate, few-flow- ered racemes, elongated and lax in fruit. Sepals erect, oblong, 3.5-4 mm long, nonsaccate, scarious at margin, usually pubescent with straight, simple tri- chomes. Petals broadly spatulate, 4.5-5.5 mm long, ca. 1.5 mm wide, rounded at apex, attenuate to distinct, clawlike base. Filaments 3—3.5 mm long; anthers oblong, ca. 1 mm long. Fruiting pedicels divaricate, slender, (6—)8—12 mm long, glabrous. Fruits linear, subterete, somewhat torulose, | 1-16 mm long, glabrous; septa hyaline; styles slender, (1.5—)2-3.1 mm long; stigmas entire. Seeds uni- seriately arranged in each locule, oblong, ca. 1.5 mm long, 0.8 mm wide, brown; funicles somewhat stout at placental end. REPRESENTATIVE SPECIMENS EXAMINED. Chile. [REGION IV, Coqumso:] Prov. Limari, Sotaqui, Gay 1029 (sco); Cordillera de Dona Rosa, Volckmann s.n., 1860/1861 (sco); Quebrada Larga, Jiles 3408 (CONC). Weberbauera suffruticosa, which is very rare and apparently known only from few collections, is easily distinguished from all other species of the genus by its semiterete, thick, linear basal leaves 0.5-0.8 mm wide. It also differs from its closest relative, W. imbricatifolia, in having a glabrous lower leaf surface. Both species have large flowers and slender styles to 3 mm lon 15. Weberbauera imbricatifolia (Barnéoud) Al-Shehbaz, comb. nov. Draba Wi Barnéoud in C. Gay, Fl. Chile 1: 158. 1846. Brava imbricatifolia (Barnéoud) A. Gray, U.S. Expl. Exped. Phan. 15(1): 58. 1854. Sisymbrium imbri- catifolium (Barnéoud) Wedd. Chloris Andina 2: t. 58B. 1857. Hesperis imbricatifolia (Barnéoud) Kuntze, Revis. Gen. Pl. 2: 934. 1891. Stenodraba imbricatifolia (Bar- néoud) O. E. Schulz, Pflanzenr. IV. cs aaa Ho 190. 1924. Type: Chile, [Regién IV, Coquimbo,] Cordillera de Coquimbo, 12,000 ft [ca. 3658 m] alt., Gay s.n. (holotype, Pp; isotype, B! (photo, Mo!); a isotypes, G!, GH!). Stenodraba imbricatifolia (Barnéoud) O. E. Schulz var. glabrata O. E. Schulz, Pflanzenr. IV. 105(Heft 86): 190. 1924. Type: Chile, Leybold 2974 (holotype, B!). Cespitose, scapose perennial. Caudices woody, much branched, the branches terminated in rosettes, covered with persistent petiolar bases of previous years. 248 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 Scapes leafless or rarely with 1 or 2 cauline leaves, (0.6—)1-4(—6) cm long, glabrous. Basal leaves petiolate, flat; blades oblong to rarely oblanceolate, 2- 7(-15) mm long, 0.5—1.5(—2) mm wide, rounded at apex, entire, ciliate, densely hispid above with simple trichomes to 1 mm long, pubescent below with much shorter, branched or rarely simple trichomes, very rarely glabrescent; petioles persistent, thick, 0.5-3(-5) mm long. Inflorescences few-flowered, ebracteate, corymbose racemes, elongated in fruit. Sepals erect, oblong, (2-)2.5-3.5 mm long, 1-1.5 mm wide, nonsaccate, scarious at margin, pubescent. Petals white, broadly spatulate, 3.5—5 mm long, |.5—2 mm wide, rounded at apex, tapered to short claw. Filaments erect, 3-4 mm long; anthers oblong, ca. 1 mm long. Fruiting pedicels divaricate to ascending, slender, 3—7(—10) mm long, glabrous. Fruits linear, terete, somewhat torulose, (7-)9-14(-16) mm long, 1.2-1.5 mm wide, glabrous; septa hyaline; styles slender, 1-2(-3) mm long; stigmas entire. Seeds uniseriately arranged in each locule, oblong, ca. 1.2 mm long, 0.8 mm wide, brown. REPRESENTATIVE SPECIMENS EXAMINED. Chile. [Region IV, Coquimso:] Prov. Choapa, Cordillera de Illapel, Vega Negra, Volekmann s.n., 1860/1861 (sGo); La Polcura, Philippi 2279 (sco). [REGION V, ACONCAGUA:] Prov. Petorca, 5 km S of Junta de Piuquenes, Rio Sobrante, Morrison 17290 (ps, uc). [REGION VI, METROPOLITANA DE SANTIAGO:] Prov. Santiago, Andes of San José, Gay 1028 (sco). Weberbauera imbricatifolia is highly variable in the size and pubescence of the basal leaves. Schulz (1924) used the leafless scapes and short (2-4 mm) basal leaves of this species to distinguish it from W. suffruticosa (both as Stenodraba). Weberbauera suffruticosa was said to have few cauline leaves and longer (6-8 mm) basal ones. The presence of cauline leaves is uncommon in both species, and different rosettes of the same plant might produce leafless stems, as well as stems with one or two leaves. Furthermore, the length of basal leaves is unreliable for the separation of these species (see descriptions). The shape, pubescence, and cross section of basal leaves provide the most reliable distinguishing characters: the leaves of W. imbricatifolia are flat, oblong to rarely oblanceolate, and almost always pubescent beneath with simple or furcate trichomes shorter than those of the upper surface; in contrast, those of W. suffruticosa are thick, semiterete, and linear, with a glabrous lower surface. 16. Weberbauera parvifolia (Philippi) Al-Shehbaz, comb. nov. Sisymbrium parvifolium Philippi, Linnaea 28: 667. 1856. Stenodraba parvifolia (Phi- lippi) O. E. Schulz, Pflanzenr. IV. 105(Heft 86): 187. 1924. Type: Chile, [Regi6én VIL, Maule,] Cordillera de Linares, Germain s.n., 1856 (holotype, sGo 63217!). Cespitose perennials. Caudices slender, much branched, the branches ter- minated in rosettes. Stems erect or subdecumbent, simple or branched at base and/or above, 6—15(—25) cm long, glabrous or sparsely pubescent with furcate and/or dendritic trichomes. Basal leaves rosulate, short petiolate to subsessile; blades narrowly oblanceolate to rarely subovate, 5—20(-35) mm long, 1-2.5(- 3.5) mm wide, obtuse to subacute, cuneate to attenuate at base, entire or rarely dentate, ciliate with simple trichomes, often with branched trichomes on sur- face, flat. Cauline leaves few, smaller and narrower than basal ones. Inflores- 1990] AL-SHEHBAZ, WEBERBAUERA 249 cences ebracteate, corymbose racemes, elongated in fruit. Sepals caducous, oblong, 1.5-2.6 mm long, 0.7—1 mm wide, glabrous or with dendritic trichomes. Petals white, oblanceolate, (1.8—)2.5-3.5 mm long, to ca. | mm wide. Filaments white, |.7-2.8 mm long; anthers oblong, 0.6-0.8 mm long. Fruiting pedicels divaricate, straight, 4-10(-16) mm long, usually glabrous. Fruits linear, flat- tened, not torulose, 8—18(-—24) mm long, 0.8-1.2(-1.5) mm wide; valves gla- brous, with conspicuous midvein; styles 0.1—0.4(—0.7) mm long; stigmas entire. Seeds uniseriately arranged, oblong-ovate, |.2—1.6 mm long, 0.7-0.9 mm wide, brown; cotyledons incumbent; funicles somewhat thickened at proximal end. REPRESENTATIVE SPECIMENS EXAMINED. Argentina. Prov. NEUQUEN: Depto. Lacar, Cha- pelco, Schajovskoy 61 (BAA); Depto. Chos Malal, cajon del ao. del Cruce, Boelcke, Correa, Bacigalupo, et al. 11279 (BAA); Depto. Minas, confluence of rivers Pichi-Neuquén and Neuquén, Boelcke et al. 13763 (BAA); upper valley of ao. Atreuco, Boelcke, Correa, Bacigalupo, et al. 11512 (BAA), 11533 (BAA, st); Cordillera del Viento, cruzada de Tricao Malal al Cajon de Butal6o, Boelcke, Correa, Bacigalupo, et al. 11565 (BAA, st); Cajon de los Chenques, Boelcke et al. 13847 (BAA); N of Varvarco Campos, Boelcke et al. 14144 (BAA); Depto. Picunches, Pino Hachado, Burkart 9738 (s1); Lago Aluminé, Kalela 1608 (H). Chile. [REGION VII, DEL LisERTADOR GENERAL BERNARDO O’HIGGINS:] Prov. Col- chaqua, San Fernando, Termas El Falco, Montero 6043 (st). [REGION VII, MAULE:] Cordillera de Maule, Germain s.n., 1856-1857 (G); Prov. Talca, Cordillera de Talca, E] Picazo, Barros 2752 (si) Weberbauera parvifolia is somewhat anomalous in the genus in having some- times erect stems that are branched above. No other species of Weberbauera has this feature. However, W.. parvifolia is related to other species of the genus that Schulz (1924) placed in Stenodraba. ACKNOWLEDGMENTS Iam most thankful to Reed C. Rollins and Neil A. Harriman for their critical reviews of the manuscript, and to Donald H. Pfister for obtaining funds from the Harvard University Herbaria that supported the SEM portion of the re- search. I am grateful to Elizabeth A. Shaw and Peter F. Stevens for correcting the Latin diagnoses, Gustavo A. Romero for helping with Spanish, Michael A. Canoso for obtaining the loans, Barbara Nimblett for typing the manuscript, Trisha Rice for the SEM work, my wife, Mona, for her continuous support, and Elizabeth B. Schmidt and Stephen A. Spongberg for their editorial advice. I am indebted to Stephan G. Beck, Clodomiro Marticorena, Barbara Ruthsatz, David N. Smith, and James C. Solomon for sending plant material. I thank the directors and curators of the herbaria (abbreviations follow Holmgren ef al., 1981), who kindly loaned the specimens on which this study is based. LITERATURE CITED BoELcKE, O. 1968. Cruciferae. Pp. 140-144 in T. W. BOcuer, J. P. HJERTING, & K RAHN, Botanical Sees in the Atuel Valley area, Mendoza Province, Argentina. Dansk Bot. Ark. 22: 1984. ae ihe on BoELCKE & M. C. ROMANczUK, FI. oan 4A: vee 544. Buenos Aires, Instituto Nacional de Tecnologia Agropecuar 250 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Driers, L. 1961. Der Anteil an Polyploiden in den Vegetationsgiirteln der Westkordillere Perus. Zeitschr. Bot. 49: 437-488. Gita, E., & R. MuscHier. 1909. Aufzahlung aller zur Zeit bekannten siidamerikan- ischen Cruciferen. Bot. Jahrb. Syst. 42: 437-487. HoLmaren, P. K., W. KEUKEN, & E. K. SCHOFIELD. 1981. Index herbariorum. Part I. The herbaria of the world. ed. 7. Reg. Veg. 106. 452 MacsripE, J. F. 1934. New or renamed spermatophytes mostly Peruvian. Candollea 5: 346-402. —. 1938. Cruciferae. Flora of Peru. Publ. Field Mus. Nat. Hist., Bot. Ser. 13(2): 937-9 RAVENNA, P. 1981. Taxonomical notes on the Chilean Cruciferae. Nordic J. Bot. 1: 142. ScHutz, O. E. 1924. Cruciferae-Sisymbrieae. Jn: A. ENGLER, ed., Pflanzenr. IV. 105(Heft 86): 1-388. 1928. Die von O. Berninger, A. Soe und besonders von E. Werder- mann in Chile gesammelten Cruciferen. 77; E. WERDERMANN, Pee zur Kenntnis der Flora von Chile. Notizbl. Bot. Gart. Berlin- Dahlem 10: 460-4 1929. Amerikanische Cruciferen verschiedener Herkunft. ai) 558- 564. 1932. Uber einige bisher ungewisse Arabis- und Sisymbrium-Arten. Ibid. 11: 641-645. . 1936. Cruciferae. Jn: H. Harms, ed., Nat. Pflanzenfam. ed. 2. 17B: 227-658. STAFLEU, F. A., & R. S. Cowan. 1976. Taxonomic literature. ed. 2. Vol. 1. Reg. Veg. 94: xl + 1136 pp. —— _ & ——. 1979. Taxonomic literature. ed. 2. Vol. 2. Reg. Veg. 98: xvii + 991 pp. 1990] STAPLES, PORANEAE pio | PRELIMINARY TAXONOMIC CONSIDERATION OF THE PORANEAE (CONVOLVULACEAE) GEORGE W. STAPLES! The tribe Poraneae Cela ecea has been Sea oy oe evaluated, and some realignment of genera is necessary. Metaporana N. E. Br. is here reas- signed to the tribe Clcee A new combination is ne for a species endemic to Socotra; its diagnostic features are discussed. Dactylostigma D. Austin 1s reduced to synonymy with Hildebrandtia Vatke ex A. Braun of the tribe Hil- debrandtieae. A new name is provided for the first Madagascan species of Hildebrandtia. Eight genera had been assigned to the Poraneae Hallier f. (Convolvulaceae) (see TABLE) when I began an evaluation of the tribe in 1983. Of these, some were questionably placed because they appeared to lack the characters used by Hallier (1893) to establish the tribe—namely, an accrescent, winglike fruiting calyx enclosing a one-seeded, indehiscent fruit with a papery pericarp (utricle). A comprehensive examination of all genera assigned to the tribe Poraneae is nearing completion. The present paper will provide an overview of the tribe and make necessary taxonomic combinations for two genera being removed from the Poraneae. The next paper will present a taxonomic revision of the enus Porana Burman f., which formed the bulk of my doctoral dissertation (Staples, 1987). The final one will recharacterize the tribe Poraneae and delimit the character states for the genera assigned to it; it will also include revisions for the genera Cordisepalum Verdc., Cardiochlamys Oliver, Rapona Baillon, and Dipteropeltis Hallier f. Within the genus Ca/ycobolus the African taxa have been studied by Lejoly and Lisowski (1985) and the Neotropical taxa by Austin and Staples (in prep.). In addition, a comprehensive palynotaxonomic study of this genus and its close relative Dipteropeltis is nearing completion (D. Vernier, pers. comm.) and will perhaps further illuminate the relationships of these genera. These works account for most of the genera that had been assigned to the Poraneae prior to 1983, when I began this study. Two genera, Metaporana N. E. Br. and Dactylostigma D. Austin, remain to be discussed. Based on morphological, palynological, and trichome characters, they are incorrectly referred to the Poraneae. They are here reassigned to other tribes, necessitating one new name and one new combination. \Harvard University Herbaria, 22 Divinity Avenue, Cambridge, Massachusetts 02138. Current address: Bernice P. Bishop Museum, Department of Botany, P. O. Box 19000-A, Honolulu, Hawaii 96817-0916. © President and Fellows of Harvard College, 19 Journal of the Arnold Arboretum 71: 251-258. ea 1990. 252 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 A conspectus of the tribe Poraneae Hallier f.* ca. 1983. D OF FI PUBLICA- | NUMBER GENUS TION OF TAXA PROVENANCE Porana Burman f. 1768 ca. 25 Asia and Australia Calycobolus Willd. 1819 ca. 20 W Par and too America Cardiochlamys Oliver 1883 2 Madaga Baillon 189] | vino Dipteropeltis Hallier f. 1899 2 W Africa Metaporana N. E. Br. 1914 7 E Africa and Madagascar Cordisepalum Verdc. 197] l Dactylostigma D. Austin 1973 Madagascar *Adapted from Austin (1973). METAPORANA Nicholas Brown (1914) established Metaporana for two East African taxa. Since that time, the genus has variously been maintained as distinct (Myint & Ward, 1968; Verdcourt, 1969) or placed in synonymy with Bonamia Thouars (Meeuse, 1957; Verdcourt, 1963). Myint & Ward (1968), in their revision of Bonamia, concluded that Metaporana was distinct from Bonamia. Reexami- nation of Metaporana in terms of the characters used to define the Poraneae, however, revealed that its persistent but nonaccrescent calyx and tardily de- hiscent capsule exclude it from that tribe. Overall morphology, trichomes, and pollen characters of Metaporana indicate that it is better placed in the tribe Cresseae Bentham & Hooker, a reassignment I here effect. The Cresseae, in need of careful systematic scrutiny, are at best loosely delimited at present by their branched style or two free styles, dehiscent capsules, nonaccrescent ca- lyces, and predominantly nonspinose, three-colpate pollen grains. Verdcourt (1969) accepted Mefaporana as a valid genus and lectotypified it with M. densiflora (Hallier f.) N. E. Br., simultaneously recognizing the first Madagascan species of the genus, M/. parvifolia (K. Afzel.) Verdc. He later (1974) identified three varieties of the latter species, M. parvifolia var. obovata, var. obtusa, and var. pilosa, and described two additional species endemic to Madagascar, MM. conica and M. sericosepala. Since his 1974 paper describes, illustrates, and lists specimens examined for the Madagascan taxa, such infor- mation is not repeated here. A description and illustration for M. densiflora are available in Verdcourt (1963, under Bonamia poranoides). | do include a list of representative specimens examined for this widely distributed African plant. This brings to four the number of species in Metaporana, one in East Africa and three from Madagascar. To these I can add another, which was described as a species of Porana. Porana obtusa Balfour f. is endemic to the island of Socotra. Only one additional collection has come to light since Balfour’s gath- ering of the type material. The slightly accrescent fruiting calyx and tardily 1990] STAPLES, PORANEAE 253 dehiscent, two-seeded capsule exclude the species from Porana and indicate a placement in Metaporana. Metaporana obtusa (Balf. f.) Staples, comb. nov. Porana obtusa Balf. f. Proc. Roy. Soc. Edinburgh 12: 83. mes Trans. Roy. Soc. Edinburgh 31: 192. ¢. 57. 1888. Type: Socotra, “inter opulos ad extremitatem campi Kadhab scandens,” Feb.—March 1880, Balfour, ee el & Scott 355 (ho- lotype, K!; isotypes, BM!, E!, LE!; fragment, GoET!). Little can be added to the 1888 description published by Balfour, which in conjunction with his plate, adequately characterizes the species. Metaporana obtusa is an erect shrub 3-10 m tall, with leathery, dull-colored, oblong leaves having an obtuse to emarginate apex and 12 to 14 pairs of secondary veins. The corolla lobes are marked with dark, elliptic (? glandular) dots, and the capsule has a glossy, resinous-looking surface, much like that of M. sericosepala Verde. Overall, M/. obtusa has the largest fruiting calyx of any species in the genus due to the slightly accrescent nature of the sepals. DIsTRIBUTION. Endemic to the island of Socotra in the Indian Ocean. ADDITIONAL SPECIMEN EXAMINED. Socotra Island: Ras Kattomahan, G. Popov So/136 (BM) Metaporana densiflora (Hallier f.) N. E. Br. Bull. Misc. Inform. 1914: 168. 1914 Porana densiflora Hallier f. Bot. Jahrb. Syst. 18: 93. 1894; in Engler, Abh. Preuss Akad. Wiss. 26, 34. 1894. Type: East Africa, 1885-1886, Fischer 284 (holotype, B, destroyed; fragment, GoET!). Bonamia poranoides Hallier f. Bull. Herb. Boissier 5: 1007. 1897, nom. nov., non B. densiflora (Baker) Hallier f. (1894). DIsTRIBUTION. Widespread in equatorial Africa, from northeastern Zaire east- ward through Uganda, Tanzania, and Kenya to the Indian Ocean. REPRESENTATIVE SPECIMENS EXAMINED. Kenya. Coast Prov. Kilifi distr.: Kilifi, G. M. Jeffrey K307 («), Graham 1923 (x); Mnarani, Lavranos 1189 9 (mM); Mtawapa, Williams Sangrai 840 (k); Sabaki, 4 mi N of Malindi, R. Polhill & Paulo 734 (x, p). Kwale distr.: behind Diani Beach, : 2 Gillett 18645 («, Pp); Diani Forest, J. B. Gillett & Kibuma 19879 (kK). Mombasa distr.: Bamburi-Shimo la Tewa, Bally & A. R. Smith 14380 (kK); Mombasa, Sacleux 2208(#1) (Pe), pene (Pp). Distr. not indicated: Jardini Forest, Foden & A. Evans 70/409 (k), Kibarani, G. M. Jeffrey K53 (kK); N of Mombasa to Lamu and Witu, A. Whyte 5.n., 1902 (Bm); Mombasa to Takaungu, A, Whyte s.n., 1902 (k); S of Mombasa, H. M. Gardner 1413 (k); Mwachi, 3 mi S of Mazeras, R. B. Drummond & J. H. Hemsley 4245 (K); Rabai Hills, Mombasa, Mlima von Riali, Rabai, W. E. Taylor s.n., Sept. 1885 (BM); Shanzu, FE. Polhill 408 (k, Pp); Takaungu, Thomas s.n., Feb, 1867 (wu). SOUTHERN PRrov., distr. not indicated, Kitui, Endau forest, J. Mbonze 31 (K). Tanzania. EASTERN Pee Dar es Salaam, Wazo Hill, Batty 37 (p), 1071 (p). Kisarawe distr.: Kazimzumbwe, Proctor ); Nyaburu, 7. H. Fundi l 54 (kK); Pugu Hills, forest reserve, Hansen 431 (c), J. H. eae 2342 (BM). TANGA Prov. Pangani distr.: cus Est, Faulkner 590 (BM, kK, P, 8); Mecca Parish, Mseko section, Mwera chiefdom, Tanner 2943 (uc); Pangani, Stuhlman 594 (HBG). Distr. not indicated: Kigombe, Geilinger 397 AS Kirindani, H. Faulkner 3612 (pr); 4 mi SE of Ngomeni, R. B. Drummond & J 254 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Hemsley 3608 = oo Mtns., Buiti, Holst 2379 (kK, M); Usambara Mtns., Duga, HAM, K, LE, M, P, S, US, W); Usambara Mtns., Maramba, Mwele, Peter 51724 (B, GOET); “Usambara: a bei Masbura, Peter 51624 (Gort); Usambara Mtns., Mkumbora, Doughty 2 (k). Prov. not indicated, Krogwe distr., Magewga Est., Faulkner 1335 (xk). Uganda. EASTERN Prov. Busoga distr.: Butembe-Bunya Co., 12 mi N of Jinja, Kagoma Forest Reserve, G. H. S. Wood 321 (k), Wandago Mutalla, ca. mi 10 on new Iganga road, G. H. S. Wood 18 («); without further locality, E. Brown 369 (k). Distr. not indicated, Aga Swamp, Teso country, P. Chandler 987 (k). WESTERN Prov. Toro distr., Ruwenzori, G. Taylor 2662 (pm). Distr. not indicated: Butiaba, plain on E shore of Lake Albert, Bagshawe 843 (Bm); 58 mi from Fort Portal, SW of Mukokya, Lock 463 (kK); Kikonongo, Ruenzori, Maitland 1037 (k). Prov. and distr. not indicated: Bwamba, R. Fyffe 19 (x); Commo, shores of Nyanza Salt Lake, Scott-Elliot 7972 (BM, k); Plucot (Usoga), R. Diimmer 2790 (sm); Waryanga-Jinja, P. Chandler 2066 (k, P). Zaire. PROV. Kivu. Beni territory: Beni, Lebrun 4599 (k, us), Bwambali, Pare Natl. Albert, De Witte 13942 (x). Territory not indicated: Bitshumbi [Vitshumbi], Bosquet, Lebrun 9269 (BR, K, MO, P); Bitschumbi, Kikongomoko, De Witte 1035 (BR, K); Kabare, Bequaert 5424 (Br); Kotonda [Katanda], Lebrun 7705 (k, Pp); NO Kongo, Semliki-Ebena, Mildbraed 2129 ies Prov. ORIENTALE, Ituri, versant Ouest, massif du Ruwenzori, H. Humbert 8809 (p). Prov. not indicated: Ruindi, Lebrun 8029 (k, Pp); prés du camp de la riviére Ruindi, De a itte 2083 (K). A comment on the typification of this species is warranted. The holotype for Porana densiflora was destroyed with the Berlin herbarium, and no isotype has come to light that might serve as lectotype. This situation prompted Verd- court (1963) to designate Holst 3205 (k) as a neotype for the name. In the material loaned for the tribal revision, I found a fragment of the original Fischer collection conserved at Goettingen. Perhaps Hallier removed a bit of the ho- lotype from Berlin for his studies at Goettingen, where he worked for some time. In any case, the application of the name is without question because Fischer 284 is the same species as Holst 3205. Although the latter collection now has no nomenclatural standing, it is useful for comparative purposes because it has been authenticated by comparison with the only fragment of the holotype known to exist. The Holst collection has been widely distributed and is more readily available to botanists than the original Fischer collection. DACTYLOSTIGMA ->d ona single poorly preserved specimen from Madagascar, Austin (1973) -oed the genus Dactylostigma, typified by D. /inearifolia. The generic name - 19 the highly divided stigmas, lobed like the fingers of a hand, which \ustin believed were unique in the Convolvulaceae. The holotype specimen is depauperate, and although stigmas are visible, the other structures inside the flowers are largely obscured by fungal hyphae. The spent flowers are en- closed in large, papery calyces, with the two outer sepals much accrescent, the third sepal less so, and the fourth and fifth ones only slightly increased in size. The fruits, which are immature on the holotype specimen, are formed from a unilocular ovary containing four ovules, only one of which develops into a seed. These features suggest a relationship to the Poraneae, to which tribe Austin referred the new genus. During a visit to the Paris herbarium in 1984, I was able to examine the 1990] STAPLES, PORANEAE 253 very rich Madagascan collections conserved there. Having studied additional, more complete material of Dactylostigma, I can now amplify Austin’s descrip- tion of the plant. The flowers are functionally unisexual, with staminate and pistillate flowers borne on separate plants. I was able to compare these speci- mens with complete material of Hildebrandtia Vatke and Cladostigma Radlk., two African genera known to be dioecious (Hallier, 1898). The stigmas of Dactylostigma proved to be identical to those of Hildebrandtia, although the Madagascan plants are clearly a different species of that genus than any of those known from Africa (Verdcourt, 1961, 1981). Austin (pers. comm.) saw no comparative material of Hildebrandtia when he described Dactylostigma and thus did not realize that his novelty was in fact the first species of Hildebrandtia from Madagascar. I here reduce Dactylostigma to synonymy with Hildebrandtia, which has been placed in the tribe Hildebrandtieae Peter (Melchior, 1964; Roberty, 1964: Willis, 1973) due to the dioecious habit, a feature unique in the Convolvulaceae. The epithet /inearifolia is preoccupied in Hildebrandtia, so a new name is required. I take the opportunity to name the first Madagascan species of the genus in honor of Daniel F. Austin, who has contributed greatly to an improved understanding of the complex relationships in the Convolvulaceae. Hildebrandtia austinii Staples, nom. nov. FIGURE. Dactylostigma linearifolia D. Austin, Phytologia 25: 426. 1973, non Hildebrandtia linearifolia Verdc. (1981). Type: Madagascar, delta de la ps: cote SW, 24-28 Aug 1928, Humbert & Swingle 5368 [2] (holotype, Mo!; isotype Slender, scandent shrub, 0.5-2 m long, rufous-golden-sericeous on young parts, glabrate with age. Stems slender, virgate, nodes laxly spaced, main axes grayish and smooth, short shoots warty due to leaf scars, tips golden-sericeous. Leave clustered in fascicles at nodes or on lateral short shoots, subsessile, folded lengthwise in dried state; blade linear to oblong or narrowly obovate, 4-20 x <1 mm, base cuneate, apex acute or obtuse and apiculate, margin entire, both surfaces shining-golden-sericeous. Flowers borne among leaves, | to few per node, sessile or on pedicels <2 mm long (in fruit lengthening to 5-6 mm), 5-merous. Staminate flowers with sepals +equal, elliptic-oblong, 2-3 x <1 mm, sericeous; corolla campanulate, 5-lobed, 5-6 mm long, white, sericeous outside, glabrous within; stamens barely exserted, unequal, 3-5 mm long, the filaments fused below to corolla tube and sparsely glandulose, free, filamentous, and glabrous above, the anthers dorsifixed, ellipsoid, introrse, whitish; pollen spheroidal, 15-18 um in diameter, 3-colpate, surface finely granulate; pistillode with rudimentary ovary, the style 1, simple, the stigma deeply palmately dis- sected. Pistillate flowers with sepals unequal (the outer 2 larger, broad-ovate, ca. 5 x 3 mm, base oblique, apex acute; median one ovate, oblique, inner 2 oblong-lanceolate), chartaceous, sericeous on both surfaces; corolla campan- ulate, 5-lobed, ca. 5 mm long, white, the lobes sericeous outside, glabrous within; staminodes unequal, <2 mm, nonpolliniferous; disc pedestallike; ovary ovoid, 4-lobed, ca. | mm tall, glabrous; ovules 4; styles 2, free, filiform, gla- brous; stigmas deeply palmately lobed. Capsule (immature) narrowly ovoid, 256 JOURNAL OF THE ARNOLD ARBORETUM “a IG NUY A NG th YY KN Ws SSS [voL. 71 Hildebrandtia austinii: A, staminate plant, habit; B, staminate flower, corolla open; C, pistillate plant, habit; D, pistillate flower, corolla open; E, fruiting calyx; F, leaf. (A, B after Rakotomama & Surveillant 3929; C, D after Humbert & Perrier de la Bathie 2423: E, F after Humbert & Swingle 5368.) Drawing by Wang Le-zhong. 1990] STAPLES, PORANEAE 207 l-seeded, tan to pale brown, glabrous, surmounted by persistent corolla and enclosed in accrescent calyx, the outer 2 sepals ovate, 13-20 x 10-14 mm, the middle one ovate-elliptic, oblique, 13 x 9 mm, the inner 2 oblong, 8-9 x ca. 2 mm, all sepals chartaceous, villous on both surfaces. Seeds ovoid, carinate, ca. 2 mm long, dark brown, smooth, glabrous, with basal hilum. DISTRIBUTION AND ECOLOGY. Endemic to Madagascar (known only from south- west coast in vicinity of Tuléar); sandy and lateritic soils 2-10 m alt. Flowering March to September, fruiting August and September. VERNACULAR NAMES. Vongo, rakiorakitz. SPECIMENS EXAMINED. Madagascar. Betioky distr.: Soalary canton, Reserve Naturelle X, Ravelonanabary 5017 [2] (Pp), Rakotomama & Surveillant 3929 [8] (p). Distr. not indicated: env. du Tuléar, delta du Fiherenana, Humbert & Perrier de la Bathie 2403 [3] (P), 2423 [2] (P); km 4 de Tuléar, Service Forestier SF 16981 [8] (p); env. de Tuléar, Perrier de la Bathie 12834 [8] (pr). The linear-oblong leaves of Hildebrandtia austinii set it apart from most other species of the genus. Only 77. /inearifolia, from Somalia, could be confused with it on the basis of leaf shape. The two species can be readily distinguished, however, by the differences in their habit and the nature of their fruiting sepals. Hildebrandtia austinii has slender, scandent stems with laxly arranged fascicles of leaves; the two outer fruiting sepals are ovate and are villous on both sides. In contrast, H. /inearifolia is a shrub with stout stems and spinescent branches, with the fascicles of leaves crowded along them; the two outer fruiting sepals are orbicular and are appressed-pubescent on the abaxial side only. The flowers of Hildebrandtia austinii provide another character that may prove taxonomically useful, once the character states are known in other species of Hildebrandtia. The pistillate flowers have a gynoecium with two free styles, each terminating in a digitate stigma, whereas the staminate ones have a pis- tillode with a single style terminating in a pair of digitate stigmas. In other species for which this character is known, such as H. africana Vatke, flowers of both sexes have the style divided to the base (Miller & Morris. 1988). One collection, Ravelonanabary 5017, has spherical hairy bodies in the ax- illary position usually occupied by flowers or fruits. These structures may be teratological byproducts (galls?) caused by infection of the reproductive organs with a pest or pathogen. ACKNOWLEDGMENTS This research was conducted while I was a doctoral student at Harvard University in the Department of Organismic and Evolutionary Biology, where I utilized the collections and libraries of the combined Harvard University Herbaria. Travel funds, which permitted me to visit British and European herbaria, were made available through the Atkins Fund and the Anderson Fund, both of Harvard University. This proved to be an exceptionally re- warding opportunity. I am indebted to Dr. Jean Leroy, Muséum National d'Histoire Naturelle, Paris, who authorized the loan of critical material on which this paper is based. Two reviewers, K. R. Robertson and B. Verdcourt, 258 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 offered their suggestions for the improvement of the manuscript. The illustra- tion of Hildebrandtia austinii was prepared by Mr. Wang Le-zhong, whose skill as a botanical illustrator I salute. The larger research program, of which this paper represents only the first part, was supported by the National Science Foundation, Dissertation Improvement Grant BSR 85 00832. I deeply appre- ciate the financial and material support provided by these institutions and individuals. LITERATURE CITED Austin, D. F. 1973. Dactylostigma, a new genus of Convolvulaceae. Phytologia 25: 425-429. Brown, N. E. 1914. Diagnoses Africanae LIX. Bull. Misc. Inform. 1914: 167-171. Hauer, H. 1893. Versuch einer natiirlichen Gliederung der Convolvulaceen auf mor- sia ae! und anatomischer Grundlage. Bot. Jahrb. Syst. 16: 453-591. 98. Uber Ae as Vatke, eine zweite didcische Convolvulaceen-Gat- tung. see 25: 510-5 Lejoiy, J., & S. a “1985. Le genre Ca/ycobolus Willd. (Convolvulaceae) en Afrique tropicale. Bull. Jard. Bot. Nat. Belg. 55: 27-60. Meeusg, A. D. J. 1957 [1958]. The South African Convolvulaceae. Bothalia 6: 641- Metcuior, H. 1964. Convolvulaceae. Pp. 427-429 in A. ENGLER, Syllabus der Pflan- zenfamilien. ed. 12. Vol. 2. Berlin, Borntraeger. Mitier, A. G., & M. Morris. 1988. Plants of Dhofar. Oman, cg of the Adviser for Conservation of the Environment, Diwan of the Royal Cou Myint, T., & D. B. Warp. 1968. A taxonomic revision of the ae Bonamia (Con- volvulaceae). Phytologia 17: 121-239. Roserty, G. 1964. Les genres de Convolvulacées Me pe Boissiera 10: 129-156. STAPLES, G. W. 1987. A revision of Porana Burman f. (Convolvulaceae) and an eval- uation of the Poraneae. Unpubl. Ph.D. ene Harvard University. VeRDcouRT, B. 1961. Notes on the African Convolvulaceae: V. Kew Bull. 15: 1-18. —, 1963. Convolvulaceae. Jn; C. E. HUBBARD & E. MILNE-REDHEAD, eds., Flora of tropical East Africa. London, Whitefriars Press, Ltd 1969. Corrections and aos to the ‘Flora of Tropical East Africa—Con- volvulaceae’: UI. Kew Bull. 23: 2 1974. The genus eae in Madagascar. Ibid. 29: 333-340 981. Anew — of Hildebrandtia (Convolvulaceae) from the Somali Re- public. Ibid. 36: 4 WILLIS, J.C. 1973. i‘ aoe of the flowering plants and ferns. ed. 8 (revised by H. K. Arry SHAw.) London, Cambridge University Press. 1990] HOOVER, BEGONIA PARVIFLORA 259 NOTES ON A NOVEL ABAXIAL LEAF EPIDERMIS IN ECUADORIAN BEGONIA PARVIFLORA W. Scott Hoover'!:? Scanning-electron-microscope observations and stomatal-density counts in- dicate a two populations of Ecuadorian Begonia parviflora have very dif- ferent “o miparee te a population near Banos, Ecuador, a population near Tena , including a mean stomatal density of 634.7 + 26.2 stomata per mm’, guard cells raised on cellular protrusions above the surrounding epidermal cells, and epidermal cells appearing indistinguishable from subsidiary cells. High stomatal density may be of adaptive significance to higher light, or exposure, intensity, although the possibility exists that raised guard cells and comparatively small epidermal cells are an architectural consequence of high stomatal density originating during early leaf development. The genus Begonia L. is recognized for its unusual leaf anatomy, with many species having a multilayered epidermis (Fellerer, 1892; Solereder, 1908; Ha- berlandt, 1909; Metcalfe & Chalk, 1950; Foster & Gifford, 1959; Esau, 1965). Numerous species of Begonia have stomatal clusters, a characteristic limited to species in only a few unrelated angiosperm families (Fellerer, 1892; Metcalfe & Chalk, 1950; Neubauer, 1967; Boghdan & Barkley, 1972; Skog, 1976; Hoo- ver, 1986). By contrast, other Begonia species have singly occurring stomata as well, which is the pattern exhibited by B. parviflora Poep. & Endl. This species is distributed at lower elevations on both the Amazonian and Pacific slopes of the Andes from Colombia to Peru. It is distinct from most species in the genus because it is a small tree (Smith & Schubert, 1941, 1946; Smith & Wasshausen, 1979). One of the most common Andean species of Begonia, B. parviflora is observed frequently as individuals or small populations and occasionally in colonies comprising many individuals. In this study two populations of Begonia parviflora are compared for char- acters of the abaxial leaf epidermis. Although the populations grow in the same region, they differ considerably in density of stomata, length of stomatal pores, structural anatomy of guard cells, and size of epidermal cells. METHODS AND MATERIALS On February 3, 1984, two colonies of Begonia parviflora were sampled on the Amazonian slope of the Andes in Ecuador. The species was frequently ‘Research Associate, Missouri Botanical Garden 2718 Henderson Road, Williamstown, Massachusetts 01267. © President and Fellows of Harvard College, 1990. Journal of the Arnold Arboretum 71: 259-264. April, 1990. 260 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 observed throughout this region, with the colonies sampled being among the largest observed. The Bafios colony was growing in moderate shade with succes- sional vegetation 12 km east of Bafios, Tungurahua Province (lat. 01°14’N, long. 78°22'W), at 1548 m altitude. The Tena population was observed at an exposed site 59 km north of Puyo, Napo Province (lat. 01°14’N, long. 77°53'W), at 906 m altitude. One leaf from each of ten healthy individual plants was collected at each population site, and clear fingernail polish was immediately applied to the lower leaf surfaces, allowed to dry thoroughly, then peeled off with forceps (Sampson, 1961). In the laboratory the epidermal replicas were projected on a screen using an Edmund Scientific microprojector. Stomatal counts and measurements were made on a single mm? area from each peel. Lengths of five stomatal pores were measured on each peel and the mean determined. One clear epidermal replica was chosen from each population for permanent fixation for scanning-electron- microscope (SEM) observations. Preparatory methods for the SEM included placing the peels on aluminum stubs, situating the stubs with the peels in a vacuum, and coating the entire peel with a layer of gold about 150 A thick. The specimens were then placed in the chamber of a Cambridge Stereoscan 100 SEM, observed from 300 to 800 x, and microphotographed at about 300 x. RESULTS AND DISCUSSION The mean stomatal density for the Bafos population of Begonia parviflora is 179.5 + 19.13 per mm?, and that for the Tena population 1s 634.7 + 26.20 per mm? (see TABLE). Data were analyzed using a ¢-test at P < .01: T,, = 44.4. Ficures | and 2 illustrate the difference between the two populations. The stomatal densities for the Tena population are among the highest reported for angiosperms (Salisbury, 1927; Carpenter & Smith, 1975; Abrams, 1987). In the literature available, species reported as having higher mean stomatal density values include (in stomata per mm?) Quercus muehlenbergii Englem. (986), Q. coccinea Muenchh. (760), Rhus copallina L. (731), Acer rubrum L, (705), Fagus grandifolia Ehth. (693), Q. rubra L. (681), Ligustrum vulgare L. (680), Q. prinus L. (662), and Q. sessiliflora Salisb. (656). Stomatal-density measurements from other species of Begonia are very limited, although several populations each of B. heracleifolia Schidl. & Cham. and B. nelumbiifolia Schldl. & Cham. have been sampled and found to have between 50 and 90 stomata per mm? (Hoover, 1986) Ranges and means for stomatal density and pore lengths for B. parviflora populations. STOMATAL-PORE LENGTH NUMBER OF STOMATA/MM? i LocATION/nh Range Mean Range Mean Baiios/10 91.7-280.0 179.5 + 19.13* 8.0-15.0 10.4 + 0.47 Tena/10 530.0-741.8 634.7 + 26.20 6.0-22.0 8.4 + 0.21 *Variance is measured as the standard error of the mean. Ficures 1,2. Scanning electron micrographs of Begonia parviflora, abaxial epidermal surfaces: 1, Banos population, x 330; 2, Puyo population, x 405. (Arrows, left to right: epidermal cell, subsidiary cell, guard cell (raised in FIGURE 2)). The stomata of the leaves from the Tena population seem to be raised above the plane of the epidermal cells (see FiGure 2), and the guard cells appear to sit on the apex of individual cellular protrusions that are built up on subsidiary cells. Raised stomata have been observed in other species of Begonia, including 262 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 B. cathayana Hemsley, B. froebelii A. DC., B. lobulata A. DC., and B. vitifolia Schott (Brouillet, unpublished data). In general this is an uncommon charac- teristic in the plant kingdom, although it has also been reported in several other genera including Pastinaca L., Prunus L., and Solanum L. (Esau, 1965). The stomata of the leaves from the Bafios population do not appear to have this structural modification; here subsidiary cells surround guard cells in the typical way (FIGURE 1). In the leaves from the Tena population, compared to those from the Banos population, the epidermal cells are nearly indistinguishable from the subsidiary cells (see Ficures 1, 2), they are greatly reduced in size and appear to be localized to depressions between the stomata. The stomata, with their asso- ciated subsidiary cells, are so densely packed that the epidermal cells on the surface are the same size as the subsidiary cells. The mean stomatal length is 10.4 + .47 um for the leaves of the Banos population and 8.4 + .21 um for those of the Tena population, a statistically significant difference (T,, = 3.83 (p = .01)). Length of the stomatal pores has been measured in many different species of Begonia grown under cultivation, with results indicating relatively little interspecific variation in this character (Fellerer, 1892). Five different population samplings of B. nelumbiifolia from along an elevational gradient in Hidalgo, Mexico, indicate rather significant variation in pore length (Hoover, 1986), suggesting that this character may not be as stable as previously reported. Rainfall along the Amazonian slope of the Andes is one of the few climatic parameters measured for this region (Schwerdtfeger, 1976). The highest annual rainfall recorded here is found at Tena, Ecuador (annual average of 6235 mm from 1965 to 1969, including the years 1968 and 1969, when accumulation totaled 8380 and 8939 mm, respectively), thus subjecting the Begonia parviflora population from this region to extreme moisture. Such high rainfall is unusual for the entire eastern region of the Andes, including Venezuela, Colombia, Peru, and Bolivia. At Puyo, 59 km south of Tena, rainfall averaged 4294 mm for the same years. No data are available for the region near the Banos pop- ulation, although according to Schwerdtfeger (1976 p. 154) “rainfall distribu- tion is highly variable, mainly because of the impact of relief.” It is important to note that precipitation generally decreases with elevation in this region, and since Bafios is 642 m higher than Tena, it likely receives less moisture and may experience somewhat of a dry season. With only two populations sampled for Begonia parviflora, it is highly spec- ulative to interpret these data from an evolutionary standpoint, even though the populations show such extreme morphological differences that they appear to represent different taxa. Limited data indicate that stomatal density increases when plants grow in exposed, sunny conditions (Jackson, 1967; Abrams, 1987, 1988). For these B. parviflora populations, the high-stomatal-density Tena plants were growing in an exposed, sunny habitat, while the low-stomatal- density Bafios individuals grew under moderately shady conditions, thus sug- gesting that the high stomatal density may be of adaptive value to the exposed conditions. Although such an interpretation conforms with existing evidence, under the circumstances this interpretation may not be entirely adequate to 1990] HOOVER, BEGONIA PARVIFLORA 263 explain the extreme morphological difference between the populations. Ac- commodation on average, of over 600 stomata per mm2, including subsidiary cells, would predictably cause compression and reduction of epidermal cells, raising guard cells above the plane of the epidermal cells and reducing pore length. It would appear that these unusual morphological characteristics may be an architectural consequence of high stomatal density resulting during early leaf development, a process in accord with Gould and Lewontin (1978), who believed that morphological evolution may, in certain instances, be determined by structural change during development. Additional expeditions are planned for Ecuador and Colombia, data from which will assist in more thoroughly describing stomatal variation in B. parviflora and perhaps helping clarify the evolutionary implications of this novel leaf epidermis. ACKNOWLEDGMENTS This work has been supported by grants from the American Begonia Society and its associated branches, and by donations from Howard and Barbara Berg, Martin Johnson, and many other A.B.S. members. I thank William Grant, of the Williams College Biology Department, for the use of the SEM, comment on the manuscript, and his understanding and help over the years. Much gratitude is extended to Marc Abrams, Luc Brouillet, Michael Loik, Bernice Schubert, and David Smith for reviewing earlier drafts of this manuscript. LITERATURE CITED AsraMs, M. D. 1987. Leaf structural and photosynthetic pigment characteristics of three gallery-forest hardwood species in Northeast Kansas. Forest Ecol. & Mana agem. 22: 261-266. . 1988. Genetic variation in leaf morphology and plant and tissue water relations during drought in Cercis canadensis L. Forest Sci. 34: 200-207. BoGuDAN, K. S., & F. A. BARKLEY. 1972. Stomatal patterns in the genus Begonia. Phytologia 23: 327-333. CARPENTER, S. B., & N. D. Smitu. 1975. Stomatal distribution and size in southern Appalachian hardwoods. Canad. J. Bot. 53: 1153-1156. Esau, K. 1965. Plant anatomy. J. Wiley and Sons, Inc., New York. FELLERER, C. 1892. Anatomie und Systematik | Begoniaceen. Unpubl. Ph.D. thesis, University of Munich, West Germany. Foster, A. S., & E. M. Girrorp, Jk. 1959. Comparative morphology of vascular plants. H. Freeman and Co., San Francisco. Gou Lb, S. J..& R.C. Lewontin. 1978. The spandrels of San Marco and the Panglossian paradigm: a critique of the adaptationist programme. Proc. Roy. Soc. London 205: 1-598. HABERLANDT, G. 1909. Physiologische Pflanzenanatomie. ed. 4. Wilhelm Engelmann, elpzig ee W. S. 1986. Stomata and ee: oa in Begonia: ecological response n two Mexican species. Biotropica 18: eee L. W. R. 1967. Effect of a on cee surfaces of deciduous tree species. Ecology 48: 498, 499. METCALFE, C. R., & L. CHALK. 1950. Anatomy of dicotyledons. Clarendon Press, Oxford 264 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Neusauer, H. F, 1967. Bemerkungen iiber den Bau der Begoniaceen. Ber. Deutsch. Bot. Ges. 80: 80-97. SALISBURY, E. J. 1927. On the causes and ecological significance of stomatal frequency with special reference to the woodland flora. Philos. Trans., Ser. B. 216: 1-65. SAMPSON, J. 1961. A method of replicating dry or moist surfaces - examination by light microscopy. Nature 191; 932, 933. SCHWERDTFEGER, W. 1976. Climates of Central and South America. ie survey of climatology. Vol. 12. Elsevier Scientific Publishing Co., Amsterda Sxkoc, L. E. 1976. A study of the Gesnerieae, with a revision of eae (Gesneriaceae: Gesnerioideae). Smithsonian Contr. Bot. 29: 1-182. SmitH, L. B., & B. G. a ERT. 1941, Flor of Peru: Begonia. Field Mus. Nat. Hist. Bot. Ser. 13: 182-2 & 1946. an sae of Colombia. Caldasia 4: 77-107. & D. C. WASSHAUSEN. 1979. Begonia of Ecuador. Phytologia 44: sit 256. sian H. 1908. Senne anatomy of the dicotyledons. Vols. 1, 2. Clarendon s, Oxford. 1990] WANG, SELAGINELLA 209 NOTES ON THE SPECIES OF SELAGINELLA FROM GUIZHOU, CHINA PEISHAN WANG! hirty species of Se/aginella are currently known from Guizhou Province in south-central China. A key to these species is given. Following the key are notes on species that are newly discovered in the province, previously unde- scribed, or otherwise in need of discussion. Selaginella Beauv. in China has not been studied thoroughly since Alston enumerated the Chinese species in 1934. In some provinces or regions, how- ever, the vascular flora, including Se/aginella, has been treated; currently about 50 species of this genus are estimated to occur in China, with 30 species known from Guizhou (Kweichow) Province in the south-central part of the country. KEY TO THE SPECIES OF SELAGINELLA IN GUIZHOU 1. Main stems erect, suberect, or scandent, or plants with densely tufted stems; plants usually rooting at base or in lower 2. Stems densely tufted or branched from base; plants xerophytic. 3. Stems red, branched from base; lateral and median leaves directed forward. Ge Ear Sees ng sas Gee Ssh mee eee ee 1. S. sanguinolenta. 3. Stems brown, densely tufted; only median leaves directed a Ts sks Soh Whee eat eed soe deena ee ces ce oak Ant pe acd . S. pulvinata. 2. Stems not branched near base, or if branched from base an not xerophytic. ants at most 6 cm high; main stems more or less zigzag. ............. ie isa tea ween fuges ties avy cesar dente eras aed aie teeny aes 3. S. kouycheensis. 4. Plants much larger; main stems not zigzag. Plants over 1 m long, scandent. ..................... 4. S. helferi. 5. Plants generally less ian 60 cm tall, not scandent. 6. Branches pubesc 7. Stems a ‘eaves wrinkled when dry; bbee ae north- eastern Guizhou. ......................0-. S. braunit. 7. vay ice leaves not wrinkled when oe suthe erm Gul Sta oe ayes Bn ua g ode oes S. flagellifera. 6. Branches alabro ante branches on main stems appressed and directed upwar 9, Leaves near base of main stem l median leaves not ee margined; lateral leaves with pasion nee STNG sa Sous uin4 eet dcomer ata 7. S. involvens. : ae near base of main stem distant; aoe leaves white margined; lateral leaves with basiscopic margin ser- WUIAIG:: & beeen aut cuesenemeuces: 8. S. moellendorffii. \O ‘Guizhou Institute of Biology, Guiyang, Guizhou, China. © President and Fellows of Harvard College, 1990. Journal of the Arnold Arboretum 71: 265-270. April, 1990. 266 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 8. Leaves below branches on main stems spreading. 10. Median leaves entire; lateral leaves entire or subentire. 11. Plants stout; lateral leaves entire; distal part of stems black when dry. ................005. 9. S. picta.? 11. Plants delicate; lateral leaves slightly serrulate at apex; distal part of stems straw colored or brownish when TY cents Lote nie ee Spee 10. S. delicatula. 10. Median and lateral leaves serrulate or ciliate. 12. Leaves white margined. ....... 11. S. sichuanica. 13. Sporophylls dimorphic, those on upper (dorsal) side of strobilus different both in shape and size from those on lower (ventral) side. 14. Plants up to 40 cm or more tall; stems in- cluding leaves up to 7 mm wide. 15. Lateral leaves with long cilia at base; median leaves acuminate. ........... 12. S. bodiniert. 15. Lateral leaves short- ia or serrulate at base; median leaves aristate. 16. Lateral leaves ovate to ovate-lan- base; species of ees wae for- dima oie ese ee . labordei. . Lateral leaves i oe eeeestinte median leaves not cordate at base; species of low hills and valleys (be- low 700 m ae Madore ees ee eee S. monospora. 14. Plants ener less ae x cm tall; stems a uding leaves up to 5 mm wide. 17. Median ee cordate at base. ...... 13. S. labordei. 17. Median pie not cordate at base. an leaves strongly aristat nes margin of lateral icone entire or subentire. ............. re eee ere . S. effusa. . Median leaves acuminate: lateral leaves with margin distinctly ser- rulate throughout. .............. 16. S. heterostachys. ro [o-e) 13. gee monomorphic. n leaves long-aristate; pe leaves ok adaxially. ...17. S. trachyphylla. . Median leaves acuminate to on lateral leaves smoot 20. Leaves cennilate: median leaves keeled. 18. S. doederleinii. 20. Leaves ciliate; median leaves not keeled. —_ Ne} This species is found in the Guangxi Zhuang Autonomous Region near the border with Guizhou, it may also grow in Guizhou. 1990] 1. WANG, SELAGINELLA 267 2 N Main stems creeping; rooting througho 22. All leaves entire and with di istinct om 1 Lateral leaves symmetrical or sub- symmetric, with cilia only at base; valley species of southern border of rovince. ....... 19. S. repanda. . Lateral leaves asymmetric, with cil- ia from base to near middle on ac- roscopic margin; species of lime- stone areas of province. ......... Lae en eee ae 20. S. omeiensis. oe eee eta 21. S. uncinata. argin 22. Median and lateral leaves serrulate or cag with or without white margin. 23. Median and lateral leaves serrulat 24. Sporophylls similar to leaves in et shape and arrangement, not form- Piao pe eee ae eee 22. S. nipponica. 24. Sporophylls different from leaves, forming strobili. 5. Plants over 20 cm long; axillary leaves oblong to sai oo ing distinct strobili. entire; strobili with | cal. er- Ss. a 25. Plant usually less than 10 cm long; axillary leaves spreading- ovate, rrulate; angia spher ceo with more than | megasporangium; microspo- 24. [oa hoes Geen tare S. liboensis. 23. Median and lateral pee eae 26. Lateral leaves white margined. 27. Plants 5-10 cm ee oe with long cilia around margin. ..... S. eee 27. oo up to 30 cm or more; leaves ciliate at base, serrulate toward 6. S. eee eth CAL aan oeaaka eens gebaueriana. 26. ier ee not white margined. 28. Sporophylls on upper (dorsal) side of strobilus smaller than those on lower (ventral) side. ......0..00.0.....00.. . prostrata. 28. Sporophylls on upper (dorsal) side of strobilus larger than or nearly equal to ones on lower (ventral) side. 29. Lateral leaves with base symmetrical, cordate. ............ 28. Hed fice wants aa ae alee ed se a oun ee 5: eee 29. Lateral leaves with base asymmetric, not corda caf margin ciliate at base, ee pd apex; lat- Ww j=) eral leaves ovate, acute to acuminate at apex va 1e: Raed ee 29.8. aan 31. Median leaves lanceolate. ......... 30. S. compta. . Leaf margin densely ciliate re lateral ee fal- 31. cate, acuminate at apex NOTES Preece S. drepanophylla. 3. Selaginella kouycheensis A. Léveillé, Repert. Spec. Nov. Regni Veg. 9: 451. 911; Alston, Bull. Fan Mem. Inst. Biol. 5: 290. 1934. I have studied the type of this species Esquirol 2158 (£) and have found Léveillé’s description of it to be incorrect. Mainly, it is not a creeping species but an erect one. Individuals that I have seen are about 1-3 cm tall (Léveillé’s original description indicates a height of 6 cm), and the strobili have only megaspores. The leaves and sporophylls are both serrulate, not entire. 268 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 Léveillé cited two collections in his original description of Se/aginella kouy- cheensis, Cavalerie 1933 and Esquirol 2158. Alston’s (1934) citation of Esquirol 2158 as the type can be construed as lectotypification, a designation that is accepted here. His other citation, Cavalerie 4156 (kK), as he had pointed out, is a mixed collection. There are two other species, S. heterostachys Baker and S. delicatula (Desv.) Alston (?) in addition to S. kouycheensis. A specimen from Yunnan, C. W. Wang 78983 (Meng-pung, Jenn-yeh Hsien (A)) is also Selaginella kouycheensis, a new record for the province. 5. Selaginella braunii Baker, Gard. ve 1867: 1120. 1867; Alston, Bull. Fan Mem. Inst. Biol. 5: 281. a S. hieronymi Alderw. Bull. Jard. Bot. Buitenzorg, ser. 2. 1: 18. a This species is known to occur only from Zhejiang, Anhui, Hubei, Sichuan, and Guizhou provinces, all located in the Changjiang River basin. From Gui- zhou the following recent collections can be cited: Tongren, East Guizhou Exped. 75325 (GAs, PE); Tongzi, P. S. Wang 77185, 77316 (both HGas), Songtao, East Guizhou Exped. s.n., June 1975 (HGAS). However, there is reason to believe that it also grows in Yunnan Province. While studying the materials in the Harvard University herbaria (A and Gu), I found a specimen, J. Delavay 339(Tsin-tan, Yunnan Province), that was cited by Alston (1934) as Selaginella elephantopus Hand.-Mazz. However, based on its pubescent stems and wrin- kled leaves, it is S. braunii. 6. Selaginella flagellifera W. Bull, Cat. no. 225: 9. 1886. S. biformis A. Br. ex Kuhn, Filic. Afr. 189. 1868, nomen nudum, A. Brown ex Kuhn, Forsch- ungsr. Ges. 4(6): 19. 1889; Alston, Bull. Fan Mem. Inst. Biol. 5: 282. 1934: Philip. J. Sci. 58: 374. 1935; Fl. Gén. Indo-Chine, 7(2): 570, 571. 1951; Ching, Flora Hainanica I: 10. 1964. This is a tropical Asian species and is at the northern limit of its range in southern Guizhou. It is similar to Selaginella involvens (Sw.) Spring and 5S. moellendorffii Hieron. but is easily distinguished from them by its pubescent branches. Collections from Guizhou examined are: Congjiang Xian, East Guizhou Exped. 75228 (ucas, PE); Anlong Xian, Zhang & Zhang 3304 (PE). Although Bull indicated that the species came from Fiji, Alston (1934) placed the name in the synonymy of Se/aginella biformis and gave the range as Burma to Sumatra and Celebes. 11. Selaginella sichuanica H. S. Kung, Acta Bot. Yun. 3: 252. 1981. Recent collections from Guizhou (Fanjingshan, Jiangkou Xian, Sino-Brit. Bot. Exped. F0673 (uGas, PE); Sino-Amer. Guizhou Bot. Exped. 480 (A, CAS, HGAS, PE)) constitute a new record for the province. The species was previously known only from Sichuan. 19. Selaginella repanda (Desv. in Poiret) Spring 77 Gaudich. Voy. Bonite, Bot. 1: 329. 1846; Alston, Bull. Fan Mem. Inst. Biol. 5: 293. 1934. Lyco- podium repandum Desv. in Poiret, Encycl. suppl. 3: 558. 1814. 1990] WANG, SELAGINELLA 269 One collection of this species (Guizhou: Luodian, by Hongshui River bor- dering Guangxi, 280 m, P. S. Wang 76321 (HGAsS)) constitutes a new record for this province. 20. Selaginella omeiensis Ching in H. S. Kung, Acta Bot. Yun. 3: 253. 1981. This is a common species in limestone areas of Guizhou. It is much like Selaginella bodinieri Hieron., only smaller in size and with uniform sporo- phylls. The sporophylls and size of S. bodinieri are variable. Possibly they are the same species, but further evidence is needed before a final decision can be made. 25. Selaginella albociliata P. S. Wang, sp. nov. Habitu S. chaetolomae Alston, sed in foliis albo-marginatis, cillis foliorum longioribus densioribus, differt. Plant 5-10 cm long; stem prostrate, creeping, ca. 3 mm in diameter including leaves, sparingly branched, rooting throughout. Vegetative leaves dimorphous: lateral leaves patent, ovate-oblong, 1.5—2 mm long, |-1.3 mm wide, round at base and blunt or acute at apex, the margin white, with numerous long cilia, these patent or ascending, up to 0.3 mm long; median leaves ovate, 1-1.3 mm long, 0.5-0.8 mm wide, round at base, aristate at apex, the margin white, with numerous long cilia. Strobili 6-10 mm long, 2—2.5 mm in diameter. Sporophylls dimorphous: larger ones on upper (dorsal) side of strobilus, oblong-lanceolate, 2-2.5 mm long, 0.7 mm wide, acuminate at apex, the margin white, ciliate; smaller ones oblong-ovate, |.8-2 mm long, 0.7 mm wide, caudate at apex, the margin white, ciliate. Megaspores light gray; microspores tangerine colored, 30 um in diameter, verrucate. Type: Guizhou, Libo Xian, on limestone surface along the Wujia River, 530 m, 20 May 1988, P. S. Wang 7798] (holotype, HGAS). ADDITIONAL SPECIMENS EXAMINED. GUIZHOU: Libo Xian, in soil on wet rock surface in forest, 660 m, P. S. Wang 7680] (cDBI, HGAS). 26. Selaginella gebaueriana Hand.-Mazz. Symb. Sin. 6: 9. 1929; Alston, Bull. Fan Mem. Inst. Biol. 5: 274. 1934. Alston (1934) listed this as a synonym of the earlier-published Se/aginella davidii Franchet. Kung (1981) stated that the two taxa are distinct and that S. gebaueriana is mainly from southwestern China, while S. davidii is from north- ern China. In its larger shoots and ciliate leaves, S. gebaueriana may easily differ from the northern species; consequently, I agree with Kung. 28. Selaginella chaetoloma Alston, J. Bot. 70: 67. 1932; Bull. Fan Mem. Inst. Biol. 5: 292. 1934. After studying types of Se/aginella chaetoloma (Cavalerie 731 (BM, E)) and additional specimens (Cavalerie 285 (kK), 517 (kK), and 1469 (g)), I find that S. prostrata H. 8. Kung (Kung, 1981) does not differ significantly, except in the 270 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 sporophylls, from SS. chaetoloma. In S. chaetoloma the sporophylls on the lower (ventral) side of the strobilus are smaller than those on the upper (dorsal) side; in S. prostrata the opposite is true. Alston (1932) cited specimens of Cavalerie 731 at both BM and E as types in the original description but did not designate either as holotype. I hereby designate the specimen at BM, the more complete specimen of the two, as the lectotype 31. Selaginella drepanophylla Alston, J. Bot. 70: 66. 1932; Bull. Fan Mem. Inst. Biol. 5: 292. 1934 I record here a second report for the species in China (Guizhou: Fanjingshan, Jiangkou Xian, Sino-Amer. Guizhou Bot. Exped. 68, 116, 367 (all A, HGAS, PE)). The type locality is in Guangxi, south of Guizhou. Alston (1934) pointed out that Selaginella xipholepis Baker, S. drepanophylla, and S. compta Hand.- Mazz. may be different forms of a single species. Recently, Dahlen (1988) treated S. drepanophylla as a synonym of S. xipholepis. After comparing the type of S. xipholepis with S. drepanophylla from Guizhou, I believe that they are very similar but separable species. Selaginella compta 1s also very similar to S. xipholepis, but additional specimens—especially those from Yunnan— should be checked before a decision is made on the status of these two taxa. ACKNOWLEDGMENTS I am grateful to the Royal Botanic Garden, Edinburgh, the Royal Botanic Gardens, Kew, and the British Museum (Natural History) for the loan of types of Selaginella. The staff of the Harvard University Herbaria provided the opportunity for me to study, and the Arnold Arboretum of Harvard University gave me financial support. I wish to express my special thanks to Drs. Alice R. and Rolla M. Tryon for their guidance and advice, and to Dr. David E. Boufford for his assistance in consulting literature and reading the manuscript. Professor H. S. Kung, from the Chengdu Institute of Biology, Academia Sinica, has generously helped me with specimen identificaiton and taxonomic prob- lems on Se/aginella during the recent years. LITERATURE CITED ALSTON, : H. G. 1932. Notes on Selaginella. Il. J. Bot. 70: 61-67. 34. An enumeration of the Chinese species of Se/aginella. Bull. Fan Mem. In st. el 5: 261-304. DAHLEN, M. A. 1988. Taxonomy of Selaginella: a study of characters, techniques, and classification in the Hong Kong species. Bot. J. Linn. Soc. 98: 277-302. Kuna, H.S. 1981. Gonbutons to the genus Selaginella Beau. from Sichuan. Acta Bot. Yun. 3: 251-256. 1990] BOOK REVIEW 271 BOOK REVIEW Genera Graminum, by W. D. Clayton and S. A. Renvoize. Kew Bulletin Ad- ditional Series XIII. Her Majesty’s Stationery Office, London. 1986. 389 pp. ISBN 0-11-250006-4. £25.00 softcover. Grass Genera of the World, by L. Watson and M. D. Dallwitz. Australian National University, Research School of Biological Sciences, Canberra. 1988. 45 pp., 5 microfiche, and 3 5.25” disks. ISBN 0-7315-0326-0. A$40.00 softcover. Grass taxonomy underwent a revolution in the mid-twentieth century, as anatomical and cytological data were incorporated into tribal and subfamilial classifications. Since that time, more and more data have accumulated, par- ticularly on anatomical characters, and an integrated worldwide treatment of the family has become not only possible but highly desirable. These two pub- lications appeared nearly simultaneously, with early versions of the Watson and Dallwitz work appearing in 1985 and 1986. They are reviewed together here because they reflect quite different approaches to the same end. The Genera Graminum represents a classical treatment of the family. It is a compact book, well laid-out and easy to use. The book is divided into two parts: Part I, The Grass Plant, includes sections on morphology, reproduction, anatomy and metabolism, classification, grasslands, and evolution, as well as a brief introduction to the descriptive treatment; Part II is an enumeration of the genera, preceded by a key to the tribes. The synonymy for each subfamily, tribe, and genus 1s presented, followed by a diagnostic description. One of the great strengths of the book 1s the nomenclature, which is detailed, complete, and with typification, and will serve as a valuable aid for future taxonomic work on the group. Clayton and Renvoize recognize six subfamilies, a welcome relief for American agrostologists, whose major source for keys has been A. S. Hitchcock’s Manual of the Grasses of the United States, which uses Robert Brown’s original division into only two subfamilies. The characters used by Clayton and Renvoize are primarily those of gross morphology: habit, inflorescence, spikelet, and floret characters. The ““modern” (i.e., micromorphological) characters are given short shrift. Anatomical char- acters are summarized briefly for each tribe, but there is no way of assessing infratribal variation, even though this is known to be considerable in some tribes (e.g., the Paniceae, which contains both C, and C, members, members with single and with double bundle sheaths, members with each of the three known decarboxylating enzymes, etc.). This means that using the Clayton and Renvoize treatment for any other analyses is virtually impossible. A diagram of relationships 1s presented for the genera of each tribe, and these are already becoming widely used, in part because they are complete. Clayton and Renvoize (p. 23) issue the disclaimer that “the diagrams are intended to © President and Fellows of Harvard College, 1990. Journal of the Arnold Arboretum 71: 271-273. April, 1990. aie JOURNAL OF THE ARNOLD ARBORETUM [vVoL. 71 give a visual impression of phenetic relationships, progressing from simple to complex structures; they obviously have phylogenetic implications, but no attempt has been made to treat these rigorously.”’ Unfortunately, I have seen these diagrams used in presentations by agrostologists as if they were phylo- genetic. This is the classic problem with nonphylogenetic descriptions of “‘re- lationships”’: authors may know the limitations of their work, but readers persist in interpreting relationships phylogenetically. I find that the greatest weakness of the book is unfortunately in the rationale for statements of similarity. For example, Cryptoch/oa (Bambusoideae) is said to be “linked to O/yra through O. longifolia” (p. 63), or Wangenheimia (Pooi- deae) is “related to Vulpia pectinella” (p. 97). There is no evidence given for any such statements, and the diagnostic descriptions rarely provide enough information for this reader to make the comparison for herself. This means that anyone who wants to use those observations as the basis for further work must reinvent the wheel—reexamine the specimens, reevaluate the characters, and try to guess the basis for Clayton and Renvoize’s assertion — before building on the observations. Watson and Dallwitz (whose early works are curiously not cited by Clayton and Renvoize) have summarized current knowledge of the genera of the grasses in the form of an automated database. Their publication is thus only 45 pages long and is simply a set of good-quality photographs illustrating the characters used in the descriptions. The generic descriptions themselves are in the mi- crofiche appended to the book. In addition, the publication comes with three floppy disks containing the database and the interactive program INTKEY (an MS-DOS program), used to access the information. The descriptions in the microfiche are complete and parallel, with the majority of the 430 characters described for most of the 761 genera. Characters are both macro- and micro- morphological, and include geographic distribution and number of species. Synonymy is included for each genus, but it is not as detailed as that in Clayton and Renvoize; type species, for example, are not cited. INTKEY is Dallwitz’s development of Pankhurst’s ONLINE and is one component of DELTA (the Descriptive Language for Taxonomy), a set of pro- grams written by M. Dallwitz and T. Paine for creating and manipulating a taxonomic database. It functions as a multiple-entry key, thus obviating the need for any other keys in the publication. It is extremely easy to use. Because the database contains so many characters (including geography), identification can be performed rapidly with only fragmentary specimens and frequently without necessitating a lot of detailed knowledge of grass morphology. This is particularly important for a family like the Gramineae, regarded by many botanists as abstruse; identifications can easily be done by nonspecialists. The database, used in conjunction with INTKEY, is a powerful tool with many helpful features that add up to tremendous flexibility. As one example, it keeps track of which characters are not recorded for particular genera, so the user can distinguish between the character states “not present” and “not known.” This 1s important not only for data analyses, but also as a guide to how well sampled a character is in general, something that is impossible with the di- agnostic descriptions of Clayton and Renvoize. Another detail that I have used 1990] BOOK REVIEW 273 extensively in cladistic studies of the family is the ability to choose any set of taxa (e.g., Pooideae of North America) and to summarize the variation within the group for each character. Thus, for example, for character 18 (culm inter- nodes solid or hollow), a single command will be able to tell you that for the particular set of 75 genera, 64 have information on the character, eight have solid stems, and 62 have hollow stems (the fact that the numbers sum to more than 64 indicates that s1x genera have both hollow- and solid-stemmed species). I could go on for some paragraphs describing the many useful aspects of the grass database plus INTKEY. Suffice it to say that I have found nothing in book form to compare with it. The data are highly manipulable and are easily updated and corrected. The database contains data amalgamated from the literature and from observations made by Watson and his colleagues. The parts relying heavily on literature surveys can be unreliable, but as the data are used by more and more people for more and more purposes, errors are gradually being corrected. Updating the database is an ongoing process. Watson and Dallwitz have not included any speculation on the evolutionary history of the family; their classification is the result ofa set of phenetic analyses, and the listing of genera assigned to each tribe is printed on the microfiche and also included in the database itself. The classification is not particularly finely resolved, and there is no discussion of possible sister-group relationships—nor indeed of any relationships below the tribal level. This is reasonable in that the authors see the database as a general-purpose tool, to be available for many different sorts of evolutionary and taxonomic studies. It is an unfortunate characteristic of systematists that more-detailed classi- fications are preferred over less-resolved ones, even if there is little support for the relationships implied by the detailed classifications. Thus, Clayton and Renvoize are already widely cited for their statements of relationship despite the fact that, as pointed out above, the reasons for these are largely undocu- mented. It seems that many workers are happy to be able to turn to a book, lift out a diagram, and assume that the relationships illustrated are accurate. The classification provided by Watson and Dallwitz, in fact, goes about as far as the available data allow; for example, neither phenetic nor cladistic analysis of the available data supports tribal oo. in the Chloridoideae, except for the well-marked Pappophoreae and Triodie To conclude, I have found the Watson a ne publication to be by far the more flexible of the two, and it gives a much better understanding of character distributions. I find its only serious limitation is that some institutions (mine included) do not have appropriate computer hardware in convenient proximity to their grass collection, so the tremendous flexibility of the auto- mated approach is outweighed by the necessity of moving the specimens to the computer. But if one is interested in identifying grasses, or in any sort of data analysis or manipulation, it is clearly the publication of choice. However, for quick reference in the absence of a computer, and for the typification of generic names, I use Clayton and Renvoize. Any herbarium and any agros- tologist will probably want to own both of these. Given the total price, it is hard to resist the combination.—E. A. Kellogg, Harvard University Herbaria, 22 Divinity Avenue, Cambridge, Massachusetts 02138. Now Available and Complete FLORA OF THE LESSER ANTILLES by Richard A. Howard and Collaborators The Flora of the Lesser Antilles, a long-term project of Dr. Richard A. Howard, former director of the Arnold Arboretum, has been brought to a conclusion with the publication of volumes 5 and 6 during 1989. Other volumes in the series are still available, either individually or as part of a complete set: VOLUME |. Orchidaceae $20 VOLUME 2. Pteridophyta 25 VoLuME 3. Monocotyledoneae 35 VOLUME 4. Dicotyledoneae, Part | q3 VoLuME 5. Dicotyledoneae, Part 2 85 VoLuME 6. Dicotyledoneae, Part 3 85 Special price for the complete 6-volume set $260* Special price for volumes 4-6 $2057 Also Available ARNOLD ARBORETUM PLANT INVENTORY A comprehensive plant inventory of the living collections of the Arnold Arboretum, together with a statement of the Arboretum’s accession policy and a fee schedule for obtaining propagating material from the collections. S21* Orders with payment in U.S. funds, including a $2 ($4 foreign) shipping and handling charge per book, should be addressed to the attention of Frances Maguire, Arnold Arboretum, 125 The Arborway, Jamaica Plain, Massachusetts 02130, U. S. A. Checks should be made payable to the Arnold Arboretum. *Prices marked with an asterisk include postage and handling. Journal of the Arnold Arboretum April, 1990 CONTENTS OF VOLUME 71, NUMBER 2 The Genera of Arundinoideae (Gramineae) in the Southeastern United States. GoOnmpon C. TUCKER dosiie iu de vdadd inaen ewes iui eaooewenes 145-177 Linnaeus, the Cortex-Medulla Theory, and the Key to His Under- standing of Plant Form and Natural Relationships. P. F. STEVENS AND S. P. CULLEN ........00.0000 000 cece ee eees 179-220 A Revision of Weberbauera (Brassicaceae). IHSAN A. AL-SHEHBAZ ..40.05c0 cusecus cuseseenyeawevaevdess 221-250 Preliminary Taxonomic Consideration of the Poraneae (Convol- vulaceae). FORGE W. STAPLES 4 cgecs oa ou Gawsed sh es sada dde an ee oesenea 25 jad oe Notes on a Novel Abaxial Leaf Epidermis in Ecuadorian Begonia parviflora. W.. SCOTT PIGOVER: oie v4-di- hatha dns denned 4G Wa eareieia bn ce HO awa we 259-264 Notes on the Species of Se/aginella from Guizhou, China. PEISHAN WANG ...........0 00 ccc cece eben eee nneeees 265-270 BOG IREVIEW o scacecucusazuaue sand dadedemcboads ausapeiecaaseneth agp puardh'eudeoseeais. atensanaw doen A 271=273 Volume 71, Number |, including pages 1-144, was issued 25 January 1990. JOURNAL OF THE ARNOLD ARBORETUM HARVARD UNIVERSITY VOLUME 71 NUMBER 3 ISSN 0004-2625 Journal of the Arnold Arboretum The Journal of the Arnold Arboretum (ISSN 0004-2625) is published quarterly in January, April, July, and October for $70.00 per year, plus $10.00 or $15.00 postage for addresses outside of the United States, by the Arnold Arboretum of Harvard University. It is printed and distributed by Allen Press, Inc., 1041 New Hampshire Street, Lawrence, Kansas 66044. Second-class postage paid at Lawrence, Kansas. POSTMASTER: send address changes to Journal of the Arnold Arboretum, % Allen Press, Inc., P.O. Box 368, Lawrence, Kansas 66044. Because publication of the Journal of the Arnold Arboretum is being suspended after Volume 71 (1990) has been completed, subscriptions can only be accepted for the sent volume. Remittances for Volume 71 should be sent to Journal of the Arnold Arboretum, 1041 New Hampshire Street, Lawrence, Kansas 66044, U.S.A. Claims will not be accepted after six months from the date of issue. Volumes I-51, reprinted, and some back numbers of volumes 52-56 are available from the Kraus Reprint Corporation, Route 100, Millwood, New York 10546, LA EDITORIAL COMMITTEE S. A. Spongberg, Editor E. B. Schmidt, Managing Editor P. F. Stevens, Book Review Editor P. S. Ashton K. S. Bawa C. E. Wood, Jr. Printed at Allen Press, Inc., Lawrence, Kansas COVER: The stylized design appearing on the Journal and the offprints was drawn by Karen Stoutsenberger. THIS PUBLICATION IS PRINTED ON ACID-FREE PAPER. JOURNAL OF THE ARNOLD ARBORETUM VOLUME 71 Jury 1990 NUMBER 3 THE GENERA OF CUPRESSACEAE (INCLUDING TAXODIACEAE) IN THE SOUTHEASTERN UNITED STATES! JEFFREY A. HART? AND ROBERT A. PRICE? CUPRESSACEAE Bartling, Ord. Nat. Pl. 90, 95. 1830, ““Cupressinae,”” nom. cons. (CYPRESS FAMILY) Aromatic, resinous, evergreen or sometimes winter-deciduous, monoecious (or dioecious in Juniperus [and Dise/ma]) trees or shrubs. Bark fibrous and 41 ran .~Tl £4} rel 1 ‘Prepared for United States, a long-term ee made possible through the support of National Science Foundation Grants BSR-8415769 (C. E. Wood, Jr., principal investigator) and BSR-8716834(N. G. Miller, principal investigator), under both ne this account was prepared. The 133rd in the series, this paper follows the format established in the first one (Jour. Arnold Arb. 39: 296-346. 1958) and continued to the present. The area covered by the Generic Flora includes North and South Carolina, Georgia, Florida, Tennessee, Alabama, Mississippi, Arkansas, nd Louisiana. The descriptions are based primarily on the plants of this area, with information about extraregional members of a family or genus in brackets [ ]. The references that we have not verified are marked with asterisks. We thank Carroll Wood and Norton Miller for the opportunities afforded by participation in the Generic Flora project and for their guidance in the study, and Rudolf Schmid for bibliographic f Callitris in Florida. Library and herbarium collections at Harvard University and the University of California, Berkeley, were consulted in this study, and we wish to thank an Stal st these institutions for ib assistance: ine illustration of Chamaecyparis thyoides (based on p n Norfolk Co. s drawn by Karen Stoutsenberger under the direction of Carroll eae while that of Juniperus cone on cultivated plants from the Sai of California Botanical Garde and Demaree 23779, ee 67-310, and McCabe 416, all at uc) was drawn by Linda Vorobik a the direction of Robert Pri his paper 1s ublished i in oan as contribution number 640 from the New York State Science Service. 21547 33rd Street, Sacramento, California 95816. Please address reprint requests S Norton Maes, noon 3132 CEC, New York State Education ss nt, Albany, New York 12230, ‘Biological Survey, New York State Museum, The State Education Department, nae. New York 2230. Current nddece Biology Department, Indiana a Bloomington, Indiana 47405. © President and Fellows of Harvard College, 1990. Journal of the Arnold Arboretum 71: 275-322. July, 1990. 276 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 separating into long strips or exfoliating in plates. Wood with axial parenchyma, lacking resin canals [sometimes with traumatic resin canals in Sequoia and related genera] and ray tracheids [reported only for Chamaecyparis nootkaten- sis]. Branches erect or spreading; branchlets erect to pendulous, terete or qua- drangular, sometimes flattened in frondlike horizontal sprays (in Cupressaceae s.s.). Foliage leaves entire to minutely serrate, linear to linear-lanceolate and spirally arranged (often apparently 2-ranked by twisting of leaf bases) [or op- posite in Metasequoia] in the traditional Taxodiaceae, or decussately opposite or in whorls of 3 [rarely 4] and scalelike or sometimes awl or needle shaped in the Cupressaceae s. s.; foliar resin canals 1 [to 3]. Pollen cones (microspo- rangiate strobil1) sessile or short stalked, solitary or variously clustered, terminal on leafy branches or short shoots (or sometimes eee microsporophylls spirally arranged, decussately opposite, or ternate, ca. 6 to 24 per strobilus; microsporangia 2 to 10 per sporophyll, globose or allloscid fone taudinailly dehiscent, pendulous, free, in | or 2 abaxial rows; pollen spheroidal, without saccae or prothallial cells; intine thick, exine with surface microverrucate, with papilla or obscure germinal aperture; male cells approximately equal in size. Ovulate cones solitary (sometimes secondarily clustered), subglobose to ovoid [to pyramidal], terminal or axillary, maturing in 1 to 3 seasons; bract and scale components largely fused in the mature cone, with bract tip protrusive or inconspicuous; bract-scale complexes spirally arranged or decussately opposite or ternate, each with | to 6 [to 10, or rarely to 20 in Cupressus] erect [or ultimately inverted], adaxial, bottle-shaped ovules; archegonia aggregated in terminal (or lateral) complexes; scales peltate or broadly ovate to triangular- ovate or oblong, thickened or strongly flattened at maturity, imbricate or val- vate, [2 to 4] 6 to ca. 20 [to 40 or more] per cone, ultimately woody and separating (or fleshy and fused into a berrylike structure in Juniperus). Seeds 1 to several per scale, with [1 or] 2 (or 3) lateral [or nearly terminal] wings derived from the seed coat, or wingless; seed coat varying from thin to very thick and woody, often with resin ducts; cotyledons 2 to 6 (to 9 in Taxodium). Chromosome number usually 2” = 22 (occasionally 33, 44 [66 in Sequoia sempervirens]). (Including Taxodiaceae Warming, Haandb. Syst. Bot. ed. 2. 163. 1884; Juniperaceae Schaffner.) TypE GENUS: Cupressus L. A family of 29 genera* and approximately 130 to 140 species concentrated in the temperate portions of the Northern and Southern hemispheres, but with some species in boreal and austral areas and with Juniperus procera Hochst. extending into tropical montane areas of eastern Africa. Of the eight genera native in North America, Chamaecyparis Spach, Juniperus L., Taxodium Rich., and Thuja L. occur in the southeastern United States, while Ca/locedrus Kurz, Cupressus L., Sequoia Endl., and Sequoiadendron Buchh. occur only in the western part of the continent. About 16 of the genera are monotypic, although several of these had much larger ranges (and presumably greater numbers of species) in the Tertiary than they do now (Florin, 1963). The largest and most Or fewer, if one treats Libocedrus Endl. in the broad sense, including some or all of the segregate genera . es Florin & Boutelje, Papuacedrus H Li, and Pilgerodendron Florin, as was variously done by De Laubenfels io 1988) and beennald (1976). 1990] HART & PRICE, CUPRESSACEAE rae | widespread genera are Juniperus (50 or more species) and Cupressus (perhaps 13 species) in the Northern Hemisphere, and Callitris Vent. (ca. 15 species) in the Southern Hemisphere. Most recent treatments of the conifers have followed Pilger in recognizing seven families, with the Cupressaceae separate from the Taxodiaceae. However, many European authors (including Emberger, H. Erdtman & Norin, and Le- breton) have united these two families in the order Cupressales, and a number of authors (Eckenwalder; Hart; Price & Lowenstein) have advocated merging the two families under the earlier name Cupressaceae, as is done here. Although Pilger and subsequent authors have circumscribed families primarily on the basis of reproductive characters, especially those of the ovulate cone, the Cu- pressaceae and Taxodiaceae are similar in development and morphology of the bract-scale complexes (Florin, 1951) and differ primarily in phyllotaxy of leaves and cone scales. The leaves of the Cupressaceae s.s. are decussately opposite or whorled, while those of the Taxodiaceae are alternate (except in Metasequoia Miki ex Hu & Cheng, in which they are decussately opposite). The shift from spiral to decussate arrangement is hardly unique to the Cu- pressaceae 5./. Both patterns are also seen in the Taxaceae (where Amentotaxus Pilger has opposite leaves) and the Podocarpaceae (where Microcachrys J. D. Hooker has opposite leaves). The taxodiaceous genera also differ from the Cupressaceae s.s. in having a papilla protruding from the germinal area of the pollen grain, although this can be very obscure in some genera (G. Erdtman, Ueno, 1960a). The two families are held together by an impressive number of morphological characters, including derived features of embryology (archegonia borne in com- plexes; free-nuclear mitotic divisions three or fewer in proembryogeny), paly- nology (pollen grains nonsaccate, lacking prothallial cells), chromosome base number (x = 11), and high degree of bract-scale fusion in the ovulate cone. Most genera have seeds with lateral wings derived from the seed coat (Singh, 1978), ovules are often more than two per cone scale, and microsporangia are usually more than two per sporophyll. Preliminary cladistic analyses of mor- phological characters (Hart) and immunological comparisons of seed proteins (Price & Lowenstein) indicate that the Cupressaceae s./. are a natural group quite distinct from the other families of conifers and that the Cupressaceae s.5. form a monophyletic group apparently derived from within the traditional Taxodiaceae. The Cupressaceae s./. appear to be only distantly related to the other extant groups of conifers except for the monotypic Sciadopitys Sieb. & Zucc., Japanese umbrella pine, which has often been treated as a morphologically isolated member of the Taxodiaceae (see Eckenwalder; Liu & Su; Pilger; Sporne) or as the separate family Sciadopityaceae Hayata (Doyle & Brennan, 1971; Hart, Price & Lowenstein; Schlarbaum & Tsuchiya, 1985). Sciadopitys is similar to the Cupressaceae s./. in having ovulate cones with substantial bract-scale fusion and several seeds per cone scale, each with two lateral wune®. derived from the seed coat, and nonsaccate pollen grains lacking prot ] s. Its chromosome base number (x = 10) is apparently derived from that of the Cupressaceae Ces by aneuploid reduction (Schlarbaum & Tsuchiya, 1985). It is unique among 278 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 conifers in having elongate “double needles” —axillary short shoots that com- bine features of leaf and stem in their development and morphology (Roth)— as its photosynthetic organs. It retains primitive states for embryological char- acters (archegonia not grouped in complexes, five sets of free-nuclear mitoses in proembryogeny— Dogra, 1980; Doyle, 1963; Liu & Su; Tahara, 1937, 1940), the derived states of which are shared by the Cupressaceae s./.; it also differs in having male cells unequal in size (Tahara, 1940), pollen grains with much more prominent verrucate sculpture (Ho & Sziklai), and wood without axial parenchyma and with much larger cross-field pits (Phillips). Sciadopitys is very distant from the Cupressaceae s./. in immunological comparisons of seed pro- teins (Price & Lowenstein), and it also has a long fossil record, dating back at least to the Jurassic (Florin, 1922, 1963; Manum). Thus it appears to be a well- separated sister-group of the Cupressaceae s./. (Hart; Price & Lowenstein), and we treat it as the monotypic family Sciadopityaceae. Most modern tribal and subfamilial classifications of the Cupressaceae/Tax- odiaceae lineage have been devised in the context of two separate families, the Cupressaceae and the Taxodiaceac, or equivalent groups of lower rank (see, for example, Gaussen, 1967, 1968, Pilger: Pilger & Melchior; Vierhapper). The Cupressaceae s.s. have been divided into tribes or subfamilies in a variety of ways (compare Endlicher,; Gaussen, 1968; Janchen; H. L. Li, 1953a; Moseley; Pilger). In perhaps the most widely utilized modern treatment, H. L. Li (1953a) divided the group into two subfamilies: Cupressoideae, with nine genera in the Northern Hemisphere (further divided into tribes Cupresseae, Junipereae Ne- ger, and Thujopsideae Endl.), and Callitroideae Saxton, including the ten South- ern Hemisphere genera plus Tefraclinis Masters in Spain and northern Africa (further divided into tribes Actinostrobeae Endl., Libocedreae H. L. Li, and Tetraclineae H. L. Li). The cupressoid genera were separated from the callitroid genera as having imbricate rather than valvate scales on the mature ovulate cones. However, the lowermost cone scales are often similar in arrangement in both groups, as was noted by De Laubenfels (1965). Moreover, the mature cone scales of Cupressus are nonoverlapping. The cones of the two subfamilies mostly differ in the reduced axis and usually fewer cone scales in the callitroid genera. The whorls of cone scales may be partially overlapping at the time of pollination in the Southern Hemisphere genera but come to lie at approximately the same level in the mature cone, rather than being separated along the cone axis. Reduction in cone-scale number has also occurred in several Northern Hemisphere genera (e.g., in Microbiota Komarov, with only two to four cone scales on a reduced axis (Kriissmann; Rushforth), in Ca/ocedrus, and within Chamaecyparis and Juniperus). The presence of derived features of embryology (laterally positioned archegonial complexes and reduction of the number of free-nuclear divisions in the early proembryo) appears to support a close rela- tionship among at least some of the Southern Hemisphere genera (Callitris, Actinostrobus Miq., and Widdringtonia Endl.—Dogra, 1984; Doyle & Brennan, 1972; Singh), although sampling has been very limited, and many of these genera have never been studied. Other classifications have associated Libocedrus and its segregate genera with the thujoid genera of the Northern Hemisphere, which share elongate cone 1990] HART & PRICE, CUPRESSACEAE 279 scales and flattened ultimate branch systems (De Laubenfels, 1953, 1988; Janchen; Pilger; Vierhapper). However, the northern genera appear to differ in wood anatomy (Boutelje; Peirce, 1937). Distribution of tropolone and bifla- vonoid compounds (H. Erdtman & Norin: Gadek & Quinn, 1985), and pre- liminary data on anatomical characters such as details of pitting of the foliar transfusion tracheids (Gadek & Quinn, 1988) only partially agree with H. L. Li’s (1953a) classification. Thus several recent authors (De Laubenfels, 1988; Gadek & Quinn, 1985, 1988) have questioned whether Li’s subfamilies are natural groups, and additional morphological and macromolecular compari- sons will be necessary to resolve their phylogenetic relationships. The nine taxodiaceous genera (with Sciadopitys excluded) comprise only about 12 or 13 species and represent remnants of a group that was larger and more widespread in the Mesozoic and Tertiary (Florin, 1963; Miller, 1977, 1988). Many of the extant genera had wider geographic distributions in the past, as is indicated by the fossil record, and several occur back to the Creta- ceous. One genus 4throtaxis D. Don, is now endemic to Tasmania, five (Cryp- tomeria D. Don, Cunninghamia R. Br., Glyptostrobus Endl., Metasequoia, and Taiwania Hayata) to eastern Asia, and three (Sequoia, Sequoiadendron, and Taxodium) to North or Central America. Classifications of the Taxodiaceae have tended to emphasize form and de- velopment of the bract-scale complexes in the ovulate cone, but there are a number of differences in the details of the treatments (see Eckenwalder; Gaus- sen, 1967: Hida, 1957, 1962; Janchen; Liu & Su; Pilger & Melchior; Vier- happer), and a rigorous tribal or subfamilial treatment seems premature. Most authors group Sequoia with Sequoiadendron and Taxodium with Glyptostrobus, and recent researchers (Eckenwalder; Gaussen, 1967; Miller, 1988; Miller Crabtree) have emphasized the similarities of Athrotaxis, Cunninghamia, and Taiwania. Preliminary cladistic analysis of a broader set of characters (Hart) and immunological comparison of seed proteins (Price & Lowenstein; Price, unpublished data) tend to support the groupings noted above, but place Ath- rotaxis in a more isolated position. These analyses differ primarily in the placement of Metasequoia and Cryptomeria relative to the other genera, as- sociating Metasequoia with the other winter-deciduous genera G/yptostrobus and Taxodium (Hart) or with Sequoia and Sequoiadendron (Price & Lowen- stein), and Cryptomeria with Taiwania and Cuaninghamia (Hart) or with Glyp- tostrobus and Taxodium (Price & Lowenstein). Chromosome numbers are often very stable within families of conifers (Eh- rendorfer; Khoshoo, 1961), in marked contrast to the situation in the angio- sperms. With the removal of Sciadopitys to its own family, the genera of Cupressaceae s./. are apparently all characterized by a base number of x = 11.° Counts have been obtained for all species of the nine taxodiaceous genera and for species of 16 of the cupressaccous genera, including all ten Northern Hemi- sphere ones (see particularly Hair; Khoshoo, 1961; L.-C. Li; Mehra & Khoshoo; Sax & Sax). The great majority of the species have the diploid number 2n = ‘Fokienia Hodginsii (Dunn) ie & Thomas has been reported by Chen to have a chromosome number of 2n = 24, but L.-C. Li & Hu have more recently obtained 2 = 22 for the species. 280 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 22. Two monotypic genera are polyploid: Sequoia sempervirens (Lamb.) Endl. is a hexaploid with 2” = 66, while Fitzrova cupressoides (Molina) Johnston has been reported by Hair to have 2” = 44. Triploid or tetraploid counts have been obtained for forms (primarily cultivated) of several species of Juniperus and also for occasional variants of the otherwise diploid Cryptomeria japonica (L. f.) D. Don and Chamaecyparis pisifera (Sieb. & Zucc.) Endl. Karyotypic comparisons of the taxodiaceous genera were reviewed by Schlarbaum & Tsu- chiya (1984), who noted some differences 1n arm lengths and ratios and presence or position of secondary constrictions. Cunninghamia and Taiwania differ from the other taxodiaceous genera in having more diversity in chromosome size and a greater number of submetacentric (vs. metacentric) chromosomes, in agreement with the relationship between these genera indicated by their sim- ilarity in cone-scale development (Hida, 1957; Liu & Su). The Cupressaceae s./. are unusual among the conifers in having archegonia grouped tightly in complexes, usually with a well-defined jacket layer (Dogra, 1984; Liu & Su; Singh). The archegonial complexes are usually located near the micropyle, but are positioned along the sides of the gametophyte in at least some of the Southern Hemisphere genera of Cupressaceae s.s. (Actinostrobus, Callitris, and Widdringtonia—see Doyle & Brennan, 1972; Singh) and in Se- quoia and Sequoiadendron (Liu & Su) and toward the chalazal end in Athrotaxis (Brennan & Doyle). There are three sets of free-nuclear mitoses in the proem- bryogeny of most genera, but the number is reduced to two in Athrotaxis and Callitris (and probably Actinostrobus and Widdringtonia). Wall formation ac- companies the first mitotic division in Sequoia, a very unusual feature among the conifers (Singh). Cleavage polyembryony of various types occurs as a regular part of development in most genera of the Cupressaceae s./. except for Ath- rotaxis, Thuja, and Thujopsis, in which it occurs only sporadically (Dogra, 1984; Doyle & Brennan, 1971, 1972). The Cupressaceae are wind pollinated, as are the other conifers, and have a pollination-drop mechanism of pollen capture, which is seen in most other gymnosperms and is evidently the primitive state among the conifers (see Doyle, 1945; Singh). Dispersal of the seeds is usually by wind or gravity, except in Juniperus, where the fleshy cones are eaten by birds or mammals, and Taxodium, where the thickened seed coat may aid in distribution by water (Fowells). Many taxodiaceous and cupressaceous genera have been shown to have vesicular-arbuscular mycorrhizae and to lack root nodules, while Sciadopitys is more similar to the Podocarpaceae and Araucariaceae in having root nodules (Khan & Valder). Chemical studies have been conducted on wood and leaves of many of the genera of Cupressaceae s./., with greatest emphasis on the Northern Hemisphere and Australian groups (see reviews in H. Erdtman & Norin; Hegnauer, 1962, 1986). Biflavonoid composition has been studied in virtually all of the genera worldwide (Gadek & Quinn, 1983, 1985; Geiger & Quinn, 1975, 1982). Com- pounds of the amentoflavone and hinokiflavone series are widespread among gymnosperms generally and have been found in most genera of the family, although the latter series 1s apparently absent in the Australian genera Callitris 1990] HART & PRICE, CUPRESSACEAE 281 and Actinostrobus (Gadek & Quinn, 1983). The cupressuflavone series is of more limited distribution, occurring in the Araucariaceae, one genus of the Podocarpaceae (Lepidothamnus Phil.), and portions of the Cupressaceae 5.5. (Gadek & Quinn, 1985; Geiger & Quinn, 1982). Cupressuflavone is a major component in Calocedrus, Cupressus, Juniperus, and Tetraclinis, and a lesser one in Platycladus Spach and individual species of Chamaecyparis and Thuja (but not others), all of these indigenous to the Northern Hemisphere. Taiwan- iaflavone, an unusual biapigenin biflavoid found only in Taiwania, Calocedrus, and Neocallitropsis Florin, is probably a convergent feature, since the three genera are very dissimilar in morphology. The cytotoxic lignans podophyllotoxin and/or desoxypodophyllotoxin, which have been used in both traditional and modern medicine as antitumor agents, are found in the leaves of a variety of cupressaceous genera, including Austro- cedrus, Callitris, Calocedrus, Chamaecyparis, Juniperus, and Thujopsis (Cairnes et al.. H. Erdtman & Norin; Hegnauer, 1986). The related but noncytotoxic compound savinin is present in the wood of several species of Juniperus and also in Taiwania (H. Erdtman & Norin; Hegnauer, 1986 Diverse terpene compounds are found in the wood, leaves, and other plant parts of the Cupressaceae s./. and are responsible for much of the aromatic nature of the plants. Most of the monoterpene compounds in the Cupressaceae s.l. occur widely in other conifers, while sesquiterpenes of the cedrane, thu- jopsane, widdrane, and cuparane types are particularly characteristic of the Northern Hemisphere genera of Cupressaceae s.s. and also of Widdringtonia (H. Erdtman & Norin). Several of these compounds have recently been reported to occur in low concentration in the foliar resin of Cryptomeria (Yatagai & Sato), and detailed studies may find them to be more widely distributed among the taxodiaceous genera. Tropolones structurally related to the terpenes are important heartwood components of several genera of the Cupressaceae s.s., particularly those of the Northern Hemisphere. These compounds, notable for their fungicidal activity, are apparently absent from the other conifers (H. Erdtman & Norin; Hegnauer 1986). Tropolones are characteristic of Calocedrus, Cupressus, Platycladus, Tetraclinis, Thuja, Thujopsis, and most but not all species of Chamaecyparis and Juniperus (H. Erdtman & Norin). They are found in Austrocedrus and the related Papuacedrus but are apparently absent in the other Southern Hemi- sphere genera. Alkaloids are relatively uncommon components in the conifers and in the Cupressaceae s./. are well documented only for Athrotaxis, all species of which contain homoerythrinane compounds (Hegnauer, 1986, 1988). Many genera of Cupressaceae 5./., especially species of Chamaecyparis, Cryp- tomeria, Cupressus, Juniperus, Metasequoia, Platycladus, Sequoia, and Thuja, are important as ornamental trees or shrubs (L. H . Bailey; Bean; Dallimore & Jackson; Ouden & Boom; Kriissmann). Sequoia and the largest trees in the world and are centers of attraction in several national and state parks in California. Wood of many genera of Cupressaceae Sas resistant to insect and fungal attack and thus has been highly sought for uses requiring durability. Wood of various species of Juniperus, Thuja, and Cha- 282 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 maecyparis has been important for shingles, and that of Sequoia for outdoor uses such as decks and fences. REFERENCES: Aase, H.C. Vascular anatomy of the megasporophylls of conifers. Bot. Gaz. 60: 277- 313. 1915. Avosi, M. C., & R. B. PARK. A survey of phloem polypeptides in conifers. Curr. — Pl. Biochem. Physiol. 2: 250. 1983. [Several genera of Cupressaceae s./. sh distinct pattern differing from those in Sciadopitys and in the other conifer families | ALVIN, K. L., & M. C. BouLTER. A Se pas method of oo ae a taxo- diaceous leaf cuticles. Bot. Jour. Linn. Soc. 69: 277-286. pls. I-S. 4. [SEM comparisons of Athrotaxis, Oona e eee es Lepr Sequoia, Sequoiadendron, and Taxodium.]} BaiLey, L. H. The cultivated conifers in North America. 404 pp. /58 pls. New York. 1933. BaiLey, N. Know your wood: cedar. Wood Working Dig. 53: 85-89. 1951.* [Calocedrus, Chamaecyparis, Juniperus, Thuja.] BAILLon, H. Coniféres. Hist. Pl. 12: 1-45. 1894. [Cupressaceae, 4-11, 26, 29, 30, 34— 40. BANTHORPE, D. V., F. S. Davies, C. GATFoRD, & S. R. WILLIAMS. Monoterpene patterns in pee and Thuja species. Pl. Med. 23: 64-69. 1973. [Juniperus nensis, J. communis, J. horizontalis, J. oxycedrus, J. phoenicea, J. Sabina, dees ee orientalis, Thuja occidentalis, T. plicata, T. Standishii.] BARTHOLOM B., D. E. Bourrorp, & S. A. SPONGBERG. Metasequoia glyptostro- ae present ae in central China. Jour. Arnold Arb. 64: 105-128. 1983. {History of discovery of the dawn redwood. and its status as of 1980.] BEAN, W. J. Trees and shrubs hardy in the British Isles. ed. 8. 4 vols. London. 1970- 1980. [Chamaecyparis, 1: 591-601. 1970; Juniperus, 2: 478-498. 1973; Taxodium, 4: 557-560. 1980; Thuja, 4: 577-586. 1980; also treated are Cryptomeria, Cupressus, Glyptostrobus, Libocedrus (here including Austrocedrus and Calocedrus), Metase- quoia, Sequoia, came es Thujopsis. ] BENTHAM, G., & J. D. HOOKER. Ger 3(1): 424-430. 1880. [Cupressaceae s./.] BouTE.E, J. B. The wood anatomy of ce Endl., s. lat., and Fitzroya J. D. Hook. Acta Horti Berg. 17: 177-216. pls. 55. BRENNAN, M., & J. Dove. The gamelan and embryogeny of Athrotaxis. Sci. Proc. Roy. Dublin ae I]. 27: 193-252. 1956. Britton, N. L. rth American trees. x + 894 pp. New York. 1908. [Cupressaceae, 86-121; Tae The swamp cypresses. Jour. New York Bot. Gard. 27: 205-207. 1926. [G/yp- tostrobus, Taxodium.]| BUCHHOLZ, J. T. The embryogeny of Sequoia SEIDEL NIECTES with a comparison of the sequoias. Am. Jour. Bot. 26: 248-256. 19 . The generic searegation of the sequoias. Ibid. 535-538. 1939b. [Sequoiadendron sningise from Sequoia. | c and subgeneric distribution of the Coniferales. Bot. Gaz. 110: 80-91. 194 8. Butts, D., & J. T. BucCHHOLz. Cotyledon number in conifers. Trans. Illinois Acad. Sci. 33: 58-62. 1940. [Original counts and literature summary, including most genera of the Cupressaceae 5. Carrnes, D. A., D. G.I. KINGSTON, & M. MADHUSUDANA RAo. High performance liquid chromatography of podophyllotoxins and related lignans. Jour. Nat. Prod. 44; 34- 37. 1981. [Occurrence in species of Austrocedrus, Callitris, Calocedrus, and Junip- erus.| 1990] HART & PRICE, CUPRESSACEAE 283 CaAmpo-DupLan, M. vANn. Recherches sur la phylogénie des Taxodiacées d’aprés leurs grains de pollen. Trav. Lab. Forest. Toulouse, Tome IJ, Sect. I, Vol. IV, Art. I: 10- 14. 1951. Recherches sur la phylogénie des Cupressacées d’aprés leurs grains de pollen. Ibid. Art. III: 1-20. 1953. CARRIERE, E. A. Traité général des coniferes. ed. 2. x11 + 910 pp. Paris. 1867. [Cu- pressinées, including Taxodiaceae, 1-236 CHAMBERLAIN, C. J. Gymnosperms: structure and evolution. xi + 484 pp. Chicago. 1935 CHANEY, R. W. Redwoods—occidentale and orientale. Science 110: 551, 552. 1949. ee Sequoia, Sequoiadendron. n of fossil Sequoia and Taxodium in western North America based on ‘the oak ae of Metasequoia. Trans. Am. Philos. Soc. II. 40: 169-263. 1951. CHANG, Y.-P. Bark structure of North American conifers. U. S. Dep. Agr. Tech. Bull. 1095. ii + 86 pp. 1954. [Cupressaceae, 47-67, figs. 29-37; sans Calocedrus, Chamaecyparis, Cupressus, Juniperus, eee Faxodium, Thuja.] CueEN, K.-Y. Chromosome number of Fokienia. (In Chinese; English summary.) A Bot. Sinica 25: 120-122. 1983. [Count of oe = 24 se ea for F. Hodginsii: ee confirmation. ] CHEN, Z.- F.-H. WANG. Development of gametophytes in Fokienia (Cupressaceae). (In Chinese: een summary.) Acta Bot. Sinica 22: 6-10. 1980a. & Lama in fertilization of Fokienia. Ibid. 221-226. 1980b. & —— arly embryogeny of the genus eters ie a note on its systematic section yen Phytotax. Sinica 19: 23-28. pls. J, 2 , CHENG, W.-C., & L.-K. Fu, eds. Gymnospermae. (In Chinese; aa portions primarily in English and Latin.) Fl. Reipubl. Pop. Sinicae 7. xiv + 544 pp. 1978. [Cupressaceae s./., 281-398; many illustrations CHOowpbHuRY, C. R. The embryogeny of conifers: a review. Phytomorphology 12: 313- 338. 1962. CutrForb, H. T., & J. CoNsTANTINE. Ferns, fern allies, and conifers of Australia. 150 pp. St. Lucia, Queensland. 1980. [Keys and descriptions for Actinostrobus, Athro- taxis, Callitris, Diselma.] CoKeErR, W. C., : OTTEN. Trees of the pra tae States. ed. 3. vil + 419 pp. 7 pls. Chapel Hill. 1945. [Cupressaceae s./., COoLTMAN-Rocers, C. Conifers and their zener 334 pp. London. 1920 Conk Le, M. T. Electrophoretic analysis of variation in native Monterey cypress (Cu- pressus eens Hartw.). Pp. 249-256 in T. S. Extas, ed., Conservation and management of rare and endangered plants. 1987. [Electrophoretic variation 1s sub- stantial fee current very restricted geographic distribution. Coutter, J. M., & C. J. CHAMBERLAIN. Morphology of gymnosperms. ed. 2. x1 + 466 pp. Chicago. 1917. DaGulILLon, A. Observations morphologiques sur les feuilles des Cupressinées. Rev. Gén. Bot. 11: 168-204. p/. 5. 1899. [Including Chamaecyparis, Cryptomeria, Cu- pressus, Juniperus, Platycladus, Sequoia, Sequoiadendron, Taxodium, Thuja. DALLIMORE, W., & A. B. Jackson. A handbook of Coniferae and Ginkgoaceae. ed. 4, revised by S. G. HARRISON. xix + 729 pp. 46 pls. London, 1966. Dark, 8. O. S. Chromosomes of Taxus, Sequoia, Cryptomeria, and Thuya. Ann. Bot. 46: 965-977. 1932. De LauBenFELs, D. J. The external morphology of coniferous leaves. Phytomorphology 3: 1-20. 1953. . The relationships of Fitzrova cupressoides (Molina) Johnston and Dise/ma Ar- cheri J. D. Hooker based on morphological considerations. Ibid. 15: 414-419. 1965. [Relationships among genera of Cupressaceae and Taxodiaceae discussed. ] — 284 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 —. Cupressaceae. /n: A. AUBREVILLE & J.-F. LERoy, eds. FI. ee. Calédonie et Dépendances 4: 144-164. 1972. [Callitris, Libocedrus, Neocallitro Coniferales. Fl. Males. I. 10(3): 337-453. 1988. [Cupressaceae, 444-447. Li- bocedrus, including Papuacedrus. | Diunoscn, H. Die Bliiten der Coniferen. II]. Entwicklungsgeschichtliche Untersuchun- gen ber die Mikrosporophyllgestaltung der Coniferen. Bibliot. Bot. 114(3): 1-24. pls. 1-7. 1936. [Cupressaceae, 9-17, pls. 3-7.] Doak, C. C. Multiple male cells in Cupressus arizonica. Bot. Gaz. 94: 168-182. 1932. Neg of Cupressus arizonica gametophytes and embryogeny. [bid. 98: 808- 815. 1937. Doara, P. D. hoe of the Taxodiaceae. Phytomorphology 16: 125-141. 1966. Embryogeny of gymnosperms and taxonomy —an assessment. Glimpses PI. Res. 5: 114-128. 1980. [Embryogeny is reasonably consistent with the proposed merger of the Cupressaceae and the Taxodiaceae.] . The embryology, breeding Bes and seed sterility in Cupressaceae—a mono- graph. Jbid. 6: 1-113. figs. 1-264. 1984. ei . Developmental lines in ar Aarie mechanisms in the Coniferales. Sci. Proc. », Dublin Soc. 24: 43-62. 1945. : a a in Pinus in relation to that in other conifers—a survey. Proc. Irish Acad. B. 62: 181-216. 1963 ENNAN. Cleavage ae yony in conifers and taxads—a survey. I. Podocarps, taxads & taxodioids. Sci. Proc. Roy. Dublin Soc. A. 4: 57-88. 1971; II. Cupressaceae, Pinaceae and ee “hid. 137-158. 1972. & W. AXTON. Contribution to the life history of Fitzroya. Proc. Irish Acad. B. 41: 19] 017. 1933. See E. Mechanism of exine rupture in hydrated type of pollen. Grana 21: 1-7. 1982. [Cupressus arizonica, sles alee commun Duncan, W. H., & M. B. DUNCAN. Trees 0 ee United States. x1 + 322 pp. Athens, Georgia, and London. 1988. eee eae S./., 40-44, 123, 124: color ee, maps of the southeastern species; C se lanceolata reported as culti- ted and reg ous escaping in the Southea FcxenwaL Dek J. A. Re-evaluation ee d Taxodiaceae: a proposed merg- . Madrono 23: 337 356 1976. Hearne cae characters tabulated for families ne and genera of Cupressaceae 5./.; taxa c a using indices of similarity; tentative classification offered for the combined fami nily EHRENDORFER, F. Evolutionary significance ofc differentiation patterns in gymnosperms and primitive angiosperms. Pp. 220-240 in C. Beck, ed., Origin and early car of angiosperms. New York. 1976. EICHLER, A. Coniferae. Jn: A. ENGLER & K. PRANTL, Nat. Pflanzenfam. II. 1: 28- 116. ol [With contributions by ENGLER & PRANTL. EMBERGER, L. Gymnospermes. Traité de botanique systématique ae 383-459. 1960. [Cupressales, including Cupressaceae and Taxodiaceae, 434—448.] ENDLICHER, S. L. Synopsis coniferarum. iv + 368 pp. St. Gall, Switzerland. 1847. ErpTMAN, G. Pollen and spore morphology/plant taxonomy. V ol. 2. Gymnospermae, Pteridophyta and Bryophyta (illustrations). Frontisp. + iv + 151 pp. 5 pls. Stock- holm. 1957. Vol. 3. Gymnospermae, Bryophyta (text). 191 pp. 24 p/s. Stockholm. 1965. ErRDTMAN, H., & T. Nortn. The chemistry of the order Cupressales. Fortschr. Chem. Organ. Naturstoffe 24: 206-287. 1966. [Important review of chemistry of Cupres- saceae and Taxodiaceae. ] FERRE, Y. DE. Cotylédons et évolution chez les Cupressinées. Bull. Soc. Hist. Nat. Toulouse 77: 71-83. 1942 . L’évolution paralléle des Taxodinées et des Abietinées. bid. 78: 71-83. 1943. —. Cotylédons et évolution chez les Taxodinées. /bid. 82: 214-224. 1947. 1990] HART & PRICE, CUPRESSACEAE 285 — & H. Gaussen. Les Cupressacées australes. Compt. Rend. Acad. Sci. Paris 267: 483-487. 1968.* . M. L. Rouane, & P. WoLtz. Systématique et anatomie comparée des feuilles de Taxaceae, Podocarpaceae, Cupressaceae de Nouvelle-Calédonie. Cahiers Paci- fique 20: 241-266. 1978. [Libocedrus, Neocallitropsis.] Fins, L., & W. J. Lippy. Population variation in Sequoiadendron: seed and seedling studies, vegetative propagation, and isozyme variation. Silvae Genet. 31: 101-148. 1982. FitscHEN, J. Handbuch der Nadelholzkunde. xv + 765 pp. Berlin. 1930. [Revised ed. 3 of BEISSNER’s Handbuch der Nadelholzkunde.] Fittinc, H. Uber die Induktion der Dorsiventralitat in den blattahnlichen Zweigsyste- men von Cupressaceen. Jahrb. Wiss. Bot. 90: 417-463. 1942. [See also Planta 37: 676-696. 1950.] Fiorin, R. On the geological history of the Sciadopitineae. Sv. Bot. Tidskr. 16: 260- 270. 1922. [Sciadopitys and related leaf fossils; see also the recent treatment by MANUM nters uchungen zur Stammesgeschichte der Coniferales und Cordaitales. Sv. Vet. -akad. Handl. III. 10: 1-588. pis. 7-58. 1931. [Classic treatment of vegetative anatomy of the conifers. ] . Evolution in cordaites and conifers. Acta Horti Berg. 15: 285-388. 1951. [Evo- lution of the coniferous ovulate cone. n Metasequoia, living and fossil. Bot. ae 1952: 1-29, 1952. [Review of i history, and relationships; range ma female reproductive organs of conifers ee taxads. Biol. Rev. 29: 367-389. 1954. The systematics of the gymnosperms. Pp. 323-403 in E. L. Kessex, ed., A century of progress in the natural sciences, 1853-1953. San Francisco. 1955. [His- torical review, mostly of morphology.] . The distribution of conifer and taxad genera in time and space. Acta Horti Berg. 20: 121-312. 1963. [Maps of current and fossil distribution for all genera of the Cupressaceae s./.] & J. B. BouTevye. External morphology and epidermal structure of leaves in the genus Libocedrus, s. lat. Acta Horti Berg. 17: 7-37. Flory, W.S. Chromosome numbers and phylogeny in the gymnosperms. Jour. Arnold Arb. 17: 83-89. 1936. [Notes the shared derived basic number, x = 11, of the Cupressaceae and the Taxodiaceae.] FoweE ts, H. A., compiler. Silvics of forest trees of the United States. U. S. Dep. Agr. Agr. Handb. 271. 762 pp. 1965. [Includes Calocedrus, Chamaecyparis, Juniperus, Sequoia, Sequoiadendron, Ta: ane Thuja.| GADEK, P. A., & C. J. QuINN. Biflavones of the subfamily Callitroideae, Cupressaceae. Phytochemistry 22: 969-972. 1583. Biflavones of the subfamily Cupressoideae, Cupressaceae. [bid. 24: 267-272. 1985. Biflavones and the affinities of Cupressus funebris. [bid. 26: 2551, 2552. 1987. (Bifavony profile does not support transfer to Chamaecyparis.| & ing of transfusion tracheids in Cupressaceae. Austral. Jour. Bot. 36: 81-92. 1988. [Unusual types are shared by certain genera of Cupressoideae and Callitroideae. ] GAUSSEN, H. Les oe In: Les gymnospermes: actuelles et fossiles. Trav. Lab. Forest. Toulouse, Tome II, Vol. I, Chap. XII: 1-64. 1967. Les Cupressacées. [bid. Chap. XIII: 1-326. 1968. serene treatment of the Taxodiaceae and Cupres- saceae.] Bibliographie. /bid. Tome II, Vol. I, Partie IIT: 1-180. 1979. [Bibliography for the entire wor GEIGER, H., & C. Quinn. Biflavonoids. Pp. 692-742 in J. B. HARBORNE, T. J. MABrRy, 286 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 & H. Masry, eds., The flavonoids. Vol. 2. London. 1975. [Cupressaceae, 720-722, 729-732.] & ——. Biflavonoids. Pp. 505-534 in J. B. HARBORNE & T. J. Masry, eds., The flavonoids: advances in research. London and New York. 1982. [An updated summary. GIvNIsH, T. J. Ecological constraints on the evolution of breeding systems in seed plants: dioecy and dispersal in gymnosperms. Evolution 34: 959-972. 1980. [Dioecy strongly correlated with seed dispersal by animals. ] Gorpbon, G. The pinetum. xxiv + pp. London. 1858. Grecuss, P. Identification of living gymnosperms on the basis of xylotomy. 263 pp. 350 pls., 8 tables. Budapest. 1955. —. Xylotomy of the living conifers. 172 + 329 pp. Budapest. 1972 [Cupressaceae, 19, 20, 86-109, pls. 8, 9, 84-114.] GULLINE, H. F. The cytology of . malas Pap. Proc. Royal Soc. Tasmania 86: 131- 136. 1952. [All three species 2 = Harr, J. B. The chromosomes of the iene 1. Tetraclineae and Actinostrobeae (Callitroideae). New Zealand Jour. Bot. 6: 277-284. 1968. [Counts for species of esha Callitris, and Tetraclinis, each 2n = 22, and Fitzroya cupressoides, 2n = 44. Harbin, J. W. Studies of the southeastern United States flora. I]. The gymnosperms. Jour. Elisha Mitchell Sci. Soc. 87: 43-50. 1971. [Cupressaceae s./., 48, 49; Callitris reported as naturalized in Florida.] Hart, J. A. A cladistic analysis of conifers: preliminary results. Jour. Arnold Arb. 68: 269-307. 1987. [Cladograms at both intergeneric and interfamilial levels; character states tabulated for the genera of the Cupressaceae s. HayatTa, B. The Sciadopityaceae represented by Sciadopitys verticillata Sieb. et Zucc., an endemic species of Japan. (In Japanese and Latin.) Bot. Mag. Tokyo 45: 567- 569. 1931. . The Taxodiaceae should be divided into several distinct families, i.e., the Lim- nopityaceae, Cryptomeriaceae, Taiwaniaceae, and the Cannioeiainiseose and fur- ther Tetraclinis should represent a distinct family, the Tetraclinaceae. (In Japanese and Latin.) Jhbid. 46: 24-27. 1932. HEGNAUER, R. Chemotaxonomie der Pflanzen. Band |. Thallophyten, Bryophyten, Pte- ridophyten und Gymnospermen. 517 pp. Basel and Stuttgart. 1962. [Cupressaceae, 342-372, 478-480; Taxodiaceae, 421-429, 481, 482.] Band 7. Nachtrage zu Band | und Band 2. x11 + 804 pp. Basel, Boston, and Stuttgart. 1986. [Cupressaceae, 49 |- 504, 801; Taxodiaceae, 535-540; extensive literature review. ] oe ast eauen and taxonomic relevance of higher plant iui Jour. Nat. Prod. 27: 2423-2427. 1988. [Alkaloids reported for Athrotaxi Henry, A., & M. ra NTYRE. The swamp cypresses, G/yptostrobus of China ona Tax- odium of America, with notes on allied genera. Proc. Irish Acad. B. 37: 90-116. pis. . 1926. = an ~ Hicke., R. Graines et plantules des coniféres. Bull. Soc. Dendrol. France 19: 13-115; 20: 134-204. 1911. [Cupressaceae, 19: 30-48 (seeds), 20: 141-159 (seedlings).] Hipa, M. The comparative study of Taxodiaceae from the standpoint of the development of the cone scale. (In Japanese; English summary.) Bot. Mag. Tokyo 70: 44-51. 1957. ected divided into five groups on the basis of pattern of cone-scale devel- opme The systematic position of Metasequota. (In Japanese; English summary.) /bid. 75: 316-323. 1962. 2 - G., & E. DE Frarne. On the seedling structure of gymnosperms. I. Ann. Bot. : 689- ee pl. 35. 1908. [Anatomy and morphology of several genera of . ag Tame ara cel eda cen and Thuja. HIRMER, M. e Bliiten der Coniferen. I. Entwicklungsgeschichte und vergleichende 1990] HART & PRICE, CUPRESSACEAE 287 senate des weiblichen Bliitenzapfens der Coniferen. Bibliot. Bot. 114(1): 1- 100. pls. A-K, 1-12. 1936. [Excellent illustrations of ovulate-cone morphology and deiclooment: in Cupressaceae and Taxodiaceae. ] Ho, R. H., & O. SziKLar. Fine : tructure of pollen surface of some Taxodiaceae and Cupressaceae species. Rev. Palacobot. Palynol. 15: 17-26. 1973. [Chamaecyparis, Cryptomeria, Cunninghamia, Sciadopitys, Sequoia, Taxodium.] Hu, H.-H., & W.-C. CHENG. On the new family Metasequoiaceae and on Metasequoia glyptostroboides, a living species of genus Metasequoia found in Szechuan a ae ae Fan Mem. Inst. Biol. 1: 153-161. 1948. Hu, Y.-S., & B.-J. YAo. eo of gymnosperm leaves. Bot. Jour. Linn. Soc. 83: ve 379. 1981. JANCHEN, E. Das System der Koniferen. Sitz.-ber. Akad. Wiss. Math.-Naturw. Wien I. 158: 155-262. 1949 [1950 KAEISER, M. The pollen of certain Taxodiaceae. Trans. Illinois Acad. Sci. 32: 94, 95. 1939. [Cryptomeria, Cunninghamia, Sciadopitys, Sequoia, Seguoiadendron, Taxo- dium. | Kent, A. H. Veitch’s manual of the Coniferae. ed. 2. 564 pp. London. 1900. Kern, F.D. A revised taxonomic account of Gymnosporangium. ix + 134 pp. University Park, Pennsylvania. 1973. [Fungal parasites of Calocedrus, dea Cu- pressus, Juniperus. | KHAN, A. G., & P. G. VALDER. Occurrence of root nodules in Ginkgoales, Taxales, and Coniferales. Proc. Linn. Soc. New S. Wales 97: 35-41. 1972. [Nodules seen in Sciadopitys Be not in the Cupressaceae s./. KuHosHoo, T. N. Polyploidy in cg veel ate Evolution 13: 24-39. 1959. Chromosome numbers n gymnosperms. Silvae Genet. 10: 1-9. 1961. [Com- prehensive summary of the iterate Konar, R. N. Investigations on the development of the male cones in Fitzroya cu- a (Mol.) Johnst. and Pilgerodendron uviferum (Dom.) Flor. Phytomor- phology 12: 190-195. 1962. . OBEROI. 2 work “ reproductive structures of living conifers and tax ads—a review. Bot. Rev. 35: 89-116. 1969. KRUSSMANN, G. aes der Nadel es ed. 2, revised by H.-D. Warpa. [6 +] 396 pp. 160 pls. Berlin and Hamburg. 1983. [Also published in English as KRUSSMANN, Manual of cultivated conifers. M. E. Epp, translator. 361 pp. /60 pls. oe Oregon. 1985.] Kuo, S.R.,T.T. WANG, & T.C. HUANG. Karyotype analysis of Formosan gymnosperms. T Awana 17: 66-80. 1972. [Karyotypes and chromosome counts for species of Chamaecyparis, Cunninghamia, and Taiwania; a = 22 Lai, Y. Z. A proposed biogenetic ee pier ee in family Cupressaceae. Cel- lulose Chem, Tech. 5: 621-623. | Lawson, A. A. The gametophvtes, rehgoni, fertilization, and embryo of Sequoia —— Ann. Bot. 18: 1-28. 1904. metophytes and eae of Cupressineae with special reference to Li- bocedrus ae Tbid. 21: 281-301. 1907. [Calocedrus decurrens.] LeBRETON, P. Les Cupressales: une définition chimiosystématique. Candollea 37: 243- 256. 1982. [Flavonoid constituents of Cupressaceae and Taxodiaceae.] LEMOINE-SEBASTIAN, C. Vascularisation du complexe bractée-écaille dans le cone femelle Cupressacées. Bot. Rhedon. A. 7: 3-27. 1969.* [See also ibid. 11: 177-187. 1971.] . Etude comparative de la vascularisation et du complexe séminal chez les Cu- pressacées. Phytomorphology 22: 246-2 972. Li, H.-L. A reclassification of Libocedrus and Cupress ssaceae. Jour. Arnold Arb. 34: 17- 36. 1953a. [Cupressaceae 5.5. split into the largely Southern Hemisphere Callitroideae and the Northern Hemisphere Cupressoideae. ] Ni3 288 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 ——_ Present distribution and habitats of the conifers and taxads. Evolution 7: 245- 261. 1953b. Li, L.-C. Preliminary report on the KaryoWpe of eae pensilis. Helongjiang Linye Kexueyuan Linye Keji 5: 5, 6. 1984.* [2n = 22.] .Hu. Karyotype analyses in Platyc ie ne ees and Fokienia ae (In Chinese: English summary.) Acta Bot. Yunnan 6: 447-451. [Both with 2” = 22.] Litrie, E. L., Jr. Atlas of the United States trees. Vol. 1. Conifers and aaa hardwoods. U.S. Dep. Agr. Forest Serv. Misc. Publ. 1146. v + 9 pp. 2 base maps, 200 species maps, 9 overlay maps. 1971. [Detailed range maps for all North American species of the Cupressaceae s./.] . Checklist of United Bee trees (native and naturalized). U. S. Dep. Agr. Agr. Handb. 531. 1v + 375 p —. The Audubon Society field guide to North American trees. Eastern region. 714 pp. New York. 1980. [Cupressaceae s./., 302-313, figs. 27, 28, 31-39. Liu, T.-S., & H.-J. Su. Biosystematic studies on 7aiwania and numerical evaluations of the systematics of Taxodiaceae. Taiwan Mus. Spec. Publ. 2: 1-113. 1983. [Genera of the Taxodiaceae compared in detail.] LonGMan, K. A., J. Dick, & C. N. PAGE. Cone induction a gibberellin for taxonomic studies in Cupressaceae and Taxodiaceae. Biol. Pl. 24: 195-201. 1982. Lotrova, L. I. Anatomic structure of the bark of Cupressaceae (In Russian; English summary.) Moscow Univ. Biol. Sci. Bull. 36: 1-7. MAHESHWARI, P., & H. SiInGH. The female ae ears of gymnosperms. Biol. Rev. 42: 88-130. 1967. ManuM, S. B. Mesozoic Sciadopitys-like leaves with observations on four species from the Jurassic of Andoya, northern Norway and emendation of Sciadopityoides Svesh- nikova. Rev. Palaeobot. Palynol. 51: 145-168. 1987. [Identity of early Sciadopitys fossils confirmed. ] Martinez, M. Las Pinaceas mexicanas. ed. 3. 400 pp. Mexico City. 1963. [Cupressaceae, Sal; —400.] Masters, M. T. Notes on the genera of Taxaceae and Coniferae. Bot. Jour. Linn. Soc. 30: 1-41. 1893. [Cupressaceae, | 1-25.] . On Taxodium and Gl/yptostrobus. Jour. Bot. London 38: 37-40. 1900. Mazzeo, P. M. Notes on the conifers of the Shenandoah National Park. Castanea 31: 240-247. 1966. [Juniperus communis, J. virginiana, Thuja occidentalis.) MeEnrRA, P. re & T. N. KHOSHOO. veles of conifers. I. Jour. Genet. 54: 165-180. pl. 5.1956. [Chromosome counts for species of Actinostrobus, "Callitris, es Barty es Se ean Taxodium, Tetraclinis, Thuja, and Widdrin tonia, all 2n J Miiiay, M. A., & ee . TAYLor. Evolutionary trends in fossil gymnosperm pollen. Rev. Palaeobot. Palynol. 21: 65-91. 1976. [Loss of saccae, as in the Cupressaceae, is the derived state in the conifers. ] Mitter, C. N., Jr. Mesozoic conifers. Bot. Rev. 43: 217-280. 1977. [Cupressaceae, 242-250: see also MILLER, 1988.] . Current status of Paleozoic and Mesozoic conifers. Rev. Palaeobot. Palynol. 37: 99-114. 1982. [Preliminary phylog pa of fossil and extant genera based on ovulate cone characters. ] . The origin of modern conifer families. Pp. 448-486 in C. B. Beck, ed., Origin and evolution of gymnosperms. New York. 1988. [Seed and ovulate-cone characters polarized by outgroup aoe seen phylogenetic analysis supports merger of the Taxodiaceae and the Cupressa . CRABTREE. A ne . eo seed cone from the Oligocene of Wash- ington. ‘Am, Jour. Bot. 76: 33. 142. 1989. [C stu aa acces Goedertii, rela- tionships of Cunninghamia to other taxodiaceous genera Mortey, T. On leaf arrangement in Metasequoia pee ee oO Natl. Acad. Sci. U.S.A. 34: 574-578. 1948 1990] HART & PRICE, CUPRESSACEAE 289 Mose ey, M. F. Contributions to the life history, morphology, and phylogeny of Wid- dringtonia cupressoides. Lloydia 6: 109-132. 1943. [Discusses relationships in the Cupressaceae S$. NeGer, F. W. Die Nadelhdlzer (Koniferen) und tibrigen Gymnospermen. 185 pp. Leip- zig. 1907 OLADELE, F. A. rae surface patterns of cuticles in Cupressaceae. Canad. Jour. Bot. 61: 1222- ke 83a. ca - electron rs study of stomatal-complex configuration in Cu- eee Ibid. 1232-1240. b. OUDEN, P. DEN, & B. K. Boom. ate of cultivated conifers. x [+ 2] + 526 pp. The Hague. 1965. Owens, J. N., & S. Simpson. Pollen from conifers native to British Columbia. Canad Jour. Forest Res. 16: 955-967. 1986. [SEM and LM photographs and descriptions: includes Chamaecyparis, Juniperus, Thuja.] PARLATORE, P. Coniferae. /n: A. D E CANDOLLE, Prodr. 16(2): 361-521. 1868. [Cupres- saceae 5./., 432-493 PaTEL, R. N. Wood anatomy of Cupressaceae and ee indigenous to New Zealand. New Zealand Jour. Bot. 6: 9-18. 1968. [Libocedru Peirce, A. S. Anatomical interrelationships of the Pexedincese. Trop. Woods 46: 1936. ——.. Systematic anatomy of the woods of the Cupressaceae. Ibid. 49: 5-21. 1937. PENHALLOW, D. P. The generic characters of the North American Taxaceae and Conif- erae. Trans. Roy. Soc. Canada, II. 2(4): 33-57. pls. I-6. 1896. [Wood anatomy of Calocedrus, Chamaecyparis, Cupressus, Juniperus, Sequoia, Sequoiadendron, Tax- odium, Thuja.] PHILips, E. W. J. Identification of softwoods by their microscopic structure. Dep. Sci Industr. Res. (London) Forest Prod. Res. Bull. 22. 111 + 56 pp. 3 pls. 1948. [A similar account appeared in Bot. Jour. Linn. Soc. 52: 59-320. 1941.] Pitcer, R. Taxodiaceae. In: A. ENGLER, Nat. Pflanzenfam. ed. 2. 13: 342- oe Cu- pressaceae. /bid. 361-403. 1926. [Detailed illustrations of the genera, figs. 179-215.] H. MEtcuior. Gymnospermae. Jn; H. MELCHIOR, A. ace rs an der Pflanzenfam. ed. 12. 1: 312-344. 1954. [Cupressaceae s./., 3 ie Price, R. A., & J. M. Lowenstein. An immunological pret of the Sciadopity- aceae, Taxodiaceae, and Cupressaceae. Syst. Bot. 14: 141-149. 1989. [Immunolog- ical data indicate that the Cupressaceae and Taxodiaceae form a single lineage, with the Cupressaceae s.s.a monophyletic subset, while Sciadopitys is a distant outgroup. ] PROPACH-GEISELER, C. Die Bliiten der Coniferen II]. Zur Morphologie und Entwick- lungsgeschichte der weiblichen Bliitenzapfen der Cupressaceen. Bibliot. Bot. 114(2): 1-56. pls. 1-13. 1936 Quinn, C. J., . A. GADEK. Sequence of xylem differentiation in leaves of Cupres- saceae. Am. Jour. Bot. 75: 1344-1351. 1988. [Species of Callitris, Chamaecyparis, Cupressus, Juniperus, and VW He ees compared. } Rapais, M. Contribution a l’étude de aad comparée du fruit des coniféres. Ann. Sci. Nat. Bot. VII. 19: 165-368. | RaprorD, A. E., H. E. AHLEs, & C. R. BELL. ese of the vascular flora of the Carolinas. Ixi + 1183 pp. Chapel Hill. 1968. [Taxodiaceae, Cupressaceae s./., 40-43. REHDER, A. Manual of cultivated trees and shrubs hardy in North America, exclusive of the subtropical and warmer temperate regions. ed. 2. xxx + 996 pp. New York. 1940. [Cupressaceae s./., 42-68. . Bibliography of cultivated trees and shrubs hardy in the cooler temperate regions of the Northern Hemisphere. xl + 825 pp. Jamaica aes, Massachusetts. 1949. [Detailed nomenclatural treatment; Cupressaceae s./., Ror, I. ropa nee nadine Deutung der Sear von Sciadopitys. Flora 152: 1-23. ROuANE, P. Etude comparée de la répartition des ramifications au cours de l’ontogenése 290 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 de quelques Cupressacées. Trav. Lab. Forest. Toulouse, Tome I, Vol. IX, Art. III: 1-277. 1973 RUSHFoRTH, K. D. Conifers. 232 pp. London. 1987. SARGENT, C.S. Silva N. Am. 10. vii + 159 pp. p/s. 497-537. 1896. [Cupressaceae s./., 69-154, pls. 516-537.) Ibid. 14. 152 pp. pls. 705-740. 1902. [Cupressaceae s./., 89- 96, pls. 738-740.] Manual of the trees of North America (exclusive eras ed. 2. xxvi + 910 pp. Boston and New York. 1926. [Cupressaceae s./., 61—90.] SAx, K., & H. J. Sax. Chromosome number and oe in the conifers. Jour. Arnold Arb. 14: 356-374. pls. 75, 76. 1933. [Chromosome numbers for species of C He ne ae Cryptomeria, Juniperus, Platycladus, Taiwania, Taxodium, and Thuja, all 2n = 22 except for a tetraploid count in J. chinensis; karyotypes for all but ae first two genera. ] Saxton, W. T. Contributions to the life history of Tetraclinis articulata Masters with some notes on the phylogeny of the Cupressoideae and Callitroideae. Ann. Bot 27: 577-605. 19 13a. . The classification of conifers. New Phytol. 12: 242-262. 1913b. [Cupressaceae and Taxodiaceae combined on the basis eS Sao characters. | Savior, L. C., & H. A. Simons. Karyology of Sequoia sempervirens: karyotype and accessory chromosomes. Cytologia 35: . 303. 1970. SCHLARBAUM, S. E., L. C. JOHNSTON, & T. TsucuryA. Chromosome studies of Metase- quoia glyptostroboides and Taxodium renner Bot. Gaz. 144: 559-565. | & UCHIYA. Cytotaxonomy and phylogeny in certain species of Taxodiaceae. PI. Syst. Evol. 147: 29-54. 1984. [Karyotypes of Cryptomeria, Cunninghamia, Meta- sequoia, Sciadopitys, Sequoia, Sequoiadendron, Taiwania, and Taxodium com- pared. & ——. Karyological derivation of Sciadopitys verticillata Sieb. et Zucc. from a pro-taxodiaceous ancestor. Bot. Gaz. 146: 264-267. 1985. SicBa, J. An international census of the Coniferae, I. Phytologia Mem. 7: 1-79. 1984. [Worldwide conspectus of the species of conifers, with several new varietal com- binations for the Cupressaceae. ] StnGH, H. Embryology of gymnosperms. /i; K. LinsBAUER, Handbuch der Pflanzen- anatomie. ed. 2. Band 10, Teil 2. x1 + 302 pp. Berlin. 1978. [Male en female gan pee pollination mechanisms; embryogeny: extensive bibliography. ] & J. TT A contribution to the life history of C ene D. Don. eons 13: 429-445. 1963. [Thorough review of cone morphology and embryology. & Oserol. A contribution to the life history of Biota orientalis Endl. Phytomorphology 12: 373-393. 1962. [Platycladus orientalis. SMALL, J. K. Manual of the southeastern flora. xxii + 1554 pp. New York. 1933. {Cupressaceae s./., 8-11; Cu ee ee Sporn_e, K. R. The morphology of gymnosperms. ed. 2. 216 pp. London. 1974. STAFFORD, H. A., & H. H. Le ESTER. Proanthocyanidins in needles from six genera of the Taxodiaceae. Am. Jour. Bot. 73: 1155-1162. 1986. [Cryptomeria, Metasequola, Sciadopitys, Sequoia, Oar Taxodium.] STEBBINS, G. L., JR. The ee omosomes and relationships of Metasequoia and Sequoia. Science 108: 95-98. 1948. STERLING, C. So ere eae the morphology of Metasequoia. Am. Jour. Bot. 36: 461- 471. 1949. [Thorough study of morphology and anatomy of M. g/yptostroboides.| Structure of the male gametophyte in gymnosperms. Biol. Rev. 38: 167-203. 1963. STRASBURGER, E. Die Coniferen und die Gnetaceen. 442 pp. Jena. 1872. [Important early accounts of development and morphology. ] Surova, T.D.,& V. KvAvApDzE. Sporoderm ul (M. 1990] HART & PRICE, CUPRESSACEAE 291 sequoia, Cunninghamia, Sciadopitys). (In Russian; English summary.) Bot. Zhur 73: 34-44, 1988. [LM, SEM, TEM comparisons; Sciadopitys is quite dissimilar to the other genera. ] SuszkA, B. The seedlings and so-called juvenile forms of the Cupressaceae family. (In Polish; ee summary.) Polsk. Towarz. Bot. Dendrol. Roczn. 11: 71-131. 1956. Suzuk!, M. The course of resin canals in the shoots of conifers. II. Araucariaceae, Cupressaceae a Taxodiaceae. Bot. Mag. Tokyo 92: 253-274. 1979a. [See also hid. 92: 333-353. 1979b. TAHARA, M. Contributions to the morphology of Sciadopitys verticillata. Cytclogia, Fujii Jubilee Vol. 1: 14-19. 1937. . The gametophytes, fertilization ng ee of Sciadopitys verticillata. Sci. Rep. Tohoku Univ. Biol. 15: 19-28. TAKASO, T., & P. B. TOMLINSON. os ie cone and ovule ontogeny in Cryptomeria (Taxodiaceae). Am. Jour. Bot. 76: 692-705. 1989. [SEM and LM studies of cone- scale development. ] — M. Studies on the germination of the pollen grains in conifers. 1. Jap. Jour. . 14: 13-21. 1953. ae O. Nuclear differentiation of male gametophytes in gymnosperms. Cytologia 47: 27-46. 1982. [Pattern of cell division in microgametophytes of Cryptomeria, Metasequoia, Sequoia, and Juniperus conferta TIEGHEM, P. VAN. Anatomie comparée de la fleur female et du fruit des Cycadées, des Coniféres et des Gnetacées. Ann. Sci. Nat. Bot. V. 10: 269-304. 1869. UENo, J. Morphology of pollen of pee aac Sciadopitys, and Taiwania. Jour. Inst. ae Osaka City Univ. 26. 1951. l observations pie Cupressaceae, and Araucariaceae. ee 10: 75. 87. 1959, [LM and TEM comparisons of pollen structure. ] Studies on pollen grains of Gymnospermae: concluding Leg to the rela- lange between Coniferae. [bid. 11: 109-136. p/. J, table 1. e fine structure oS cell ig of some gymnosperm olla Biol. Jour. Na oo 10: 19-25. 196¢ VIERHAPPER, F. eae neuen ee chemics der Coniferen. Abh. Zool.-Bot. Ges. Wien 5(4): 1-56. 1910. Von Rup torr, E. Volatile leaf oil analysis in eed stematic studies of North Amer- ican conifers. Biochem. Syst. Ecol. 2: 131-167. 1975. [Terpene composition of several species of a ee Pannen and 7huja compared. | Wana, F.-H., S.-C. Lee, & Z.-K. CHEN. The embryogeny of Taiwania in comparison with that of es genera of ims (In Chinese; English summary.) Acta Phytotax. Sinica 18: 129-137. pl. 1. Witscn, E. H. The conifers and taxads en Publ. Arnold Arb. 8. xi + 91 pp. 59 pls. 1916. WobpeHousE, R. P. Pollen grains. xv + 574 pp. New York. 1935. [Cupressaceae, s./., 247-250, 268-273.] Woycickl, S. On the ise of the Retinospora forms in Thuja, Biota, and Chamae- cyparis. Acta Soc. Bot. Polon. 23: 443-458. 1954. [Reproductive plants cei the juvenile leaf form (Retinospora types) apparently result from genetic mutatio plants propagated vegetatively from seedlings revert to adult leaf form.] YAMAZAKI, T., & M. TATEOKA. On the identifi Hee a pollen of Taxodiaceae. (In Japanese.) Saikyo Univ. Fac. - r. Sci. Rep. 8: YATAGAI, M., & T. Sato. Terpenes of leaf oils oes ee Biochem. Syst. Ecol. 14: 469-478. 1986. (Crptomera, Sc pan AKAHASH erpenes of leaf oils from Cupressaceae. Biochem Sy st. Ecol. ‘13: a7) 385. L985. [Profiles from species of Chamaecyparis, Juniperus, Tit ee ZAVARIN L. Lawrence, & M. C. THOMAS. Compositional variations of leaf mono- 202 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 terpenes in Cupressus macrocarpa, C. py. ee C. Goveniana, C. Abramsiana, and C. Sargentii. Phytochemistry 10: 379-393. 1971. ——. & J. G. BIcHo. Tropolones of Cupressaceae—III. Phytochemistry 6: 1387- 1394. 1967. [Chemosystematic investigation of Cupressus. ] KEY TO THE GENERA OF CUPRESSACEAE (INCLUDING TAXODIACEAE) IN THE SOUTHEASTERN UNITED STATES General characters: Monoecious (or dioecious) evergreen or winter-deciduous trees or shrubs; foliage leaves alternate (often appearing 2-ranked) and linear to linear-lanceolate, or opposite to whorled and needlelike, awl-shaped, or scale- like; pollen cones with spirally arranged, opposite, or whorled microsporophylls, each sporophyll with 2 to 10 globose abaxial microsporangia, pollen nonsaccate, lacking prothallial cells; ovulate cones subglobose to ovoid or oblong; bracts and ovuliferous scales strongly fused in mature cones; scales peltate to ovate or oblong, alternate or opposite or whorled, bearing 1 to 6 [to 20] erect [or inverted] adaxial ovules, archegonia clustered; seeds with 2 (or 3) lateral wings [or 1 nearly terminal one] or wingless; cotyledons 2 to 6 (to 9); chromosome base number x =]] A. Foliage leaves alternate, linear to linear-subulate; branchlets winter pee ws US fea ep here tee tense add ie toe Gu atede hs Pah adele Gnas baeaveue apne xodium. A. eg leaves opposite or whorled, mostly reduced and scalelike; spe ever- ee panei forming flattened sprays. C. Cones globose; cone scales peltate. ...............0.. 2. Chamaecyparis. C. Cones ovoid or ellipsoid; cone scales not pelta D. Branchlets flattened in horizontal plane; na laterally winged; immature cone scales only slightly fleshy. ..................00...0000. 3. Thuja. D. ae flattened in vertical plane: seeds wingless; immature cone scales Platycladus.® 7 CTY MESHVe © ooulns nia acne Fase e aon FSi oe ee es eau oe pear flattened sprays. E. ne ces less than | cm long, fleshy, fused into indehiscent, berrylike STCUICTULGSS, & s5e4.5.57 atesiea ataety gear © sept pe pap aha ats es eRe renee RA eS uniperus. . Cone ae more than | cm long, woody at maturity, with evident borders, ear to release seeds. oo. tee . Callitris. x 1. Taxodium Richard, Ann. Mus. Hist. Nat. Paris 16: 298. 1810. Winter-deciduous [to evergreen] trees, pyramidal to narrowly conical when young, the crown often broad in older individuals; trunk much enlarged at base, often buttressed. Bark light- to reddish-brown, fibrous, ridged, often apex of the shoot persistent and with prominent axillary buds, those lower on the shoot without evident axillary buds and deciduous; winter buds globose, scaly. Juvenile leaves linear-lanceolate, whorled or spirally arranged; adult 6Platycladus orientalis (L.) Branee, malty to seater lil Is commonly planted tr in our region and nay occasionally escape (Little, | Ib) t g sme te (Wunderlin, pers. comm.). 1990] HART & PRICE, CUPRESSACEAE 293 leaves spirally arranged, either 2-ranked by twisting of leaf bases, thin and linear-lanceolate, or 5- to 8-ranked, linear-subulate and keeled, closely ap- pressed to the branches (the 2 leaf forms on the same or different trees). Pollen cones small, ovoid, in long, drooping racemes or panicles terminating the previous year’s shoots; microsporophylls 6 to 10 [to 15], broadly ovate to peltate; microsporangia (2 to) 4 to 9 (or 10), globose, pendulous, in 2 rows at base of the abaxial side of the microsporophyll; pollen with evident papilla. Ovulate cones terminal on short, scaly branchlets near ends of previous year’s branchlets, maturing in | year; immature bract-scale complexes spirally im- bricate, each with (1 or) 2 (or 3) erect, bottle-shaped ovules; mature ovulate cones subglobose to somewhat ovoid; cone scales thick, woody, peltate, 4-sided, with mucronulate umbo; resin vesicles with blood-red resin prominent on interior portions of cone scale. Seeds usually 2 on the adaxial side of each cone scale, erect, attached laterally to the stalk of the scale by a large, pale hilum; seed coat thick, woody, lustrous, with 3 small, unequal corky wings; cotyledons 3 to 9. Chromosome number 27 = 22. Type species: Taxodium distichum (L.) Rich. (Name from Greek, in allusion to the yewlike leaves of the type species.)— BALD CYPRESS, SWAMP CYPRESS. A genus of two closely related species or perhaps a single polymorphic one, native to swampy and riverine areas of the southeastern and central United States and from extreme southern Texas south through much of Mexico to Guatemala. The genus was widespread in Europe and western North America in the Tertiary, becoming extinct in these areas with climatic deterioration in the Pliocene (Florin, 1963). There has been considerable disagreement concerning the number of species to recognize within 7axodium. Britton (1926), Dallimore & Jackson, and Reh- der (1940) each recognized three species (7. distichum Rich. (bald cypress), T. ascendens Brongn. (pond cypress), and 7. mucronatum Ten. (Tule tree, Mon- tezuma cypress), although substantial intergradation has been reported, par- ticularly between pond and bald cypresses. Watson (1983, 1985), after reviewing morphological, anatomical, biochem- ical, and cytological data, concluded that the differences between the pond and bald cypresses are minor, showing considerable overlap and being subject to environmental modification. Watson thus suggested varietal status (as var. imbricarium (Nutt.) Croom) for the pond cypress, and this treatment is adopted here. Montezuma cypress is Somewhat more distinct in morphology and ecology and is allopatric in distribution (reaching extreme southern Texas from Mexico and Guatemala) and 1s thus usually treated as the separate species Taxodium mucronatum, although it may be more appropriately treated as a third variety, T. distichum var. mexicanum (Carr.) Gordon. It differs in being semievergreen, retaining its annual leafy shoots until after the new shoots have leafed out in winter or spring, and being considerably less cold hardy in cultivation than the other taxa. It has sometimes been reported to have larger, more glaucous ovulate cones (Henry & McIntyre) and longer pollen-cone-bearing branches, but more thorough sampling indicates considerable overlap for these features 294 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 (Brown, 1984; Little, 1980; Martinez, 1963). The smaller branches tend to be more drooping in the Montezuma cypress, and knees are only occasionally present (Martinez, 1950, 1963). Kaieser (1953) reported that the horizontal walls of the ray parenchyma cells are thinner in the Mexican taxon, while there is little difference in the wood anatomy of the bald and pond cypresses. Isozyme electrophoresis of population samples would be very useful in assessing the degree of genetic differentiation of Montezuma cypress from the other taxa. Taxodium distichum var. distichum, bald cypress, swamp cypress, 2 = 22, is a large tree (to 50 m) with the trunk angular at the base and the leaves two- ranked on the annual branchlets. The bark is usually thin and smooth, and the knees are usually slender. Taxodiuim distichum var. distichum 1s native on the Coastal Plain from southern Delaware to Florida, west to the valley of the Devil River in Texas, and northward in the Mississippi Embayment through Louisiana and Arkansas to southeastern Missouri and Tennessee, western and northwestern Kentucky, southern Illinois, and southwestern Indiana. The species occurs in riverine swamps that are usually inundated for several months of the year and in low, saturated stream-bank habitats and wet depressions in pine barrens. It attains its largest size in the Gulf and south Atlantic coastal areas, where it tends to form pure stands in the great river swamps. The species commonly occurs with water tupelo or tupelo (Nyssa aquatica L., N. sylvatica Marsh.) and also grows in drier habitats with red maple, water ash, and sweet gum. Taxodium distichum var. imbricarium (Nutt.) Croom (7. ascendens Brongn., T. imbricarium (Nutt.) Harper, 7. distichum var. nutans auct., non (Aiton) Sweet), pond cypress, 27 = 22, is a smaller tree (to ca. 25 m in height) that occurs on the Atlantic and Gulf Coastal plains from southeastern Virginia to southeastern Louisiana. Its trunk has rounded ridges, and its appressed, subu- late leaves are ca. 5-12 mm long on both annual and perennial branchlets. The bark tends to be thicker and more strongly furrowed than in var. distichum, and the knees, if present, are short and rounded. Watson (1983) summarized the ecological differences between the bald and pond cypresses. Pond cypress tends to grow in pine-barren ponds, often un- derlain by limestone, whereas bald cypress generally occurs 1n riverine swamps. Additionally, the habitat of the pond cypress has a decreased water flow and is more prone to drought and fire; the pH is lower, and nutrients are less available. Neufeld and Watson (1983) have suggested that the pond cypress 1s a recent derivative of the bald cypress that evolved in response to harshening environments along the — Plain, with the smaller, closely appressed leaves helping to reduce water los Taxodium has often re considered to be closely related to the East Asian genus G/yptostrobus, which is similar in its habitat, possession of knees, pattern of cone-scale development, relatively large number of cotyledons (averaging ca. 5 or 6), and winter-deciduous branchlets (Britton, 1926; Henry & McIntyre; Hida, 1957; Pilger & Melchior). \Wetasequoia is also similar to Taxodium in being winter deciduous and having the pollen cones in a racemose arrangement, and it groups with 7axodium and Glyptostrobus in Hart’s preliminary cladistic analyses. In contrast, immunological comparisons of seed proteins by Price & 1990} HART & PRICE, CUPRESSACEAE 295 Lowenstein indicate that 7axodium is most similar to G/yptostrobus and Cryp- tomeria and that Metasequoia is most similar to Sequoia and Sequoiadendron. 7axodium is a very distinct genus, differing from the other winter-deciduous members of the family in a number of vegetative and reproductive characters. The knees of G/yptostrobus are curved and bent rather than conical and erect as in Taxodium (Henry & McIntyre). The older perennial branchlets of G/yp- tostrobus bear scalelike leaves that remain green for several years and have rows of white stomatal dots on the surface, whereas those of Taxodium bear elongate linear-subulate leaves that become brown and corky in the second year. The cones of G/yptostrobus are pyriform and terminal on the branchlet, have scales that are elongate and imbricate at maturity and lack prominent resin pockets, and have the bract and cone-scale united at the base but free at the tip. The body of the seed is small and ovoid, bearing a single long, nearly terminal wing. In contrast, the mature cones of Taxodium are globose to el- lipsoid, are borne laterally on the major branchlets, and have peltate scales with the edges meeting but not overlapping. The seed body is larger, much thicker walled, and three-angled, with only small, corky wings in the angles. Metasequoia differs from Taxodiuim in a number of features, notably in its smaller, more flattened cone scales, 1ts opposite rather than spirally arranged leaves and cone scales, and its strongly compressed seeds that are more than two (usually five to eight) per scale, each with two lateral wings and only two cotyledons (Florin, 1952: Stebbins Taxodium develops three unusual structures in response to flooding: butt- swells, buttresses, and knees. The term “buttswell” refers to the enlarged basal portion of the trunk, while buttresses are longitudinal ridges on the buttswell. Variation in size and shape of these structures is associated with fluctuations in exposure of the areas to air or water during the early part of the growing season, with the height of the buttresses tending to correspond to the average depth of flooding. According to Kurz & Demaree, buttress development results from the simultaneous presence of water and air, while individuals grown under permanently flooded conditions or in well-drained soils not subject to flooding fail to develop buttswells or buttresses. The well-known “cypress knees” are usually emergent, cone-shaped struc- tures produced as extensions of the roots and may be as tall as 3—4 m (Brown, 1984). They may arise as small swellings on the upper surfaces of shallow adventitious roots (Shaler; Whitford) or may be formed from shallow roots growing upward and then bending sharply downward, with the geniculate point becoming the “knee” (Brown & Montz). Various functions have been attributed to the knees. One suggestion has been that they act as pneumatophores (“breath- ing organs’) that funnel oxygen into the root system (Dickeson & Brown; Mattoon; Shaler). More recent experiments have shown, however, that little if any gas exchange occurs (Kramer e/ a/.), and that no stomata, lenticels, or conducting tissues are found on the surface of the knees (Brown, 1984). Other possible functions include the storage of starch, which accumulates there in significant amounts (Brown & Montz), and stabilization of the plant during severe storms (Lamborn, 1890a). Removal of the knees appears to have little adverse affect on the trees, at least over the short term (Mattoon). 296 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 Despite considerable variation in the extent of buttressing, buttswells, and knees, some authors have attempted to use these features to distinguish bald cypress from pond cypress. Watson (1983) pointed out that their magnitude appears to be ecophenically determined since they generally do not develop if the stem is not exposed to flooding; they are thus of limited taxonomic utility. Many bald-cypress trees reach considerable size and age. One in Louisiana measured ca. 13 m in circumference and 26.7 m in height (Brown & Montz). Some individuals of Montezuma cypress also attain great size and have multiple stems at the base. At Santa Maria de Tule in Oaxaca, the famous “‘Tule tree 1s reported to have the largest diameter of any tree (ca. 12 m), although it appears to represent three individuals that have grown together (Cronemiller; Little, 1980). This tree has been estimated to be anywhere from 1000 to 3000 years old (Kriissmann). The wood of Taxodium is known for its durability in contact with soil and severe weather. The heartwood is so durable that it has been called “the wood eternal” and has been used in several types of heavy construction. The Seminole Indians in Florida have also used the wood in construction and for fenceposts, stockades, and dugout canoes. More recently, it has been valued for interior woodwork, cooperage, fence posts, and railroad ties (Little, 1980). The wood of pond cypress 1s purportedly heavier and stronger than that of var. distichum (Harper, 1902). Wood production from 7. distichum peaked at more than one bilhon board feet in 1913 but has subsequently declined substantially because of heavy cutting (Mattoon; Sternitzke). In recent years second-growth bald cypress has again built up in abundance in the southeastern United States. The resin of Taxodium mexicanum has been used in Mexico as a cure for wounds, ulcers, and toothaches (Henry & McIntyre). Taxodium is of some horticultural importance in the eastern United States and Europe and is frequently planted along watercourses and farms as a wind- break in China. Cultivars of 7axodium are hardy as far north as Massachusetts, New York, and Michigan and grow best in deep, sandy loam with plentiful moisture and good drainage (Fowells). REFERENCES: Under family references see ALvIN & Bouter; L. H. BAILEY; BEAN; Britton, 1908, 1926; CHANEY, 1951; CokER & ToTTEN; DALLIMORE & JACKSON; G. ERDTMAN; H. ERDTMAN & Norin; FLorin, 1931, 1952, 1963; FoweLis; GAUSSEN, 1967: HARDIN: HART; HEGNAUER, 1962, 1986; HENRY & McIntyre; Hipa, 1957; KHosHoo, 1961; KRUSSMANN; LITTLE, 1971, 1979, 1980; Lru & Su; MARTINEZ, 1963; MEHRA & KHOSHOO; MILLER, 1977: PEIRCE, 1936: PHILLIPS; PILGER; PILGER & MELCHIOR; PRICE & LOWENSTEIN; REHDER, 1940, 1949; SARGENT, 1896, 1926; Sax & SAx; SCHLARBAUM, JOHNSON TSUCHIYA; SMALL: STEBBINS; UENO, 1960B; AND WODEHOUSE. BEAVEN, G. F., & H. J. OostinGc. Pocomoke Swamp: a study of a eae swamp on the eastern shore of Maryland. Bull. Torrey Bot. Club 66: 367-389. 9. BERNARD, J. M. The status of 7axodium are i ) Richard (bald cores) in New Jersey. Bull. Torrey Bot. Club 92: 305-307. 5. [Extant stands not nativ BLANck, C. E. An ecological study of eo ae distichum (L.) Nee 1990] HART & PRICE, CUPRESSACEAE 297 in eastern North Carolina. ae Uapubl, Master’s thesis, East Carolina University, Greenville, North Carolin Bowers, L. J. Tree ring Se of baldcypress growing in varying flooding regimes in Barataria Basin, Louisiana. 175 pp. Unpubl. Ph.D. dissertation, Louisiana State University, Baton Rouge, Louisiana. 1981.* a A. Note sur quelques coniféres de la tribu des Cupressinées. Ann. Sci. t. I. 30: 176-191. 1833. [Pond and bald cypresses distinguished. eee . i Cypress—the tree unique: the wood eternal. Jour. New York Bot. Gard. 1: 36-39. 1951. Morphology and biology : as trees. Pp. 16-24 in K. C. Ewer & H. T. OpuM, eds., Cypress swamps. 4. [T. distichum.] & G. N. Montz. une ve tree unique, the wood eternal. xvi + 139 frontisp., 179 ae & white, 9 color photographs. Baton Rouge, Louisiana. 1986. [Reviewed by W. D. ReEsE, Castanea 52: 128. 1987.] Cain, S. A. Bald cypress, es distichum (L.) Rich., at Hovey Lake, Posey County, Indiana. Am. Midl. Nat. 16: 72-82. 1934. CLEVENGER, S. V. The causes +s of cypress knees. Am, Nat. 24: 581. 1890. Coker, W. C. On the gametophytes and embryo of Taxodium. Bot. Gaz. 36: 1-27, 114-140. pls. J-1/. 1903. [7. distichum.] The bald cypress. Jour. Elisha Mitchell Sci. Soc. 46: 86-88. p/. 7. 1930. [An extremely large tree from Seminole County, Florida, illustrated.] Coutter, S. Histology of the leaf of Taxodium. I and II. Bot. Gaz. 14: 76-81, 101- 107. pl. 17. 1889. ara: F. P. El sabino, the national tree of Mexico. Jour. Forestry 55: 461, 462. 957. [7. mucronatum.] pes G.L. A note on the morphology pee deciduous shoot of Taxodium distichum. Bull. cae Bot. Club 66: 167-172. 9a. . structure and ete ev apical meristem in the shoots of Taxodium Ficiant Ibid. 431-452. evelopment of the ie leaves of Taxodium distichum. Am. Jour. Bot. 27: 471-482. 1940. DeMAREE, D. Submerging experiments with Taxodium. Ecology 13: 258-262. 1932. [Seeds and seedlings of 7. distichum do not survive extended submerge DENUYL, D. on e observations on bald cypress in Indiana. Ecology 42: 841-843. 1961. Ewe, K. C., & H. T. Opum, eds. Cypress swamps. xviii + 472 pp. Gainesville, Florida. 1984. ae of 7. distichum swamps.] FAULKNER, S. P. Genetic variation of cones, seed, and nursery grown seedlings of baldcypress (Taxodium distichum (L.) Rich.) provenances, Unpubl. Master’s thesis, Louisiana State University, Baton Rouge. GeIcer, H., & W. DE GROOT-PFLEIDERER. Die ica von Taxodium distichum. Phytochemistry 12: 465, 466. 1973. Gour.ay, W. B. The Mexican Si cypress (7axodium mucronatum Tenore). Quart. Jour. Forestry 34: ae 61. - HALL, G. W., G. M. GS, ie E. So.tis, & P. So_tis. Genetic uniformity of El Arbol del Tule ae Tule re Madrofio 37: 1-5. 1990. [Electrophoretic analysis of enzymes of leaf material from each of 8 major segments and of 2 nearby trees. “The results are consistent with the hypothesis that the Tule Tree is one genetic Harper, R. M. Taxodium distichum and related species, with notes on some So aa sage influencing their distribution. Bull. Torrey Bot. Club 29: 383-399, 1902. rther observations on Taxodium. Ibid. 32: 105-115. 1905. eee J. W. The Mexican cypress. Forest Leaves 11: 24. 1907 KaArEsER, M. Mo rphology and embryogeny of the bald cypress, Taxodium distichum (L.) Rich. Unpubl. Ph.D. dissertation, University of Illinois, Champaign-Urbana. 1940.* 298 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 b ‘the pond cypress (7 axodium ascendens Brongn.). Trans. Ilinois Acad. Scl. 42: 63-67. 1949. . Microstructure of the wood of the three species of Taxodium. Bull. Torrey Bot. Club 80: 415-418. 1953. [Bald and pond cypresses not separable on the basis of wood anatomy, while Montezuma cypress has thinner horizontal walls on its ray- parenchyma cells. KRAMER, P. J., W.S. Ritey, & T. T. BANNISTER. Gas exchange of cypress knees. Ecology 33: 117-121. 1952. [Little evidence for gas exchange between roots and knees.] Kurz, H., & D. DEMAREE. Cypress buttresses and knees in relation to water and air. Ecology 15: 36-41. 1934. . WAGNER. The role Bae roots in survival of cypress. (Abstract.) Jour. Tennessee Acad. Sc1.:27: 201.1952 & Factors in cypress ice devel opment. Ecology 34: 157-164. 1953. [Cypress domes are formed in depressions by the confluent tops of crowded 7. distichum trees, with the shorter trees at the outer and drier edges apparently pruned by fire.] LAMBORN, R. H. A knees of the bald cypress: a new theory of their function. Garden Forest 3: 21, 22. 1890a. [Knees help to anchor the tree ae a . storms. } The ia Taxodium distichum. Am. Nat. 24: 333-340. pl. J2. LANGDON, O. G. Silvical characteristics of baldcypress. U. S. Dep. Agr. at Sei: Southeast. Exper. Sta. Pap. 94. 7 pp. / fig. 1958. MArTinez, M. El ahuehuete (7axodiuim mucronatum Ten.). Anal. Inst. Biol. México 21: 25-82. 1950. saa G. N., & A. CHERUBINI. An ecological study of a baldcypress swamp in St. Charles Parish, Louisiana. Castanea 38: 378-386. 1973. NeuFELD, H. S. Ecophysiological implications of tree architecture for two cypress taxa, Taxodium distichum (L.) Rich. and 7. ascendens Brongn. Bull. Torrey Bot. Club Rao, A. R., & J. P. Tewari. On the foliar sclereids of Taxodium distichum Rich. Proc. Natl. Inst. Sci. India 27: 41-45. SHALER, N.S. Notes on 7axodium distichum, or bald cypress. Mem. Mus. Zool. Harvard SHARMA, G. K., & L. MADSEN. Variation in bald cypress from different habitats. Jour. Tennessee Acad. Sci. 53: 115, 116. 1978. SLtavin, A. D. Our deciduous conifers. Am. Hort. 12: 48-53. 1933. [T. distichum: excellent photographs of vars. distichum and imbricarium.] SMALL, J. K. A botanical expedition to the Big Cypress. Nat. Hist. 20: 488-500. 1920. The cypress, southern remnant of a northern fossil type. Jour. New York Bot. Gard. 32: 125-135. 1931. STAHLE, D. W., M. K. CLEAVELAND, & J. G. HEHR. North Carolina climate changes oe enie oe tree rings: A.D. 372 1985. [Based on living trees of T. distichum to 1700 years eee H. S. Bald ¢ cypress: endangered or expanding species. Econ. Bot. 26: 130- 134. 1972. TomLinson, P. B. The biology of trees native to tropical Florida. ix + 480 pp. Allston, Massachusetts. 1980. [Taxodium, 68-73; excellent illustrations of 7. distichum. Differences between 7. distic hum and T. ascendens ‘are maintained in cultivated trees grown oe by side. VasiL, V., & R. K. SAHNI. Morphology | embryology of Taxodium mucronatum Tenore. See 14: 369-384. 1964. VeRNOoN, R. O. Cypress domes. Science 105: pea 99. 1947. Watson, F. D. A taxonomic study of pondcypress and baldcypress. 214 pp. Unpubl. Ph.D. dissertation, North Carolina State University, Ralei 3. The nomenclature of pondcypress and baldcypress (Taxodiacese’: Taxon 34: 1990] HART & PRICE, CUPRESSACEAE 299 506-509. 1985. [7. distichum var. imbricarium is the correct name at the varietal level for pond cypress. We cu, W. H. An aa study of the bald cypress in Indiana. Proc. Indiana Acad. Sci. 41: 207-213. 1931. WuitForp, L. A. A theory of the formation of cypress knees. Jour. Elisha Mitchell Sci. Soc. 72: 80-83. 1956 2. Chamaecyparis Spach, Hist. Nat. Veg. Phan. 11: 329. 1841. Pyramidal monoecious evergreen trees (rarely shrubs) with nodding leading shoots and a slender, spikelike crown from a thickened trunk. Branchlets slen- der, flattened (gradually becoming terete in later years if not shed), distichous and forming horizontal sprays. Bark reddish- [to grayish-]brown, irregularly ridged, often peeling in strips or scales. Wood soft [or hard], whitish to pinkish [to yellowish], aromatic [or not aromatic]. Leaves entire, aromatic when crushed; juvenile leaves whorled, linear-lanceolate, acuminate; adult leaves decussate, scalelike, dimorphic, the facial pair flattened, ovate to rhombic, acuminate to obtuse, with a central gland [or eglandular]; the lateral pair rounded or strongly keeled. Cones borne terminally on lateral branchlets, opening in early spring from buds formed the previous year; pollen and ovulate cones borne on separate branches. Pollen cones ovoid to oblong, quadrangular; microsporophylls ca. 8 to 12 (to 20), decussate; microsporangia 2 to 4 (to 6); pollen microverrucate, with obscure germinal aperture. Ovulate cones globose [to ellipsoid], maturing in | year [2 in C. nootkatensis], bearing 4 to 8 [rarely to 16] decussate, peltate scales; ovules bottle shaped, erect, [1 or] 2 [to 8] per scale. Maturing cones more or less erect, glaucous, ultimately red-brown. Seeds | or 2 [rarely to 8] per scale, ovate, with 2 broad lateral wings; cotyledons 2. Chromosome number 2n = 22. Lecrotrype species: Chamaecyparis thyoides (L.) BSP.; see Britton, N. Am. Trees, 102. 1908. (Greek name from chamai, on the ground, and Ayparissos, cypress, alluding to the affinity of the genus to Cupressus, true cypress.)— FALSE CYPRESS, CYPRESS. A genus of six species in North America and eastern Asia. Three species are native to eastern Asia and three (Chamaecyparis thyoides (L.) BSP., C. noot- katensis (D. Don) Spach, and C. Lawsoniana (A. Murray) Parl.), to North America, with the latter two restricted to the Pacific Coast region. The Asian species are C. formosensis Matsum. in Taiwan, C. obtusa (Sieb. & Zucc.) Endl. in Japan and Taiwan, and C. pisifera (Sieb. & Zucc.) Endl. in Japan. Chamaecyparis is generally thought to be closely related to the widespread Northern Hemisphere genus Cupressus L., which also has globose cones with peltate scales, and has sometimes been treated as a subgenus or section of the latter genus. It differs in having smaller ovulate cones (ca. 6-15 vs. ca. 15—40 mm in diameter), more heavily flattened branchlets, entire (vs. minutely ciliate) foliage leaves, and a shorter reproductive cycle (usually one vs. two years until ovulate-cone maturation). A small-coned species native to China has some- times been treated as Chamaecyparis funebris (Endl.) Franco, but its biflavo- 300 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 micropyle wears = fees fee MB f rol MS ARI een Co “SE micropyle cone scale Chamaecyparis. a—n, C. thyoides: a, branchlet tip, showing lateral and moe leaves, < 5; b, detail of branchlet tip with two terminal microsporangiate strobill, x 12; c, microsporophyll with two sporangia (abaxial view), x 25; d, microsporophyll a view), showing dehisced sporangia, < 25; e, two ovulate cones at time of pol- lination, x 12; f, cone from “e,” seen from above to show arrangement of cone scales m, diagrammatic longitudinal section of mature seed, with embryo unshaded, gameto- phytic storage tissue dotted, seed coat hatched, x 10; n, embryo dissected from seed, showing two cotyledons, x 10. 1990] HART & PRICE, CUPRESSACEAE 301 noid pattern supports its original placement in Cupressus (Gadek & Quinn, 1987 Chamaecyparis nootkatensis is unusual in the genus in having less-flattened branchlets and a longer reproductive cycle; it is also divergent in its biflavonoid profile and tropolone composition (H. Erdtman & Norin; Gadek & Quinn, 1985). This species also is notable for producing intergeneric hybrids with Cupressus in cultivation. Leyland cypress (x Cupressocyparis Leylandii (Jack- son & Dallim.) Dallim. = Chamaecyparis nootkatensis x Cupressus macro- carpa Gordon) apparently originated in England from spontaneous crosses in both directions (Osborn). More recently, hybrids between Chamaecyparis noot- katensis and Cupressus glabra Sudw. and C. /usitanica Miller, respectively, have also been reported from cultivated plants in England (Mitchell). Our sole species, Chamaecyparis thyoides (including C. Henryae H. L. Li, C. thyoides var. Henryae (H. L. Li) Little), Atlantic white cedar, white cedar, swamp cedar, “juniper,” occurs in swamps and wet woods near the Atlantic and Gulf coasts from southern Maine to northern Florida and westward to southeastern Mississippi. In the northern part of its range, it often occurs in ure stands, while in the south it frequently grows with bald cypress. It can be distinguished by its adult leaves that are usually glandular and not conspicu- ously whitened below, and by its branchlets that are irregularly arranged rather than held in the horizontal plane. In segregating the southern populations of the species (from Florida through Mississippi) as C. Henryae, H. L. Li cited a number of morphological differences (e.g., smoother bark with twisting ridges, lighter-colored microsporophylls, and less-glaucous ovulate cones), but the divergence between northern and southern populations is apparently clinal rather than abrupt (Hardin; E. L. Little, 1966). Thus, E. L. Little (1966) reduced C. Henryae to varietal status under C. thyoides and later (1979) placed it in synonymy. Chromosome counts of 2n = 22 have been reported for four species of Chamaecyparis (Khoshoo, 1961; Kuo et al.; Sax & Sax); evidently no count has ever been made for C. thyoides. Natural interspecific hybridization has not been reported, but artificial crosses have been attempted in several combina- tions (Fukuhara; Yamamoto, !981a, b). Meiotic irregularities have been re- ported in hybrids between C. obtusa and C. pisifera (Fukuhara), while a high percentage of nonviable seedlings was obtained in crosses of C. Lawsoniana and C. pisifera (Yamamoto, 198 la). Among species of Chamaecyparis, significant differences have been reported in distributions of foliar terpenoids (Von Rudloff; Yatagai et al.) and heartwood tropolones (H. Erdtman & Norin), but more comprehensive population studies are needed. Chamaecyparis nootkatensis appears to be unique in the genus in having the tropolone nootkatin and the biflavonoid cupressuflavone, which are widespread in the genus Cupressus (H. Erdtman & Norin; Gadek & Quinn, 1985). Most species of Chamaccyparis are utilized as ornamentals, with C. Law- soniana (Port Orford cedar), C. pisifera (Sawara cypress), and C. obtusa (Hinoki cypress) being of particular importance. Hinoki cypress has religious signifi- cance in Japan and was often planted outside Shinto temples and used for 302 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 construction of temples and palaces (Wilson). The durable wood of the North American species has been valued for construction of boats and houses and for cooperage and shingles (Sargent, 1896, 1926). Chamaecyparis thyoides has historically been an important timber tree in the eastern United States, but large trees have been greatly depleted by logging (Jarvis; Tangley). Submerged logs of the species are so resistant to decay that they have been ‘‘mined” from swamps (Jarvis). REFERENCES: Under family references see L. H. BAILEY; BEAN; BRITTON, 1908; DALLIMORE & JACKSON; H. ERpTMAN & Norn; FITSCHEN; FLORIN,. 1931, 1963; FOWELLS; GADEK & QUINN, 1985, 1987: GAussEN, 1968: GreGcuss, 1955, 1972; HARDIN; HART; KHOSHOO, 1961; KRUSS- MANN: Kuo ef a/., LEBRETON; LITTLE, 1971, 1979, 1980; OUDEN & Boom: Owens & Simpson: REHDER. 1940, 1949: ROUANE; SARGENT, 1896, 1926; SAx & SAx; SMALL; UENO, 1960b; Von RUDLOFF; WILSON; and YATAGAI ef al. ANDERSEN, H.E. Silvical piavadien sues of Alaska-cedar (Chamaecyparis nootkatensis). U.S. Dep. Agr. sbesae Serv. Alaska Forest Res. Center Pap. 11. 10 pp. 1959. BANNAN, M. Abnormal xylem rays in Chamaecyparis. Am. Jour. Bot. 37: 232-237. 1950. nce Br rays in C. Lawsoniana, C. ee and C. thyoides.] ——. The microscopic wood oe of North American species of Chamaecyparis. Canad. Jour. Bot. 30: 170-187. Beck, G. F. Two newly pine see 7Tsuga and Chamaecyparis) among the coniferous woods of the Teruiary. Northwest Sci. 18: 9, 1944.* BELLING, A. J. Postglacial migration of Chamaecyparis ihyoides (L.) B.S.P. (southern white cedar) in the northeastern United States. 220 pp. Unpubl. Ph.D. dissertation, New York University. New York. 1977.* Brusu, W. D. Knowing your trees: Atlantic whitecedar— Chamaecyparis thyoides (L.) Britton, Sterns & Poggenberg. Am. Forests 53: 218, 219. 1947. [Uses, map, illustra- tion. BUCHB ON Af ae The embryogeny of Chamaecyparis obtusa. Am. Jour. Bot. 19: 230- 238. 1932 — CHENG, Y. 5. & E. Von RUDLOFF. oe volatile oil of the leaves of Chamaecyparis eine Phytochemistry 9: 2517-2527. 1970. CuesNoy, L. Sur Vorigine ee i. orarilies du proembryo du Chamaecyparis Lawsoniana A. Murr. (Cupressaceae). Caryologia 25: 223-232. 1973. ELeuTerius, L.N., & S. B. JONEs. ec of white cedar in Mississipp1. Castanea 37: 67- lies 1972. [Includes notes on ecology and uses of C. thyoides.] FUKUHARA, N. Meiotic observations in the pollen mother cell of masa hybrid between eee mene obtusa and C. pisifera. Jour. Jap. Forestr : 43 441. 1978. [Irregular meiotic pairing in the hybrids; both species ee 2n = 221] GIANARDOLI, M. Recherches cytologiques sur la reproduction a oa Chamaecyparis Lawsoniana. Compt. Rend. Acad. Sci. Paris 254: 4499-4501. Hayes, G. L. Silvical characteristics of the Port Orford ve (C fen paris Lawsont- ana). U.S. Dep. Agr. Forest Serv. Silvical Ser. 7. 11 p 8.* Jarvis, G. “Juniper,” the versatile white cedar (C ee eo thyoides). Nature Mag. 38: 543, 544. 1945. KorsTIAN, C. F. Natural regeneration of southern white cedar. Ecology 5: 188-191. 1924. [C. thyoides.] . Southern white cedar. U.S. Dep. Agr. Tech. Bull. 251: 1-76. 1931. [C. thyoides.] Li, H. L. A new species of Charnaeespas Morris Arb. Bull. 13: 43-46. 1962 Henryae = C. thyoides var. Hen 1990} HART & PRICE, CUPRESSACEAE 203 Li, S.-J. Female reproductive organs of Chamaecyparis. Taiwania 17: 27-39. 1972. [C. Jormosensis, C. obtusa var. formosana. Littte, E. L., Jk. Varietal transfers in CRTESSYS and Chamaecyparis. Madrono 18: 161- 167. 1966. [C. thyoides var. Henryae.]} LittLe, S. Silvical sla e of Atlantic white cedar. U. S. Dep. Agr. Forest Serv. Northeast. Forest Exper. Sta. Pap. 118. 11 + 16 pp. 1959. McDona pn, C. B., & A. N. Aso. Structure and er trends of an Atlantic white cedar (Chamaecyparis thyoides) forest in Tyrell eye North Carolina. (Abstract.) ASB Bull. 29: 71. 1982. [Forms pure stands in mpy areas of the Carolinas and Virginia; ae tend to succeed to Nyssa- Taxodium forest if left undisturbed. } MITCHELL, A. A note on two hybrid cypresses. Jour. Roy. Hort. Soc. 95: 453, 454. 1970. i hybrids reported involving cultivated plants of Chamaecyparis nootkatensis, Cupressus glabra, and Cupressus lusitanica.] NEAL, O. M., Jr. The status of Chamaecyparis thyoides in Maine. Rhodora 42: 343, 344, 1940. Osporn, A. An interesting hybrid conifer: Cupressocyparis Leyvlandii. Jour. Roy. Hort. Soc. 66: 54, 55. 1941. [Cupressus macrocarpa x Chamaecyparis nootkatensis. | Owens, J. N., & M. MoLperR. Cone initiation and development before dormancy in yellow cedar (Chamaecyparis nootkatensis). Canad. Jour. Bot. 52: 2075-2084. 1974. . Pollination, female gametophyte, and embryo and seed development in yellow cedar (Chamaecyparis nootkatensis). [bid. 53: 186-199. 1975. Simpson, & M. Motper. The pollination mechanism in are cypress (Chamaecyparis nootkatensis). Canad. Jour. Forest Res. 10: 564-572 TANGLEY, L. Taking stock of white cedar wetlands. Becene 34: 682- ae 1984, Warp, D. B. Southeastern limit of Chamaecyparis thyoides. Rhodora 65: 359-363. 1963. [Locally abundant in northwestern Florida, with distribution ending in Marion ty.] YAMAMOTO, C. Xantha seedlings from interspecific crosses between Chamaecyparis Lawsoniana and Chamaecyparis aie a as a remarkable case of hybrid inviability. (In Japanese; English summary.) r. Jap. Forestry Soc. 63: 64-67. 198la Possibilities of interspecific hy eae between Chamaecyparis fawsontane and nee other Chamaecyparis species. (In Japanese; English summary.) /bid. 31 1- 319, ZOBEL, D. a Ti elongation patterns of Chamaecyparis Lawsoniana. Bot. Gaz. 144: 92-103. 1983. = 3. Thuja Linnaeus, Sp. Pl. 2: 1002. 1753: Gen. Pl. ed. 5. 435. 1754. Evergreen, pyramidal, monoecious trees or sometimes shrubs; leading shoot erect; trunk often lobed and buttressed, sometimes dividing into 2 or more upright secondary stems. Bark reddish brown [to grayish], thin, fissured on older trees, peeling in irregular patches and fibrous shreds. Wood soft, pale. with light brown [to dark brown] aromatic heartwood. Branches horizontal at first, becoming ascendent; branchlets slender, pendulous, forming flattened, frondlike, horizontal sprays, gradually becoming terete, the smaller leafy branchlets deciduous after several seasons. Foliage fragrant; juvenile leaves in spirally arranged whorls, linear-lanceolate, acuminate, spreading or reflexed, retained on adult plants of some cultivars; adult leaves decussately opposite, scalelike, closely 1mbricate (except on rapidly growing shoots), the facial leaves ovate, acute tipped (to ovate-lanceolate on rapidly growing shoots), with a central gland [or eglandular], the lateral leaves folded over the facial ones, keeled. Buds small, naked, hidden by the leaves. Cones terminal and solitary, appearing in early spring; pollen and ovulate cones usually produced on dif- 304 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 ferent branchlets. Pollen cones nearly sessile, cylindrical or globose; microspo- rophylls ca. 4 to 6 [to 12], decussate, short stalked and more or less peltate, each with 2 to 4 microsporangia; pollen with obscure germinal aperture. Ovu- late cones maturing in | season, terminal on short lateral branchlets, ovoid or oblong, with 8 to 12 imbricately arranged, oblong [to broadly ovate] scales, the central 4 to 6 fertile and bearing 2 (sometimes to 4) erect, bottle-shaped ovules; mature cones more or less erect; scales brownish, somewhat woody, with a minute [or more prominent] spine near the apex. Seeds (1 or) 2 (or 3) per scale, thin and flattened, with resin blisters in the thin seed coat, the membranaceous wings nearly — the whole seed, notched at the micro- pylar end; cotyledons 2. Chromosome number 2” = 22. LECTOTYPE SPECIES: Thuja occidentalis L.’? (From - Greek name of a resin-bearing conifer.)— ARBORVITAE, WHITE CEDAR. A genus of five species, two in North America and three in Asia. Thuja plicata D. Don, western red cedar, is native to northwestern North America; T. occidentalis L. to the eastern deciduous forest area. Thuja Standishii (Gor- don) Carr., 7. koraiensis Nakai, and 7. sutchuensis Franchet are native to portions of eastern Asia. Thuja occidentalis, arborvitae, northern white cedar, white cedar, swamp cedar, 2” = 22, 1s found from Nova Scotia westward to southeastern and central Manitoba, southward to the Great Lakes States and very locally south, barely reaching our region in mountainous areas of North Carolina and Tennessee. The species 1s common in the northern portion of its range, occurring over great areas of swampy forest land, where it forms largely impenetrable forests, as well as along rocky stream banks and drier limestone ridges, with best growth on neutral to alkaline substrates (Fernald; Fowells; Sargent, 1926). At the southern end of its distribution in western Virginia and Tennessee, arborvitae is less abundant, occurring only at higher elevations, usually on limestone or dolomitic cliffs. A number of floras (Britton, 1908; Coker & Totten; Little, 1980; Sargent, 1896, 1926) report the species from North Carolina, but there are apparently no natural populations remaining in the state (Clebsch; Radford et al.) Thuja occidentalis differs from the other North American species (7. plicata) in usually having four rather than six fertile scales in the ovulate cone; these are also less prominently spine tipped. Thuja plicata is a larger tree, reaching 30-—40(—70) (vs. 15-20(—25)) m in height, with more lustrous leaves that are more prominently whitened below The generic relationships of Thuja have not yet been determined with cer- tainty. It has been suggested that it is most closely related to the monotypic eastern Asian genus Thujopsis Sieb. & Zucc. (Hart), which is similar in its tropolone profile (H. Erdtman & Norin) and embryogeny (Dogra, 1984) but differs in having subglobose ovulate cones with three to five (vs. usually two) seeds per scale and more spreading, hatchet-shaped lateral leaves. Platycladus orientalis (L.) Franco (Biota orientalis (L.) Endl., Thuja orientalis L.) has often ! pified when Spach (Hist. Nat. Veg. Phan. 11: 333. 1841) transferred the any other Egactane ues . Thule dhenialx to Platycladus Spach. 1990] HART & PRICE, CUPRESSACEAE 305 been treated in the genus 7huja and may also be closely related, but is distinct from Thuja and Thujopsis in having fleshier immature cone scales, hard-coated, unwinged seeds, regularly occurring cleavage polyembryony, and vertically rather than horizontally oriented sprays of branchlets (Dallimore & Jackson; Singh & Oberoi). Chromosome counts of 2” = 22 have been reported for three of the five species of Thuja (T. occidentalis, T. plicata, and T. Standishii) by Sax & Sax. All of the chromosomes were more or less isobranchial in a cultivar of T. occidentalis studied by Mehra & Khoshoo Several preliminary comparisons have been made of terpenoid profiles from foliage of Thuja species (Banthorpe ef a/.; Von Rudloff, 1975; Yatagai et a/.). Thuja Standishii is quite different from 7. occidentalis and T. plicata in its monoterpene profile, while Platycladus orientalis is apparently quite similar to the latter two species based on the preliminary data of Banthorpe and col- leagues. Thuja plicata shows very limited variation in leaf terpenoids and isozymes (Copes; Von Rudloff & Lapp; Von Rudloff et a/.; Yeh) and oT went through a genetic bottleneck during Pleistocene ae while 7. o cidentalis seems to be more variable at the isozyme level (Walker). Thuja occidentalis differs from three of the other species a the genus in having cupressuflavone present in its leaves, in addition to biflavonoids of the amentoflavone and hinokiflavone series (Gadek & Quinn, 1985). The rate and pattern of root development in Thuja occidentalis were found to vary in relation to swampy versus calcareous substrate (Habeck). An in- creased rate of stem growth and greater wood pete were also found in plants growing on relatively dry limestone substrates by Harlow. Thuja occidentalis and T. plicata are valued for ee light, durable wood, which has been used for construction and to make fenceposts, railroad ties, and shingles. American Indian tribes have used the thick sapwood layers of both to make woven baskets and logs of 7. plicata to make canoes and totem poles. The bark is rich in tannin, and the leaf oil has been used medicinally. Thuja occidentalis was one of the first North American trees cultivated in Europe, having been planted in Paris by about 1536. It was named arborvitae (tree of life) after tea brewed from the bark (which 1s rich in vitamin C) saved the crew of the French explorer Jacques Cartier from scurvy (Little, 1980). Several species of Thuja are cultivated as ornamentals, and diverse cultivars of T. occidentalis have been selected (Dallimore & Jackson; Kriissmann). REFERENCES: Under family references see L. H. BAILEY; BANTHORPE et a/.; BEAN; BRITTON, 1908; COKER & TOTTEN; DALLIMORE & JACKSON; DAGUILLON; DoGRA, 1984; G. ERDTMAN; H. ERDTMAN & Norn; FITSCHEN; FLoRIN, 1931, 1963; FOWELLS; GADEK & QUINN, 1985; GAUSSEN, 1968; GreGcuss, 1955, 1972; HARDIN; HART; HEGNAUER, 1962, 1986; KHOSHOO, 1961; KRUSSMANN; LITTLE, 1971, 1980; MEHRA & KHOSHOO; OWENS & SIMPSON; PEIRCE; PHILLIPS; RADFORD ef a/.; REHDER, 1940, 1949; SARGENT, 1896, 1926; SAx & SAX: SINGH: SINGH & OBEROI; Von RuDLorF, 1975; and YATAGAI ef a AVITABILE, A. The eastern white cedar, 7 ees orcienians L., an early pollen source for honeybees. Am. Bee Jour. 122: 261. 306 JOURNAL OF THE ARNOLD ARBORETUM [voL. 7] BANNAN, M. W. Vascular rays and adventitious root formation in Thuja occidentalis L. Am. Jour. Bot. 28: 457-463. 1941a. [Unusually wide rays are the sites of initia- tion of adventitious roots.] . Wood structure in He occidentalis. Bot. Gaz. 103; 295-309, . Ring ie tracheid s ane - volume in the wood of Thuja conn L Canad. Jour. Bot. 32: 466— ‘479. . Girth increase in white cedar ae occidentalis) stems of irregular form. [bid. 35: 425-434. 1957. CAPLENOR, D., & H. Sperr. Thuja occidentalis L. on the Eastern Highland Rim in Tennessee. Jour. Tennessee Acad. Sci. 50: 74, 75. 1975. [Distribution and ecology of 7. occidentalis in Tennessee. | CLesscuH, E. E. C. Was arbor vitae (Thuja occidentalis a native to North Carolina in historic times? Am. Jour. Bot. 76 (Suppl.): 157. 1989. [T. occidentalis apparently present in low numbers in fae early 1900 s but now repa a & G. t discoveries of ran xtensions of vascular plants associated with disjunct eee vitae (Thuja a. L.) in the Southeast. (Ab- stract.) ASB Bull. 33: 69. 1986. [Native occurrence of arborvitae in North Carolina is undocumented.] Cook, P. L. A morphological comparison of two Se of Thuja. Unpubl. Ph.D. dissertation, Univ. of Hlinois, Champaign—Urbana. 1939.* Cores, D. L. Isoenzyme uniformity in western red a aa a Oregon and Washington. Canad. Jour. Forest Res. 11: 451-453. 1981. [7.1 FERNALD, M. L. Lithological factors a the range of Pinus aay and Thuja occidentalis. Rhodora 21: 41-67. 1919. [7. occidentalis largely restricted to basic soils; in outlying areas found on Bene soils. Haseck, J. R. White cedar ecotypes in Wisconsin. Ecology 39: 457-463. 1958. [Local ecotypic differentiation in root development in 7. occidentalis linked to substrate conditions. ] Hanpa, M. R. ie life history of Thuja occidentalis. Jour. Burma Res. Soc. 16: 214- 219. 1926 Har.ow, W. M. effect of site on the structure and growth of white cedar, Thuja occidentalis L. a ogy 8: 453-470. p/. 7. 1927. [Stem growth is faster and wood is stronger and heavier in trees from limestone outcrops than in those from wetland siles. Lanp, W. G. A morphological study of Thuja. Bot. Gaz. 34: 249-259. pls. 6-8. 1902. [Embryology of 7. occidentalis. Martin, P. C. A morphological comparison of Biota and Thuja. Proc. Pennsy Acad. Sci. 24: 65-112. 1950. [Platycladus (Biota) occidentalis and Thuja ses compared. | Owens, J. N.. & M. MoLperR. Sexual a aa in western red cedar (Thuja plicata). Canad. Jour. Bot. 58: 1376-1393. 1980. PoLHEIM, F. Thuja gigantea hore sete —ein Haplont unter dem ie cael Biol. Rundsch. 6: 84-86. 1968. [An unusual haploid variant of 7. plicat VAARTAJA, O. ae ie in photoperiodism of trees with special a to Pinus resinosa and Thuja occidentalis. Canad. Jour. Bot. 40: 849-856. 1962. [7. occidentalis 1s unusual in having growth rate little affected by day length. Von RupLorr, E. Gas liquid chromatography of the terpenes. VI. The volatile oil of Thuja apt Donn. Phytochemistry 1: 195-202. 19 & M.S. L Populational eae in the leaf oil terpene composition of western red cedar, Thuja plicata. Canad. Jour. Bot. 57: 476-479. 1979. [Very little terpene variation among populations y J& . YEH. Chemosystematic study of Thuja plicata: multivariate analysis of leaf oil terpene composition. Biochem. Syst. Ecol. 16: 119-125. 1988. 1990] HART & PRICE, CUPRESSACEAE 307 [Discriminant analysis shows small differences between coastal and interior popu- lations. Wacker, G. L. Breeding systems and genetic variability in the population center and disjunct ranges of northern white cedar, Thuja occidentalis. (Abstract.) ASB Bull. 33: 49. 1986. [Substantial electrophoretic variability present within and among populations; both clonal and sexual reproduction occur within populations. ] & E. C. Ciesscu. The ecology of northern white cedar, Thuja occidentalis L., in its southern disjunct range. (Abstract.) ASB Bull. 31: 89. 1984. [New local stands in Tennessee, Kentucky, and Virginia found by using the distributions of other disjunct plant species.] Wo re, F. Annual rings of Thija eee in relation to climatic conditions and movement of sand. Bot. Gaz . 1932. Yen, F. C. Isozyme variation of es Sie (Cupressaceae) in British Columbia. Biochem. Syst. Ecol. 16: 373-377. 1988. [Low levels of variation seem to imply a Pleistocene genetic bottleneck. ] 4, Juniperus Linnaeus, Sp. Pl. 2: 1038. 1753; Gen. Pl. ed. 5. 461. 1754. Dioecious (rarely monoecious), evergreen, erect to prostrate shrubs, or py- ramidal to open-crowned trees. Bark reddish-brown, usually thin and scaly, falling off in longitudinal strips [rarely thick and broken into plates]. Wood fragrant, close grained, with heartwood brownish to reddish-brown. Branches spreading or upright; branchlets rounded to nearly quadrangular or triangular, grooved and somewhat flattened. Leaves aromatic, entire or minutely dentic- ulate; juvenile leaves in whorls of 3, linear-lanceolate to subulate, spreading; leaves of mature plants either also needlelike and in whorls of 3 (having the spreading portion linear-lanceolate, with rigidly pointed apex, eglandular, en- tire, abscising at the juncture with the stem [or retained on the decurrent leaf base], the abaxial surface concave and grooved, with | or 2 whitened stomatal bands) or in sect. SABINA mainly scalelike and decussately opposite (acute to acuminate or sometimes blunt tipped, closely appressed, imbricate, entire or minutely denticulate, glandular or eglandular, firmly attached to the decurrent base). Buds naked or covered with scalelike leaves, ovate to acute. Cones axillary or terminal on short branchlets from buds of the previous autumn. Pollen cones (microsporangiate strobili) solitary [or in clusters of 3 to 6 in J. drupacea], ovoid to oblong; microsporophylls decussate or ternate, (6 to) 10 to 20 per cone, ovate to peltate, entire to denticulate; microsporangia (2 or) 3 to 6 [rarely to 8] per sporophyll, globose, attached to the abaxial edge of the sporophyll; pollen with obscure germinal aperture. Ovulate cones ovoid to globose, maturing in 1, 2, or 3 years, subtended by several whorls of persistent scalelike bracts, with 3 to 8 [or 9] decussate or ternate fleshy scales, these alternating with or bearing on their inner surfaces | or 2 [rarely 3] erect ovules; mature ovulate cones berrylike, succulent or ultimately dry and fibrous, blue, blue-black [or reddish to brownish], resinous, with scales strongly fused and their suture lines seldom evident [rarely conspicuous], obscurely or conspic- uously umbonate. Seeds | to 3 [rarely to 14] per cone, ovoid, terete or angled, wingless, often grooved or pitted by pressure from resin vesicles in the cone; seed coat thick and bony. Cotyledons 2 [to 6]. Chromosome number 2” = 22 308 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 (sometimes 37 = 33 [4n = 44]). (Including Sabina Miller, Arceuthos feb & Kotschy.) LECTOTYPE SPECIES: Juniperus communis L.; see Britton, N. Trees, 107. 1908. (Classical Latin name for juniper.)— tone. RED CEDAR. The largest genus of Cupressaceae, with 50 or more species in North and Central America, Eurasia, and northern and eastern Africa, very widely dis- tributed from the arctic to the tains of the subtropics, and with one species, Juniperus procera Hochst., occurring on tropical mountains from 14°N to 12°S in eastern Africa (Florin, 1963). Junipers are often dominant plants in subdesert vegetation—for example, in the Great Basin area of the western United States. Of the 13 species native to the United States (Zanoni), Juniperus Ashei Buchh., J. communis L., and J. virginiana L. are native in the Southeast. Three sections are usually recognized in the genus, with sects. JUNIPERUS and SABINA (Miller) Spach represented by native species in our area. Section CARYOCEDRUS Endl. contains only a single species, J. drupacea Labill., native to Greece and Asia Minor and cultivated as an ornamental in Europe and the United States. This species, occasionally segregated in the monotypic genus Arceuthos Antoine & Kotschy, has very large (ca. 2-2.5 cm in diameter) and relatively woody ovulate cones with the coats of the three seeds connate, as well as clustered pollen cones, but is otherwise quite similar to sect. JUNIPERUS, having all leaves ternate and needlelike and the cones axillary Gaussen’s (1967, 1968) informal classification treats the three sections as subgenera and divides them into ten sections (which were not formally de- scribed), largely on the bases of geography, cone color, number of seeds per cone, and presence or absence of teeth on the leaf margin. The genus has also been divided into eight informal species groups, generally following the ‘‘sec- tions” of Gaussen, by Rushforth. Given the probability of repeated convergence for cone color and seed number, it is likely that several of the species groups in sect. SABINA are artificial, and no subsectional groups are recognized in Zanont’s treatment of the North and Central American species. Section JUNIPERUS (sect. Oxycedrus Spach) (foliage leaves needlelike, ternate, eglandular, with blade jointed to the leaf base and abscising at the juncture with the stem; winter buds evident; cones solitary, axillary, with microspo- rophylls and cone scales ternate; ovulate cones one- to three-seeded, ripening in two or three years) comprises cight species (Dallimore & Jackson: Rushforth), primarily from northern or mountainous areas of North America, Eurasia. northern Africa, the Canary Islands, and the Azores. The only North American species is the very widespread Juniperus communis L. (J. nana Willd., illeg.; J. svbirica Burgsd.), common juniper, ground juniper, mountain juniper, 2n = 22, a prostrate or spreading shrub or sometimes a columnar to irregularly branched tree with a single white stomatal band generally wider than the adaxial leaf margins. It is native to much of northern North America, occurring south in the higher mountains to North and South Carolina and Georgia in the east and New Mexico and California in the west, and from coastal Greenland across Europe and northern and central Asia. Franco divided the species into four subspecies on the basis of habit and leaf morphology; he referred the eastern North American forms to subsp. depressa (Pursh) Franco (var. depressa Pursh) 1990] HART & PRICE, CUPRESSACEAE 309 because the stomatal band is narrower than the leaf margin. In our region the species is found primarily on barren rocky slopes and is a matted, prostrate shrub or very rarely a dwarfed tree (Coker & Totten). Section SABINA Spach (adult leaves mostly scalelike and decussately opposite for sometimes all ternate and needlelike], often with an abaxial gland, with the leaf base clearly decurrent on the stem and not jointed and abscising at the stem juncture; winter buds indistinct; cones terminal on elongating branchlets, usually solitary; ovulate cones maturing in one [or two] years, usually one- or two- [more rarely to 14-]seeded) comprises 40 or more species widely distrib- uted in North and Central America and the Caribbean, southern and central Europe to eastern Asia, and northern and east-central Africa. The North Amer- ican species have often been divided into entire- and denticulate-leaved groups (Gaussen, 1967, 1968; Hall, 1952c,; Rushforth; Zanoni & Adams, 1976). The majority of them are denticulate leaved, while Juniperus horizontalis Moench, J. scopulorum Sarg., J. virginiana, the Mexican J. Blancoi Martinez, and several Caribbean species have entire scale leaves and may form a natural group (Zanoni & Adams, 1976). The denticulate-leaved species in North America tend to occur in more xeric habitats (Hall, 1952c). Juniperus virginiana L. (Sabina lg (L.) Antoine), eastern red cedar, red cedar, savin, 2” = 22 (rarely 3n = 33, Stiff), is named in allusion to its red heartwood. It is very widely distributed in the eastern half of the United States and southeastern Canada. It is a tree reaching 10-15(-30) m in height, with the trunk occasionally up to | m in diameter. In the Atlantic States the species often occurs on dry, gravelly slopes and rocky ridges, especially on calcareous soils. In Kentucky, Tennessee, northern Alabama, and Mississippi it covers great areas of rolling limestone hills, forming nearly pure stands of small bushy trees. The “cedar glades” in the Nashville Basin of Tennessee, noted for their unusual flora (Quarterman), are dominated by the species. Eastern red cedar is also often found in abandoned fields and along fence rows. In coastal areas of the eastern Gulf States, it often grows in deep swamps (where it tends to become a large tree), as well as on coastal sands. In southwestern Texas, Ar- kansas, and Louisiana it attains its largest size on rich alluvial bottomlands. The populations of the southeastern Coastal Plain (from eastern North Car- olina west to southeastern Texas) have often been treated as a separate species, Juniperus silicicola (Small) Bailey (Sabina silicicola Small; Juniperus barba- densis auct., non L.), but detailed comparisons of morphology and terpenoid chemistry by Adams (1986) indicate that they are better treated as var. silicicola (Small) Silba. Varietas silicicola has been distinguished as having shorter scale leaves, longer pollen cones, smaller ovulate cones, and more slender twigs, but there is considerable overlap in these characters, and multivariate comparisons fail to separate the geographic groups cleanly (Adams, 1986). Multivariate comparison of terpenoid profiles gives discrete but closely adjoining inland and coastal groups. Adams noted that coastal populations tend to have cin- namon-colored rather than brownish bark and a rounded rather than pyramidal crown, On the basis of both morphology and chemistry, populations from Texas and Louisiana, previously mapped as the coastal form by Little (1971), appear to fit into var. virginiana better than into var. silicicola. JOURNAL OF THE ARNOLD ARBORETUM yy, Coc - NN, 459 FiGure 2. Juniperus. a-j, J. virginiana: a, branchlets with only scale leaves, bearing inature ovulate cones, x ¥4: b, branchlet with scale and needle leaves, x 34; c, detail of branchlet with needle ae showing decurrent leaf bases, x 5: d, microsporangiate strobilus before shedding of pollen, subtended by numerous scale leaves, x 5; e, mi- crosporophyll (abaxial view), showing dehisced Tn x 10; f, branchlet with ovulate cone near time of pollination, x 7; g, cone scale (adaxial view) with 2 erect ovules near time of pollination, x 10; h, mature ovulate cone a fused cone scales, x 3; 1, cross section of mature cone, only 2 seeds maturing—note resin vesicles outside seeds, x 3; j, seed, showing pits and ridges, x 5. k-q, /. communis: k, branch, showing ternate leaves and axillary ovulate cones, x ¥4: 1, detail of abscised portion of leaf in adaxial view, showing broad stomatal band, x 5; m, microsporangiate strobilus after shedding of pollen, x 5; n, microsporophyll (abaxial view), x 10; 0, short axillary shoot with young 1990] HART & PRICE, CUPRESSACEAE 311 The geographic distribution of Juniperus virginiana adjoins or partially over- laps those of several other entire-leaved species: J. horizontalis to the north, J. scopulorum to the west, and a taxonomically complex Caribbean group of species (Adams, 1983a; Adams & Hogge; Adams, Jarvis, Slane, & Zanoni) to the south. Hybridization with J. horizontalis and J. scopulorum has been sug- gested on the basis of morphological variation patterns (Fassett, 1944b, 1945b; Hall, 1952c; Schurtz). Terpenoid profiles are suggestive of past introgression from J. scopulorum into J. virginiana (Adams, 1983b; Comer et al.; Flake, Urbatsch, & Turner). Hybridization of J. virginiana and J. horizontalis on the edge of the Driftless Area in Wisconsin has been well documented by multi- variate analyses of morphology and terpenoid chemistry, as well as electro- phoretic-banding patterns of peroxidase enzymes (Palma-Otal ef a/.) Juniperus virginiana is most similar in morphology to J. scopulorum and the Caribbean species complex. Juniperus scopulorum tends to differ from J. vir- giniana and the Caribbean species in having its cones mature in the second rather than the first year. Morton noted some variation for this character in J. scopulorum and suggested that it be treated as a variety of J. virginiana, but this view has not been accepted by most later authors. Other characters that have been used to separate these species (Fassett, 1944a, b; Hall, 1952c; Rehder, 1940) include the degree of overlap of the mature scale leaves (much greater in J. virginiana), the width of the leaf epidermal cells (greater in J. scopulorum), the shape of the leaf glands (more elongate in J. scopulorum), and the number of sporophylls per pollen cone (lower in J. scopu/orum), although the variation patterns are complex in both and comparisons based on wide sampling are needed. The terpenoid distribution in the two taxa is distinctly bimodal and provides support for their treatment as separate species (Adams, 1983b; Comer et al., Flake, Urbatsch, & Turner), partially intergrading in the northern and southern Great Plains. As many as six species closely related to Juniperus virginiana have been recognized from the islands of the Caribbean (see discussions in Adams, 1983a, 1986; Adams & Hogge; Adams, Jarvis, Slane, & Zanoni), although some au- thors (Dallimore & Jackson; Silba) have placed most of these taxa in synonymy under J. barbadensis L. Juniperus bermudiana L., a rare species endemic to Bermuda, differs from J. barbadensis and J. virginiana in its stouter ultimate branches, averaging 1.5 mm or more in width. Preliminary studies (Adams, 1983a; Adams & Hogge) have indicated significant differences in terpenoid distribution among some of the Caribbean taxa and between the Caribbean taxa and the morphologically similar J. virginiana var. silicicola. A detailed monographic treatment of this group 1s much needed. Juniperus Ashei Buchh. (J. mexicana Spreng., nom. illeg.), Ashe juniper, rock cedar, mountain cedar, 2” = 22, occurs on upland limestone or dolomite outcrops in the Ozark Mountains of northwestern Arkansas, southwestern Mis- ovulate cone at apex, showing 3 ovules near time of pollination, x 10; p, portion of branchlet with mature ovulate cone—note remnant leaf bases fused to larger stem, x 3; q, apical view of ovulate cone, showing suture lines between 3 fused cone scales, x 3. Ww — JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 sourl, and adjacent Oklahoma, in the Arbuckle Mountains of southern Okla- homa, in extreme southwestern Arkansas, and more broadly in west-central Texas (where it often forms dense stands on the Edwards Plateau) and north- eastern Mexico. It is a small, bushy tree differing from J. virginiana in its more irregular or rounded branching habit, its minutely serrulate rather than entire leaves, and its typically rounded rather than elongate leaf glands. There have been a number of reports of hybridization or morphological intergradation between J. Ashei and J. virginiana (e.g., Hall, 1952a, c; 1955), but these have not been substantiated by subsequent comparison of terpenoid profiles (Adams, 1975a, 1977, Adams & Turner; Flake, Von Rudloff, & Turner). The disjunct northern populations of J. Ashei, well within the distribution of J. virginiana, are very similar in terpenoid profiles to populations of J. Ashei in the main portion of its range (Adams, 1975a). Juniperus Ashei is apparently most similar to the Mexican species J. sa/tillensis Hall in both morphology and terpenoid patterns (Adams, 1975a, 1977; Zanoni & Adams, 1976). The terpenoids of J. Ashei appear to represent a subset of those in the latter species, which may be its progenitor (Adams, Von Rudloff, Zanoni, & Hogge, 1980). Juniperus 18 a very distinct genus in the Cupressaceae on the basis of its unique fleshy cones with fused cone scales: it has sometimes been placed in a monogeneric tribe or subfamily. Its affinities to other genera are not well es- tablished on the basis of morphology, but the genus is similar to Cupressus in its distribution of tropolones (H. Erdtman & Norin) and biflavonoids (Gadek & Quinn, 1985 Juniperus has been the subject ofan increasing number of detailed taxonomic studies using multivariate analyses of terpenoid chemistry, as well as mor- phology (see, for example, Adams, 1983a, b, 1986; Adams & Hogge; Adams, Von Rudloff, Hogge, & Zanoni; Adams, Von Rudloff, & Hogge; Zanoni & Adams, 1975, 1976), and preliminary revisions have been presented for the Mexican and Guatemalan (Zanoni & Adams, 1979) and North American (Za- noni) taxa of sect. SABINA. Species delimitation is often very difficult in the genus, and overall monographic treatment is much needed. Problems in dis- tinguishing species are partially due to the relatively cryptic characters of leaves, stems, cones, and seeds used to separate them and to our poor knowledge of certain species. To a large degree, however, they are due to the complexity of variation within species and the limited divergence or convergence among them. Fassett’s (1944a, b; 1945a, b) work on J/. virginiana, J. horizontalis, and J. scopulorum illustrates the problems in attempting to discriminate among closely related species by the use of univariate comparisons of morphological characters— problems that have been obviated in part by multivariate analysis and the use of independent chemical data sets. Fassett found that quantitative variation within individual populations can be substantial, and that characters yielding statistically significant differences among species often show a consid- erable degree of overlap. Zones of past or present eal in the areas where the species meet also add to the taxonomic complexit As in the rest of the family, Juniperus most frequently sac a chromosome number of 2n = 22. Counts have been obtained for at least 23 species (see especially Hall, Mukherjee, & Crowley, 1973, 1979; Khoshoo, 1961: Mehra: 1990] HART & PRICE, CUPRESSACEAE 313 Mehra & Khoshoo; Sax & Sax) of which 19 are reported to be diploid or preponderantly diploid. Triploid or tetraploid plants have been found predom- inantly in horticultural variants (e.g., in J. chinensis L. and J. squamata Buch.- Ham.); their frequency in wild populations is unclear. Some species differ in karyotype; e.g., one chromosome pair is markedly heterobrachial in J. hori- zontalis and J. procera, while all of the chromosomes are more or less isobrachial in J. communis, J. virginiana, and several other species (Mehra; Mehra & Khoshoo; Mujoo & Dhar; Ross & Duncan). As discussed by Lemoine-Sébastian (1968), evolution within the genus Ju- niperus has been marked by repeated reductions in the numbers of both cone- scale whorls and ovules, culminating in several species with a single apparently terminal ovule enveloped by one whorl of cone scales. Seeds with very hard coats contained in fleshy, “berrylike” cones are effective adaptations for seed dispersal by birds or sometimes mammals (Holthuijzen & Sharik, 1984, 1985; Phillips). As in a number of other groups of conifers, animal dispersal of seeds is coupled with a dioecious breeding system (Givnish). Several species of junipers (e.g., Juniperus virginiana) are hosts for cedar- apple rust (Gymnosporangium spp.), which produces conspicuous gall-like growths on the plant (illustrated in Coker & Totten) and is a serious pathogen of cultivated apples and other woody Rosaceae. The wood of Juniperus is fragrant, very durable, and little damaged by insects. Wood and bark of several North American species have been found to be very effective termiticides (Adams, McDaniel, & Carter). The wood of J. virginiana has been much used for pencils, although incense cedar (Ca/ocedrus decurrens (Torrey) Florin) is now much more widely employed for this purpose because of heavy exploitation of red cedar (Hemmerly). Red cedar wood has been widely used for fenceposts and furniture, especially moth-resistant cedar chests for storage of clothing. In the western United States junipers are also highly valued for their aromatic firewood. The essential oil (cedarwood oil) from the heart- wood of Juniperus species has been widely utilized in compounding fragrances for soaps, perfumes, and industrial uses, as well as in microscopy as a mountant (Adams, 1987; Hemmerly). Leaf or fruit oils of Juniperus have been used medicinally and can possess powerful diuretic properties. Many species of junipers are grown as ornamentals, and a large number of habit and color variants have been selected, particularly in J. chinensis, J. communis, and J. virginiana (Dallimore & Jackson; Kriissmann; Ouden & Boom). Juniper “ber- ries” from J. communis are used as flavoring agents in cooking and in the production of the alcoholic beverage gin (the name shortened from Dutch jenever, traceable back to Latin juniperus). REFERENCES: Under eae references see BAILEY; BEAN; pees 1908; Coker & TOTTEN; DAL- LIMORE & JACKSON; G. ERDTMAN; H. ERDTMAN & NorIn; FITSCHEN; FLoRIN, 1931, 1963; HEGNAUER, 1962, 1986; KHosHOoOo, 1961; KRUSSMANN; LITTLE, 1971, 1980; MEHRA & KHOSHOO; OUDEN & Boom; Owens & SIMPSON; REHDER, 1940, 1949; RUSHFORTH; SAR- GENT, 1896, 1926: SAx & SAx: SILBA: SINGH; UENO, 1960b; and VON RUDLOFF. 314 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 ADAMOVICH, E. I. Resin receptacles of Juniperus communis L. (In Russian.) Lesn. Zhur. 4: 159-161. 1961.* ApAMS, R. P. Seasonal variation of terpenoid constituents in natural populations of Juniperus Pinchotii Sudw. Phytochemistry 9: 397-402. 1970. Chemosystematic and numerical studies of natural populations in Juniperus Pine hotii Sudw. Taxon 21: 407-427. 1972. [/. Pinchotii and J. monosperma com- p . Reevaluation of the biological status of Juniperus Deppeana var. Sperryi Correll. Brittonia 25: 284-289. 1973. [Terpenoids of J. Deppeana, J. flaccida, and J. Pin- chotii.] Gene flow versus selection pressure and ancestral differentiation in the com- position of species; analysis of eae variation of Juniperus Ashei Buch. using terpenoid data. Jour. Molecular Evol. 5 5. 1975a. Numerical chemosystematic cies zr infraspecific variation in Juniperus Pin- chotii Biochem. Syst. Ecol. 3: 71-74. 1975b. Chemosystematics—analysis of populational differentiation and variability of ancestral and recent populations of Juniperus Ashei. Ann. Missouri Bot. Gard. 64: 184-209. 1977. —. Diurnal variation in the eae of Juniperus scopulorum (Cupressaceae) — summer versus winter. Am. Jour. Bot. 66: 986-988. 1979. [Diurnal and day-to-day variation within trees is lower in winter when plants are less active metabolically. ] The effects of gases from a burning coal seam on morphological and terpenoid characters in Juniperus scopulorum (Cupressaceae). Southwest. Nat. 27: 279-286. 1982. [A local columnar form of J. scopulorum is environmentally induced: hy- bridization with J. horizontalis in North Dakota is indicated by multivariate com- parisons . The junipers (Juniperus, Cupressaceae) of Hispaniola: comparisons with other Caribbean species and among collections from Hispaniola. Moscosoa 2: 77-89. 1983a Infraspecific terpenoid variation in Juniperus scopulorum: evidence for Pleis- tocene refugia and recolonization in western North America. Taxon 32: 30-46. 3b ———. Geographic variation in Juniperus silicicola and J. virginiana of the southeastern United States: multivariate analyses of morphology and terpenoids. /bid. 35: 61- 75. 1986. [/. silicicola is best treated as a variety of J. virginiana. Investigation of Juniperus species of the United States for new sources of ce- darwood oil. Econ. Bot. 41: 48-54. 1987. [Composition of volatile oils from the heartwood of ten species.] HAGERMAN. A comparison of the volatile oils of mature versus young leaves of Juniperus scopulorum: chemosystematic significance. Biochem. Syst. Ecol. 4: 75-79. 1976. [Hydrocarbon ee enoids show highest levels in younger leaves, oxygenated terpenoids in older Diurnal aaatee! in the volatile terpenoids of Juniperus scopulorum (Cupressaceae). Am. Jour. Bot. 64: 278-285. 1977 & L. HOGGE. ay ete: studies of the C aribbean j junipers based on their volatile oils. Biochem. Syst. Ecol. 11: 85-89. 1983. ——, C. E. Jarvis, V. SLANE, &T. A. ZANONI. Typification of Juniperus barbadensis . and J. bermudiana L. and rediscovery of J. barbadensis from St. Lucia, BWI (Cupressaceae). Taxon 36: 441-445. 1987. [Typification of J. barbadensis and J. bermudiana is clarified; these are apparently distinct species with narrow ranges in the Caribbean. ] C. A. McDaniel, & F. L. Carter. Termiticidal activities in the heartwood, bark/sapwood and leaves = eee species from the United States. Biochem. Syst. Ecol. 16: 453-456. 1988. 1990] HART & PRICE, CUPRESSACEAE 315 L. TURNER. Chemosystematic and numerical studies of natural populations of Juniperus es Buch. Taxon 19: 728-751. 1970. ; FF, & L. HoGce. Chemosystematic studies of the western North American ee based on their volatile oils. Biochem. Syst. Ecol. 11: 189-193. 1983. ———_.,, ———__, ———., & T. A. ZANonI. The volatile constituents of Juniperus Blancoi and its affinities with other entire leaf margin junipers of North America. Jour. Nat. Prod. 44: 21-26. 1981. [J. Blancoi compared to J. horizontalis, J. scopulorum, and J. virginiana. | ,T. A. ZANONI, & L. Hocce. The terpenoids of an ancestral/advanced species pair of ae Biochem. Syst. Ecol. 8: 35-37. 1980. [J. Ashei and J. saltillensis have similar profiles, with a more restricted set of compounds in the former. ] ———,, ———, & ———.. The south-western USA and northern Mexico one- seeded junipers: their volatile oils and evolution. Jbid. 9: 93-96. 1981. [J. erythro- arpa, J. monosperma, J. Pinchotii.] henner? M., & M. CREss. Changes within the seeds of Juniperus Hee during the processes of after-ripening and germination. Jour. Forestry -801. 1942. AGRAMONT, F., R. BUSKING, J. MITCHELL, & E. ENZINGER. The red pee Missouri Bot. Gard. Bull. 36: a 92. 1948. Wanation of J. virginiana in the St. Louis area.] ALEKSANDROVSKY, E. S. Biology of blooming and fruiting of Juniperus turcomanica Fedtsch. (In Russian; English summary.) Lesovedenie 1972(3): 76-84. 1972 Anpre, D. Contribution 4 l'étude morphologique du céne femelle de quelques gym- nospermes (Cephalotaxacées, Juniperoidées, Taxacées). Nat. Monspel. Bot. 8: 3-35. 1956. ANTOINE, F. Die Cupressineen-Gattungen: 4rceuthos, Juniperus and Sabina. 71 pp. Vienna. 1857.* AREND, J. L. An early eastern red cedar plantation in Arkansas. Jour. Forestry 45: 358- 360. 1947. LJ. virginiana. ] Ayaz, M. Anatomy of juniper (Juniperus excelsa) seed. Pakistan Jour. Forestry 30: 99- Bertscu, A. Untersuchungen an rezenten und fossilen Pollen von Juniperus. Flora 150: 503-513. pl. 72. 1961. Biross, C. G. The water conductivity and growth habits of Juniperus horizontalis Moench and Juniperus virginiana L. Ecology 28: 281-289. 1947. [No appreciable differences between species in water conductivity of the stem, but stem growth of J. virginiana is much higher under favorable conditions. ] Boyp, H. Eastern red cedar —— us virginiana ae a list of references. U. S. Dep. Agr. Library, Louisiana Branch, N Le eans. 5 pp. 19 BUCHHOLZ, J. T. The Ozark white pee t. Gaz. ne 326 332. 1930. [Juniperus Ashei described and distinguished from J. monosperma CHATURVEDI, M. Studies on the pollen grains of Juniperus L. Curr. Sci. Bangalore 50: 548, 549. 1981.* Cuesnoy, L. Sur le développement des cones femelles du Juniperus communis L., dans : région parisienne, de la oo a la graine. Compt. Rend. Acad. Sci. Paris, D. 262: 2018-2021. pl. J. 1966. . Nature et évolution a formations dites ‘“‘astéroides” de la cellule centrale de l'archégone du Juniperus communis L. Etude en microscopie photonique et élec- tronique. /bid. 264: 1016-1019. pls. 1-6. 1967. CiaMPI, C. Processi post-fecondativi nel genere Juniperus. Osservazioni in Juniperus oe L. e Juniperus macrocarpa S. 8. Caryologia 11: 334-347. pls. 17-20. 1959 — COCKER ee R. Junipers of the Mediterranean area. Arb. Bull. 20: 15, 16, 35. CoLumnGwoon, G. H. Eastern red cedar. Am. Forests 44: 30, 31. 1938. [/. ee 316 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Comer, C. W., R. P. ADAMs, & D. F. VAN HAVERBEKE. Intra- and interspecific variation of Pek virginiana and J. a aa seedlings based on volatile oil compo- sition. Biochem. Syst. Ecol. 10: 297-306. 1982. [Separate species status and probable past hybridization indicated by multiv ane comparisons. Cuccut, C. C. Indagine geobotanica sul ginepri europel. Delpinoa 11: 171-222. pls. J- 7. 1958. [Morphology and distribution of J. communis and J. Sabina; key to Eu- ropean junipers, DoGRA , & §. TANDON. Observations on the embryology of Juniperus procera Hochst. Glimpses Pl. Res. 6: 114-124. figs. ]-53. 1984. [Early embryogeny compared to that of other species of Juniperus. | DunHoux, E. Evolution structurale de la paroi du grain de pollen du Juniperus communis L. (Cupressacées), cultivé in vitro, au cours de . aes d’hydration. Compt. Ren Acad. Sci. Paris, D. 274: 2767-2770. pls. I-3. Les déplacements du noyau végétatif ne : tube pollinique du Juniperus communis L. (Cupressacées) cultivé in vitro. Rev. Cytol. Biol. Veg. 4: 311-330. 1981.* ENGELMANN, G. The American junipers of the section Sabina. Trans. St. Louis Acad. Sci. 3: 583-592. 1877. [Nine species recognized. ] Evans, G. E., & H. P. RASMUSSEN. Chromosome counts in three cultivars of Juniperus L. Bot. Gaz. 132: 259-262. 1971. LJ. horizontalis, J. Sabina, 2n = 22; J. chinensis, 3n = 33.] Fassett, N.C. The validity of Juniperus virginiana var. crebra. Am. Jour. Bot. 30: 469- 477. 1943. [The northern var. crebra is generally more narrow crowned than var. virginiana, degree of acuteness of the leaves and degree of seed pitting show little consistent difference. ] . Juniperus virginiana, J. horizontalis, and J. scopulorum. 1. The specific char- acters. Bull. Torrey Bot. Club 71: 410-418. 1944a. I]. Hybrid swarms of /. virginiana and J. scopulorum. Ibid. 475-483. 1944b. [Individuals with intermediate character states found in areas of geographic overlap.] III. Possible hybridization of J. hori- zontalis and J. scopulorum. Ibid. 72: 42-46. 1945a. [Where the two species occur together, J. scopulorum tends to approach J. horizontalis in several characters; a depressed form apparently intermediate between the two species is described as J. scopulorum var. patens Fassett.] IV. Hybrid swarms of J. virginiana and J. hori- zontalis. Ibid. 379-384. 1945b. [Where the ranges of the two species overlap, in- termediates (described as J. Ya var. ambigens Fassett) are found.] V. Taxo- nomic treatment. /bid. 480-482. FerGuson, E. R. Eastern red cedar: an ae bibliography. U. S. Dep. Agr. Forest Serv. Res. Pap. SO-64. 21 pp. 1970. FLAKE, R. H., L. UrBAtscH, & B. L. TURNER. Chemical documentation of allopatric introgression into Juniperus. Syst. Bot. 3: 129-144. 1978. [Terpenoid profiles suggest introgression from J. scopulorum to J. virginiana.) . E. Von Ruptorr, & B. L. TURNER. Quantitative study of clinal variation in Juniperus virginiana using terpenoid data. Proc. Natl. Acad. Sci. U.S. A. 64: 487- 494. 1969. [Clinal variation a northeast to southwest reported, with no evidence of hybridization with J. Ashe & 7S eon of a clinal pattern of chemical differentiation in “Juniperus virginiana from terpenoid data obtained in successive years. Recent Adv. Phytochem. 6: 215-228. 1973.* Franco, J. po A. Taxonomy of the common juniper. Bol. ae Brot. 32: 101-120. 1962. [Four geographic subspecies of J. communis recogniz Fretz, T. A. Effect of photoperiod and nitrogen on the cae of foliar monoter- penes Soe horizontalis Moench cv. P/lumosa. Jour. Am. Soc. Hort. Sci. 101: 611-613. 1976. [Increased photoperiod altered the relative concentration of several terpene pee ee 1990] HART & PRICE, CUPRESSACEAE aif GALL, E. C. Juniperus at the Morris Arboretum. Morris Arb. Bull. 13: 3-10. 1962. [Keys and notes for 13 Species illustrations. | GAUSSEN, H. La classifica (Juniperus). Compt. Rend. Acad. Sci. Paris, D. 265: 954-957. ye aaa ae H. J. Juniperus communis L. (In Dutch.) Dendroflora 5: 29-34. 1968.* Groves, G. R. The Bermuda cedar. Unasylva 9: 169-172. 1955. [Utilization of J. ten em ] HALL, M. T. A hybrid swarm in Juniperus. Evolution 6: 347-366. 1952a. [/. Ashei and J. virginiana in the cee ckle sili of Oklahoma; see also later chemical com- parisons (e.g., FLAKE, VON RUDLOFF, & TURNER), which do not support the hy- gees of eau hy ae ] ation in native junipers. Cranbrook Inst. Sci. ee Letter 21: 62-64. 1952b. Variation and hybridization in Juniperus. Ann. url Bot. Gard. 39: 1-64. 1952c. {Morphological evidence for purported saan between J. Ashei and J. virginiana. . Nomenclatural notes concerning Juniperus. Rhodora 56: 169-177. 1954. [Com- parison of J. Ashei and J. monticola. . Comparison of juniper populations on an Ozark glade ne old fields. Ann. — Bot. Gard. 42: 171-194. 1955. [Variation in J. virginia w your cultivated junipers. Arb. Bull. 20: 10-12. 1957. pee new species of Juniperus from Mexico. Fieldiana Bot. 34: 45-53. 1971. [/. salillensis, ] &C. J. CARR. ee of Juniperus in the Palo Duro Canyon of western Texas. Southwest. Nat. 13: 7 1968. _J.F. McCormick, & G.G. Foca: Hybridization between J. Asheiand J. Pinchotii in southwestern Texas. Butler Univ. Bot. Stud. 14: 9-28. 1962.* . MUKHERJEE, & W. R. CROWLEY. Chromosome counts in cultivated junipers. Jour. Arnold Arb. 54: 369-376. 1973. [Diploid counts (all 2” = 22) obtained for seven species, including J. communis and J. virginiana; tetraploid counts for two varieties of /. Fee and one of J. squamata. | “hromosome numbers of cultivated junipers. Bot. Gaz. 140: 364- 370. 1979, [Counts for ten species, including J. communis (2n = 22) and J. virginiana (2n = 22, 3n = 33), with the triploid cultivar intermediate to J. chinensis in morphology HARPER, R. M. The diverse habitats of the eastern red cedar and their interpretation. mae 12: 145-154. 1912. L/. virginiana largely absent from fire-controlled habi- Ere T. E. Economic uses of eastern red cedar. Econ. Bot. 24: 39-41. 1970. [/. virginiana. | Ho.THulyZzeENn, A.M.A., & T. L. SHARIK. Seed longevity and mechanisms of regeneration of eastern red cedar (Juniperus virginiana). Bull. Torrey Bot. Club 111: 153-158. 1984 & ———. The avian dispersal e! stem of eastern red cedar (Juniperus virginiana). Canad. Jour. Bot. 63: 1508-1515. 1985. IRVING, R. S. A chromosome count . Juniperus Ashei (Cupressaceae) and additional numbers for Hedeoma (Labiatae). Sida 8: 312, 313. 1980. [2m = 22 for J. Ashei.] Jack, J. G. Fructification of Juniperus. Bot. Gaz. 18: 369-375. pl. 33. 1893. [Ovulate cones of J. virginiana mature in one year, of J. horizontalis in two years, of J. communis in three years JoHNsoNn, T. N. Longevity of stored juniper seeds. Ecology 40: 487, 488. 1959. [Seeds of J. Deppeana, J. monosperma, and J. osteosperma can survive long periods in dry storage KAEISER, M. Microscopic anatomy of the wood of three species of Juniperus. Trans. Illinois Acad. Sci. 43: 46-50. 1950 318 JOURNAL OF THE ARNOLD ARBORETUM [voL. 7] —. Microstructure of the wood of Juniperus. Bot. Gaz. 115: 155-162. 1953. [Most species of sect. SABINA have large intercellular spaces between the tracheids; these are absent in the other sections.] KELLEY, W. A., & R. P. AbAms. Seasonal variation of isozymes in Juniperus scopulorum: systematic significance. Am. Jour. Bot. 64: 1092-1096. 1977. [Some bands are uae ce seasonally. is of 1 rete variation in natural populations of Juniperus Ashei. Rho- dora $0: 107- 13 78. Ken, S. Cedar species: their geographical distribution and uses. Am. Perfumer Essent. Oil Rev. 51: 137-140. 1948.* [Juniperus.] Kiem, W. M. Cotyledon variations in Juniperus occidentalis Hook. Aliso 4: 129. 1958. [Geographic differences in number of cotyledons.] LeBRETON, P., & S. THIVEND. Sur une sous-espéce du genévrier de Phénicie, Juniperus phoenicea L., définie a partir de critéres biochimiques. Nat. Monspel. Bot. 47: 1- 12. 1981. [Two subspecies are separated on the basis of high vs. low levels of prodelphinidin. ] LEMOINE-SEBASTIAN, C. Anatomie ae plantules de quelques Juniperus. Trav. Lab. For- est. Toulouse, Tome I, Vol. VI, Art. XXI: 1-12. p/. 5. 1964. . Appareil reproducteur a des Juniperus. Ibid., Art. XXIV: 1-35. 1967. [Pol- len-cone morphology, including that Be a coma ae . hese na. a L’inflorescence femelle des Junipere e, phylogenése. [bid., Tome I, Vol. VII, Art. V: 1-460. pis. 229-247, 1968. (Detailed asda afoul ate- cone development i in 21 species, including J. communis and J. virginiana.] MarTiNEz, M. Los Juniperus mexicanos. Anal. Inst. Biol. site . 17: 1-121. 1946. [A new classification of the Mexican taxa of sect. Cae pro J Massey, A. B. Ecology of the red cedar, Juniperus virginiana. (Abstract.) ASB Bull. 11: 261. 1954. Masters, M. T. Bermuda juniper and its allies. Jour. Bot. (London) 37: 1-11. 1899. MatTHews, A. C. The morphological and cytological development of the sporophylls and seed of Juniperus virginiana L. Jour. Elisha Mitchell Sci. Soc. 55: 7-62. pls. I- 9, 1939, McCormick, J. F., & R. B. PLarr. Ecotypic differentiation in southeastern Juniperus. (Abstract.) ASB Bull. 9: 30. 1962. [Purported introgression between J. Ashei and J/. virginiana discuss Menra, P. N. Co nifers of the Himalayas with particular reference to the Abies and Juniperus complexes. Nucleus 19: 123-139. 1976. [Morphology and distribution of nine species of Juniperus, new chrom counts for J. Fargesii, J. pseudosabina both 2n = 22), and J. Wallichiana (4n = MINCKLER, L.S., & R. A. RyKeER. Color, form, ie growth variations in eastern redcedar. Jour. Forestry 57: 347-349. 1959. [Common garden ts with J. virginiana. | Morton, C. V. Notes on Juniperus. Rhodora 43: 344-348. 1941. Mujoo, S., & G. L. Doar. Cytology of Juniperus communis var. nana Syme. Nucleus 24: 46-48. 1982. [Chromosome number 2” = 22; all chromosomes with median or submedian centromer NicHois, G. E. A morphological study of Juniperus communis var. depressa. Beih, Bot. Centralbl. 25: 201-241. 1910. OrrLtey, A. M. The development of the gametophytes and fertilization in Juniperus communis and Juniperus virginiana. Bot. Gaz . 1909. Pack, D. A. After-ripening and germination of Juniperus seeds. Bot. Gaz. 71: 32-60. 1921. LJ. communis, J. virginiana. | PALMA-OTAL, M., W. S. Moore, R. P. ApAms, & G. R. JoswiAk. Morphological, chem- ical, and biogeographical analyses of a hybrid zone involving Juniperus virginiana and J. horizontalis in Wisconsin. Canad. Jour. Bot. 61: 2733-2746. 1983. [Multi- 1990] HART & PRICE, CUPRESSACEAE 319 variate analyses of morphology and terpene chemistry and peroxidase banding pat- terns used to document hybridization. ] Puts, F. J. The dissemination of junipers by birds. Forestry Quart. 8: 60-73. 1910. Powe LL, R. A., & R. P. ADAMS. Seasonal variation in the volatile oo of Juniperus scopulorum (Cupressaceae). Am. Jour. Bot. 60: 1041-1050. 1973. QUARTERMAN, E. Major plant communities of Tennessee cedar area Ecology 31: 234- 54. 1950. Ross, J. G., & R. E. DuNcAN. Cytological evidences of hybridization between Juniperus virginiana and J. horizontalis. Bull. Torrey Bot. Club 76: 414-429. 1949. [Karyotypic variation and meiotic irregularities in areas of apparent hybridization. ] SAN FELICIANO, i M. MEDARDE, J. L. Lopez, J. M. MIGUEL DEL CORRAL, P. PUEBLA & RERO. Terpenoids from leaves of Juniperus thurifera. Phytochemistry 27: 3341-2248. 1988. [Diverse diterpenoids of the labdane, pimarane, and abietane types were isolated from the leaves of J. thurifera. SARGENT, C. S. Notes on cultivated conifers. 1V. Garden Forest 10: 420, 421. 1897. [J/. scopulorum and J. virginiana distinguished. ] Scuurtz, R. H. A taxonomic analysis of a triparental hybrid swarm in Juniperus L. 90 pp. Unpubl. Ph.D. dissertation, University of Nebraska, Lincoln. 1971.* [J. hort- zontalis, J. scopulorum, J. virginiana. ] SEBASTIAN, C. Essais de germination de quatre espéces du genre Juniperus. Bull. Soc. Sci. Nat. Phys. Maroc 38: 115-122. 1958a. . Sur la haan de Juniperus phoenicea. Ibid. 187-194. 1958b. StirF, M. L. A naturally occurring aoe (Abstract.) Virginia Jour. Sci. II. 2: 317. 1951. i as from V TATRO, V. E., R. W. Scora, F. C. VASEK, “& J. AMOTO. Variation in the leaf oils of three species of en Am. Jour. ae 60: 236-241. 1973. [¥. californica, J. occidentalis, J. osteosperma. VAN HAVERBEKE, D. F. A population analysis of Juniperus in the Missouri River Basin. Univ. Nebraska Studies 38: 1-82. 1968.* [Variation of J. scopulorum and J. virgin- iana in the northern Great Plains VAN MELLE, P. J. Juniperus texensis sp. a Texas juniper in relation to J. ee ee J. Ashei et al. Pilon. i: oe 35. 1952. cism in Juniperus meee Tbhid. os 1953. VASEK, F. C., R. W. Scora. Analysis of the oils of western North American junipers by gas-liquid chromatography a Jour. Bot. 54: 781-789. 1967. [/. oe J. deppeana, J. monosperma, J. occidentalis, J. rims J. scopuloru Von Rup torr, E. Gas-liquid chromatography of terpenes. X VI. The Sra oil of the leaves of Juniperus . pare Buchholz. Canad. Jour. ao 46: 679-683. 1968. F. M. CoucHMAN. The volatile a leaves of Juniperus scopulorum Sarg. Canad. Jour. Chem. 42: 1890-1895. ZANONI, T. A. The American junipers ar en Sabina (Juniperus, Cupressaceae) — a century later. Phytologia 38: 433-455. 1978. [Synoptic taxonomic review of the group.] & R. P. ApAms. The genus ieee oe amend in Mexico and Guatemala: numerical and morphological analysis. Bol. Soc. Bot. México 35: 69-92, 1975. & genus Juniperus (Cc ‘upressaceae) in Mexico and Guatemala: merical and chemosystematic analysis. Biochem. Syst. Ecol. 4: 147-158. 1976. [Species relationships based on terpenoid profiles.] & The genus Juniperus (Cupressaceae) in Mexico and Guatemala: syn- nymy, key, and distributions of the taxa. Bol. Soc. Bot. México 38: 83-121. 1979. Zvi H. E. A novel juniper tree. Am. Bot. 23: 130. 1917. [A prostrate juniper a. 0.5 m high and 40 m wide. ] sy 320 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 71 5. Callitris Ventenat, Dec. Gen. Nov. 10. 1808. Evergreen monoecious trees or shrubs with spreading or erect branches. Bark grayish, furrowed [brownish and fibrous in C. Macleayana]. Wood very dense, not distinctively colored. Juvenile leaves needlelike, in whorls of four; adult leaves scalelike, triangular, in whorls of 3, with bases decurrent and fused to the stem. Pollen cones solitary or clustered near branchlet tips; microsporo- phylls in whorls of 3 [rarely 4], each with (2 or) 3 (or 4) sporangia. Ovulate cones globose [to ovoid or conical], terminal on short, thickened stalks, bearing 6 [sometimes 8 in C. Macleayana] cone scales in 2 alternating and unequal whorls of 3, appearing to form a single whorl at maturity; cone scales thickened, triangular-ovate, valvate and opening out from the very reduced cone axis (columella); ovules [6 tO} 18 to 36 [to 54] per cone, in 3 intersecting rows arranged around the lla and at the base of the cone scales. Seeds flattened, irregularly tetrahedral, bearing [1 or] 2 or 3 lateral wings; cotyledons 2 (rarely 3). Chromosome number 2 = 22. (Including Octoclinis F. Mueller; Frenela Mirbel.) Lecrotype species: Callitris rhomboidea (R. Br.) A. & L. Rich.; see Bullock, Taxon 6: 227. 1957. (From Greek ka//istos, beautiful, and treis, three, in reference to the arrangement of leaves and cone scales.)—CYPRESS PINE. A genus of approximately 15 species, 13 in Australia and two in New Cal- edonia. Ca/litris 1s divided into sect. Ocroc.inis Bentham, including only C. Macleayana (F. Mueller) F. Mueller, and sect. CALLitTRis (sect. Hexaclinis Bentham). Ca/litris Macleayana 1s unusual in having some ovulate cones (those on shoots retaining the juvenile leaf type) with eight rather than s1x scales, both numbers often occurring on the same tree; loosely fibrous rather than dense, vertically furrowed bark; and only one elongate wing on the seed rather than two or three smaller ones (Baker & Smith; Clifford & Constantine; Garden). Callitris columellaris F. Mueller var. campestris Silba (C. glaucophylla J. Thompson & L. Johnson; C. glauca R. Br. ex Baker & Smith, nom. illegit.), white cypress pine, 27 = 22, has escaped from cultivation and has become locally naturalized in sand-pine (Pinus clausa) scrub in Brevard, Indian River, Orange, Osceola, and Seminole counties in eastern Florida (Judd, pers. comm.; Little, 1979; Wunderlin, pers. comm.). It 1s characterized by usually glaucous, unkeeled leaves and generally solitary ovulate cones with rugose but not ver- rucate scales that separate to near the base at maturity The nomenclature and species circumscription in the Callitris columellaris complex has been controversial. Franco chose a specimen referable to C. col- umellaris var. columellaris as a lectotype for C. Hugelii, a new combination based on Frenela Hugelii Carr., which would be an earlier name for the species. Blake has contended, however, that the specimen chosen was an inappropriate neotype probably representing a different species from that described in the protologue. Several authors (Baker & Smith: Garden; Lacey; Thompson; Thompson & Johnson) have recognized three ecogeographic species in the Callitris colu- mellaris complex: C. glaucophylla (C. glauca) in inland areas of the southern two thirds of Australia, C. intratropica Baker & Smith in the tropical zone of northern Australia, and C. columellaris s.s. in coastal areas of Queensland and 1990] HART & PRICE, CUPRESSACEAE 321 New South Wales. Other authors (Blake; Clifford & Constantine; Dallimore & Jackson; Venning, 1979, 1986) have emphasized morphological intergradation among these taxa, however, and they are treated here as varieties following Silba. There appear to be genetically based differences in plant habit and in foliage color and density among the three varieties (Lacey; Thompson & John- son). In particular, the inland var. campestris usually has more glaucous foliage than the other varieties (hence the name “‘white cypress pine’’), while var. columellaris has denser, dark green leaves and a more irregular branching pattern. Varieties campestris and columellaris are quite similar in cone size, while var. intratropica (Baker & Smith) Silba has been distinguished on the basis of its smaller cones (usually less than 1.8 cm wide) with narrower upper cone scales (Thompson & Johnson). It is uncertain whether plants from south- ern Queensland, between the generalized geographic ranges given for vars. intratropica and campestris by Thompson & Johnson, were included in the morphological comparisons (see Blake). Differences in leaf and wood chemistry among the taxa have been reported based on limited sampling (see Baker & Smith; Lacey; Thompson & Johnson), but thorough range-wide comparisons are needed to assess their validity. The widespread south Australian species Callitris Preissii Miq. can also approach C. columellaris in morphology and leaf-oil chemistry (Adams & Simmons), and the taxa have been reported to produce fertile hybrids (Thomp- son & Johnson). Chromosome counts for six species of Callitris, all 2n = 22, were given by Mehra & Khoshoo. Similar karyotypes are seen 1n these species, with the chromosomes having median or submedian centromeres. Natural hybridiza- tion has been reported for three of these taxa, C. Preissii, C. verrucosa (A. Cunn. ex Endl.) F. Mueller, and C. co/umel/laris var. campestris, in all possible combinations (Adams & Simmons; Garden; Thompson & Johnson). Morphologically, Ca/litris shows the greatest similarity to the western Aus- tralian genus Actinostrobus Miq., which differs most prominently in having its ovulate cones subtended by a number of closely imbricate bracts. These are lacking in Cal/litris. Both genera have decurrent leaves in whorls of three and ovulate cones of six basally fused cone scales; they also have similar biflavonoid profiles, lacking in cupressuflavone and hinokiflavone derivatives (Gadek & Quinn, 1985). The southern African Widdringtonia Endl., which is similar to Callitris and Actinostrobus in embryology, differs in having decussately opposite foliage leaves, only four scales per ovulate cone, and a much more complex biflavonoid profile. The sesquiterpene alcohol guaiol is a very characteristic component of the heartwood of Callitris, often crystallizing from cut stumps (Baker & Smith); within the Cupressaceae s.|. it has otherwise been reported only from the New Caledonian genus Neocallitropsis (H. Erdtman & Norin). Several species of Callitris are important timber trees in Australia, furnishing hard, durable, termite-resistant wood for construction. The bark is rich in tannin. Resin exuding from the inner bark of cut stumps is similar to sandarac (obtained from Tefraclinis) and has been used in the manufacture of varnishes and incense (Lacey). In Australia various species of Ca/litris are widely planted S22 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 as windbreaks and ornamental trees and are particularly valuable for their drought resistance. REFERENCES: Peed seg’) references see iy Cee ate DALLIMORE & JACKSON; DOYLE & BRE N, 1972: H. ERpTMAN & NoriIn; GADEK & QUINN, 1985; GAUSSEN, 1968; ae Hage: HEGNAUER, 1962, 1986; eee LITTLE, 1979; MEHRA & KHOSHOO; and SILBA. Apams, R., & D. Simmons. A chemosystematic study of Callitris (Cupressaceae) in south-eastern Australia using volatile oils. Austral. Forest Res. 17: 113-126. 1987. [C. columellaris, C. Endlicheri, C. Preissii, C. rhomboidea, C. verrucosa.) Bairp, A. M. The life history of Ca/litris. Phytomorphology 3: 258-284. 1953. A research on the pines of Australia. xiv + 458 pp. 28 dnnumbeped pls. 3 maps. Sydney. 1910. [Extensive treatment of Ca/litris, 13-290, including morphology, chemistry, wood and leaf anatomy, and uses; excellent il- lustrations BLAKE, S. T. New or noteworthy plants chiefly from Queensland. Proc. Roy. Soc. oe 70: 34-39. 1959. [C. Hugelii treated as a name of uncertain application. ] BoLAND, D. J., coord. Forest trees of Australia. ed. 4. xvi + 687 pp. Melbourne. 1984. “IC Callitris, 48-56: C. columellaris (as C. glauca), C. Macleayana, C. Preissii, maps, photographs including SEM’s of leaves and pollen cones. ] Buttock, A. A. The typification of the generic name Callitris Vent. Taxon 6: 227, 228. 1957. [C. rhomboidea proposed as lectotype species. COSTERMANS, L. Native trees and shrubs of southeastern Australia. vi + 422 pp. Dee Why West, New South Wales. 1986. [Ca//itris, 142, 381; maps, illustrations of cones, photographs of habit.] Franco, J. po A. Nomenclatura de algumas coniferas. Anais Inst. Super. Agron. 19: 12-15. 1952. [C. Hugelii lectotypified as an earlier name for C. columellaris.] GARDEN, J. A revision of the genus Ca//itris Vent. Contr. New S. Wales Natl. Herb. 2: 363-392. 1957. Kuan, I. U., & W.H. Ansari. Flavonol glycosides from Callitris glauca. Phytochemistry 6: 1221, 1222. 1987. Lacey, C. J. Silvicultural oe ua a white cypress pine. Res. Note Forestry Commiss. New S. Wales 26. 51 pp. 3. [C. glauca = C. columellaris.| Loony, W. J., & J. DoyLe. New ohservaton on the life history of Ca/llitris. Sci. Proc. Roy. Dublin Soc. 22: 241-255. RupMAN, P. Causes of natural durability in timber. Holzforschung 18: 52-57. 1964.* [Chemistry of Callitris specie SAXTON, W. T. Contributions . the life history of Callitris. Ann. Bot. 24: 557-569. 1910 THompson, J. Cupressaceae. Contr. New S. Wales Natl. Herb., Flora Ser. 1-18(5): 46- 55. 1961 —— & L.A. S. Jounson. Callitris glaucophylla, Australia’s “white cypress pine” — w name for an old species. Telopea 2: 731-736. 1986. [Replaces C. glauca when C. columellaris is treated as aye ] VENNING, J. Character variation in Australian species of Callitris Vent. (Cupressaceae). Unpubl. Ph.D. dissertation, Cnivensty of Adelaide. 1979.* [Recognizes 13 species in Australia. . Callitris Vent. Pp. 105-108 ee Jessop & H. R. ToELKEN, eds., Flora of South Australia, part I. ed. 4. Adelaide. 1986. 1990] CANTINO, STOMATA AND TRICHOMES 323 THE PHYLOGENETIC SIGNIFICANCE OF STOMATA AND TRICHOMES IN THE LABIATAE AND VERBENACEAE Puitie D. CANTINO! Epidermal anatomy was surveyed in leaves of 127 genera of Labiatae and 59 of Verbenaceae sensu /ato, with emphasis on the morphology of the stomatal complexes and the minute subsessile glandular trichomes that are found in nearly all members of both families. The phylogenetic significance of the data above the genus level was analyzed, using the subfamilies of Verbenaceae as outgroups to the Labiatae, and the Scrophulariales as outgroup to the Ver- benaceae. Many of the oo S exhibited such a large amount sei variation that they have little the spec level. In general, the presence/absence ‘of ‘stomatal types varied less aera genera than presence/absence of glandular trichome types. Although the phy- in light of other characters, the derived states in parentheses suggest the ex- istence of clades comprising the following taxa: Brazoria, Macbridea, and Physostegia (absence of anomocytic stomata); all Labiatae except the Pros- tanthereae and Amethystea, Tetraclea, Tinnea, and Trichostema of the Ajugeae (presence of diallelocytic stomata); Phyla (presence of two-armed unicellular trichomes; parallelocytic stomata); subfam. Verbenoideae, with the possible exception of four genera (absence of ates ‘hairs’ in nonglabrous species; i.e., only unicellular hairs occur); and all Chloanthoideae except Nesogenes Grecenee of branched, multicellular trichomes). The suite of stomatal types found in tribe Prostanthereae and Tetrac/ea and Trichostema of tribe Ajugeae differs markedly from that found in the rest of the Labiatae and resembles that in some Verbenaceae, particularly subfam. Chloanthoideae and tribe Clero- dendreae. However, difficulty in assessing polarity of two of the relevant char- acters makes evaluation of cladistic relationships difficult. In an earlier paper (Abu-Asab & Cantino, 1987a) the leafanatomy of subtribe Melittidinae (Labiatae) was surveyed, and an attempt was made to evaluate the phylogenetic significance of anatomical variation in the group. This effort was hindered by a scarcity of published data on the leaf anatomy of other subgroups of the Labiatae. Assessment of character polarities within the ingroup was based on outgroup comparison, but the outgroups comprised a mere scat- tering of labiate genera for which anatomical data happened to be available. This is not an unusual problem since there are few comprehensive surveys of the leaf anatomy of large families. The present survey is comprehensive to the extent that all major groups of the Labiatae and nearly all those of the Verbenaceae sensu lato have been ‘Department of Botany, Ohio University, Athens, Ohio 45701. © President and Fellows of Harvard College, 1990. Journal of the Arnold Arboretum 71: 323-370. July, 1990. 324 JOURNAL OF THE ARNOLD ARBORETUM [voL. 7] included. However, the effort to maximize the breadth of the survey with regard to genera and suprageneric groups has resulted in a rather scanty sample of intrageneric variation, particularly within subfam. Nepetoideae (Labiatae). Moreover, not all aspects of epidermal anatomy were studied; emphasis was placed on those characters that Abu-Asab and Cantino (1987a) found to be of systematic significance. Despite these limitations, this survey is the only one available for the Labiatae or Verbenaceae as a whole and one of few available for a major angiosperm family. It is hoped that the data provided here will facilitate evaluation of the phylogenetic significance of leaf-epidermal variation within genera of both families. TAXONOMIC BACKGROUND It is widely accepted that the Labiatae evolved from the Verbenaceae (i.e., the immediate ancestors of the Labiatae, if alive today, would be assigned to the Verbenaceae). The two families form the core of the order Lamiales of Dahlgren (1980), Cronquist (1981), Thorne (1981), and Takhtajan (1986). The boundary between the Labiatae and the Verbenaceae is unclear. Traditionally, they have been distinguished on the basis of stylar position—terminal in the Verbenaceae and gynobasic 1n the Labiatae. However, in the members of tribes Ajugeae and Prostanthereae of the Labiatae as well as some Verbenaceae, the gynoecium is intermediate in structure, the ovary being shallowly four-lobed and the style neither terminal nor fully gynobasic. Thus the taxonomic limits of the Labiatae are unclear, and there is no discrete character state (let alone a clearly derived one) supporting its monophyly. On the contrary, pollen mor- phology suggests that the Labiatae may be polyphyletic, with the more primitive genera having arisen from different subgroups of the Verbenaceae (Abu-Asab & Cantino, 1987b). Consequently, if a character survey of the Labiatae is to be useful in phylogenetic inference, it should include representatives of a wide variety of Verbenaceae as well. The classification of the Verbenaceae used here is that of Moldenke (1971), except that Cronquist’s (1981) broader circumscription of the family is adopted. Thus, the segregate taxa Avicennioideae, Chloanthoideae, Nyctanthoideae, Phrymoideae, Stilboideae, and Symphorematoideae, recognized as families by Moldenke, are treated as subfamilies here. This is done for convenience of data tabulation only and is not intended as a judgment on the relative merits of familial versus subfamilial rank for these taxa. For the Labiatae Erdtman’s (1945) subfamilial classification (see Cantino & Sanders, 1986) has been adopt- ed. Within subfam. Nepetoideae Bentham’s (1876) tribes are used (with cor- rected nomenclature of Sanders & Cantino, 1984), with the exception that those Pogostemoneae with tricolpate pollen are removed to the Lamioideae. Within subfam. Lamioideae five tribes are recognized here: Ajugeae sensu Bentham (1876), Prostanthereae sensu Bentham (1876), Lamieae sensu Abu-Asab and Cantino (1987a), Pogostemoneae (comprising Colebrookea Smith, Coman- thosphace S. Moore, Fusteralis Ral, Leucosceptrum Smith, Pogostemon Desf., and Rostrinucula Kudo), and Scut g Scutellaria L., Salazaria Torrey, and Harlanlewisia Epling). Although it would be simpler to adopt 1990} CANTINO, STOMATA AND TRICHOMES 225 Bentham’s classification in its entirety, modifying it as is done here increases the proportion of the tribes for which there is evidence of monophyly. None- theless, certain infrafamilial taxa (designated with quotation marks in the tables) are recognized here in spite of their probable nonmonophyly because they have not yet been sufficiently studied to subdivide them in a way that better reflects phylogeny. Their use facilitates tabulation and summary of the data, but in recognition of their questionable status, their monophyletic component taxa are treated as separate units in the analysis. MATERIALS AND METHODS Leaf mounts were prepared from herbarium material by a procedure devel- oped by Jon Hamer, modified from Abu-Asab and Cantino (1987a). Dried leaves were soaked overnight in a weak solution of Alconox soap in water, then transferred to five percent sodium hydroxide for twelve hours to three days, depending on leaf thickness. After being bleached in a 30 percent solution of household bleach (30 minutes to four hours, depending on the material), the leaves were placed in 50 percent ethanol for at least ten minutes, then stained with ferric tannate (2.5 percent tannic acid in 50 percent ethanol, followed by 2.5 percent ferric chloride in 50 percent ethanol; modified from Berlyn & Miksche, 1976) and mounted in surface view. A set of permanent slides is on deposit in the Bartley Herbarium of Ohio University (BHO). The study set included representatives of 59 genera of the Verbenaceae sensu latoand 127 of the Labiatae. Within the latter, 69 genera of subfam. Lamioideae and 58 of subfam. Nepetoideae were included. Six additional genera of the Lamioideae and one of the Nepetoideae were examined by Abu-Asab and Cantino (1987a). When the two data sets are combined (see TABLES 1, 2), the total represents approximately 60 percent of the genera of Verbenaceae sensu lato, 73 percent of subfam. Lamioideae, and 36 percent of subfam. Nepetoideae (estimates of the number of genera in the Verbenaceae sensu /ato and in the subfamilies of the Labiatae are derived from Moldenke (1971) and Cantino & Sanders (1986), respectively). A much higher proportion of the genera of the Verbenaceae and the Lamioideae were sampled than of the Nepetoideae be- cause variation in the former groups may be particularly helpful in understand- ing the origin and early evolution of the Labiatae. (Subfamily Nepetoideae represents a single large clade that arose from within the paraphyletic or poly- phyletic subfam. Lamioideae; see Discussion.) The list of voucher specimens was excluded from this report because of its length, but copies have been deposited in the libraries of the Harvard University Herbaria, the Missouri Botanical Garden, the New York Botanical Garden, and the United States National Herbarium. The vast majority of the specimens from which leaves were obtained are at A, BHO, and GH, but a few are at MO, Ny, and us (abbre- viations follow Holmgren et a/., 1981). In the examination of the prepared slides, emphasis was placed on two sets of characters that Abu-Asab and Cantino (1987a) found to be of systematic significance in the Labiatae: the morphology of the stomatal complexes and the structure of the minute, subsessile glandular trichomes that are character- 326 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 TABLE |. Stomatal characters in Labiatae and Verbenaceae. Taxa® Stomatal Ty Stom. ANO ACT CYC ANI HEL ome 312 DI3 DI4 Pos.4 ee "Lami a "Ajugea ee ajugiflora Prain - = = - + * hypo Ajuga chamaepitys (L.) Schreber t+ + - - - * + amph A. ae nsi o to = = - - * + = amph A. la Ce Be + + - - - * + = amph A lineerifelia a, + + - = - * + - amph rer eae A. aS easerr + + - = - * + - amph A. iaphila W.W.Smith + + - = - * + amph Amethystea coerulea L. (#1) w* + - 4 + + - hypo A. coerulea L. (#2) * + - - + + - - hypo Cymaria dichotoma Benth. t+ + = - Ge or hypo ET ngpoensis (Hemsley) Kudo i aS - + * + hyp K atum (ienstey 2 Kudo (#1) t+ - = = - * + + hypo K. ear (Hemsley) Kudo (#2) t - = = - + e+ hypo K. ornatum (Hemsley) — oa + + = = - + * + hypo K. pernyi (Franchet) K r - - a eo hypo Rubiteucris palmata (Benth. ex Hook.f.) Kudo we + 0-0 - - + + - hypo Schnabelia oligophylla Hand. -Mazz. 7? 7? =? ? + 2? ? hypo Tetraclea coulteri A.Gray (#1) a i + - amph T. coulteri A.Gray (#2) a eee 4 ce amph coulteri A.Gray (#3) wf f+ + Yr - amph Teucrium arduinii L * + + rr - amph T. canadense L to - = = - * + - hypo T. chamaedrys L.§ oo. ie ee ae a ee hypo T. integrifolium F.Muell. ex Benth. wee + + + xr - amph T Jaciniatum Torrey + + - - yr o* + - amph T. marw SY ee a 2? oe + hypo T. Poe ial olden Schreber + - = = - * + - amph qinnea aethiopica Kotschy ex Hook.f. (#1) a + - - hyp T. aethiopica Kotschy ex Hook.f. (#2) eof 4 t+ - - hypo Te antiscorbitica Welw. x + - - - x a hypo T. apiculata me . obyns & Lebrun xe ot - - + 2 = hypo T. ealoiati 1 i ro+o- - hypo T. Sees 3. Soene (#1) ne a 2 t+ + - - amph T. rhodesiana S.Moore — a - + - - hypo T. somalensis Gurke ex Ch We ke Se 2 he hypo Trichostema arizonicum A.Gray (#1) wf e+ + - amph T. arizonicum A.Gray (#2) , oe a oe es amph T. brachiatum L. + 2? 2? * 2 2% - 92 amph T. dichotomum L.§ + + = * oe ES “vere amph T. lanatum Benth + - = * eis, Fat, oe amph T. lanceolatum Benth.$ eo ot. Ea er amph T. lanceolatum Benth + ? + ? + 2? ? amph T. setaceum Bel: teh a Ae ses i Se amph pane eee mum parviflorum S.Moore + - - - * + 4 hyp A. schimperi Perkin Yorof s - + * 4x hypo A. wa nce Benth: ex Hook.f. t+ to - = - + eK + hypo Acrotome angustifolia G.Taylor +o - + = - * + - amph A. flecki as is ) Launert 43, ae ae og OR amph A. ISTE: en + r- - - * f+ amph A. inflata Be oe to = + = - * + + amph 1990] CANTINO, STOMATA AND TRICHOMES 327 TABLE 1. (continued). Taxa? Stomatal Types? Stom,. ANO ACT CYC ANI HEL PAR roe pi3 014 os.4 pallota frutescens (L.) J.Woods i a amph B. nigra L. (#1) of ue ea hypo B. nigra L. (#2) 5 a inte B. pseudodictamnus (L.) Benth. ee a hypo Brazoria arenaria Lundell a a a amph B. pulcherrima Lunde 118 - bo ees, Ee er oe Se = amph B. scutellarioides Engelm. & A ray! a amph B. truncata (Benth. ) Engelm, & ee Gray® a ea ee amph Chamaesphacos ilicifolius Schrenk wo oF - = =e ROU amph Chelonopsis forrestii ey i hypo C. lichiangensis W.Smith RP I Pe hypo C. longipes Makino Hee af. Sh. Ben” ah SR GS SE, we hypo C. moschat + ? ? 2? 2? ? + ? ? hyp Cc. moschata 8 * 2 7 2 2? 7? + 2? ? ~~ hypo Cc. odo Scalia. Diels eo oh es ee ae ie hypo Colquhounia coccinea Wallich = -- - 4 hypo C. seguinii Vaniot 2 ae” a ce de hypo Craniotome furcata (Link) Kuntze a ee 2 hyp C. versicolor Reichb. (#1) $2 2 2 2 2 + + + hypo C. versicolor Reichb. (#2) eo hypo Eremostachys bachardenica B.Fedtsch. Ss a amph E. ut iliensis Regel oe ke et, A amph E. isochila Pazij & Vved. ee ae amph E. eos Bunge Boa 6 os 2S. a a « amph E. regeliana Aitch. & Hemsley * + - = = * + - amph E. speciosa Rupr. ee ee ee amph E. tuberosa (Pallas) Bunge Kw + - - = + + 2 = amph Eriophyton wallichianum Benth. wo + - = = ef - amph Galeobdolon luteum Hudson? +o = f+ ee eh Ow hypo Galeopsis ladanum L. ee oe ee . amph G. ochroleuca Lam a a, ae ee a amph G pubescens Besser a a or ae hypo G. tetrahit L. ae a ey ee hypo Go onphostenma chinense Oliver i a A GY OO SY hypo G. nitum Wallich + 2? 2? 2? 2? 2? &* + ? hypo G SCE Wallich a hypo G. velutinum Benth. Se ow oe Oe Oe TE ae oe hypo pagochilus Gabulieus Benth. eo Se ss = =. We Bs os amph ilicifoli t+ - - - - - * + 2 amph 7: plat Seale ee & C.Meyer (#1) So a amph L. platycalyx Fischer & C.Meyer (#2) a amph Lagopsis supina (Stephan) Ikonn.-Gal. i amph Lamiophlomis rotata (Benth.) Kudo i ae amph Lamium album L to 2 2 ek oe e+ inte L. maculatum L. (#1) + + 2 = = = * + | hypo L. maculatum L. (#2) se i ee eG. ep hypo L. moschatum oe es e we Se OM ae og amph L. pictum Boiss, & Heldr a de tl em me, OD B.S amph L. purpureum § ae Se Se ae & amph 328 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 TABLE 1. (continued). Taxa? Stomatal Types” Stom. ANO ACT CYC ANI HEL PAR om DI3 DI4 = Pos.4 Leonotis latifolia Gurke a + - inte L. leonitis R.B + + - - - - * Y = amph L. leonurus (L.) ’R. Br. (#1) + 7 7 7 % + + 7 F amph L. leonurus (L.) R. ae — * + - = = = + | inte L. nepetaefolia (L.) R.Br. a amph Leonurus cardiaca L. § ee ee hypo L. heater i: “sweet ee eS OS OH inte L. macranthus Max eee es ae a Pd” hypo Leucas eRe Eng + ¢ | 2 2) 2 * F @ amph L. capen (Benth. y pore K+ - = = = + + | amph i. ¢1 es Benth t+ + - = = = * + = hypo L. decemdent ne a ee hypo L. mildbraedii Perkins i a ee ne Oe a ee hypo L. mollissima Wallich an a a ee ae Sa inte Loxocalyx urticifolius Hemsley + + - - = = * Y - hypo Macbridea alba Chapman Ss eh 2h fa = oe eae amph M. caroliniana oo Blake —_ Ss = SS 2 oe ee ee amph Marrubium desertii Noé ex Cosson wo Fo - - = = + + amph M. peregrinum L a a ae rae a a amph M. vulgare L. Kop eS Ue Oe Ce OR ES amph M. vulgare L.§ ie eS ee et amph Melittis melissophyllum L.3 + + - - - YF * + | hypo Metastachydium sapittatum (Regel) C.Y.Wu & Li i amph Microtoena insuavis (Hance) Prain ex Briq. (#1) + - - - - - * + = hypo M. insuavis (Hance) Prain ex Briq. (#2) ee hypo M. moupinensis oes t ex Prain * + 7 2? 2 PP + 2? ? hypo M. robusta He * + ds: Se hypo M. urticifo Lia a aeiey oe sk ek os Me ee hypo Moluccella laevis L. a 2 amph M. spinosa L. oe oe pF amph Notochaete hamosa Benth. a hypo re eee Boiss. +e ei SoS. SE amph oO. egrifoli nth. : an CO (EY MY Ca, MY Gd inte oO. Tita ae ) pea ex Hook.f. + + - ko + amph QO. persica (Burm.f.) Bo i amph Panzeria argyracea Kuprian. * £°9 9 FP > F 2 F inte Paraphlomis Sea (Blume) Prain ex Backer & Ba | ce an ea i, a hypo P. rugosa ra ) ‘Prain + + = = = eK He hypo Phlomidoschema parviflorum (Benth.) Vved. + - 5 = YT * + - amph phionis _2etatia Bunge Se co Fk ee amph B. bra Royle i amph os : nee L + + - = - = * $+ | hypo P. maximoviczii Rege t+ + - - = = * + hypo P. pratensis Karelin & Kir Ce a ee amph P. setigera Fale. ex Benth. ee eee ee inte P. taurica Se eae ex Bunge t+ - = = = eK HF amph P. tuberosa L. Se ee ee hypo P. umbrosa ae eer een. i dee ge oe fe hypo 1990] CANTINO, STOMATA AND TRICHOMES TABLE 1. (continued). Stom ACT CYC ANI HEL PAR a7 p13 D014 Phyllostegia breviden Ps grandiflora ice aia ) on, (#1) P. grandiflora (Gaudich.) Benth. (#2) P. hispida H Aillebrand physostegia an usttfolia Fern. rotroro trols ro on le lo 5 Spa foun (0 n “10 |o ity leas > |p 3 virginiana (L. Prasium majus L. (#1) P. majus L. (#2) Roylea calycina (Roxb.) Briq. Sideritis ambigua Fenzl hirsuta IA [111 [A 1 IX IM Gs. pu ies Vent: Stachyopsis oblongata (Schrenk) Popov & Vved. St tachys acerosa Boiss. Ss. ua ey Li s. Be tonica nth. Sy betonieaeflora Rupr. SB: ccine acq Ss. aa in S. labiosa Bertol. Ss. xiddeliit House’ S. spathulata Burchell — ee s. SyTeeya Boiss. ex S. tenuifolia Willd i a Stenogyne diffusa A.Gray (#1) S. di Saaaat ee oa ay (#2) s. kameham e Wawra Ss pur ea = ied s. Hess Benth Synandra hispidula (Michx. ) Baillon’ a ee eas aa ) Briq. . pe a (Boiss.) Wiedemannia multifida (L.) Benth. Pogostemonea Colebr cokes 0 oppositifolia Smith (#1) C. oppositifolia each (#2) Comanthosphace stellipil C. sublanceolata (Miq.) = ee “GiL) C. sublanceolata (Miq.) S.Moore (#2) Eusteralis cruciata (Benth. ) Panigr. E. sampsoni E. stellata (Lour.) ee (42) +eett+eet *% t+etteettetes + +44 %4 fee XXX HX HH + H+ Hee KH HH e+ HH 1 Ret tee te terete eeteeet +%*#*+% + K+ e+ % 330 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 TABLE |. (continued). Taxa? Stomatal Type Stom. ANO ACT CYC ANI HEL PAR 70 DI3 DI4 ~=Pos.4 Pogostemon cablin Sea Benth. 5 a a inte P. elsholtzioides Be a a ee a ee hypo PB. ber Benth. t+ - - - = = + * £ hypo P. glaber Benth. (#2) fe we: ie ee inte P. heyneanus Benth. 5 a ae) ee hypo P. plectranthoides Desf. Se eo hypo Rostrinucula dependens (Rehder) Kudo a hypo Prostanther Hemiandra Bonen R.Br. ee Wee Se ke. amph Hemigenia incana Benth. (#1) OS Se ee ae Se ce oe amph H urpu .Br. Oe ow Oke, fe +e Se amph H. saligna Diels ae Bo ees, SP ae es tee Fe Se amph agnen pete brevidens Benth. Me a es es me amph ericifolia Benth. ee ee amph Prostanthera aspalathoides Cunn. ex Benth. ee amph P. cuneata Benth. ee a. a ae ce inte P. euphrasioides Be os oS ES Se en amph P. lasianthos Labill. (#1) * 2. Kk Yr + Yr - - hypo P. lasianthos Labill. (#2) + - - * + vr re - hypo P. niv . ex Benth. (#1) ff ne Bk ate a a, amph P. ni C ex Benth. (#2) a amph P. ovalifolia R ee ae er et Je hypo P. rotundifolia R.Br. wo + - + = y + = = amph P. rotundifolia R.B Oo fe Oe & 4B & =e 2 inte P. saxicola R i os 2 Ae Oe Oe amph Westringia amabilis J.Boivin * yo - + & - = hypo ae brevi folia Benth ee, ee, ee ee hypo eelii Maiden & Betche w+. - = ee amph fruticosa (Willd. ) ee * + - yg - + = = = hypo Scutellarieae Salazaria mexicana Torrey (#1) i amph S. mexicana Torrey (#2) Po a - - * + - amph Scutellaria amoena C. a Wright + a amph Ss. elli iptica Muhlenb. + - - - - < + & + hypo S. gardoquioides Benth MS ken OU hypo S. hirtella Juz. 2? 2? 2? 2? 2? 2 *&* + 2? inte S. incana Biehler$ a ek - oo. k + hypo S. integrifolia L,3 a a inte S. lateriflora L.3 i. Ry es ee ce DR hypo S. multicaulis Boiss t+ + = = = = * + Gg amph S. nervosa P 5 4 i « S 2) se oe & hypo S. oriental L ae ee ee eo int S. ovata Hill + a ee ee ee © hypo S. serrata Andrz.$ + oo Awe Se ae OE hypo eens Tribal Af ‘ Ajugoides humilis a iq.) Makino (#1) t+ + - = = = * + = hypo A. ree is (Miq.) Makino (#2) a ae. ee hypo Anisomeles heyneana Soe ek a we es oe ee hyp A. ee (L.) ea G re + ? ? 7? 7? 2? * + ? amph A. dica (L.) Kuntze Re + 9? 2? 2? 2 PP * + ? amph A. qeTEbe ee (L.) R.Br. ex Sims (#1) Se ee hypo urysolen gracilis Prain (#1) Mee hypo 7 gracilis Prain (#2) 2! ee ee hypo Hypogomphia turkestana Bunge ce ce 2 eth oe: = amph Suzukia shikikunensis Kudo t+ + = = = = * + = hypo 1990] CANTINO, STOMATA AND TRICHOMES TABLE |. (continued). 331 Taxa® Stomatal Types®¢ Stom, ANO ACT CYC ANI HEL PAR 7) 013 DI4 Pos. Nepetoideae "Mentheae" Ceratominthe odora (Griseb.) Hauman + + a hypo gol linsonta canadensis L. (#1) + + - = * + hypo . canadensis L. (#2) + + - = * + - hypo Cunila origanoides (L.) Britt. + - - - * + = inte Cyclotrichium origanifolium (Labill.) Manden. & Scheng. + - - = * + - amph Elsholtzia patrinii (Lepechin) Garcke ~ - 2 ® + = inte Hedeoma drummondii Benth. - - = FOR + amph H. graveolens Cha apman ex A.Gray + - Se amph H. nanum (Torrey) Briq. + - - - * + - amph Keiskea japonica Miq. + = - - * + - hypo Lepechinia hastata (A.Gray) Epling + - - = *® + - amph Lycopus 7 Muhlenb. + - - = K+ inte L. tubellus iy nch + + - = e+ amph L. virginicu L ro a ee eo hypo conese officinalis L. (#1) + - - = * + - hypo officinalis L. (#2) + + a a hypo Mentha arvensis L. Gs mee OR RS oe hypo M. citrata Ehrh. r- - KO FOF hypo iperita L rr = = oe HR bs hypo Micromeria biflora Benth. + - = Ko + = hypo M. punctata ith 4 es ee ee hypo Monardella odoratissima Benth. + - a a amph M. villosa Benth. + - - = * + 24 amph perillula reptans Maxim. (#1) + - - - + * 4 hypo . reptans Maxim. (#2) + + - - * + 27 hypo Pogogyne zizyphoroides Benth. + - - = + * - amph Poliomintha glabrescens A.Gray ex Hemsl. + - - - * + 27 amph Fyenanthemum atbesceris A.Gray + ? 2? Po +t + + hypo P. floridanum E.Grant & ete - - > + FS hypo P aff. incanum (L. . a chaux + - a a a hypo Rhabdocaulon strictum (Benth.) Epling - - - + * - hypo Rhododon ciliatus (Benth.) Epling ro: - 2 *® + - amph Satureja arkansana (Nutt.) Briq. + - i a ae amph S. douglasii (Benth.) Briq + - a a hypo S. parvifolia (Philippi) Epling + + - = + = hypo S. popovii B.Fedtsch. & tch = a a amph S. vulgaris F - - * + 827 hypo Thymus serpyllum L. ; - = * + = inte Nepetea ee breviflora (A.Gray) Epling - = i es amph A. cana (Hook.) Wooton As Standley + - + * + - mph A. nepetoides (L.) Kun + - - r * + 4 ypo A. pallidiflora (A.A eee k + - - = &* + = amph A. scrophulariaefolia (Willd.) ae + + i a a 2 hypo 2 JOURNAL OF THE ARNOLD ARBORETUM TABLE |. (continued). [voL. 71 Taxa® Sto tal Types? ANO mat St ACT CYC ANI HEL PAR DI2 DI3 DI4 Pos. Cedronella canariensis (L.) W C. canariensis (L.) W Dracocephalum hens leyanum (Prain) Marquand D. heterophyllum Benth. D. parvifLorun Rake D. ruysce L. Glechoma hederacea L. Lophanthus chinensis Benth. Meehania cordata (Nutt.) Britton Nepeta cataria L. N. clarkes Hook. i N. curviflora Boi N. ciseslee Royle ¢ ex Benth. N. nepetella Prunella vulgaris L. Ocimeae Acrocephalus fruticosus Dunn A. indicu s Kuntze Asterohyptis stellulata (Benth.) Epling bb & Berth. (#1) + Webb & eee. (#2) + +HenK + Catopheria capitata Benth. ex Hemsley (#1) xr ° C. capitata Benth. ex Hemsley (#2) Eriope crassipes Benth. Eriopidion strictum (Benth.) R.Harley Fuerstia africana T.C.E.Fries Haumaniastrum coeruleum (Oliver) Duvign. & Plan cke Hemizygia canescens (Gurke) Ashby Holostylon strictipes G.Taylor Hyptis alata (Raf.) Shinn. H. emoryi Torrey H. mutabilis oa : oe H. oblongifolia Be Icomum lineare Burkill Lavandula multifida L, L. stoechas L Nautochilus labiatus (N.E.Br.) Bremek. Ocimum americanum L. i 0. ea i (#1) QO. gratissimum L. (#2) Orthosiphon affinis N.E.Br OQ. ar shetus (Blume) hie QO. spi (Lour.) Me t++et+ Plectranthastrum clerodendroides T.C.E Fries - +1 RS +e e+ eee HK ++ +++ ++ 44H mn — 1990] CANTINO, STOMATA AND TRICHOMES TABLE 1. (continued). Taxa? Stomatal Types™® tom. ANO ACT CYC ANI HEL PAR DI2 DI3 DI4 Pos.4 prectranthus forsteri Benth. Kor es se 5 KK amph P. utellarioides (L.) R.Br. Ce a ee ee hypo Rabdosia excisa (Maxim.) H.Hara + + - = = = * Y = hypo R. inflexa (Thunb.) H.Hara i hypo R. nervosa (Hemsley) C.Y.Wu & Li i inte Solenostemon scutellarioides L.Codd (#1) + = = eee RK amph Syncolostemon densiflorus Benth. Se amph "Salvieae" Arischrada bucharica (Popov) Pobed. Yo- 5 t= = 5 fF Oe + amph Blephilia hirsuta (Pursh) Benth. ® + YY = = = = * + = hypo Monarda clino pola Lis to - = = ee KH hypo M. See ee Dee a IS I Re RE. OF, hypo M. pun ta i i i i on amp Perovskia abrotanoides Karelin Se amph P. atriplicifolia Benth. t+ + = = = = K+ amph Salvia carnosa Douglas sae ne a Cs CO amph S. farinacea Benth. oo Vee. ee, Re amph S§. lyrata L ee a amph S. reflexa Hornem. to - = = = = & * - amph Verbenaceae Avicennioide Avicennia aieida Jacq. a a a a hypo aryopteridoidea Ceogeeniacae Caryopteris grata Benth a Se ee ee es hyp C. incana (Thunb.) Miq. (#1) + + ? 2? 2? ? 2? ? ? hypo C. mongholica Bunge (#1) ew + = + - amph C. mongholica Bunge oe ee oY amph C. nepetaefolia (Benth.) Maxim i sa hypo C. odorata (Ham.) Rob 5 a a hypo C. terniflora Maxim + rr - - + * + - hypo Glossocarya siamensis Craib eK + - + - - - hypo Petraeovitex kinabaluensis Munir to - = = = + KK + hypo P. multiflora (Smith) Mer a hypo Teijsmanniodendr Teijs Eee tod ende ron ahernianum (Merr.) Bakh. K+ = = eee hyp T. subspicatum (H.Hallier) Kosterm. ee a hypo Chloanthoideae Achariteae Nesogen dupontii Hemsley i i a ee ee a amph N. eapieaaieldes (Hook, & Arn.) A.DC. a amph Spartothamnella puberula (F.Muell.) Maiden & Betche (#1) Me ee ed St ee Se a tes hypo S. puberula (F.Muell.) Maiden & Betche (#2) Kf = + = + YC = - hypo S. puberula (F.Muell.) Maiden & Betche (#3) i inte Chloanthe Ghilganthes stoechadis R.Br. too- = we eee hypo Cyanostegia angustifolia Turcz. a a ee 2 a amph €. microphylla S.Moore 5 i i a a i. amph JOURNAL OF THE ARNOLD ARBORETUM TABLE 1. (continued). [VoL. 71 Taxa® Types? ANO Stomatal re Stom ACT CYC ANI HEL PAR DI2 DI13 DIS Pos. Physops oer exsuccosa (F.Muell.) Druce Newcastelia cephalantha F.Muell. Nyctantho Nyctanthes neues tristis L. Phrymoidea Phryma Latest Symphorem Congea C ide atoide orbesii g. tomentosa eee Mo ii oe ng & te oxb. Sphenodesme ferruginea (Griffith) Briq. S. pentandra J Symphorema Sn rae (Blanco) Fernandez-Villa Ve rbenoidea (oH lipustri ein Van Houtte G res Greenm an Duranta cia Mold. D liga. mutisii DD cern Mold. (#1) D. repens L. Rehdera trinervis (S.F.Blake) Mold. Rhaphithamnus spinosus (A.L.Juss.) Mold. Lantaneae Aloysia gratissima (Gill. & Hook.) Tronc. Bouchea fluminensis lee Conc.) Mold. B . prismatica (L.) Ku Diostea juncea (Gillies & Hook.) Miers Lantana L. hor L. involucr Lippia graveolens Kunth rida Kunth (#1) ho aa a aes a Phy ia incisa Sma P. la P. nodi 1l olata gona ae Greene (L.) fl lor E.Gre § Priva aspera Kunth P. grandiflora (Ortega) Mold. a eta frantzit Polak. hl jamaicensis (L.) Monochileae Amasonia campestris (Aublet) Mold. th. A. hirta Ben + % + 1990] CANTINO, STOMATA AND TRICHOMES TABLE 1. (continued). 335 Taxa? Stomatal Type Stom. ANO ACT CYC ANI HEL PAR a DI3 DId Pos.4 Petreeae Casselia hymenocalyx Brig. - suet ge - hypo Petrea amazonica Mold. + - - ee + + hypo P. arborea Kunt + ? Yaa oe ? hypo P. volubilis L. (#1) + - 7 ae. Ae _ hypo P. volubilis L. (#2) + + Sy wat, “Ge SE - hypo Verbe Glandularta bipinnatifida (Nutt.) Nutt. xe - fe - amph G. nadensis (L.) Small e+ - eee - amph Hierobotana inflata (Kunth.) Briq. + ? Dt ee ? hypo Junellia ligustrina (Lagascana) Mold. e+ - eee - amph Tamonea curassavica (L.) Pers. + ? ? 2? ? ? amph Verbena hastata L.® * or ae. es as 2 ypo V. litoralis Kunth w o+ + - rr = amph V. macdougalii A.A.Heller * + fe . amph V. pumile Rydb. * + & op e amph V. icifolia L.§ * + = 2. ap ss hypo Viticoideae Callicarp Aegiphila “aculeifers Mold. + * a 7 hypo A. de ana el 4o. b . a : hypo A. Seen o i x + + 7 7 + 5 hypo Callicarpa americana x oF - ee + - hypo C. dichotoma (Lour.) K ‘Koch * + (i a aes ? hypo C. mollis Siebold & Zu x + ee ep = hypo Clerodendreae Clerodendrum aculeatum (L.) Schldl * + - 2 5 - hypo C. anafense tton & P.Wilson * + t+ - + - - hypo Cc. capitatum (Willd.) Schum. & Thonn. * + - eee . hypo Cc. cuneatum Gurke + + a te Se ae r amph C. floribundum R.Br x + ro- + : amph Cc. abrum E.Me w+ fe ok A inte Cc. inerme (L.) ner * e «© os os : hypo C. myricoides (Hochst.) R.Br ex o+ - = * “ hypo C. philippinum Sch * + 7 et re 2 hypo C. squiresii x + - - + - - hypo C. trichotomum Thunb, w+ S pe; 22 = hypo Far radaya amicorum (Seemann) ae e+ - = = = - hypo By: oval ifolia i a he Seeman * - a as noes 2 hyp F. splen F.M * + an a - hypo Holmskioldia sanguinea Retz. (#1) w+ to - = = - hyp H. sanguinea Retz. (#2) * 4 +o - oc + = hypo H. sanguinea Retz. (#3) ye : ee : hypo H. tettensis (Klotzsch) Vatke + * ro- o- - hypo Kalaharia spinescens Gurke * + - + - - amph Karomia fragrans Dop eh te Se s hypo Oxera morierii Vieill. oes Met ed he = hyp OQ. neriifolia Beauv. + + * + + - : hypo sulfurea Dubard + + * + + = hyp Teucridium parvifolium Hook.f. w+ to - - + - hypo 336 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 TABLE 1. (continued). Taxa® Sto al Typ ANO ACT CYC an HEL oe = 013 014 Abe Vit Cornutia grandiflors (Cham. & Schldl.) Schauer * + - - - - + = = hypo Cc. pyramida * 7? 2? 2? ? ? + ? ? hypo Garrettia siamensis Fletcher ig sis EE RS SP hypo Gmelina delavayana D K+ = ee ee hypo G. moluccana (Blume) “backer 5 hypo G. racemosa (Lour.) Me Sane ae a A ss hypo Premna barbata Wallich t+ - = = 2 - F + - hypo P. corymbosa Willd a hypo P. foetida Reinw. 5 a a ee ae hypo P. japonica Miq. i hypo P. octonervia Merr. & Metcalf a hypo Esau eae (A.Rich.) Millsp. * + - - += - + = - hypo . wrightii Mills i hypo Tsoongia axillariflora Merr. Se hypo Vitex agnus-cas 48 SA ae ve, ee OS hypo V. cannabifolia er & Zucc. eK + - + - ro - - = hypo "If name is followed by the symbol §, data are ries eg -AS - uh Sonatas (1987a). All other data are from the present survey. #1, #2, efe different specimens. of a species; the numbers correspond to entries in the nee oe jeeouieed in fou major botanical libraries (see Materials and Methods). >stomatal ie ANO, anomocytic ae actinocytic; crc, cyclocytic; ANI, anisocytic; HEL, helic eee PAR, paracytic , dia oo (2 subsidiary cells); 013, 3-celled diallelocytic; 014, 4-celled een aaa ee *, commonest type(s); +, pr are (no more than two examples found); nt; ?, unknown whether eoeae or eae (stomata are poorly stained or Speed by dense trichomes). cer Position: amphistomatic, hypostomatic, or intermediate (i.e., a few stomata the adaxial surface). "Rare eee eg staurocytic, in Lavandula multifida; parallelocytic, in the three species of Phyla istic of the Labiatae. The fundamental structure of the nonglandular trichomes (1.e., unicellular vs. multicellular and simple vs. branched) was also noted. The classification system developed by Abu-Asab and Cantino (1987a) for the subsessile glandular trichomes 1n subtribe Melittidinae was found to apply well to the Labiatae as a whole and 1s used here (see the APPENDIX). For it to extend to the Verbenaceae, an additional gland type (type 11) was added. RESULTS TABLES | and 2 require some introductory comments. Because the leaves of some species stained poorly, and the subsessile glands and stomata in others were obscured by a dense layer of nonglandular trichomes, the tables are heavily laden with question marks. Even when the data are incomplete, however, some information may be inferred. For example, type 5 glands were definitely present 1990] CANTINO, STOMATA AND TRICHOMES 337 on the leaves of 7innea apiculata (see TABLE 2, under ‘‘Ajugeae’’), and more complex glands were present as well, but it was unclear whether they were type 8,9, or 10. On the other hand, types 1-4, 6, 7, and 11 were definitely absent. A species has been included in a table only if there were clear observations to tabulate with regard to the characters of concern. Each table therefore in- cludes some species not found in the other. For example, Acrotome angustifolia is present in TABLE | but not TABLE 2 because the stomatal complexes stained sufficiently for their configurations to be discerned but the subsessile glands did not stain well enough to be classified as to type. For subsessile glands and stomatal complexes I have indicated relative abun- dance by designating with an asterisk the most common gland and stomatal types in a species. If no asterisk is present, it may be because subsessile glands were so infrequent that a meaningful estimate of relative abundance could not be made or, alternatively, because such a high proportion of the glands (or stomata) were poorly stained or hidden by nonglandular trichomes that relative abundance could not be estimated. If two or more types of stomata or glands are marked with an asterisk, they were roughly equal in abundance on the leaf surface (or, occasionally, one type was most abundant on the abaxial surface and another on the adaxial). In order to provide a more comprehensive survey and facilitate comparison of taxa, data published by Abu-Asab and Cantino (1987a) are included in the tables. Because actinocylic stomata were not distinguished from anomocytic in the earlier study (discussed below), the slides used by Abu-Asab and Cantino were reexamined to determine the distribution of actinocytic stomata. In the process, a few other errors of omission were discovered. When discrepancies exist between the earlier paper and this one, the data here should be assumed to be correct. In TABLES | and 2 the names of a few suprageneric taxa of Labiatae (La- mioideae, Ajugeae, Mentheae, Salvieae) are placed in quotation marks to in- dicate that these taxa are probably not monophyletic (see Taxonomic Back- ground). No attempt was made to do the same for the Verbenaceae, in which phylogenetic relationships are less well understood. STOMATA Ten types of stomatal complexes were observed (definitions adapted from Payne, 1970, Dilcher, 1974, and Wilkinson, 1979): actinocytic (stoma sur- rounded by a single ring of five or more radially elongate cells enclosing the guard cells); anisocytic (stoma surrounded by three subsidiary cells, one of which is markedly smaller than the other two); anomocytic (stoma surrounded by cells that are indistinguishable from other epidermal cells); cyclocytic (stoma surrounded by a single ring of small subsidiary cells); diacytic (stoma enclosed by a pair of subsidiary cells whose common walls are perpendicular to the guard cells); diallelocytic (stoma enclosed by an alternating complex of three or more C-shaped subsidiary cells of graded sizes oriented perpendicular to the guard cells); helicocytic (stoma surrounded by a helix of four or more cells); paracytic (stoma bordered on either side by one or more subsidiary cells whose 338 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 la 2a 2b Anomocytic Paracytic BODO a 3b 3c Anisocytic (@) ‘ Sa \ 6a \ 5b a 4c Diacytic 6c Diallelocytic Diallelocytic 1990] CANTINO, STOMATA AND TRICHOMES 339 long axes parallel those of the guard cells; subsidiary cells may or may not meet over the poles); parallelocytic (stoma with an alternating complex of three or more C-shaped subsidiary cells of graded sizes oriented parallel to the guard cells); staurocytic (stoma surrounded by three or four similar subsidiary cells with anticlinal walls arranged crosswise to the guard cells). he actinocytic type, scored as anomocytic by Abu-Asab and Cantino (1987a), is recognized here with reservations. The term has been used differently by different authors (discussed by Baranova, 1987) and is considered by Stace (1965) to be a mere modification of the anomocytic type. As a rule of thumb, I have scored a stomatal complex as actinocytic if there are at least five radially elongate subsidiary cells that are longer than the other epidermal cells. Using this definition, however, many stomatal complexes were scored as actinocytic that closely resemble what Wilkinson (1979, fig. 10.3a) considered to be anomo- cytic. Others scored here as actinocytic might be classified by some workers as stephanocytic, a newly described stomatal complex (Baranova, 1987) that is intermediate between the actinocytic and cyclocytic types. Wilkinson (1979, p. 99) noted that “giant or hydathodic stomata” are frequently actinocytic. Those observed in the present study were frequently, but not invariably, larger than the other stomata on the leaf. As discussed by Abu-Asab and Cantino (1987a), two kinds of diallelocytic stomata occur in the Labiatae, one with three subsidiary cells and the other with four. Because they do not always occur together (the latter is much rarer than the former), they have been listed separately in TABLE I. Stomatal ontogeny was not systematically studied, but ontogenies could sometimes be inferred from the morphology of mature stomatal complexes. The ontogenetic pathways of most stomatal types that are common in the Labiatae and Verbenaceae are shown in FiGurRE | Anomocytic and diacytic stomata were the most frequently encountered types in both the Labiatae and the Verbenaceac (see TABLE 1). The former were observed in all Verbenaceae and the vast majority of Labiatae examined, the latter in slightly more than half the genera of Verbenaceae and all genera of Labiatae except in tribe Prostanthereae, where they were rare. Diallelocytic stomata are far more frequent in the Labiatae than in the Verbenaceae. In the Labiatae diallelocytic stomata with three subsidiary cells were found in nearly all species of subfam. Nepetoideae and tribes Pogoste- moneae and Scutellaricac, in most genera of Lamieae, and in six genera of Ajugeae, but they have not been found in tribe Prostanthereae. In the Ver- benaceae three-celled diallelocytic stomata were found in all examined species of Amasonia L. f., Bouchea Cham., Phyla Lour., Stachytarpheta M. Vahl, and Clerodendrum L. subg. Cyclonema (Hochst.) Gurke, three species of Premna Ficure |. Stomatal ontogenetic pathways: anomocytic (1a), paracytic (2a, b), aniso- cytic (3a-c), diacytic (4a-c), diallelocytic with 3 subsidiary cells (5a-c), diallelocytic with 4 subsidiary cells (6a—c). (M = meristemoid, sensu Fryns-Claessens & Van Cotthem, 1973.) (Ontogenies of diacytic and 4-celled diallelocytic stomata adopted from Payne, 1970. Figure originally published in Abu-Asab & Cantino, 1987a.) 340 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 L., two species of Carvopteris Bunge, and four other scattered species. Dial- lelocytic stomata with four subsidiary cells are very rare in the Verbenaceae. In the Labiatae they are widespread, occurring in every tribe except the Pros- tanthereae, but their distribution within genera is very inconstant: although observed in 44 genera of the family, in only four were they found in all examined species (excluding those in which only one species was studied). Anisocytic stomata are widespread in the Verbenaceae and particularly com- mon in the Chloanthoideae and the Clerodendreae. In the Labiatae they are essentually restricted to the two tribes that lack a gynobasic style—the Ajugeae and the Prostanthereae. Particularly characteristic of the Prostanthereae, aniso- cylic stomata were found in nearly every species and were the most common type in three of the five genera examined. In the Ajugeae they were found in all examined specimens of Tetraclea A. Gray and Trichostema L. but were rare or absent in the other genera. Anisocytic stomata can be derived via several ontogenetic pathways (Payne, 1970). The ontogeny documented by Abu-Asab and Cantino (1987a) and shown in FiGure |, which if extended can give rise to helicocytic stomata (see below), is responsible for most or all of the stomata scored here as anisocytic. Superficially similar stomatal complexes that ap- peared to have been derived via other ontogenetic pathways (e.g., see Payne, 1970, figs. 25-27) were scored as anomocytic rather than anisocytic. Helicocytic stomata, found in five genera of the Verbenaceae and two of the Prostanthereae but rarely in all species of a genus, occur only when anisocytic stomata are present and usually when they are the most common type. This association was also noted by Wilkinson (1979), and the ontogenetic connection between the two types has been documented by Payne (1970). Paracytic stomata are widespread in the Verbenaceae, somewhat less wide- spread in subfam. Lamioideae, and very rare in subfam. Nepetoideae. The are particularly characteristic of Prostanthera Labill. and Trichostema in the Labiatae and of Duranta L., Petraeovitex Oliver, Phyla, Oxera Labill., and subfam. Symphorematoideae in the Verbenaceae. Payne (1970) commented that the paracytic type is ontogenetically the most variable stomatal complex in the dicotyledons. In the Labiatae and the Verbenaceae stomata scored as paracytic develop through at least two and perhaps more ontogenetic pathways. In the majority of genera where it is a common type, it appears to have an ontogeny whose initial steps are shared with the anisocytic stoma (see FIGURE 1). In other genera, in which anisocytic stomata are absent and diacytic and diallelocytic types are common, paracytic stomata may have an ontogeny sim- ilar to that of diacytic stomata (see FiGure |), differing only in the final division, the guard-cell mother cell dividing parallel to the subsidiary cells instead of perpendicular to them. This hypothesis is supported by the occurrence of occasional intermediates between diacytic and paracytic stomata, in which the guard cells lic at an oblique angle to the subsidiary cells (Pant & Kidwai, 1964: Inamdar & Bhatt, 1972). In Sphenodesme Jack and Symphorema Roxb. para- cyltic stomata are abundant and neither anisocytic nor diacytic types were found, suggesting that paracytic stomata may arise via a third, unknown pathway in these genera. Actinocytic st ta are widespread in both the Labiatae and the Verbenaceae 1990] CANTINO, STOMATA AND TRICHOMES 34] but are seldom particularly common on a given leaf. The remaining three stomatal types were found in very few species: cyclocytic in three species of Duranta, staurocytic in Lavandula multifida, and parallelocytic in the three examined species of Phyla. Amphistomatic leaves are slightly more frequent than hypostomatic ones in the Labiatae, but both conditions occur in every tribe (see TABLE |). In contrast, over 70 percent of the species of Verbenaccae investigated, including all ex- amined members of the Symphorematoideae, Callicarpeae, Petreeae, and Vi- ticeae, and all but one species of the Caryopteridoideae and Citharexyleae, had hypostomatic leaves. Amphistomatic leaves predominate in the Chloanthoi- deae, Lantaneae, and Verbeneae. Variation within genera is common, both conditions being found in 20 genera of the Labiatae and five of the Verbenaceae. The difference between the Labiatae and the Verbenaceae in the proportion of species with each condition may be an ecological correlate. The Labiatae are much better represented than the Verbenaceae in arid and semiarid regions, and amphistomatic leaves tend to occur more commonly in xeric habitats (Parkhurst, 1978). SUBSESSILE GLANDULAR [TRICHOMES Subsessile glandular trichomes have been widely reported in the Labiatae and the Verbenaceae under a variety of names, including peltate hairs, glandular dots, and glandular scales (Solereder, 1908; Metcalfe & Chalk, 1950, Huang & Cheng, 1971; Bosabalidis & Tsekos, 1982: Werker, Ravid, & Putievsky, 1985a). The adjective “subsessile” was applied by Abu-Asab and Cantino (1987a) because the glands may appear sessile in surface view but can be seen in cross section to have a short, usually discoid stalk cell (Fahn, 1979, fig. 92; Abu- Asab & Cantino, 1987a, fig. 3). The same term was employed for similar glands in the Acanthaceae by Ahmad (1978) and Karlstrém (1978, 1980). In many Labiatae the subsessile glands function in the secretion and storage of the essential oils (volatile terpenoids) that characterize the family (Fahn, 1979, Bosabalidis & Tsekos, 1982; Bruni & Modenesi, 1983; Werker, Ravid, & Putievsky, 1985a). Their ontogeny has been well documented (Bosabalidis & Tsekos, 1982, 1984; Bruni & Modenesi, 1983). Because subsessile glandular trichomes occur in nearly all Labiatae but vary in structure, they would seem to offer considerable potential as taxonomic characters. Abu-Asab and Cantino (1987a) developed a classification of subsessile gland types based on the number of cells and the cell-wall configurations in the head of the gland. This classi- fication, modified to include a gland type found only in the Verbenaceae, has been adopted here (see FIGURE 2, APPENDIX). Capitate glandular trichomes 16 . those whose stalk is long in relation to the size of the head —are also widéspread | in the Labiatae but were not included in this survey because it was clear at the outset that they exhibit too much intrageneric variation to be of much use as phylogenetic indicators above the species level. Their presence complicated the scoring of subsessile gland types, however, because they occasionally intergrade. The intermediates have a stalk that is elongate rather than discoid but shorter than the head. As a rule of [VoL. 71 JOURNAL OF THE ARNOLD ARBORETUM Zz 3 10 1 Examples of the | 1 types of subsessile glandular trichomes in the Labiatae and Verbenaceae, as seen in surface view (see also APPENDIX). FIGURE 2. 1990] CANTINO, STOMATA AND TRICHOMES 343 thumb, these were counted as subsessile and included in TABLE 2 if the length of the stalk was no more than half the height of the head. Subsessile glands are present in nearly all the Labiatae and Verbenaceae (see TABLE 2) but vary greatly in abundance. They are generally more common on the abaxial than the adaxial surface of the leaf. In some specimens only one or two glands were found in spite of much searching, and they were not found at all in three species of Verbenaceae (Gmelina moluccana, Nesogenes dupontii, Premna octonervia) and one of Labiatae (Trichostema lanceolatum). The glands occur in other species of each of these genera, and it is possible that they simply are very sparse in these four species and would be found if more leaves were examined. In Trichostema lanceolatum capitate glandular trichomes are abun- dant on the leaves and have been shown by Heisey and Delwiche (1984) to contain a phytotoxic essential oil. Thus the leaves of all examined species of Labiatae had glandular trichomes of some kind. Types 4 and 5 were the most frequently observed subsessile glands in both families. Within the Verbenaceae type 4 was recorded in 78 percent and type 5 in 71 percent of the genera; within the Labiatae these figures were 76 and 72 percent, respectively (genera for which presence or absence could not be de- termined are excluded from percentage calculations here and elsewhere). In view of the frequency of type 4 glands in both families, their rarity in the Mentheae 1s noteworthy. Type 2 glands are the third most frequent type in both families, occurring in 41 percent of the genera of Verbenaceae and 48 percent of the Labiatae, but at the tribal level they are less uniformly distributed than types 4 and 5. Type 2 glands are most common in subfam. Nepetoideae (except tribe Mentheae), tribe Lantaneae, and the genera Ajuga L., Anisomeles R. Br., Eusteralis, Kinoste- mon Kudo, Leucas R. Br., Phlomis L., Pogostemon, Sideritis L., and Teucrium L. in subfam. Lamioideae. They were infrequently encountered in subfam. Viticoideae and tribes Citharexyleae. Prostanthereae, and Mentheae and were not found at all in the Symphorematoideae or the Viticeae. Type | glands were present in only six genera of Verbenaceae, four of them in the Lantaneae, but a wider distribution in the Verbenaceae was reported by Robert (1912; discussed below). In the Labiatae they were encountered fre- quently in tribe Mentheae and moderately so elsewhere in subfam. Nepetoi- deae. In subfam. Lamioideae they were found only in Anisomeles, Eusteralis, and Pogostemon. Type 3 glands were observed in about a third of the genera of both families but were never very common when present, were rarely found in all examined species of a genus, and were only found when type 2 or 4 was present as well. Type 3 is probably an occasional derivative of the ontogenetic pathways that lead to types 2 and 4 glands and, as such, may be expected to occur irregularly in any species in which type 2 or 4 is common. Type 6 glands are much more frequent in the Labiatae than in the Verbena- ceae, where they were found in only ten genera, usually in only one species per genus. In the Labiatae they were most frequently present in subfam. Nepeto- ideae, although rare in tribe Ocimeae. In subfam. Lamioideae they were ob- 344 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 3a O-OD- - 1 2 Ficure 3. Hypothesized transformation series of subsessile glandular trichome types 1-10. (Figure originally published in Abu-Asab & Cantino, 1987a.) served in over a third of the examined genera, but usually in only a single species. Type 7 glands were found in nine species of Labiatae scattered through six tribes, no two species in the same genus. None were found in the Verbenaceae. Never common, type 7 glands occur only when type 6 is present as well and are best viewed as a rare derivative of the ontogenetic pathway that produces type 6 glands. Type 8 glands were found in 11 genera of Verbenaceae and 20 of Lamioideae but in only one genus of Nepetoideae. They were seldom observed throughout a genus. The only genera in which type 8 glands appear to be characteristic (1.¢., present in most or all examined species or, if only one species was ex- amined, then the most common gland type in that species) are Callicarpa L. and Citharexylum L. in the Verbenaceae and Brazoria Engelm. ex A. Gray, Craniotome Reichb., Cymaria Bentham, Prostanthera, Scutellaria, and Tinnea Kotschy & Peyr. in the Labiatae. Type 9 glands are also much more common in the Labiatae than in the Verbenaceae. In the Labiatae they occur most frequently in the Mentheae and least frequently in the Ajugeae, the Ocimeae, and the Pogostemoneae. Based on their structure, it appears that type 9 glands may arise via two different ontogenetic pathways: through the development of one or more tangential walls in what would otherwise be a type 8 gland, or the development of tertiary radial walls in what would otherwise be a type 6 gland (see FiGurE 3). Based on the co-occurrence of type 6 or 8 glands with type 9 it appears likely that type 9 glands have arisen via the type 6 pathway in subfam. Nepetoideae, where type 8 glands are very rare, as well as in some genera of the Lamieae. They have apparently arisen via the type 8 pathway in a scattering of Verbenaceae and Lamuoideae (¢.g., Duranta, Tinnea). However, there are many species in which type 9 glands are associated with both types 6 and 8 or neither, preventing indirect inference of the developmental pathway. Type 10 glands were found in five species of Verbenaceae scattered among four genera. In the Labiatae they were encountered in about 20 percent of the 1990] CANTINO, STOMATA AND TRICHOMES 345 genera of both subfamilies but usually not throughout a genus. The only genera in which type 10 glands are “characteristic” (defined above in discussion of type 8 glands) are Brazoria, Hemiandra R. Br., and Macbridea Elliott ex Nutt. in the Labiatae and Ho/mskioldia Retz. in the Verbenaceae. In general, type 10 glands probably develop as an extension of the type 9 pathway (see FIGURE 3), but they may also develop from type 7 glands through the addition of tertiary radial walls. This may be the case in Hemigenia saligna, where types 7 and 10 glands occur but not type 9. Type 11 glands are broad and scalelike, varying greatly in size even on a single leaf but always much larger than the other subsessile glands. Viewed with a dissecting microscope, they are yellowish or brownish and sometimes glistening (and therefore presumably glandular), but it is unclear whether they are fundamentally similar to the other subsessile gland types (hence their ex- clusion from FiGurE 3). They occur only in the Verbenaceae, where they were encountered in 30 percent of the genera but often in only one species per genus. They are most widespread in the Viticoideae and the Citharexyleae and were not found in the Chloanthoideae or the Verbeneae. Because they are sparse when present at all (rarely more than three seen on a slide), they may have been overlooked in some species and thus be of wider occurrence than TABLE 2 suggests. Glands of similar construction were documented by Robert (1912) in some species of Clerodendrum, Duranta, Faradaya F. Mueller, Lippia L., and Stachytarpheta. Their functional significance is unclear, but Fedorowicz (1916) referred to similar structures in Me/lampyrum L. (Scrophulariaceae) as extrafloral nectaries. Metcalfe and Chalk (1950) stated that extrafloral nectaries are common in Clerodendrum. NONGLANDULAR [TRICHOMES Nonglandular trichomes were found in most species (see TABLE 2). Species with glabrous leaves (subsessile glandular trichomes are ignored in the defi- nition of “glabrous” used here) were encountered in nearly a quarter of the genera of Verbenaceae examined but only one tenth of the genera of Labiatae. In the latter glabrousness is commonest in the Prostanthereae and rarest in the Ajugeae and the Nepetoideae. The leaves of the vast majority of Labiatae and nonverbenoid Verbenaceae bear simple, multicellular (i.c., uniseriate) “hairs.” The rarity of these trichomes in the Verbenoideae and the Prostanthereae is, in contrast, noteworthy. Ex- cluding these two groups, nearly all other Verbenaceae and Labiatae that lack uniseriate hairs are either glabrous or bear branched, multicellular trichomes, which presumably evolved from uniseriate hairs. In most nonglabrous Pros- tanthereae and Verbenoideae, however, the hairlike trichomes are unicellular. Unicellular hairs are widespread in other groups of the Labiatae and the Ver- benaceae as well; thus it is the absence of multicellular hairs in nonglabrous species rather than the presence of unicellular ones that characterizes the Ver- benoideae and the Prostanthereae. Branched, multicellullar trichomes were found in 20 percent of the genera of Verbenaceae and 16 percent of the genera of Labiatae examined. They are 346 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 TABLE 2. Distribution of nonglandular trichomes and subsessile glands in Labiatae and Verbena ceae. Taxa? Subsessile Glands? Nonglandular‘® 123456789 10 11 su sm bu bm Lamiaceae i vueeane alk "Ajugeae Acrymia ajugiflora Prain - e+ + + - ee + - = Ajuga a (L.) Schreber - +t t+ eo ae Pa car ae A. genevensis ee eee ee ee, a a noe Ss & A. danmannit ee a Won pe Fe eae = ae A. linear ole Pampan. ao pb oe os 4c 7 A. see ea or AE ae es a eee aS ee A. ona W. W.Smith ok ee ee ae S gb oe 38 thystea coerulea L. (#1) - - - * + ee ae A. CTT L. (#2) ee i Be ae ope 4) yp ekg Cymaria dichotoma Benth. - = - ++ * - - f+ = « eas ningpoensis (Hemsley) Kudo - ++ * - - > - - + = = K. atum (Hemsley) Kudo (#1) -* + * - Se ae +a hs K. ornatum (Hemsley) Kudo (#2) - t+ * + - +e + + - - K. ornatum (Hemsley) Kudo (#3) - ++ * - - 2: t+ + -) - K. pernyi (Franchet) Kudo - + + * + -- - + + - - Rubiteucris palmata (Benth. ex Hook.f.) Kudo oe = eee ye aap is “ Schnabelia oligophylla Hand. -Mazz. - - + * - eee “fs Tetraclea coulteri A.Gray (#1) -- - + % ss eae dc ee T. coulteri A.Gray (#2) - - - + i Gas + 4s 4, os T. coulteri A.Gray (#3) at + * ene SE ee ee Teucrium a Liz -++%* - - 2 - + + - - T. canadense g - * + ¥* Sg ed 3S oe eee as chamaedrys as Be, ae he oa Ge ASS r erg : eae ex Benth. -+-%*- - - - - + - - T. laciniatum To Se ee A 2. + boos uy marum L. Se Safes cir, AIR ar ee + + = = DY reer ere Schreber - ee + + - - - - + - - ae aethiopica Kotschy ex Hook.f. (#1) - ee e+ w+ - - +t = - T. aethiopica Kotschy ex Hook.f. (#2) -+% ==; S ae ee T pieiecorhae ee a Welw. . eee we ee eee ys) T. apiculata W.Robyns & Lebrun a oe ee ime das, oa T. galpinii Briq. a es se OS Pa oa 2 ay yer ne phe chodesiana S.Moore (#1) -- - ++ * +4 - a oak wa oS rT. alensis Gurke ex Chiov. 7 ae + *& + ce Se eee arizonicum A.Gray (#1) - ++ 8 + - - - + + - - T. arizonicum A.Gray (#2) - ++ %* - --- a 4s GES - Sera Ls - + - & + ares oe a Ce ae T. dichotomum L.3 ee a = es So° Hi- er T. lanatum Benth SP ah me eee es cee ® a2 ° 4 2 T. lanceolatum Benth.$ : ee oe oa = De. ils latum Benth - + 2 & os ris} a se T. setaceum Houtt Se 2 a ae eee Lamie Achyrospe rmum parviflorum S.Moore + * ap ae a ae ee ae . schimperi Perkins ah ee ae gee Se . ss a wallichianum Benth. ex Hook.f. ve Pawn? eae eae + + - = Acrotome Fheckit (Gurke) Launert - + - 4+ %* ++ - + + - - A. hispida Ben ero ha ee OE, 22.9 i ss es OE A. inflata Bent . 6 a ep ee te , a ae 1990] CANTINO, STOMATA AND TRICHOMES 347 TABLE 2. (continued). Taxa® Subsessile Glands? Nonglandular® 12345678910 11 su sm bu bm Ballota frutescens (L.) J.Woods a 2 a ae B. nigra L. (#1) a + t+ - - B. pseudodictamnus (L.) Benth. 777747777? =? er “et, a Brazoria arenaria Lundell! ae a ee ee ee, B. pulcherrima Lunde 118 ay Se Re eS = Ae Ss B. scutellarioides Engelm. & A. Gray$ ee ee + + - - B. truncata (Benth.) Engelm. & A.Gr ay’ - - - tReet t+ - + - = Chamaesphacos ilicifolius Schrenk ee - + - = aaa ee fo er are _Anthony® - + - * + - - a Ae a ve Cc. ichiangensis W.Smith £ Sec, OR ee ee eS ae c. cee Makino a oe: eee C. odontochila Diels Ee alone even ee ee ee o “se Ge 2 Co olquhounia coccinea Wallich - - -+4+77777 - S yer “a <2 C. seguinii Van oO oe Mae Se ce a rs oe furcata (Link) Kuntze -- - ++ 77747 - es Cc. ver olor Reichb. (#3) eee oak Gy ee eR ae ce Sa. a Soe bachardenica B.Fedtsch. a - + - - E. iliensis Regel eRe eee a See ee 4 “a & E. 1 labiosa Bu nge De DO OF OE gee 6 se E. regeliana Aitch. & Hemsley -~ + -*% + - = = ee = SB ee Se E. speciosa Rupr a a a oe ee: . oe oe ae Eriophyton wallichianum Benth. 2 a tee je 3d Galeobdolon luteum Hudson? fay. Bo ge We HE eee Soo &. se er Galeopsis ladanum L. oD See Sa eS & Ge 4er 4e G. ochroleuca Lam. a, Se ee a et oe fe ee G. BEBeeeric Besser oa: joe BY a oe ce ae G. tetrahit L. - 2 - HH ee ee 7 ae. 4k +e . Gomphostemma lane 8 cor ee + + - + G. javanicum (Blum nth. -- = +77777 7 - - + - + G. lucidum Wallic ‘aa 2 aes te TS, fet ae Ges Hip a ae G. velutinum Benth. 7272272472? 2? ? ee oe ee haplostachya (A.Gray) John - *®¥ +--+ -82 - = “ ie a = He eee (Drake) Sherff 7 ?P+++7777 7? - - +t - - Lagochilus cabulicus Sy dey he th fae 5 #- He 12, 3 Lg iacanthophy}lus | pene Benth. 7774+ 7744+ 7? - + + - - Le aos - ete K+ Bea Ae a ae Lis ee Te a 2 See, ee, ee OS e~ ar es L. platycalyx ee & C.Meyer (#2) - 7 7 w+ --*%- - $2 St Lagopsis supina (Stephan) Ikonn.-Gal. a ee a + + - = Lamiophlomis rotata (Benth.) Kudo a a a Lamium album a ee ee ee er fe Sep L. maculatum (#1) ee OR Ss ee SS Pr oa maculatum L. (#2) Dp te 2 a ee we Me ek L. moschatum Mil Se ey ee Se eS oe L. pictum Boiss. & Heldr oe fa et ee a ey ee . a ee purpureum 3 -++%*%% ++--- ra + + % 2 Leonotis Lee Gurke a - + - - L. leo 4 ee Rk Oe ee “te Ss & L. ie rus ee ) R, Br. (#1) - - - * + aS ee es ye “ee L. leonurus (L.) R.Br. (#2) a ae ep ee et ayes a 8 348 JOURNAL OF THE ARNOLD ARBORETUM TABLE 2. (continued). [VoL. 7] Taxa® Subsessile Glands? 234567891011 su Nonglandular sm bu bm c nurus cardiaca Leonurus L. heterophyllus cae eucas altissima Engl Irie ie ie it Ls a. capensis ora ) Engl. ci mollissima Wallich Loxocalyx urticifolius —— Macbridea alba Chapm M. caroliniana rine Blake$ Marrubium desertii Noe ex Cosson M. peregrinum L M. vulgare M. vulgare L. L. § Melittis melissophyllum L.§ Metastachydium sagittatum (Regel) C.Y.Wu & Li ef B ba e-gia-dic4 u sae cone sta He guelnensis Franchet ex Prain sle rticifo lia Hemsley Mo eer paevis Lis M. spin Notochaete hamosa Benth. Otos stegia aucheri Boiss. persica (Burm £. ) Panzeria argyracea Kuprian. Paraphlomis javanica (Blume) Prain ex B Phlomidoschema parviflorum (Benth.) Vved. acker & Bakh.f. P. rugosa (Benth.) Prain P. c P. herba- L P. maximovicz P. pratensis P. setigera c P. taurica Hartw P. tuberosa P. umbrosa Taree. Phyllostegia b Phlomis agraria Bunge bracteosa Royle revidens Drevidens ray P. grandiflora TTS ) ee P. grandiflora P hispida Hilleb 1 a antanoides She | ch.) Benth. nd crotoena insuavis (Hance) Prain ex 1 ex Hook.f. s (#1) (#2) ? ++ 45 ? Set HKHHHK wt te et *+ % % + % % e+ + et + % + + ~~ + tired +4 t+ettg ++ Feet + ++ee t++et + t++eeteegGgi 1990] CANTINO, STOMATA AND TRICHOMES 349 TABLE 2. (continued). Taxa? Subsessile Glands? Nonglandular® 1234567891011 su sm bu bm Physostegia angustifolia Fern.$ eos Efe eb tHE a P. digitalis Small pes, hp pS MR Se eh ee ee a, ho oe P. godfreyi Cantin ee ee eee o a oe P. leptophylla Small Sik ee a iy: Sek es Ge oe. oh =e P. longisepala 2 ean Gk 2.8 ee: Ss ee ua,’ te PB purpurea (Walter) Blake?’ aiden ot Ae a S. tenuifolia Willd. be teeta GN he Ee Lee. Tee Ca Stenogyne diffusa A.Gray (#1) ecu prada? et Sh ee 8 Be es oe S. di fase. A.Gray (#2) ee ay PORE ee ee ei oe ae go as es. he S. kamehamehae ee a oe. a a oe Saepuree H.Mann ae ys a ae ee, ae Pr a ee S. rugosa Benth. ee ee ee ee ee ee SS a ae Ss Synandra hispidula (Michx.) Baillon’ - - = KH tH - + - = - “ s Sa a brahuica — ) Briq. a - + - = T. persica (Boiss.) Br a a ee Wiedemannia multifida (L.) Benth. ee + + - = Pogostemoneae Colebrookea oppositifolia Smith (#1) 72777477474 - + + - = C. oppositifolia Smith (#2) oa Calg OI tate ak Se et a i a a oa Comanthosphace stetiipila S. ac Sek rer a Ve eS S oa & Cc. sub}anceotata (M oe Si eee i) - - - te 2 2 - = = - + - = C. sublanceolata (Miq.) S.Moore (#2) el lap. Tet RE aes al, a Ge Ges e- Yeo ot. oh fusteralis cruciata (Benth.) Panigr. i WA et ae ee, SS ee See oe E. sampsonii (Hance) Panigr. 94299 99 2 FO - Be E. steltata (Lour.) Panigr. (#1) i - - 7 a oe E. ellata (Lour.) Panigr. (#2) 4 Oe ae. Be Se ae 250 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 TABLE 2. (continued). Taxa® Subsessile Glands? Nonglandular® 1234567891011 su sm bu bm Pogostemon auricularius aie ) Hassk., th [v) he nn r > eee ee Lt H+ HK : ' ' ‘ ' ' ' i ae Se t+etetete ; Rostrinucula dependens ae Kudo - 2 - tH He - - - eK no R. sinensis (Hemsley) C.Y.W aay Ge’ cia ame ae’ Ses ae Ae ke as - - - + Prostanthereae Hemiandra pungens R.Br. a a a ae Hemigenia incana Benth. (#1) ee ae ee eee ee ae incana Benth. (#2) Se AOE Che ct ee an: ae Perr Diels hi tee ee ey SR ene eR oe ee ee Microcorys brevidens Benth. ee es oe! ee ee ae a ee oe Se rostanthe cuneata Be - - - ? hea Ene Labill. a - oe oe lasianthos Labill. (#2) - 7 nivea Cunn. ex Benth. (#1) - - nivea Cunn. ex Benth. (#2) - - Br. Poser t Westringia amabilis J.Boivin ee ee - ee W. cheelii Maiden & Betche ei ae Meee a OS et ee W. fruticosa (Willd.) Druce Behe ie De ee, yee Scutellari Salazaria mexicana Torrey (#1) a a Se S. mexicana ee (#2) Pn ee ea ea + + ' cutellaria amoena C.H.Wright - + . elliptica Muhlenb.® os e : on = lay ct o — — > ~ ~ ) Daoud Lael Lae oy io |s + foo aa tI< @ |o . aint Hn = [R = Ios Ph Ie |e - olor jo Cie Ie oO Cle nm ros a > HN |o eS: (my ow ' eet e+ tte H 3 fag fa ct ie. fe) D> Ee he Le wn w a) s n n \ + ' ; ee a bot ' ' ' ' tet teetetee ; ; om + + ata Andrz. - Uncertain Tri : Ajugoides eee oe iq.) Makino (#1) - tte - - - ee ee + + - = A. humilis (Miq.) Makino (#2) ae ee ee a Pee Se Anisomeles heyneana Benth A ic . abari i .Br. ex Sims (#1) : SE TTIOe (L.) R.Br. ex Sims (#2) [> [> I> | a (5 Pealoe he ie) > |f> |R Hen or ist) —- e x ae ing = se NR ee ee t++ttt : P , : : ; : \ \ ++ete et : Eurysolen gracilis Prain (#1) Bek eee Hee Pane ee Oe 3s ee E. gracilis Prain (#2) 7777447777 - to + - - Hypogomphia turkestana Bunge Se le es fe ee Se Se ee S.-ae e Leucosceptrum canum Smith CU Pee 22 2 2 - = = + Suzukia shikikunensis Kudo 27? +4+77777? - - + - = 1990] CANTINO, STOMATA AND TRICHOMES col TABLE 2. (continued). Taxa? Subsessile Glands” Nonglandular® 12 3 6789 1011 su sm bu bm Nepetoideae "Mentheae" Ceratominthe odora (Griseb.) Hauman Ke ee Pe ee a ee oe Cyclotrichium origanifolium (Labill.) Manden. & Scheng. - - = - - & = = +H = % @ ce Elsholtzia fruticosa (D.Don) Rehder -* +7777? 4+? - + + - + patrinii (Lepechin) Garcke -*+-72777FF4+? - a A Aes Se Hedeoma graveolens — ex A.Gray ee + + = = H. nanum (Torrey) Br oS ee IN PD BF De = $e we se Keiskea japonica Miq. So ok ROR BO Qe 2 eS ae des Lepechinia hastata (A.Gray) Epling t- - - © $e ee ee a er Lycopus ame erfeanus Muhlenb. - +--+ *® - - = ee . Sp: sks ~ L. rubetius Moench OF Dive: OR es ee ay Be Se 1 virginicus L. aOR a Bese oe eS Sk do he a, Melissa officinalis L. (#1) ene a ee ee ees. 4 ye 8 Mentha arvensis L. 299777?+7777 - es a, ey M. citrata Ehrh. BP BED BD DE = me we oY M. piperita L. 2727727+4+4+7777 - a Monardella odoratissima Benth. a ee a oa M. villosa Benth. fees PREP T+? = Bi se Saree reptans Maxim. (#1) ee ee ee oo reptans Maxim. (#2) poe eee a Sy (ay “a ae Pogogyne zizyphoroides Benth. i t+ - - = Poliomintha glabrescens A.Gray ex Hems]. wee - 272 2 2? - + b= = ae floridanum E. aa & Epling - - - - * 2772772? - qe Ss aff. incanum (L.) Micha Sen Ae eS ek’ ey ay : a & -s Rhododon ciliatus (Benth.) Epling * ee 2 +7777? = ee! See Satureja arkansana (Nutt.) Briq. *- - - 47777 ~«-- 2 S. douglasii (Benth.) Briq. a er He ik ae OS s arvifolia (Philippi) Epling Same OU any ny ae a a He s i B.Fedtsch. & Gontch eS Se, Jo oop 2 roe ane ee Thymus serpyllum L. 4+$77774+7777 - te ee Ge: eee Aga poeactis ae ees (A.Gray) Epling -+--+4+ 7477? - - + = = A. cana (Hook.) Wooton & Standley -* - - 7? +77?+? - + + - = A. nepetoides (L.) Kuntze 7277747777? - + + - = A. pallidiflora (A. A.Heller) Rydb. 7 ae Cy Oa Ja CR a a Ca ga > gd ue 8 Cedronella canariensis (L.) Webb & Berth. (#1) ne a t+ + - = pracocephatua oe Benth. - * - + - ++ - - - - + + = = D. pa Ear i faa ae OY Ga RY ct Ge? Cae ca + + - - Bo ecgee ‘ sot, We Hey as Sd A ke, ac es OR, Ve qe J. uh Glechoma hederacea L. rE os te, FA eee LS ne er Lophanthus chinensis Benth. a fee 2 Sd Meehania cordata (Nutt.) Britton Vai On cae <6 Oc Acrocephalus fruticosus Dunn -~ +--+ *%® + - H+ - + + - + indicus Kuntze a ee +. oho Ge Becium obovatum (E.Meyer) N.E.Br. a ee ee ee ee - + = = Capitanya otostegioides Gurke 72777? 447777 - ee ee ee Catopheria capitata Benth. ex Hemsley (#1) oe eee See eS Sse ae c capitata B Benth. ex Hemsley (}2) tee + - - ee ee a C. chiapensis A.Gray ex Benth. WE SAO ode dee ks Me Ge 4 ya Endostemon obtusifolius (E.Meyer) N.E.Br. os eh we Sees a a ee Eriope crassipes Benth. aaa aR ae A a Ca e/a Se ee: See Eriopidion strictum (Benth.) R.Harley 2? * - + - - - - ee + + - - Fuerstia africana T.C.E.Fries 2+ 27 be? 2222? = - +t + eo as callianthum (Briq.) arley Sas EAGkY ee eg See Ae = “A a; ce Hi, ee (Oliver) Duvign. & Plancke - *®& - k*& - 2 ee ee ee dh, 2g. Hemizygia canescens (Gurke) Ashby -* - 2 ee ee ee + + - - Holostylon katangense W.Robyns et Lebrun a + + - - H. strictipes G.Taylor te Se AR a ee ces es = 2 @ = -« Hyptis alata (Raf.) Shinn. Gives Se ME ok ay ra a oe H. emoryi Torre \ an aan ae te ae ee a ae — «BOM H. mutabilis (Rich.) Briq. 4s UE es ee as = ope oblongifolia Benth. oe ee OE Ae re eS ms Tcomum lineare Burkill +*¥ +777 72777? - Ser el 4S Lavandula multifida L. 6 4 eA a es Bee a <2. 4 L. stoechas L ake te, “Oe Ge ies leh ae, te ie 2 2 ae, ES Nautochilus labiatus (N.E.Br.) Bremek. a - + - - Ocimum STS Saas L. a ee ee Chk £ af os oS O. basilicum L. BN GS ele se Se eS oie QO. SeSen rn L. (#1) ae Ae sok) or kre oe % if. Ss OQ. gratissimum L. (#2) ie ee a erthesfphen affinis N.E.Br. Sy ite nek gels iB ge Eee So Se 1c aristatus a Miq ef A ises Speech, as Met We ek kee 8 ce rs armen usus oe Pee Se Ay a ee oe Plectranthastrum clerodendroides T.C.E Fries ane ee a! ee? a a a a ie ee Plectranthus forsteri Benth. a a ae ee ee ee S- & P. scutellarioides (L.) R.Br. ~ t+ + - = - = = = a 1990] CANTINO, STOMATA AND TRICHOMES 353 TABLE 2. (continued). Taxa® Subsessile Glands? Nonglandular® 234567891011 m bu bm Rabdosia excisa (Maxim.) H.Hara ee ee Be Ses Yer <5 R. inflexa (Thunb.) H.Hara Se het Se oy es es Ss Ge Ss a 2 R. nervosa (Hemsley) C.Y.Wu & Li - t+ *¥% - - - - - - + - = Rabdosiella calycina (Benth.) Codd sel oe ge NE Se ee Ge ee ae Solenostemon scutella ides (L.) Codd (#1) - + - * + - - - - - 7 a” a hie Stern ETT EEE (L) "ada (#2) OS dock eh te? Gee a Syncolostemon densiflorus Benth. -+-*¥ - - - - - - a ee "Salvieae" Arischrada bucharica (Popov) Pobed. a ee eo - +t - = Blephilia hirsuta (Pursh) Benth. ° os 2 PEP 2 ee + + - - Monarda cline podia L. ?+774+7777 7? - + - - perute 777747777? + + - - M. pun i 27??? +7722727? + +t - - Perovskia abrotanoides Karelin +++ 44+ * - - * * + - = + P. atriplicifolia Benth. -*¥t++tet- - + - = ss Rosmarinus officinalis L. 222? 2?2?°72?7?2 2? ? 7? 2 - + Salvia carnosa Douglas eee ee ee a qo ww ee < Ss. farinacea Benth. Sa ee Be eel a S. reflexa Hornem SV a oe eS S ste ta [5 Verbenaceae Avicennioide Avicennia eee Jacq. re ee aryopteridoideae ees terideae aryopteris divaricata (Siebold & Zucc.) Maxim 27? + +7777? - + - - Cc. forrestii Diels (#1) 779%*272272772? 4. as ak C. forrestii Diels (#2) 777% 27777? + + - - C. grata Benth. See a a ee Ss es: 2 C. incana (Thunb (#2) 2747477722? es ee C. mongholica Bunge (# oa a ae Sa’ a a ck’ ok’ ae 2 (ae C. nepetaefolia (Benth.) Maxim Sys cn ee TR Ho ee ase - a eee C. odorata (Ham. ) b ee ee ee ee a. Sheu es, Cc. terniflora Maxim + x. 2 oe eee Glossocarya siamensis Craib - +++ * + - - - = + ae Peronema canescens Jack PD ie PD eo ee 4 ah - ee Petraeovitex kinabaluensis Munir - 2 eo t+ B® ee ee oe 4 ee P. multiflora (Smith) Merr er a ene ae ee ae a ee geleuaneaen es endr Tei Ais mannio a endeon ahernianum (Merr. ) : -~+t+---+ %* a ke ee (H.Hallier) Kosterm. a ee te Se OE ee. Se a) se oe pene ean ae Acha “Nesosenes dupontii Hemsley wilco, Sie Be ae ac oa ee ee ie euphrasioides (Hook. & Arn.) A.DC. a ae ee Pityrodia atr iplicin na (F.Muell.) Benth. DP chee: BOP Be Pee oo eh ee Sb P. dilatata (F.Muell.) Benth. 277 +4+77277? - = - + P. paniculata . Mae: ) Benth. - *& + - - - 2 eee ee ee 354 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 TABLE 2. (continued). Taxa? Subsessile Glands? Nonglandular® su sm bu b 1234567891011 —— puberula (F.Muell.) e M n & Betche (#1) ? Pe Bi BP Pa). 2 + + - + s npeniie ( uell.) Maiden & Betche (#2) eos ete, Se ee eR + + - + S ber (F.Muell.) Mai n & Betche (#3) Ge Reh Ane al, “Wee ls eae hee S - . + Chloantheae Chloanthes stoechadis R.Br. ROR RR BP ER 2 a a Cyan egia angustifolia Turcz. 2?2?2?2? +2772??? ? - oe ee c. eee S.Mo 2277? +727727? ? a ae Physopsi ees Phe exsuccosa (F.Muell.) Druce - * +477? - - - = - a 2 Lachnostachys eriobotrya (F.Muell.) Druce ? 2??? 2? ?7777?7 =? a Mallophora globiflora Endl. PPR Pee ee Re ee 8 - oe t+ Newcastelia cephalantha F.Muell. he Seon tee ey a tee ee ee eee baemndetnee Nyc peanthee: oe Cristis Lb. +++ %4+-----~ =- a gs & Phrymoideae Phryma leptostachya L. ee ee = 5 & Aegina nge ee Mold Bi veat nee eather Bete ee se ae ae cask Cc forbeets oa ng & Gamble i fae fee I as, vee a ee EE 2 4b & 5 Cc. mentos ages, OR ye ae eee, oe wy ces. AE Sphenodesme ferruginea (Griffith) Briq. 9 A ant er Sat Pa amg SEN ke Col CoM - + - + S. pentandra Jack Se eee oe ee ee ee o a & Symphorema luzonicum (Blanco) Fernandez -Villar CED See a ote = i Ss Verbenoide eae 1 ' ‘Giz ligustrinum Van Houtte Sear oe C€. punctatum Greenman Bet i ie lol ie) + 1D a. js) 5 ' + eH + a +o Heer \ +1 +e : f ; Duranta mandonii Mold. oe mutisii L.f. eee Sere Mold. a ets ee SEY repens L. (#1) et Ok i. repens L. (#2) ee IMIS le Rehdera trinervis (S.F.Blake) Mold. a - 2 + - Rhaphithamnus spinosus (A.L.Juss.) Mold, no + - - - Lantan Aloysia gratissima (Gill. & coe Tronc. a + - - = A. wrightii (A.Gray) A.A.Helle Kier see OS ee eS oe om Os Bouchea prismatica (L.) Kuntze - *t+ttee - - - - e + - - = Diostea juncea (Gillies & Hook.) Miers 72:7 2 P4279 ee + = = = Lantana horrida Kunth (#1) Rit AS eee leer e, Le a Ne ee te L. horrida Kunth (#2) Se ee Ce a ee a 4 ea ae - Je L. involucrata L. te ee ee ee oe 4 = & 1990] CANTINO, STOMATA AND TRICHOMES TABLE 2. (continued). Taxa® Subsessile Glands Nonglandular® 1234567 9 10 11 su sm bu bm Lippia graveolens Kunth OR Shes ont fee ct Jes ey, i Se, Phyla incisa ll i ous Caan ees Gas Cn me VV PB; ee (Michaui) ee Greene! e+ - +t - - - - - ee PB. nodiflora (L.) E.Gre 7474777777 | priwa aspera Kunt ee GA ees ep ver Ss P. grandiflora en Mold. oR eae ee Stachytarpheta frantzii Polak. ~- *®& - ee ee eee S. jamaicensis (L.) Vahl Neuter St ye cee ee Ser YS Monochileae Amasonia campestris (Aublet) Mold. i irta Benth. | Pa ie ae ee ee ae a CA Petreeae Casselia hymenocalyx Briq. i a a Petres arborea Kunth 1 CUR ae lie et amet Ga ca ca GR Ca P. volubilis L. (#1) nt et TR Oh Re ce eS Pi caleba lis L. (#2) ha, eh ee Ce ea ie ay Recordia boliviana Mold. ye ay oR eet age (as a ae et Verbe Glandularia bipinnatifida (Nutt.) Nutt. a G: nadensis (L.) Small oe, Mee Se ee ee Hierobotana inflata (Kunth.) Briq. 2 BBD PED DROS 2 Junellia ligustrina (Lagascana) Mold. - + - - = ee ee ee Tamonea curassavica (L.) Pers. SRS ae ge ee Qe es ee Verbena hastata L.! ie nia ee ee ee V. litoralis Kunth ai ole ane a 7 ay Sas V. macdougalii A.A.Heller a eee ee ee ee V. pumila Rydb. Se ae ee ae eee V. urticifolia L. § moe i i es a Be tae Viticoideae Callicarpeae Aegiphila aculeifera Mold. oe 2 Gh eS A. deppeana Steudel ote ee Se oe oe A. pendula Mold. oe aoe ee ee ee ey Callicarpa americana L. ee ee ee a | c. dichotona (Lour.) K.Koch oe eae se gt Be aia C. mollis Siebold & Zucc. am ie Sey Se es) a a Geunsia farinosa Blume Cae dle Cae lia A Coy ae a ae a Clerodendre Clerode fecon actleatum (L.) Schldl. a ee a a G. anafense Br n & P,Wilson Sa eddyi ge Sh OR i ey cesta Cc erin cn wee ) Schum. & Thonn. oo RS To Cc. cuneatum Gurke ae * _ 4 C. floribundum R.Br. ose Ee ee Re eS C. glabrum E.Meyer sah Ree er Ba od, Ga toe See ae! C. inerme (L.) Gaertner se et ee AE ie es ao C. philippinum Schau eee ee ee a ee C. squiresii Merr. he Se See ed a va od + ' ' 1 Faradaya amicorum ee Seemann - sos F. splendida F.Mu BR Hee go ++ ttt 356 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 7] TABLE 2. (continued). Taxa® Subsessile Glands? Nonglandular® 123456789 10 11 su sm bu bm Holmskioldia oe Retz. (#1) -- +--+ tee eke = ks x H. sanguinea Retz. (#2) a a + - = H. nguinea Se (#3) 2.5 5. eb et aoe a. 2 Sa « Hi eee (Klotzsch) Vatke - = - - te - - HH | + + - = Kalaharia spinescens Gurke - 2 - - -- + - + + ar 7 Karomia fragrans Dop Sif oe eae, oe oo se Se ae morierli Sema o oS oo Oe MH ee ce a er rlifolia Bea Sb) as as ESB se ses Se . = = & auteurs Dubar oe ee do ee er = fs wow Teucridium parvifolium Hook.f. -++%* - - - - = - - + - - Tecto Pe cies. domingensis Jacq. ee A ROU ee et A er tee eee er, SE ev. = Vitic Sane = Erandiflore (Cham, & Schldl.) Schau 72727477777 - ae Coe Cc. canines L. eS ne ee 4s. fi ns Garrettia siamensis Fletcher ne - + - = Gneline delavayana Do ??++ +77? 1 ee 2 Ss moluccana (Blume) paciee SS eet ee i, 12s + ae Premna barbata Wallich ey ee A ed ee a abe & ts P. corymbosa Willd ee ee et eee ee Pe ee P. foetida Reinw. . SS eee Se ee OS ee i a, PB. Japontca Miq. i Je. oe, Ge te ee hy eis? «fe ae! ee P. octonervia Merr. & Metcalf Sica aa tay ee eae ve ee AB 4 aks Se avicennioides (A.Rich.) ills 7724777777 - a 6 4. 46 P. mined Millsp. at eh kee dan ee ee Tsoongia axillariflora Merr. a + + - | Vitex agnus-castus So gis. ee ea ee ay ae . a ee” V. cannabifolia ie & Zucc. i a + + - = "If name is followed by the oo ol a species; the n major botanical Libr ries (see §, data are ria ge Asab and ; alae to in the ruguenes list deposited in Materials and oe — Bead ares in Appendix 1. , presen unknown whether eee or oa < aes trichomes). El Sasiaee roti richom su bu, #/2, refer Methods). Symbols: *, commonest type pre meee (1987a). All different specimens of four e(s); rx present or eck (glands are poorly simple unicellular anched uunfeel ular; bm, branched qietceliui ar. Symbols ir nce +, pre abse simple ja aragniera a Sisal most widespread in the Chloanthoideae and were not oe in ia Rela deae, Ajugeae, Prostanthereae, Scutellarieae, or Nepe (Conn, . However, Robe (1912) reported them in some species of a eae ee and within the Prostanthereae they are known from two species of Prostanthera 1984) and at least two species of Hemigenia R. Br. (Bentham, 1870; rt B. Conn, pers. comm.). Outside the Chloanthoideae, genera whose leaves are ; 1990] CANTINO, STOMATA AND TRICHOMES 357 characteristically clothed with branched, multicellular trichomes include Co- manthosphace, Gomphostemma Bentham, Leucosceptrum, Marrubium L., Pe- rovskia Karelin, Phlomis, and Rostrinucula (see TABLE 2; Bokhari & Hedge, 1971: Azizian & Cutler, 1982; Press, 1982). Branched, unicellular (two-armed) trichomes were found only in Phy/a. DISCUSSION SAMPLE SIZE LIMITATIONS Because this survey was motivated by an interest in phylogenetic relation- ships among genera and suprageneric groups, particularly within subfam. La- mioideae and related Verbenaceae, taxonomic breadth of coverage was em- phasized at the expense of depth. Very few species are represented by more than a single specimen. Consequently, the data in the tables should not be used to characterize species or infer interspecific relationships, although they may suggest avenues for further research in certain genera. However, the sample sizes for many genera of Lamioideae and for suprageneric groups in both families are sufficient to provide a general picture of the distribution of trichome and stomatal types in the Lamiales. In most genera of subfam. Nepetoideae, the number of species sampled 1s too low to be any more than suggestive about the epidermal anatomy of the genus. Although perhaps disappointing to those whose primary interest lies in this subfamily, the shallow sampling of this group 1s justifiable in relation to the ultimate objective of the survey—an improved understanding of the origin of the Labiatae and of relationships among its basal clades (which lie within the paraphyletic or polyphyletic subfam. Lamioideae). Although subfam. Ne- petoideae includes well over half the genera of Labiatae, 1t represents but a single clade (Cantino & Sanders, 1986) whose closest relatives lie somewhere within subfam. Lamioideae. Knowledge of its character-variation pattern is therefore no more nor less critical to an understanding of the origin and early evolution of the Labiatae than is that of any single genus in the Lamioideae. COMPARISON WITH PUBLISHED DATA There has been no broad survey of the morphology of trichomes in the Labiatae or the morphology of stomata in the Verbenaceae. El-Gazzar and Watson (1968) investigated stomatal configurations in a wide range of Labiatae but listed only the predominant type in each genus. Furthermore, their obser- vations regarding many genera of the Lamioideae conflict markedly with my own (discussed below). Robert’s (1912) study of trichome morphology in the Verbenaceae provides extensive data on the nonglandular trichomes and cap- itate glandular trichomes of some 55 genera, but the descriptions and illustra- tions of the more complex subsessile glands (except type | 1) are, for the most part, inadequate to classify them according to the system used here. Nonethe- less, the study 1s a useful complement to the present one in that the nonglandular trichomes are described in far more detail than they are 1n this paper. Moreover, 358 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Robert included 11 genera in his survey that are not covered here, most of them in the Chloanthoideae and the Stilboideae. In addition to these two surveys, some information 1s available on individual genera or groups of genera (Labiatae— Mayer (1909), Bech (1963), Kaleva (1967), Wieffering (1970), Bokhari & Hedge (1971, 1976), Inamdar & Bhatt (1972), Heinrich (1973), Rudall (1980, 1986), Azizian & Cutler (1982), Press (1982), Sharma & Shome (1982), Bruni & Modenesi (1983), Manzanares et al. (1983), Shah & Naidu (1983), Bosabalidis & Tsekos (1984), Werker, Putievsky, & Ravid (1985), Werker, Ravid, & Putievsky (1985a, b); Verbenaceae— Mullan (1931), Pant & Kidwai (1964), Kundu & De (1968), Inamdar (1969a), Ramayya & Rao (1969), Inamdar et al. (1976), Fahn & Shimony (1977), Puff (1978), Trivedi & Upadhyay (1978), Bhatt ef al. (1979), Mathew & Shah (1981, 1983), Shah & Mathew (1982a, b By and large, the data provided in these publications are consistent with those reported here, but some disagreements and additions to the data base warrant discussion. Robert (1912) reported type | subsessile glands in nine genera of the Verbenaceae 1n which I either failed to find them (Avicennia L.., Casselia Nees & C. Martius, Dicrastylis J. L. Drumm. ex Harvey, Petitia Jacq., Priva Adanson, Stachytarpheta, and Verbena L.) or could not determine wheth- er they were present or absent (Peronema Jack), or that I did not study (/fem- phora F. Mueller). Type | glands were reported in vicennia by Mullan (1931) as well, but not in the very thorough study by Fahn and Shimony (1977), so this discrepancy probably represents genuine intrageneric variation rather than error. In the Labiatae type | glands have been reported from some species of Phlomis (Bech, 1963: Avizian & Cutler, 1982). a genus in which I failed to find them, as well as from numerous genera of subfam. Nepetoideae (see below). Studies of the subsessile glands of tribe Mentheae (Mayer, 1909; Kaleva, 1967; Bruni & Modenesi, 1983: Bosabalidis & Tsekos, 1984: Werker, Putiev- sky, & Ravid, 1985: Werker, Ravid, & Putievsky, 1985a), including five genera not examined here, confirm that the group 1s characterized by the absence of type 4 glands and the presence of both type | glands and those with tangential walls (usually types 6 and/or 9). Type | glands were also reported from the majority of the Ocimeae studied by Shah and Naidu (1983) and in subtribe Hyptidinae (Ocimeae) by Rudall (1980). In tribe Salvieae glandular trichomes that appear to be type | have been reported in Dorystaechas Boiss. & Heldr. ex Bentham, /forminiun L.. Aferiandra Bentham, Perovskia, Salvia L., and Zhumeria Rech. f. & Wendelbo (Bech, 1963; Bokhari & Hedge, 1971, 1976; Sharma & Shome, 1982). Thus type | glands occur widely in subfam. Nepe- toideae and may be better thought of as characterizing this more inclusive group rather than tribe Mentheae alone. It is difficult to evaluate the apparent conflicts in the distribution of stomatal types in the Verbenaceae—among published works and between some of them and the present study—because of differences in how authors classified the stomatal complexes. For example, Mathew and Shah (1981) reported anisocytic stomata in more genera of Verbenaceae than I have, but upon examination of their illustrations, it appears that they have classified as anisocytic many sto- mata that I would have scored as anomocytic—stomata that happen to be in 1990] CANTINO, STOMATA AND TRICHOMES 309 contact with three surrounding cells, but these (some or all of them) not dis- tinguishable from the other epidermal cells. Several other stomatal types (e.g. haplocytic, tetracytic) recognized by Shah and Mathew in this and other papers (1982a, b) were also treated as anomocytic in the present study. On the other hand, stomata that I would have scored as actinocytic are treated by Mathew and Shah as anomocytic in some cases and cyclocytic in others. In spite of these classification problems, which apply mainly to the papers by Shah and Mathew, published illustrations provide some genuine additions to the data base. Stomata that I would have scored as anisocytic are documented for seven genera in which I did not record them (Bouchea, Gmelina L., Petrea L., Phyla, Premna, and Stachytarpheta) or that I did not study (7ectona L. f.) (Pant & Kidwai, 1964; Inamdar er a/., 1976; Bhatt et al., 1979; Mathew & Shah, 1981). In addition, diacytic stomata have been recorded from Citharexy- lum, Gmelina, and Nyctanthes L. (Trivedi & Upadhyay, 1978; Bhatt et al, 1979: Mathew & Shah, 1981), paracytic from Nyctanthes, Premna, and Tectona (Inamdar et al., 1976; Trivedi & Upadhyay, 1978; Mathew & Shah, 1981), and cyclocytic from a few species of Clerodendrum (Shah & Mathew, 1982a). With regard to the Labiatae, there is universal agreement that the diacytic types of stomata (including diallelocytic) predominate in subfam. Nepetoideae, but El-Gazzar and Watson’s (1968, 1970) observations on subfam. Lamioideae conflict markedly with my own. The data in TaBLe | do not support their assertion (1970, p. 476) that if Prunella L., Cleonia L., and the North American Melittidinae are excluded, anomocytic stomata ‘are the rule” in Bentham’s “Stachydeae” (Lamieae). Both diacytic and anomocytic stomata occur in all examined genera of the group thus circumscribed by El-Gazzar and Watson (i.e., excluding Prunella, etc.), and diacytic stomata are as or more common than anomocytic in about two thirds of them. Furthermore, El-Gazzar and Watson (1968) listed Achyrospermun Blume, Acrotome Bentham ex Endl., Ajuga, Anisomeles, C Binaniosphace Craniotome, Galeobdolon Adanson, Gomphostemma, Moluccella L., Phyllostegia Bentham, Rostrinucula, Sideritis, Thuspeinanta T. Durand, and Wiedemannia Fischer & C. Meyer (not all in Bentham’s Lamieac) as re Seeley anomocytic and/or anisocytic stomata, whereas | found diacytic an ytic stomata to be commonest in these genera and did not find anisocytic onal in any of them. — PHYLOGENETIC SIGNIFICANCE This is one of several character surveys being conducted in preparation for a cladistic analysis of the Lamiales. Although the phylogenetic significance of the data in TABLES | and 2 can best be assessed in the context of such an analysis, Hennigian reasoning can be applied to single characters to yield ten- tative suggestions about their significance. Thus if it can be shown that a character state is derived in a particular group, it may be treated as a potential synapomorphy for a clade within that group unless the distribution of other characters indicates that such a conclusion is unparsimonious. This approach will be employed here. Outgroup comparison (Watrous & Wheeler, 1981; Maddison et al., 1984) 360 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 will be used to assess character polarity. It 1s generally recognized that the Labiatae arose from the Verbenaceae (Cronquist, 1981), which is consequently paraphyletic, but the affinities of the Labiatae are difficult to identify more precisely. Pollen morphology (Raj, 1983) and gynoecial structure (Junell, 1934) suggest that the closest relatives of the Labiatae lie within the Viticoideae and/ or the Caryopteridoideae and that the Verbenoideae are unlikely to be closely related. The other subfamilies, except perhaps the Chloanthoideae, seem to have little in common with the Labiatae. In evaluating the polarity and phy- logenetic significance of characters within the Labiatae, one must therefore pay particular attention to their occurrence in the Viticoideae and the Caryopteri- doideae. The Scrophulariales is the best-supported sister group of the Lamiales (Can- tino, 1982; Frohlich, 1987) and will be treated here as the sole outgroup. More distant outgroups will not be considered because higher-level relationships in the Asteridae are poorly resolved. The distributions of stomatal and trichome types are reasonably well documented in the Acanthaceae (De, 1967; Ahmad, 1974, 1978: Karlstr6m, 1978, 1980) and the Gesneriaceae (Rosser & Burtt, 1969; Van Cotthem, 1971; Sahasrabudhe & Stace, 1974; Herat & Theobald, 1979; Yuen & Dehgan, 1982; Wiehler, 1983), but extensive surveys have not been published for the other families of Scrophulariales. In the discussion that follows, statements about the epidermal anatomy of these families are based on the summaries provided by Solereder (1908) and Metcalfe and Chalk (1950), the distributions of stomatal types tabulated by Wilkinson (1979), and studies of particular genera or groups of genera (Bignoniaceae— Paliwal (1970), Jain (1978), Elias & Newcombe (1979), Henrickson (1985); Lentibulariaceae—Cas- per (1966), Komiya (1972), Trinta (1979), Fineran (1985); Myoporaceae— Dell & McComb (1977); Pedaliaceae (including Martyniaceae)— Mullan (1936), In- amdar (1969b), Karatela & Gill (1984): Plantaginaceae—Andrzejewska-Golec & Swietoslawski (1987); Scrophulariaceae — Fedorowicz (1916), Spoerri (1930), Bhatt & Inamdar (1975), Stefanova-Gateva & Boeva (1979), Doaigey & Harkiss (1982), Henrickson & Flyr (1985)). Anomocytic stomata, found in all examined species of Verbenaceae and nearly all Labiatae, are the most widespread stomatal type in the Scrophular- iales as well, occurring commonly in every family except the Acanthaceae (absent) and the Plantaginaceae (rare). Hence the hypothesis that their presence is ancestral (and their absence derived) in the Labiatae is parsimonious, re- gardless of whether the Labiatae are monophyletic or polyphyletic and of which groups of Verbenaceae are most closely related to the Labiatae. It is also worth noting that presence/absence of anomocytic stomata exhibits less intrageneric variation than most other characters in this study and is therefore of particular value in assessing intergeneric phylogenetic relationships. In tribe Lamieae anomocytic stomata are universally present in 42 of the 45 investigated genera and universally absent in the other three (Brazoria, Machridea, and Physostegia Bentham). Absence of anomocytic stomata is thus a corroborating synapo- morphy for a clade proposed by Abu-Asab and Cantino (1987a) comprising these three genera. It may be of some phylogenetic significance in subfam. 1990] CANTINO, STOMATA AND TRICHOMES 361 Nepetoideae as well, but the much poorer sample of this subfamily makes evaluation difficult Diallelocytic stomata are common in the Acanthaceae but rare to absent in the other families of the Scrophulariales. If the Acanthaceae were the basal clade of the Scrophulariales, polarity of this character within the Lamiales would be equivocal. However, assuming the Acanthaceae are not basal (a reasonable assumption; see Cronquist, 1981), it 1s most parsimonious to hy- pothesize that absence of diallelocytic stomata is the ancestral condition in the Lamiales and that their presence is derived. The three-celled diallelocytic type is of special interest as the only epidermal feature whose distribution parallels relatively closely the traditional boundary between the families; 1.¢., 1t is very common in the Labiatae but infrequent in the subfamilies of the Verbenaceae that appear to be most closely related to them. Its phylogenetic significance 1s discussed below (see Circumscription of Labiatae). Four-celled diallelocytic stomata can be hypothesized to represent a derived state as well, again assuming that the Acanthaceae is not the basal family of the Scrophulariales, but because this character exhibits far more intrageneric variation than the previous one, it is of little phylogenetic significance above the genus level. In the Scrophulariales cyclocytic stomata have been reported only from the Bignoniaceae, where they are apparently infrequent. Their presence in three of the four examined species of Duranta is therefore hypothesized to be derived. The other species (D. repens) has quite a different set of stomatal types (anisocy- tic and helicocytic but not cyclocytic), a distinction that might be of systematic value within the genus and warrants further study. It should be noted, though, that many stomata in the former three species of Duranta have an incomplete or irregular circle of subsidiary cells and thus seem to be intermediate between anomocytic and cyclocytic. Study of stomatal ontogeny in this genus is needed before any systematic conclusions are drawn. Two other rare stomatal types, staurocytic and parallelocytic, also appear to be derived states and of possible phylogenetic significance, the former at the species level within Lavandula L. and the latter as a synapomorphy for Phyla. The distributions of the other stomatal types are of limited phylogenetic value above the genus level because of difficulties in assessing polarity, ques- tions of homology, and/or high intrageneric variability. Actinocytic stomata have been reported only a few times in the Scrophulariales (in a few members of the Bignoniaceae, Gesneriaceae and Lentibulariaceae), but it is unclear whether they are genuinely rare or have been classified as anomocytic by most authors. Moreover, the ontogeny of actinocytic stomata in the Lamiales is unknown; thus they cannot necessarily be considered homologous in the taxa where they occur. A more definite homology problem exists in the case of paracytic sto- mata, which appear to develop via several different ontogenetic pathways in the Lamiales (discussed above). Anisocytic stomata are the predominant type in the Gesneriaceae; they also occur at least occasionally in the Bignoniaceae, Scrophulariaceae, Pedaliaceae, and Myoporaceae but never in the other families of Scrophulariales. Hence it is unclear whether presence or absence is ancestral in the Lamiales. Similarly, 362 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 diacytic stomata are common in the Acanthaceae and the Plantaginaceae, at least occasional in the Lentibulariaceae, the Pedaliaceae, and the Scrophular- iaceae, and rare to absent in the other families of the Scrophulariales. Again, polarity in the Lamiales cannot be assessed without knowing something about cladistic topology within the Scrophulariales. Nor can polarity be assessed within the Labiatae, because both anisocytic and diacytic stomata occur widely, but far from universally, in the two subfamilies of the Verbenaceae that are thought to be closest to them. This is unfortunate, because the two characters exhibit relatively little intrageneric variation within the Labiatae and an in- triguing, inversely correlated distribution pattern. In the Prostanthereae and two genera of the Ajugeae (Tetraclea and Trichostema), diacytic stomata are rare while anisocytic stomata are nearly universally present and often the most common stomatal type. In the rest of the Ajugeae and all gynobasic-styled Labiatae, the reverse is true. From a phenetic standpoint at least, these char- acters neatly distinguish the Prostanthereae from the gynobasic-styled Labiatae and also give cause to question whether 7etraclea and Trichostema belong in the family (further discussed below). The inverse correlation between anisocytic and diacytic stomata 1s weaker in the Verbenaceae, where many genera have neither, and a few (e.g., Holmskioldia and Lantana) have both. The group of Verbenaceae 1n which anisocytic stomata are most consistently present is subfam. Chloanthoideae. Their absence in Nesogenes A. DC., which also lacks the branched trichomes characteristic of the Chloanthoideae (see below), is con- sistent with its exclusion from this group (as by Marais, 1981). Stomatal position (leaves hypostomatic vs. amphistomatic) is of little value as a phylogenetic indicator because of its high intrageneric variability, the existence ofan intermediate condition with a few stomata on the adaxial surface ofthe blade, and the apparent correlation with environmental factors (discussed above). This correlation appears to hold in the Scrophulariales as well: hy- postomatic leaves predominate in the families best represented in the humid tropics (Acanthaceae, Gesneriaceae, Bignoniaceae), while amphistomatic ones are commonest in those well represented 1n xeric habitats (Myoporaceae, Peda- liaceae, Plantaginaceaec). Both conditions occur widely in the Scrophulariaceae. Subsessile glands, nearly universally present in the Labiatae and the Ver- benaceae, are equally characteristic of the Scrophulariales, where they are fre- quent in all families except the Myoporaceae and the Plantaginaceae. Type 4 is the most common type in the Scrophulariales and therefore probably rep- resents an ancestral condition in the Lamiales, where it was found in more than three quarters of the genera of both families. The absence of type 4 glands in nearly all examined Mentheae can thus reasonably be hypothesized to be a derived state and may be of help in circumscribing this poorly defined tribe. The character’s value as a phylogenetic indicator 1s reduced, however, by its relatively high intrageneric variability elsewhere in ie Labiatae and the Ver- benaceae. Subsessile glandular trichomes of types |, 2, and 5 are also quite widespread in the Scrophulariales but are not so universal as to permit polarity assessment in the Lamuales. e more complex gland types (6-1 1) are generally rare in the Scrophular ee (except that type 11 is widespread in the Bignoniaceae). It is reasonable = 1990] CANTINO, STOMATA AND TRICHOMES 363 hypothesize that all are derived in the Lamiales. However, these characters exhibit so little constancy within genera (and sometimes even within species; e.g., types 8 and 9 in Tinnea aethiopica, types 8 and 10 in Holmskioldia sanguinea) that their phylogenetic rane above the species level is prob- ably minimal. Moreover, types 10 may arise through more than one ontogenetic pathway (see FiGuRE 3 a associated text) and are thus not nec- essarily homologous where they do occur. In view of these problems, the presence of type 10 glands in all species of Brazoria and Macbridea and some species of Physostegia, invoked by Abu-Asab and Cantino (1987a) in support of a clade composed of these three genera, must be considered very weak evidence. This clade is more convincingly supported by the absence of anomo- cytic stomata, as discussed above. Nonglandular trichomes, both simple-unicellular and uniseriate, occur widely in the Scrophulariales, the latter type being found in nearly all nonglabrous members of the order. The presence of both kinds of “hairs” is thus probably ancestral in the Lamiales. There is a great deal of intrageneric variation in presence vs. absence of unicellular trichomes in the Lamiales, but much less so for uniseriate trichomes. The rarity of uniseriate trichomes in the Verbenoi- deae is therefore of interest. Presence of on/y unicellular hairs (1.e., absence of uniseriate hairs in nonglabrous species) may represent a synapomorphy, and the taxonomic position of the few genera of Verbenoideae in which uniseriate trichomes were observed (Amasonia, Duranta, Recordia Mold., Stachytarphe- ta) should be examined. Indeed, the pollen morphology of Amasonia (but not of the other three genera) suggests that it belongs with the Viticoideae or the Caryopteridoideae rather than the Verbenoideae (Raj, 1983). The rarity of multicellular foliar hairs in the Verbenoideae has also been noted by Robert (1912) and El-Gazzar (1974). This character may be of phylogenetic significance in the Prostanthereae as well, but with over half the examined species being glabrous, the sample size of those with hairs of any kind was too small to draw any conclusions Branched, multicellular trichomes occur in a scattering of genera in the Scrophulariales but are not common in the major families; they are unlikely to be ancestral in the Lamiales. Their presence in most Chloanthoideae may be a synapomorphy for a large subgroup composed of all genera except Nesoge- nes. (Although the foliage of Cyanostegia Turcz. 1s glabrous or nearly so, branched hairs are present on the ovaries and fruits (Munir, 1978).) Presence of branched, multicellular trichomes is also a probable synapomorphy uniting the species of individual genera such as Ph/omis and Perovskia—genera that, on the basis of other characters, are unlikely to be closely related to each other. Unicellular, two-armed trichomes occur in one genus of Acanthaceae (Solereder, 1908) but are clearly a derived condition in the Lamiales, where they represent a syn- apomorphy of Phyla. CIRCUMSCRIPTION OF THE LABIATAE This study was originally undertaken in the hope that it might contribute to an understanding of phylogenetic relationships among the basal clades of the 364 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Labiatae. Several stomatal characters have turned out to be of interest in this regard. As discussed above, the presence of diallelocytic stomata can reasonably be hypothesized to be a derived state. As such, it would seem to support a clade composed of the gynobasic-styled Labiatae (1.e., subfam. Nepetoideae, subfam. Lamioideae tribes Lamieae, Pogostemoneae, and Scutellarieae, and the genera of “uncertain tribal affinities” in TABLE |) plus Acrymia Prain, Ajuga, Cymaria, Kinostemon, Rubiteucris Kudo, and Teucrium of the Ajugeae. In the Verbenaceae diallelocytic stomata were found in seven genera of subfam. Ver- benoideae, two of Caryopteridoideae, and two of Viticoideae (see TABLE 1). On the basis of pollen and gynoecial morphology (Junell, 1934; Raj, 1983), the Verbenoideae are unlikely to be closely related to the Labiatae, so the occurrence of diallelocytic stomata in this group is probably due to parallelism. A close phylogenetic relationship between one or more of the other verbena- ceous genera in which diallelocytic stomata occur (i.e., Carvopteris (sect. Pseu- docaryopteris Briq. only), Clerodendrum (subg. Cyc lonema only), Petraeovitex, and Premna) and the Labiatae that share this derived state is somewhat more plausible but may be unparsimonious when other characters are considered. Although diallelocytic stomata were not found in all species of the Lamieae, Scutellarieae, and Nepetoideae, they occur in the vast majority of the members of these three groups, each of which 1s defined as a clade on the basis of other (nonepidermal) characters. The presence of diallelocytic stomata is thus a “non- universal derived state” shared by a set of clades and, as argued by Cantino (1985), provides evidence for a larger clade composed of them. The critical requirements are that the state be derived and that the groups united on the basis of it each be supported as a clade by other characters. A shared, nonuniver- sal derived state constitutes weaker evidence of phylogenetic relationship than a shared, universal derived state—1.e., one found throughout the groups it unites (Cantino, 1985). The distributions of anisocytic and diacytic stomata exhibit a pattern that is highly congruent with that of the diallelocytic stomata, although it is not known whether presence or absence of the former two types is derived. In the Labiatae with a gynobasic style and most Ajugeae, anisocytic stomata are rare and diacytic stomata are nearly universally present. All three stomatal char- acters (two of them used only in a phenetic sense, since their polarity cannot be assessed) delimit a group composed of all gynobasic-styled Labiatae plus some Ajugeae. Tribe Prostanthereae 1s consistently excluded from this group, as are Tetraclea and Trichostema of tribe Ajugeae. Among the other Ajugeae, all three characters support the inclusion of Acrymia, Ajuga, Cymaria, Kinoste- mon, and Rubiteucris. The stomatal characters conflict or are ambiguous with regard to the other four genera. Schnabelia Hand.-Mazz. has diacytic stomata, but the slide was too poor to determine whether anisocytic or diallelocytic stomata were present as well. In 4methystea L. diacytic stomata are present, diallelocytic stomata absent, and anisocytic stomata rare. Teucrium consis- tently has diacytic and diallelocytic stomata, but anisocytic ones were fou in two species. 77nnea consistently has diacytic stomata sae lacks diallelocytic; anisocytic stomata were found in one species only. The strong positive cor- relation between the occurrences of diacytic and eee stomata 1s hardly 1990] CANTINO, STOMATA AND TRICHOMES 365 surprising since the initial steps in their ontogenies are the same (Payne, 1970), but there is no reason to expect the observed negative correlation between diacytic and anisocytic stomata. Indeed, this correlation is much weaker in the Verbenaceae. The relationships suggested by these characters must be evaluated in the light of others, and recircumscription of the Labiatae at this time would be highly premature. However, those taxa whose epidermal anatomy is divergent from the rest of the Labiatae should be carefully examined. One of these (Tetraclea) has been assigned to the Verbenaceae by some authors (e.g., Mol- denke, 1971), and the flowers of another (7richostema) bear a striking resem- blance to those of Carvopteris sect. Pseudocaryvopteris in the Verbenaceae. The supratectal spinelike projections on the pollen of Tetraclea (Raj, 1983) and Trichostema (Abu-Asab & Cantino, 1989), as well as a few other genera of Ajugeae, constitute a derived feature shared by many Verbenaceae (Raj, 1983) and entirely unlike the exine sculpturing of the gynobasic-styled Labiatae (Abu- Asab & Cantino, 1987b). The Prostanthereae have traditionally been placed in the Labiatae because the ovary is moderately lobed (with the style somewhat sunken but not gy- nobasic) and matures to form four nutlets. However, the characteristic suite of stomatal types in the Prostanthereae, with anisocytic stomata abundant in all but one genus (they are present but uncommon in Westringia Smith), diacytic stomata rare, and diallelocytic stomata absent, differs markedly from that in the gynobasic-styled Labiatae and resembles certain groups of Verbenaceae— particularly subfam. Chloanthoideae and tribe Clerodendreae. This 1s intriguing from a biogeographic standpoint since the Prostanthereae and most genera of Chloanthoideae are Australian endemics, while the Labiatae (other than Pros- tanthereae) are rather poorly represented in Australia (Bentham, 1870). In a numerical phenetic analysis of Labiatae and Verbenaceae (El-Gazzar & Watson, 1970), the Prostanthereae, Chloanthoideae, and Stilboideae clustered together tightly. It is not clear that the similar suites of stomatal types in the Prostanther- eae and Chloanthoideae reflect cladistic relationship, since absence of dialle- locytic stomata 1s an ancestral state and the polarity of the other two characters cannot be assessed without knowing more about cladistic topology of the Ver- benaceae and/or the Scrophulariales, but the possibility that these two Aus- tralian groups may be more closely related than generally thought warrants study. ACKNOWLEDGMENTS I gratefully acknowledge the technical assistance of Jon Hamer, who 1m- proved upon the clearing and staining procedure previously used in our lab and prepared the vast majority of the slides. Some slide preparation was also carried out by Gregory Rhinehalt. FiGuRE 2 was drawn by Rebecca Samson, and Ficures | and 3, reprinted from an earlier paper in the Journal of the Arnold Arboretum, were originally prepared by Mones Abu-Asab. I thank the curators of the Harvard University Herbaria, the Missouri Botanical Garden, the New York Botanical Garden, and the U. S. National Herbarium for permission to 366 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 remove leaf materia! for study. | am grateful to Paula Rudall for introducing me to the Royal Botanic Garden’s computerized data base of anatomical lit- erature, which greatly facilitated my literature survey of the Lamiales an Scrophulariales, and to Ray Harley for his generosity with his time and facilities during my stay at Kew, where the manuscript was prepared. This research was supported by National Science Foundation Grant BSR 83-06878 (Principal Investigator, P. D. Cantino) and Ohio University Baker Fund Award 88-10. LITERATURE CITED Asu-AsaB, M.S., & P. D. CANTINO. 1987a. Phylogenetic pRueeea tes of leaf anatomy in subtribe See : abiatac) and related taxa. J. Arnold Arbor. 68: 1-34. & . Palynological evidence for a te origin of tribe Augen (Laie) oe J. Bot. 74: 722. . 1989. Pollen morphology of 7richostema (Labiatae) and its systematic implications. Syst. Bot. 14: 359-369. AHMAD, K. J. 1974. Cuticular and epidermal structures in te a of Eranthemum and Pseuderanthemum (Acanthaceae). Bot. Not. 127: at 97 — hairs of Acanthaceae. Blumea i 10l- Aceneeieu Gotec, E., & J. SwiETOSLAWSKI. 1987. The Gees of hairs in species of eae L. sectio Coronopus DC. Acta Soc. Bot. Poloniae 56: 367-379. AziziaAn, D., & D. F. CuTLer. 1982. Anatomical, cytological and eae ety studies on Pilon L. and Eremostachys Bunge (Labiatae). J. Linn. Soc. t. 85: 249-2 BaRANOVA, M. A. 1987. Historical so epnien of the nese Sane of mor- phological types of stomates. Bot. ’, (Lancaster) 53: 53-79. Beco, T. D. 1963. The anatomical structure of the hairs in ae prickly Sees sage (Phlomis pungens Willd.). (In Russian.) Pee k Bot. Zurn. 20(5): 96-98. BENTHAM, G. 1870. Flora Australiensis. Vol. 5. Reeve and Co., London. 1876. Labiatae. Pp. 1160-1223 in G BENTHAM & J. D. Hooker, Genera plantarum. Vol. 2. Reeve and Co.. Lon BERLYN, G. P., & T. P. MIKSCHE. ese Rota microtechnique and cytochemistry. Iowa State Univ. Press, Ames, Iow Byatt, D.C _A.INAMDAR. 1979. Structure and development of stomata in some eeiciaca. ean. J. Bot. 22: 173-182. & J. A. INAMDAR. 1975. Structure and development of stomata in some Scroph- ulariaceae. Botanique (Nagpur) 6: 209-218. BoxHarl, M. H., & I. C. HepGe. 1971. oe on the tribe Meriandreae of the Labiatae. Notes Roy. Bot. Gard. Edinburgh 31: 53-67. 1976. Zhumeria (Labiatae): anatomy, taxonomy and affinities. Iran. & 1-10. Pe aru a & I. TsEKos. 1982. Glandular scale development and essential oil secretion in Orica ‘eae L. Planta 156: 496-504. & 19 landular hair formation in Origanum species. Ann. Bot. (London) 53: “559- 563. Brunt, A., & P. MopENEs!. 1983. Development, oil storage and dehiscence of peltate trichomes in Thymus vulgaris (Lamiaceae). Nordic J. Bot. 3: 245-251. CANTINO, P.D. 1982. Affinities of the Lamiales: a cladistic analysis. Syst. Bot. 7: 237- y/ —_ “1985. En opetete inference from nonuniversal derived character states. Syst. Bot. 10: 119-12 & R.W. ne 1986. Subfamilial classification of Labiatae. Syst. Bot. 11: 163-185. 1990] CANTINO, STOMATA AND TRICHOMES 367 CASPER, - J. 1966. Monographie der Gattung Pinguicula L. Biblioth. Bot. 31(127, 128): 1-209. Conn, B. J. 1984. A taxonomic revision of Prostanthera Labill. section Klanderia (F. v. Muell.) Benth. (Labiatae). J. Adelaide Bot. Gard. 6: 207-348. CoTTHeM, W. R. J. vAN. 1971. eases morphologische Studien awe oe und eine neue Klassifikation ihrer Typen. Ber. Deutsch. Bot. Ges. 84: 14 8. CronaguisT, A. 1981. An _ a of classification of flowering ae Co- lumbia Univ. Press, New Y DAHLGREN, R. 1980. A ee sy stem of classification of the angiosperms. J. Linn. Soc., Bot. 80: 91-124. De, A. 1967. Cytological, anatomical and palynological studies as an aid in tracing affinity and phylogeny in the family Acanthaceae. Trans. Bose Res. Inst. Calcutta 30: 51-65. DELL, B., & A. J. McComs. 1977. Glandular hair formation resin secretion in Eremophila fraseri F. Muell. (Myoporaceae). Protoplasma 92: 71-86. Ditcuer, D. L. 1974. Approaches to the identification of sie eets leaf remains. Bot. Rev. ae 40: 1-157 Doaicey, A. R., & K. J. HARKISS. 1982. Comparative anatomy of the leaf and stem of ee, meonanthum Hoffmans. & Link and A. braun-blanquetii Rothm. J. Coll. Sci. King Saud Univ. 13: 249-257 EL-GaAzzAr, A. 1974. Numerical taxonomy of the Verbenaceae: a re-assessment. Egypt. & L. Watson. 1968. Labiatae: taxonomy and susceptibility to Puccinia menthae Pers. New i 67: 739-743. 70. A taxonomic study of Labiatae and related genera. [bid. 69: 451-486. a T. S., & L. F. Newcombe. 1979. Foliar nectaries and glandular trichomes in Catalpa (Bignoniaceae). Acta Bot. Sin. 21: 215-224. ERDTMAN,G. 1945. Pollen morphology and ter eee IV. Labiatae, Verbenaceae and Avicenniaceae. Svensk Bot. Tidskr. 39: 279-285. Faun, A. 1979. Secretory tissues in plants. ‘onen Press, New York. & C. SHimony. 1977. Development of the glandular and non-glandular leaf hairs of Avicennia marina (Forsskal) Vierh. J. Linn. Soc., Bot. 74: 37-46. Feporowicz, S. 1916. Die Driisenformen der Rhinanthoideae— Rhinantheae. Bull. cad. Sci. Cracovie, Ser. B., 1916: 286-322. cee B. A. 1985. Glandular trichomes in Utricularia: a review of their structure and function. Israel J. Bot. 34: 295-330. FROHLICH, M. W. 1987. Common-is-primitive: a partial validation by tree counting. Syst. Bot. 12: 217-237. FRYNS-CLAESSENS, E., & W. VAN COTTHEM. aes ee classification of the ontogenetic types of stomata. Bot. Rev. (Lancaster) 39: 71-138. HernricnH, G. 1973. Entwicklung, vere und eee der Driisenschuppen von Mo- narda oe Pl. Med. 23: 154- Helsey, R. M., & C. C. DELWICHE. a Phytotoxic volatiles from Trichostema lan- ane diab aiacy. Amer. J. Bot. 71: 821-828. HENRICKSON, J. 1985. A taxonomic revision of Chilopsis (Bignoniaceae). Aliso 11: 179- 197, . Fryer. 1985. aE of Leucophyllum and Eremogeton (Scroph- iinrigesde). Sida 11: 107-172. Herat, R. M., & W. L. THEOBALD. 1979. Comparative studies of vegetative anatomy and morphology of the Gesneriaceae of Sri Lanka. J. Linn. Soc., Bot. 78: 285-298. Hotmaren, P. K., W. Keuken, & E. K. Scuortetp. 1981. Index herbariorum. Part |. Herbaria of the world. ed. 7. Bohn, Scheltema, & Holkema, Utrecht. 368 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Huana, T. C., & W. T. CHENG. 1971. A preliminary revision of Formosan Labiatae (1). Taiwania 16: 157-174 INAMDAR, J. A. 1969a. Epidermal structure and ontogeny of stomata in some Verbena- ceae. Ann. Bot. (London) 33: 55-66. Epidermal structure and stomatal ontogeny in vegetative and floral organs of Martynia annua L., Pedalium murex L. and Sesamum indicum L. Flora 158: 526-537. D.C. BHatr. 1972. Structure and development of stomata in some Labiatae. sa Bot. (London) 36: 335-344. . GANGADHARA, R. B. BHAT, & M. A. PATHAN. 1976. Studies on the foliar vichomes Bea and stomata of Tectona grandis Linn. f. Geobios 3: 73-75. Jain, D. K. 1978. Studies in Bignoniaceae. II. Foliar epidermal structures. J. Indian 74, JUNELL, S. 1934. Zur Gy eee Nae und Systematik der Verbenaceen und Labiaten. Symb. Bot. Upsal. 4: 1-219 KALEVA, K. 1967. Notes on some eee characteristics in Thymus L. (Labiatae). Ann. Bot. Fenn. 4: 95-101. KARATELA, Y. Y., & L. S. Gitt. 1984. Phytodermology and ontogeny of stomata in some Pedaliaceae. J. Econ. Tax. Bot. 5: 237-240. KARLSTROM, P. 1978. Epidermal leaf structures in species of Strobilantheae and Petali- dieae (Acanthaceae). Bot. Not. 131; 423-433. 1980. Epidermal leaf structures in species of ere Pseuderanthemeae, Graptophylleae and Odontonemeae (Acanthaceae). /bid, 133: 1-16. Komrya,S. 1972. Systematic studies on the Lentibulariaceae. Bok of Biology, Nippon Dental College, Tokyo. (Typescript, available at Royal Botanic Gardens, Kew.) Kunbu, B.C., & A. De. 1968. Taxonomic position of the genus Nyctanthes. Bull. Bot. Surv. India 10: 397-408. Mapopison, W. P., M. J. DonoGuug, & D. R. MAppison. 1984. Outgroup analysis and parsimony. os Zool. 33: 83-103. MANZANARES, P., C. GOMEz-CAmpo, & M. EsTRELLA ToRTOSA. 1983. Estudios sobre el indumento oi las especies ibéricas y balearicas del género Teucrium L. (Lami- aceae). Anales Jard. Bot. Madrid 40: 93-106. Marais, W. 1981. Twonew gamopctalous families, Cyclocheilaceae and Nesogenaceae, for extra-Australian *“Dicrastylidaceae.” Kew Bull. 35: 797-812 MATHEW, L., & G. L. SHAH. 1981. Observations on the struc ture and ontogeny of stomata in some Verbenaceae with a note on their taxonomic significance. Feddes Repert. 92: 515-526. 1983. Structure, development, organographic ey and taxo- nomic significance of trichomes in nine species of Verbena. Ibid. 94: 323-333. Mayer, F. 1909. Systematisch- ‘anatomische Untersuchung der Pogostemoneae Rei- chenb. unter besonderer Beriicksichtigung der inneren Driisen von Pogostemon u. aa ae sowie der Patschuli-Droge. Ph.D. dissertation, Univ. Erlan Mercatrg, C. R., & L. CHALK. 1950. Anatomy of the dicotyledons. Vol. 2. Oxford & Mo.penke, H. N. 1971. A fifth summary of the Verbenaceae, Avicenniaceae, Stilba- ceae, Dicrastylidaceae, Symphoremaceae, Nyctanthaceae, and Eriocaulaceae of the world as to valid taxa, geographic distribution, and synonymy. 2 vols. Published by the oe MuLLAn, D. 1931. On the occurrence of glandular hairs | glands) on the leaves of some a halophytes. J. Indian Bot. Soc. 10: 184— . 1936. On the mucilage glands and absorbing hairs a Pedalium murex Linn. Ibid. 15: ee Munir, A. A. 1978. A taxonomic revision of the genus Cyanostegia (Chloanthaceae). oe. 45-67 1990] CANTINO, STOMATA AND TRICHOMES 369 PALIWAL, G. S. 1970. Epidermal structures and ontogeny of stomata in some Bignonia- ceae. Flora 159: 124-132. Pant, D. D., & P. Kipwar. 1964. On the diversity in the devel t and organization of stomata in Phyla nodiflora Michx. Curr. Sci. 33: 653, 654. PARKHURST, D. F. 1978. The adaptive significance of stomatal occurrence on one or both surfaces of leaves. J. Ecol. 66: 367- Payne, W. W. 1970. Helicocytic and allelocytic stomata: unrecognized patterns in the Dicotyledonae. Amer. J. Bot. 57: 140-147. Press, J. R. 1982. Taxonomic studies in the Labiatae tribe Pogostemoneae. Bull. Brit. Mus. (Nat. Hist.), Bot. 10: 1-89 Purr, C. 1978. The peculiar leaf trichomes of Petrea volubilis (Verbenaceae). J. S. African Bot. 44: 119-125. Ray, B. 1983. ea pollen morphology of Verbenaceae. Rev. Palaeobot. RamayyA, N., & J. RAO, 1969, Range of structural and ontogenetic stomatal variations in three species of Ocimum (Labiatae). Curr. Sci. 38: 79-82. RospertT, G. 1912. Recherches sur l’appareil pilifére de la famille des Verbénacées. Lucien Declume, Lons-le-Saunier. Rosser, E. M., & B. L. Burtr. 1969. Studies in the Gesneriaceae of the Old World XX: ae characters in the tribe Trichosporeae. Notes Roy. Bot. Gard. Edinburgh 29: 39-58. RuDALL, P. 1980. a anatomy of the subtribe Hyptidinae (Labiatae). J. Linn. Soc., Bot. 80: 319-340. 1986. Leafanatomy of Hyptis sect. Pachyphyllae (Labiatae) and related species. Kew Bull. 41: 1017-1025. SAHASRABUDHE, S., & C. A. Stace. 1974. Developmental and structural variation in the trichomes and stomata of some Gesneriaceae. New Botanist 1: 46-62. SANDERS, R. W., & P. D. Cantino. 1984. Nomenclature of the subdivisions of the Lamiaceae. Taxon 33: 64-72. SHAH, G. L., & L. MATHEW. 1982a. Structure and ontogeny of stomata on vegetative and floral organs in ten species of Clerodendrum. Sardar Patel Univ. Res. J. 1: I-11. . 1982b. The structure, ontogeny and organographic distribution of stomata and trichomes on the vegetat ie - floral organs in four ornamental taxa of Lantana. J. Econ. Tax. Bot. 3: 407-4 & A.C. Narpu. 1983. Trichomes on — of some Lamiaceae. Geophytology 76; SHARMA, H. P., & U. SHome. 1982. Leaf surface studies of some medicinal salvias. Proc. Indian Acad. Sci. 91: 449-463 SOLEREDER, H. 1908. Systematic anatomy of the dicotyledons. Clarendon Press, Oxford. SPOERRI, W. 1930. Beitraége zur Anatomie des Blattes pharmazeutisch gebrauchlicher Scrophulariaceendrogen. Inaugural dissertation, Univ. Basel. Stace, C. A. 1965. Cuticular studies as an aid to plant taxonomy. Bull. Brit. Mus. (Nat. Hist.), Bot. 4: 1-78 STEFANOVA-GATEVA, B., & A. N. Boeva. 1979. Anatomical studies of folia verbasci. Pl. Med. 37: 20-31. TAKHTAJAN, A. L. 1986. Floristic regions of the world. Univ. California Press, Berkeley. ee R.F. 1981. Phytochemistry and angiosperm phylogeny: a summary statement. 233-295 in D. A. YouNG & D. S. SEIGLER, eds., Phytochemistry and angiosperm eee Praeger, New Y a E. F. 1979. Revisdo das espécies do género Genlisea St.-Hil. (Lentibulariaceae) das regides sudeste e sul do Brasil. Rodriguésia 31(49): 17-139. Triven, B. S., & N. Upapuyay. 1978. Epidermal structures and ontogeny of stomata of Nyctanthes arbor-tristis L. Geophytology 8: 22-31. 370 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Wartrous, L. E., & Q. D. aire ‘oie The out-group comparison method of character analysis. Syst. Zool. 30: WeRKER, E., E. PuTievsky, & U. ee a The essential oils and glandular hairs in different chemotypes of Origanum vulgare L. Ann. Bot. (London) 55: 793-801. , U. Ravin, & E. Putievsky. 1985a. Structure of glandular hairs and identifi- eaten of the main components of their secreted material in some species of the Labiatae. Israel J. Bot. 34: 31-45 & . 1985b. Glandular hairs and their secretions ” the vegetative and reproductive organs of pale sclarea and §. dominica. Ibid. 239-252. WIEFFERING, J. H. 1970. Enige op g de beharing van het geslacht Galeopsis . Gorteria 5: 248-255. WieHLerR, H. 1983. A synopsis of the Neotropical Gesneriaceae. Selbyana 6: 1-219 WILKINSON, H. P. 1979. The plant surface (mainly leaf). Pp. 97-165 in CLR. METCALFE & CHALK, eds., Anatomy of the dicotyledons. ed. 2. Vol. 1. Clarendon Press, Oxford. Yuen, C.K. K.H., & B. DEHGAN. 1982. Comparative morphology of the leaf epidermis in the genera Codonanthe (M aes) ee and A s Schrader (Gesneria- ceae). J. Linn. Soc., Bot. 85: 283-2 APPENDIX L. Pascua of subsessile erandulat trichomes the Labiatae and Verbenaceae. Type |. Head composed of one en Type 2. Head composed of two Type 3. Head Secon by three beer walls to form three cells, one twice as large as the other Type 4. Head neat of four cells. Type 5. Head of more than four cells. usually divided by four primary radial walls that are more or less perpendicular to each other; tertiary and tangential walls absent, no more than one secondary radial wall arising on a given side of any primary radial wall. TYPE 6. fs in type 5, but with tangential walls present; partial and tertiary radial walls a Ts TYPE 7. AS i in type 6, but with partial radial walls present. Tyre 8. Head of more than four cells; tertiary radial walls present and/or more than one secondary radial wall aeove: on the same side of at least one primary radial wall; tangential walls absent. Type 9. Asin type 8, but with eae walls present; partial radial walls absent. Type 10. As in type 9, but with partial radial walls present. Type 11. More complex than type 10, with many tangential walls forming concentric layers of cells. *IIlustrated in FiGure 2; adapted from Abu-Asab and Cantino (1987a). 1990] ZHUGE, BURRETIODENDRON ar ON THE GENUS BURRETIODENDRON SENSU LATO (TILIACEAE)! REN ZHUGE? Burretiodendron Rehder is a small tiliaceous Jee with six species, limitedly distributed in southwestern China, northern Vietnam, Burma, and Thailand. Although much research has been ee out on this taxon since 1936, there are still a number of problems both in taxonomy and morphology requiring further clarification. The new evidence weighs against the separation of E.x- centrodendron from Burretiodendron. Burretiodendron 1s an ee taxon that appears to be related to Sicrea Hallier f. and Schoutenia Korth. on the basis of pollen morphology. The distribution pattern of the group a that it may have originated near the China-Vietnam border Burretiodendron, a small genus of six species, 1s distributed from south- western China through Tonkin to northern peninsular Thailand and Burma. Most species of the genus are restricted to the border area of southern China and northern Vietnam, where they grow primarily on limy soils in rain forests or dry deciduous woods. Burretiodendron is most readily distinguished from the other genera of the Tiliaceae by its characteristic flowers and fruits, but due to the diversity of its habitat and flower types, mistakes in observation are easy to make, particularly when material is limited. Chang and Miau (1978) divided the genus into two genera, se ees sensu stricto and E.xcentrodendron Chang & Miau, on the basis of incon bservations. Excentrodendron was described as having er flowers and 2 an evergreen habit and as being distinct from Burretio- dendrons.s. | have checked almost all the specimens of both taxa kept in Chinese herbaria, as well as photographs of some specimens housed in other countries. The specimen cited under E.xcentrodendron tonkinense (A. Chev.) H. T. Chang & R. H. Miau as evidence of its bisexual flowers is China—Soviet Union Ex- pedition 2618, from Jinping county of southwestern Yunnan; it bears juvenile fruits with several stamens at the base. Although the flower really is bisexual, the specimen is definitely B. esquirolii. Additional studies on flower type 1n- dicate that most species of both Burretiodendron s.s. and Excentrodendron are monoecious or dioecious, with only B. esquirolii being monoecious or poly- gamous. My observations show that B. Ayvdiifolium Hsu & Zhuge retains its old leaves until the new ones unfold the following spring. It is possible that the habitats of B. siamense Kosterm. and B. brilletii Kosterm. are similar to ‘Prepared for the International Symposium on Plant Resources, 4-7 October 1988, Kunming, omanan, People’s Republic of China Box 48, Southwest Forestry College, Kunming, Yunnan, People’s Republic of China. © President and Fellows of Harvard College, 1990. Journal of the Arnold Arboretum 71: 371-380. July, 1990. Se JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 that of B. kydiifolium because the leaves of all three species are more or less the same (see FiGuRE |). However, habitat is unsuitable for use as a taxonomic basis because it is a character extrinsic to the plant. The evidence provided above indicates that there are complex evolutionary relations among the taxa of the genus, and it 1s probably incorrect to divide these taxa into two groups. When we consider flower structure and fruit mor- phology, it appears that the genus should be accepted 1n its broad sense. he systematic position of the genus has long been a subject of controversy. Rehder (1936) considered that the taxon did not seem to be closely related to any of the genera of the Tiliaceae or the Sterculiaceae. He thought that it was perhaps best placed in the Tiliaceae near Luehea Willd. because both lack androgynophores and have five bundles of stamens, although they differ mark- edly in flower sex, sepals, and fruits. Kostermans (1961) believed the genus to be most closely allied to Colona Cav. because both have a winged septicidal capsule. Craigia W. Smith & W. E. Evans, usually placed in the Sterculiaceae, was considered by Chang and Miau (1978) to be most closely related to Bur- retiodendron s.s. and Excentrodendron and had been included in their new subfam. Excentrodendroideae. In Hutchinson’s (1967) system Burretiodendron was treated as one of the six genera under the tribe Enteleeae, based on the stamens all bearing fertile anthers and the fruits being capsules, but this is apparently not so. The five-winged capsules of Craigia resemble the fruits of Burretiodendron. For this reason the two genera are often confused; for example Craigia yun- nanensis has been misidentified as B. combretoides Chun & How (Chun & How, 1956) and B. yunnanense Kosterm. (Kostermans, 1961). Craigia yun- nanensis can be most readily distinguished from Burretiodendron spp. by its petaloid staminodes and its lack of petals. I have examined the fruits of Craigia yunnanensis carefully. The species does have a loculicidal capsule rather than a septicidal one, which can be seen in the original drawing in Smith and Evans (1921). Evidence collected from palynology and wood anatomy also shows that the two groups have few similarities. Burretiodendron differs from the other genera of tribe Enteleeae in having winged fruits that divide septicidally into five cocci. It differs markedly from Colona in having slender-clawed petals without a glandular portion at the base, and stamens connate into five bundles at the base. So far the pollen morphology of over forty genera in the Tiliaceae has been studied by scanning electron microscopy. I have examined the pollen grains of several species of Burretiodendron under the SEM and have found three pollen patterns in this group: B. Asienmu with globular, panaperturate, coarsely reticulate grains (see FIGURE 2c); B. esquirolii with globular, panaperturate, papillate, echinate ones (see FiGurE 2b); and B. Aydiifolium with oblate, mono- porate, corrugate ones not previously recorded for the family (see FIGURE 2a). If palynology 1s considered, Burretiodendron seems to be more closely allied to Schoutenia Korth. and Corchoropsis Sieb. & Zucc. (see FiGURE 2). Mor- phologically it resembles Schoutenia and Sicrea Hallier f. in having an oblong, basifixed anther and two ovules in each locule and is much less similar to Corchoropsis. The latter is considered by many to be most closely related to 1990] ZHUGE, BURRETIODENDRON a73 DE ee if he Rant ae ory \ ) Ficure |. a, B. esquirolit: a, leaf; b, fruit; c, staminate flower. d-g, B. kydiifolium: d, leaf; e, fruit; f, carpellate flower; g, staminate flower. h, i, B. Asienmu: h, leaf; 1, fruit. j, k, B. pbeonicunie j. leaf; k, fruit. 1, B. brilletii, leaf. m, n, B. siamense: m, leaf, n, fruit. yy ae NES Aah ay lot wna Sava tia try n ry te * a “at Aly, ayy DN oe ; "a ative eeu wn Bee a stae a Artery PEE Sey atts Sais. t5> a : SS Ficure 2. Pollen grains: a, Burretiodendron kydiifolium: b, B. esquirolii; c, B. hsien- mu: d, Tilia tuan Szyszyl.. e, Craigia yunnanensis W. Smith & W. E. Evans; f, Colona floribunda Craib; g, Corchoropsis crenata Sieb. & Zucc.; h, Luehea speciosa Willd.; i, Schoutenia ovata Korth. 1990] ZHUGE, BURRETIODENDRON Bee) the Sterculiaceae. According to Erdtman’s (1952) description, Sicrea 1s also similar to Burretiodendron in having globular, panaperturate pollen. The paly- nology hints that the systematic position of Burretiodendron may be near Sicreaand Schoutenia. Further studies will be necessary to resolve this question. Five species of Burretiodendron have been reported from the border area between China and Vietnam. These exhibit almost all of the typical features of the group: for example, B. Avdiifolium has the staminodes and the solitary carpellate flowers; B. esquirolii, the gynophore, the glandular sepals, and the two bracteoles enveloping a unisexual or bisexual flower bud; and B. hsienmu, the glands on the nerve axils and the three bracteoles enveloping a unisexual bud. Orly one species, B. siamense, has a disjunct distribution, occurring in northern peninsular Thailand and the Mergui Archipelago of Burma. It there- fore seems possible that Burretiodendron was derived in the border area between China and Vietnam. Both Burretiodendron hsienmu and B. kydtifolium produce a very hard wood that can be used in building houses and boats. The most famous chopping blocks in the Kwangdong and Hong Kong regions are made of the ““hsienmu” wood. Today the lumber resources of the genus are limited, and some species have been listed in the state catalogue of rare and endangered plants. Burretiodendron Rehder, J. Arnold Arbor. 17: 47. 1936. Parapentace Gagnep. Bull. Soc. Bot. France 90: 70. 1943, nomen nudum. Excentrodendron H. T. Chang & R. H. Miau, Acta Sci. Nat. Univ. Sunyatsen 1978(3): 21. 1978. Type species: B. esquirolii (Léveillé) Rehder. Trees with leaves concentrated at ends of branchlets. Leaves alternate, sim- ple; petiole usually long and slender, slightly swollen toward apex; stipules early caducous; blade palmati- or triplinerved, sometimes trilobate at apex, sym- metrical or asymmetrical at base, entire. Flowers unisexual and sometimes bisexual (and plants polygamous), solitary or in small cymes, racemes, or pani- cles; bracteoles 2 or 3 enveloping flower bud, caducous; sepals 5, free, valvate, with or without glandular part inside at base; petals 5, free, more or less unguiculate. Staminate flowers with stamens numerous, connate at base into 5 phalanges; anthers basifixed, linear-oblong, bilocular, longitudinally dehis- cent; staminodes absent or rarely present; ovary reduced. Carpellate and bi- sexual flowers with ovary 5-angular, 5-locular, 2-ovular in each locule; styles , free, clavate. Capsules 5-winged, splitting septicidally into 5 cocci. Seeds without endosperm; cotyledons large, foliaceous. KEY TO THE SPECIES OF BURRETIODENDRON A. Leaves with upper surface glabrous, margin entire but with apex sometimes 3-lobed; flowers unisexual, staminate ones in small panicles, carpellate ones solitary or in B. mean coriaceous or subcoriaceous, 3- to 5-basinerved:; sepals usually glandular ase. C. Leaves oblong-elliptic or ee acute or rotundate at base, 3-basi- nerved, nerve axils glandul 376 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 D. Capsules 3-4 cm long, eee ee Barer oa oe dace B. hsienmu. Capsules ca. 5 cm long, obconical. ................. 2. : obconicum. C. Leaves broadly ovate, ee. or truncate at base, 5-basinerved, nerve axils eglandular. . 2.000... eee eee eee SOB. brilfetiz. B. Leaves chartaceous, 5- to 9-basinerved; sepals eglandular at base. Bracteoles 3; staminodes 5, linear; leaves cordate at base, : e basinerve so Aodeueast ips sd neta ace a guint ds estan ategasieits Gast oseeauad ae Uaaaet ads ede ae 3 kydiifolium. E. pracleoles< 2; staminodes absent; leaves truncate or a at base, 5- to Wie ead Sand Oo edonontdea tre waded Rares eu wes B. siamense. A. Leaves with both surfaces shortly haley pilose, margin denticulate; plants polyga- mous, flowers in cymes; Ovaries Stipitate. 20.00 .0.0..0.0.000005. . B. esquirolii. i? ies air hsienmu W. Y. Chun & F. C. How, Acta Phytotax. at 5: 1956. Excentrodendron hsienmu (W. Y. Chun & F. C. H T. Con & R.H. Miau, Acta Sci. Nat. Univ. Sunyatsen 1978(3): 23. 1978. Type: China, Kwangsi [Guangxi] Prov., Lungchow [Longzhou], Wu Lien Hsiang, April, 1955, C.F. Liang 31523b (holotype, scsi). FiGure l/h, 1. Pentace tonkinensis A. Chev. Bull. Econ. Indoch. 20: 803. 1918, nomen nudum, Kos- termans, Reinwardtia 5: 239. 1960. Parapentace tonkinensis (A. Chev.) Gagnep. Bull. Soc. Bot. France 90: 70. 1943. Burretiodendron tonkinense Kosterm. Rein- wardtia 5: 239. 1960. Tyee: Vietnam, Hoa-binh, eas 8 (Pp; photo!). Excentrodendron rhombifolium H. T. Chang & R. H. Miau, Acta Sci. Nat. Univ. ae 1978(3): 23. 1978. Type: China, epee Prov., Longzhou, Qing-shan g, S. Q. Chen 11852 (holotype, scis!; isotype, 1BG!). Tree up to 40 m high. Leaves with petiole 4-10 cm long; blade orbicular- ovate to subrhomboid, 10-18 by 7-12 cm, long-acuminate at apex, acute or rotundate at base, coriaceous, glabrous, 3-basinerved, glandular in nerve axils. Flowers imperfect, staminate ones 6 to 12 in small panicles, carpellate ones solitary or 2 or 3 per racemule; bracteoles 3, caducous; sepals lanceolate, ca. I-1.5 cm long, densely stellate-pilose, glandular inside at base; petals spathu- late, clawed, about as long as sepals: stamens 25 to 35. Capsules ellipsoid, 3- 4 cm long, glabrous. DistTRIBUTION. China (Yunnan, Guangxi) and northern Vietnam. ADDITIONAL SPECIMENS EXAMINED. China. YUNNAN Prov.: Hekou, R. Zhuge 10276, 10277 (both swec), C. J. Wang 1023 (swec); Jinping, China-Soviet Union Exped. 1073 (KUN, PE); Malipo, R. Zhuge 10274 (swrc), C. J. Wang 807 (swrc). GUANGXI PRov.: Longzhou, Y. K. Li 00243 (pc, PE); Jinx1, 7. Af. Li & Z. J. Lt 1368 (1BG); Ningming, S. K. Lee 200409 (18G), Baise, Baise Exped. 01960 (pe); Debao, Z. Y. Wei 00244 (1BG); Longan, J. X. Zhong s.n. 25 Feb. 1955 (scat). Vietnam: near Lao Cai, China: Vietnam E: rs 717 Gan ). Pentace tonkinensis, proposed by Chevalier in 1918, was briefly described in French. In 1943 the species was transferred by Gagnepain to his new genus Parapentace, which was merged by Kostermans in 1961 into Burretiodendron, there, Kostermans recognized it to be conspecific with Burretiodendron hsien- mu. However, in 1978 Chang and Miau distinguished the species from B. Asienmu again, placing both in the new genus /-xcentrodendron. They believed 1990] ZHUGE, BURRETIODENDRON ag 7 B. tonkinense to be characterized by fewer flowers in a raceme and a nodiferous pedicel, and B. Asienmu by numerous flowers ina panicle and an enodal pedicel. hen I examined specimens of the two species in several larger Chinese her- baria, I found that the floriferous specimens were always identified as B. hsien- mu and the fructiferous ones as B&B. tonkinense. Field observations revealed that staminate trees of B. hsienmu bear six to twelve flowers in small panicles, while carpellate trees have one to three flowers per raceme; the nodes on the pedicel are merely the cicatrices of the caducous bracteoles, as becomes apparent with the growth of the fruit. Furthermore, in Gagnepain’s original description Para- pentace tonkinensis was noted as “fl. 6 groupées en glomérules de 10 environ.” For this reason we can be sure that B. fonkinense only represents the carpellate trees of B. Asienmu. Although B. hsienmu was proposed later than Parapentace tonkinensis, it 1s considered to be the correct name of this species because Parapentace 1s an invalid name without Latin description or diagnosis. Excentrodendron rhombifolium, reported from the same area as Burretioden- dron hsienmu, was based on a single specimen with rhomboid leaves. Since I have collected similar leaves from trees of B. Asienmu, and the rhomboid leaf is only one of various leaf shapes characteristic of the species, it is better to reduce EF. rhombifolium to synonymy under B. hsienmu. 2. Burretiodendron obconicum W. Y. Chun & F. C. How, Acta Phytotax. Sin. : t. 4. 1956. Excentrodendron obconicum (W. Y. Chun & F. C. How) H. T. Chang & R. H. Miau, Acta Sci. Nat. Univ. Sunyatsen 1978(3): 24. 1978. Type: China, Guangxi Prov., Longzhou, Wu-lian Hsiang, July 1955, C. F. Liang 31537 (holotype, scsi!; isotype, 1Bc!). FIGURE 1j, k. Tree 20 m tall. Leaves with petiole 4-8 cm long; blade elliptic to oblong- elliptic, 9-15 by 4.5-5.5 cm, long-acuminate at apex, acute at base, subcoria- ceous, glabrous, 3-basinerved, nerve axils glandular. Flowers unknown. In- fructescences racemose or dichotomous. Capsules obconical, ca. 5 cm long, glabrous. DISTRIBUTION. China (Guangx1). Burretiodendron obconicum 1s very similar to B. Asienmu in having three- basinerved, glandular, coriaceous leaves but differs in having obconical capsules and oblong-elliptic leaves. It is confined to southern Guangxi Province close to the border of China and Vietnam. 3. Burretiodendron brilletii Kosterm. Reinwardtia - 240. 1960. Parapentace brilletii Gagnep. Bull. Soc. Bot. France 90: 71. 1943. Type: Vietnam Hoa-binh, Brillet 19 (holotype, Pp; isotype, - FIGURE I. Tree; branchlets densely and minutely gray-stellate-tomentose. Leaves with petiole slender, 1.5—4 cm long, densely minutely stellate-pilose; blade broadly ovate to suborbicular, 7-11 by 4-10 cm, acuminate at apex, subcordate or truncate at base, subcoriaceous, both surfaces densely reticulate, 5-basinerved, nerve axils eglandular but with scattered stellate hairs present on nerves below. Inflorescences axillary fascicles. Staminate flowers in lax panicles; sepals lan- 378 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 71 ceolate, glabrous inside with small nectariferous basal part; petals spathulate, slightly longer than sepals; stamens 15 to 18; pistillode ovoid, 1 mm in diameter, stellate-pilose, apex tricuspidate. Carpellate flowers and fruits unknown. DistTRIBUTION. Northern Vietnam. ADDITIONAL SPECIMEN EXAMINED. Vietnam: Tudin, Nghien s.n. 29 Jan. 1961 (KUN). Burretiodendron brilletii, first reported by Gagnepain in 1943 under Para- pentace (based on a specimen bearing staminate flowers), 1s still poorly known. It is rare and confined to northern Vietnam. It differs from the other species of Burretiodendron in having broadly ovate to suborbicular, five-basinerved leaves with the base subcordate and the nerve axils eglandular. 4. Burretiodendron kydiifolium Hsu & Zhuge, sp. nov. FIGURE |d-g. Hacc species a speciebus ceteris generis bene distincta; differt a B. esquirolii Rehder foliis glabris, bracteolis 3, floribus é staminodiis 5, linearibus, ovariis sessilibus, a B. Asienmu foliis crasso-chartaceis, basi late cordatis, 7-9 palmi- nervatis. Tree up to 15 m high; branchlets densely brown-stellate-lepidote. Leaves with petiole 3.5-10 cm long, glabrous; blade chartaceous, suborbicular, 7-15 by 7-13 cm, sometimes 3-lobed at apex, cordate at base, entire, 7- to 9-basinerved. Staminate flowers 3 to 7 per raceme, carpellate ones solitary or rarely 2 or 3 per raceme; bracteoles 3 enveloping flower bud; sepals oblong- elliptic, 6-8 cm long, not glandular inside near base; petals flabelliform, apex praemorse, base cuneate; stamens 25 to 30; staminodes 5, linear, longer than stamens; ovary 5-angular with 5 free clavate styles. Capsule ellipsoid, 3-4 cm ong. Type. China, Yunnan Prov., valley of Yuanjiang River, 480 m alt., 12 May 1986, R. Zhuge 10418 (holotype, SwFc; isotype, KUN) DISTRIBUTION. Known only from Yunnan, China. ADDITIONAL SPECIMENS EXAMINED. China. YUNNAN Prov.: valley of Yuanjiang R., 820 malt., R. Zhuge 90669, 90670 (both swec). Burretiodendron kydiifolium corresponds to a set of flowering and fruiting specimens that I collected from the valley of the Yuanjiang River in Yunnan Province. It resembles B. siamense somewhat in having chartaceous, broadly ovate, five-basinerved leaves that are often three-lobed at the apex, but the two species are quite distinct. The flowers of B. Aydiifolium have three (vs. two) bracteoles and five staminodes. 5. Burretiodendron siamense Kosterm. Reinwardtia 6: 4. ¢. 2. 1961. Type: northern peninsular Siam, Khao Chawng Kachok, common in deciduous forest, Thaew Sinthiphongse 27 (holotype, BKF). FIGURE Im, n. Burretiodendron esquirolii auct. non Rehder: Smitin. Nat. Hist. Soc. Bull. Siam 19: 88. 1958 1990] ZHUGE, BURRETIODENDRON 319 Tree 8-12 m tall; branchlets densely very shortly brown-stellate-lepidote- pilose. Leaves with petiole up to 5 cm long, glabrous; blade broadly ovate, 8- 15 by 5.5-10 cm, sometimes 3-lobed at apex, truncate or subcordate at base, entire, chartaceous, 5- to 7-basi-nerved, glabrescent. Flowers solitary, sub- tended by 2 large, broadly ovate bracteoles; sepals lanceolate-ovate; petals broadly elliptic, slightly longer than sepals, membranous; stamens numerous. Capsules oblong-ellipsoid, 5-6 cm long, glabrous. DistTRisuTION. Northern peninsular Thailand, Mergui Archipelago of Burma. Kostermans’ (1961) original description of Burretiodendron siamense was based on fruiting specimens collected from northern peninsular Thailand and the Mergui region of Burma. The description of the flowers was added in 1965. This species can be readily distinguished from the others in the genus by its solitary flower with two bracteoles and its much bigger fruits. 6. Burretiodendron esquirolii (Lévl.) Rehder, J. Arnold Arbor. 17: 48. 4. 178. 1936. Pentace esquirolii Léveillé, Repert. Sp. Nov. 10: 147. 1911. Type: China, Kweichow [Guizhou] Prov., west of Lo-fou, Nov. 1905, J. Cay- alerie 2648 (holotype, E; photo and fragment of holotype, A). FIGURE la-c. 25. 1978. Type: China, Guangxi Prov., Long-lin, 26 Oct. 1957, C. C. Chang 4755 (holotype, 1BG!). Tree up to 20 m high; branchlets densely minutely stellate-pilose. Leaves with petiole 3-9 cm long, slender, densely stellate-pilose; blade ovate to sub- rotundate, 10-25 by 8-20 cm, acute at apex and often 3-lobed, cordate at base, denticulate, chartaceous, palmately 5- to 7-nerved, both surfaces densely stel- late-pilose. Plants polygamous, flowers 3 to 11 in small cymes; bracteoles 2, caducous, broadly ovate to elliptic; sepals ovate-elliptic, glabrous inside with 1 or 2 basal, oblong, elevated, glandular areas; petals broadly obovate, slightly longer than sepals, margin ciliolate: stamens ca. 30; ovary ovoid, 5-angular, with 5 clavate styles and gynophore. Capsules oblong in outline, 4-5 cm long. DistRIBUTION. China (southeastern Yunnan, southern Guizhou, and Guangxi provinces). ADDITIONAL SPECIMENS EXAMINED. China. YUNNAN Prov.: Yuanjiang, R. Zhuge 90607, 90608 (both swec), G. D. Tao 38162, 38164 (both ytBG); Shibing, Yunnan Forestry Institute 77153 (yNFI); Jinping, China-Soviet Union Exped. 2617, 2618 (both Kun). GuizHoU Prov.: Ceheng, Z. Y. Chao 1208 (KuN, PE); Luodian, 7. J. Luo 2661, 3667 (both Gzr1); Wengan, Li-po Exped. 1930 (GzF1). GUANGX1 Prov.: Longlin, C. C. Chang 10236 (BG, scBI); Tiane, 7. 7. Li 601284 (1BG). The first species of Burretiodendron, originally described as Pentace esquirolii, was based on three specimens collected by J. Cavalerie and J. Esquirol at Lofou and Yangli, Guizhou Province, China. Cavalerie 2648 was the holotype of the 380 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71] species. Esquirol 817 and 2717 were both cited in Flore du Kouy-Tchéou (Léveillé, 1915) under P. esquirolii, and the latter again under Eviolaena es- quirolii. They were all recognized by A. Rehder (1936) to be different from Pentace and other genera of the Tiliaceae in having unisexual flowers, five- winged fruits dehiscing into free cocci, and sepals with glands. Within Burre- tiodendron, B. esquirolii is easily distinguished from the other species by its pubescence and its longer fruit pedicels. Burretiodendron longistipitatum was based on a single specimen with fruit pedicels a little longer than those of B. esquirolii. This is quite a variable character, and the same variability is also found in specimens of B. esquirolii. This species is therefore reduced to synonymy under B. esquirolil. EXCLUDED SPECIES Burretiodendron umbellatum Kosterm., which was reported in 1962 from Thailand, was regarded as a doubtful species by Chang and Miau in 1978. As early as 1965 Kostermans, relying on additional and better materials from the same location, had revealed that the species was identical with Mansonia gagei Prain (Sterculiaceae). ACKNOWLEDGMENTS I would like to thank my advisers, professors Hsu Yung-chun (Southwest Forestry College, Kunming, China) and Chen Cheih (Kunming Botanical In- stitute, Academia Sinica) for comments and criticisms on the contents of this paper; Dr. J. E. Vidal (Muséum National d’Histoire Naturelle, Laboratoire de Phanérogamie, Paris) for kindly sending photographs of specimens; and Mr. Li Nang (Southwest Forestry College, Kunming) for the artwork. LITERATURE CITED CHANG, H. T., & R.H. MiAu. 1978. The taxonomy of Excentrodendroideae, Tiliaceae. (In Chinese.) Acta Sci. Nat. Univ. Sunyatsen 1978(3): 19-26. Cuun, W. Y., & F.C. How. 1956. phe novae arborum utilium Chinae meridionalis. (In Chinese.) Acta Phytotax. Sin. 1-18. ERDTMAN, G. 1952. os oe and plant taxonomy: angiosperms. Hafner Publ. Co., New Yo GAGNEPAIN, F. 1943. Tiliacées nouvelles d’Indochine. Bull. Soc. Bot. France 90: 70, 71 Hutcuinson, J. 1967. Tiliaceae. Pp. 471-497 in The genera of flowering plants. Vol. 2. Clarendon Press, Oxford, England. KosteRMAnNS, A. J. G. H. 1961. The genus Burretiodendron Rehder (Tiliaceae). Rein- wardtia 6: 1-16. 1965. ease meue ease notes. Bull. Bot. Surv. India 7: 130. LEVEILLE, H. 191 Flore du Kouy- cou REHDER, A. 1936, any meena a new genus of Tiliaceae. J. Arnold Arbor. 17: 47-49, SmitH, W. W., & W. E. Evans. 1921. Craigia, a new genus of Sterculiaceae. Trans. Bot. Soc. Edinburgh 28: 69-71 1990] OHASHI, DESMODIUM SCHUBERTIAE 381 DESMODIUM SCHUBERTIAE (LEGUMINOSAE), A NEW SPECIES FROM CAMBODIA AND VIETNAM HiIROYOSHI OHASHI! A new species of Desmodium (Leguminosae- Siege D. schubertiae Ohashi, is found in Cambodia and Vietnam. It resembles D. rubrum (Lour.) DC. and 1s thought to have evolved from D. heterocarpon (L.) D While preparing a treatment of Desmodium Desv. for the Flore du Cambodge, du Laos et du Vietnam at the Harvard University Herbaria in | I found a new species of Desmodium among the specimens collected by Poilane in Cambodia and Vietnam and kept in the herbarium of the Laboratoire de Phanérogamie, Muséum National d’Histoire Naturelle (p). It is named in honor of Dr. Bernice G. Schubert, of the Arnold Arboretum of Harvard University, in recognition of her distinguished contribution to taxonomic research in Des- modium. In the present paper, the new species is described and its taxonomic relationships are discussed. Desmodium schubertiae Ohashi, sp. nov. FIGURE. Species haec a Desmodio rubro differt foliolis inferioribus dense sericeis, anguste ellipticis vel oblongo-ellipticis; calycibus lobis adaxialibus apicis pro- funde bifidis; ovarus uncinulato- et glandulifero-puberulis; leguminibus ad su- turas inferiores uncinulato-puberulis, cetero glabrescentes, ad suturas superiores incrassatis. Type: Vietnam (Annam), Prov. du ean massif du Chu Yang Siuh, 1500 a 1800 m alt., sol granitique médiocre, 22 April 1941, E. Poilane 32489 (ho- lotype, P; isotype, TUS; photo of serene A). Shrubs ca. 2 m high, freely branched; young branches densely covered with appressed, silky, white or ferrugineous hairs and spreading, minute, hooked at base, ciliate, scarious, striate, glabrous inside, densely appressed-pubescent outside; petiole 8-15 mm long, densely strigose with appressed, ferrugineous hairs. Leaflets narrowly elliptic or oblong-elliptic, 2-4 cm long, 8-18 mm wide, obtuse to emarginate at apex, obtuse or rounded at base, entire and densely ciliate, subcoriaceous, the upper surface thinly pubescent with appressed soft hairs ca. 0.5 mm long, the lower surface densely so with ascending silky hairs ‘Biological Institute, Faculty of Science, Tohoku University, Sendai 980, Japan. © President and Fellows of Harvard College, 1990. Journal of the Arnold Arboretum 71: 381-384. July, 1990. 382 JOURNAL OF THE ARNOLD ARBORETUM Desmodium schubertiae: a, holotype; b, pods, x 2. [VOL. a 1990] OHASHI, DESMODIUM SCHUBERTIAE 383 to 1.5 mm long, the principal lateral nerves 6 to 8 on each side of midrib, prominent and not directly reaching margin, the cancellate veins inconspicuous below; stipels subulate, 4-5 mm long, silky outside. Inflorescences terminal and occasionally axillary, pseudoracemose, 7-14 cm long, 2-flowered at each node: rachis with dense spreading, hooked hairs ca. 0.3 mm long. Pedicels 5- 6 mm long, spreading-hairy with straight, rigid hairs less than 0.5 mm long, glandular hairs ca. 0.3 mm long, and minute, hooked hairs ca. 0.1 mm long. Primary bracts ovate with acuminate apex, 6-7 mm long, 2-2.5 mm wide, scarious, striate, glabrous inside, with dense minute, hooked hairs and dense subspreading, straight hairs outside; secondary bracts and bracteoles absent. Calyces 3.5-4 mm long, pubescent with straight hairs and minute hooked hairs, 4-lobed, lobes longer than tube, the upper one 2.5-3 mm long, bifid at apex, teeth ca. | mm long, the lateral ones narrowly triangular, the lowest one longest. Flowers ca. 6 mm long, violet-rose; androecium diadelphous, the filaments glabrous; ovary and lower part of style covered with minute hooked hairs and glandular hairs. Pods ascending, sessile, narrowly oblong, 1. 5-2.5 cm long, ca. 4 mm wide, straight, with 3 to 5 articles, dehiscent along lower suture, this undulate, with minute thinly uncinate hairs, the upper suture almost straight, much thickened, glabrescent, the lateral surfaces with prominent reticulate veins, glabrescent; articles 4.5—-5 mm long. Seeds compressed-reniform, ca. 3.5 mm long, 2.5 mm wide, reddish purple, rim-arillate around hilum DDITIONAL SPECIMENS EXAMINED. Cambodia: Bokor, montagnes de lEléphant, 1000 m alt., E. Poilane 23059 (p, TUS; photo A) Desmodium schubertiae belongs to sect. Nicolsonia of subg. Sagotia because its pods, calyces, and bracts agree with those characterized for this section. It is a group of 12 Asiatic species (Ohashi, 1973). Of these, D. ferrugineum Wallich ex Thwaites, D. heterocarpon (L.) DC., D. jucundum Thwaites, D. nemorosum F. Mueller ex Bentham, and D. rubrum (Lour.) DC. are similar to the new species in general appearance. These six species can be distinguished from each other by the following key: 1. Articles of pod at least twice as long as broad. ................---- D. nemorosum. 1. Articles of pod less than twice as long as broad. 3. Pods stalked, reflexed when mature. .................005- D. ferrugineum. atu 4. Leaflets coriaceous, with lateral nerves adel below; leaves 3-foliolate; stipules broadly ovate, 5-6 mm wide; articles 5-6 mm long. ........... i Synapse Sen ta cs ekg onder ard eet ae gee eer eg D. jucundum. 4. Leaflets not as above; leaves 1- and 3- or all 3-foliolate; a Ns narrowly triangular, less than 3 mm ae articles less than 4.5 mm long. ........ 2. Leaves |-foliolat 5. Leaflets coer below, elliptic to nearly orbicular, with 4 to 6 pairs of lateral nerves; upper calyx-lobes minutely bifid at apex; ovary glabrous; oat glabrous, with upper suture only slightly thickened. ........... Tu 5. Leaflets densely silky-hairy below, narrowly elliptic or oblong- elliptic, oa 6 to 8 pairs of lateral nerves; upper calyx-lobes distinctly bifid at apex; ovary 384 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 with minute hooked hairs and glandular pee glabrescent except on lower suture, with upper suture much-thickened. .......... D. schubertiae. Although Desmodium schubertiae 1s most similar to D. rubrum in the char- acters indicated in the foregoing key, it is most closely related to D. hetero- carpon. There are fundamental similarities among these three species in habit, leaflets, flowers, and pods. However, D. schubertiae is found in mountainous regions between 1000 and 1800 m elevation, while D. rubrum is found at the seaside or in open places at low altitudes. There is a significant difference in distribution between these two species. Desmodium heterocarpon occurs in the tropical, subtropical, and warm regions of India, southeastern and eastern Asia, the Pacific islands, and Australia. It is common on sunny roadsides through forests, in thickets, or in grasslands, from sea level to more than 1000 m altitude, and is possibly sympatric with both D. schubertiae and D. rubrum in Indochina. In Indochina and neighboring regions are concentrated Desmodium harmsii Schindler (southern Vietnam), D. rubrum (Vietnam; Hainan and Kwangtung, People’s Republic of China), D. schubertiae (Vietnam, Cambodia), and D. strigillosum Schindler (Burma (rare), Laos, Cambodia, and Vietnam), all en- demic species of sect. Nico/sonia. This distribution pattern is a characteristic feature of the section. Moreover, D. griffithianum, which occurs in India (As- sam), Burma, Thailand, Laos, Vietnam, and southwestern China, and D. top- pinil, endemic to Burma, show related patterns of distribution. All these species are considered to be most closely related to D. heterocarpon (Ohashi, 1973) and perhaps are derived from it. Desmodium schubertiae is therefore thought to be derived from D. hetero- carpon in a mountainous region of Cambodia and Vietnam, and D. rubrum may also be evolved from D. heterocarpon, having adapted to open places near the sea or at low elevations. LITERATURE CITED Onasui, H. 1973. The Asiatic species of Desmodium and its allied es ied nosae). 318 pp., 76 pls. Ginkgoana 1. Academia Scientific Book, Inc., Tokyo 1990] AL-SHEHBAZ, DRABA 385 A NOTE ON THE CHILEAN ENDEMIC DRABA THLASPIFORMIS (BRASSICACEAE) IHSAN A. AL-SHEHBAZ! new combination based on Ludema thlaspiforme is proposed in Draba. The species is described and illustrated. Draba thlaspiformis (Philippi) Al-Shehbaz was originally described by Phi- lippi (1872) as a questionable species of Eudema Humb. & Bonpl. Gilg and Muschler (1909) transferred it to Lesquerella S. Watson but gave SA at for that transfer. Schulz (1924) maintained it in the latter genus and later (Schulz, 1929) gave it a new name, Draba philippii O. E. Schulz, which he based on a type different from that of £. thlaspiforme Philippi. The species is a perfectly good representative of Draba and has nothing to do with either Eudema or Lesquerella. Philippi’s name is the earliest legitimate name for the species, and there are no nomenclatural obstacles preventing its transfer to Draba. Therefore, a new combination based on EF. thlaspiforme is needed in Draba. A detailed description and an illustration are provided below in order to assist in the determination of this very rare and obscure species. Draba thlaspiformis (Philipp1) Al-Shehbaz, comb. nov. FIGURE. Eudema thlaspiforme Philippi, Anal. Uniy. Santiago 41: 675. 1872. Lesquerella thlas- piformis (Philippi) Gilg & Muschler, Bot. Jahrb. Syst. 42: 466. 1909. Type: Chile, [Region hy a nee de Sandee: ] valle de Maipo, mina Cristo, Davila s.n., 1870 holotype, sGo; isotypes, B!, GH!). Draba philippii O. E. Schulz, ‘Notiz ‘bl. Bot. Gart. Berlin-Dahlem 10: 599. 1929, Type: Chile, [Region Metropolitana de Santiago,] Cordillera de Maipu, Philippi s.n. (ho- lotype, B!). Perennial herbs. Trichomes short stalked, stellate, to 0.3 mm in diameter; rays basically 4, | to all furcate or trifid at variable distances from stalk, very rarely all unbranched. Stems several from base, ascending, 7-12 cm tall, sparsely pubescent. Basal leaves rosulate, short petiolate, oblong, obtuse; cau- line leaves sessile, oblong to oblong-elliptic, 10-15 mm long, 4-6 mm wide, entire, sparsely pubescent. Inflorescences corymbose racemes, elongated in fruit, lowermost flowers bracteate. Sepals erect, somewhat persistent, broadly oblong, 2.5-3 mm long, |.5-2 mm wide, sparsely pubescent, rounded at apex, mem- branaceous margin to ca. 0.4 mm wide. Petals clawed, broadly obovate, 6-7 mm long, 2.5-3 mm wide, white. Filaments erect, conspicuously dilated at ‘Arnold Arboretum, Harvard University, 22 Divinity Avenue, Cambridge, Massachusetts 02138; current address, Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166-0299. © President and Fellows of Harvard College, 1990. Journal of the Arnold Arboretum 71: 385-387. July, 1990. 386 A Dobbs Ria thlaspifor ia ae Philippi s.n.): a, infructescence and portion of stem: b, sepal: petal; d, stamen; e, fruit. Scale bars = 1 cm (a), | mm (b-e). base, 3-4.5 mm long, white; anthers ovate, 0.6-0.8 mm long. Nectar glands 4, | on each side of lateral stamen, low. Fruiting pedicels divaricate, straight, (3-)5-10(-15) mm long, glabrous on upper surface, sparsely pubescent on lower. Fruits sessile, suborbicular, strongly flattened parallel to septum, 5-8 mm long, 4-6 mm wide; valves thin, rounded at apex and base, glabrous, obscurely nerved; septa complete; style slender, 2-3 mm long, glabrous; stigma slightly 2-lobed. Mature seeds not seen. Draba thlaspiformis is apparently known only from the two type collections that were made some 120 years ago. It is closely related to D. gilliesii W. J. Hooker & Arn., from which it is easily distinguished in having untwisted, suborbicular fruits and bracteate lower portions of inflorescences, as well as 1990] AL-SHEHBAZ, DRABA 387 lacking median nectaries. In contrast, D. gilliesii has oblong to lanceolate or linear, usually twisted fruits, ebracteate inflorescences, and median nectaries that subtend the bases of paired filaments. ACKNOWLEDGMENTS I am grateful to Oanh Tran for typing the manuscript and to Elizabeth B. Schmidt and Stephen A. Spongberg for their editorial advice. LITERATURE CITED Gita, E., & R. MUSCHLER. 1909. eu eae stidamerikanisch- en Cruciferen. Bot. Jahrb. Syst. 42: 4 i, Puiuiprr, R. A. 1872. Cruciferae. [n: a de las nuevas plantas incorporadas ltimamente en el herbario chileno. Anales Univ. Santiago 41: 666-6 Scuutz, O. E. 1924. Cruciferae-Sisymbrieae. Jn: A. ENGLER, ed., Pflanzenr. IV. 105(Heft 86): 1-396. 929. Amerikanische Cruciferen verschiedener Herkunft. Notizbl. Bot. Gart. Berlin- Dahlem 10: 558-564. BOOK REVIEW 389 BOOK REVIEW ~— Same and Woody Vines of Northern Florida and Adjacent Georgia Alabama, by Robert K. Godfrey, with the majority of illustrations by Mee Darst. University of Georgia Press, Athens and London, 1989 ix + 734 pp. ISBN 0-8203-1035-2. $50 hardcover. As might be expected from his previous splendid publications and his ex- traordinary field knowledge, Robert Godfrey’s new book on the woody plants of northern Florida and adjacent areas is a superb addition to the botanical literature of the Southeast. Having done extensive field work myself in the area many years ago, I perused this book very carefully to see which of my woody species he had omitted and which I might have overlooked in southwestern Georgia. Of the former, Corylus americana Walter and Crataegus brachy- acantha Englem. & Sarg. apparently have not been found in northern Florida. Agave virginica L. and Erythrina herbacea L. surely have been found there but arguably lack sufficient woodiness in northern Florida to qualify for the book, although in peninsular Florida FE. herbacea does become a sizable shrub or small tree. I had failed to find, or at least to recognize, perhaps ten species attributed by Godfrey to southwestern Georgia. In 1946-1949 most of them were not generally recognized as distinct species or were not known from southern Georgia. I should like to look for them there now. I was particularly delighted to note that Bumelia thornei Cronq., of the Sapotaceae, has not only been rediscovered in Georgia but has been added by Godfrey to the Florida flora. Godfrey’s taxonomy ts rather traditional. His species and generic concepts and nomenclature are most acceptable and up to date, but his disinterest in phylogeny is apparent. Nandina is still retained in the Berberidaceae, Sambucus and Viburnum in the Caprifoliaceae, Phoradendron in the Loranthaceae, Nyssa in the Nyssaceae, and Decumaria, Hydrangea, Itea, Philadelphus, and Ribes in the Saxifragaceae. On the other hand, the family name Avicenniaceae Is accepted for Avicennia, the Black Mangrove. For ready reference the families are arranged alphabetically in the major groups, but unfortunately four of the larger families must be sought under the older names Gramineae, Palmae, Guttiferae, and Leguminosae. I should have preferred that species, like genera, be consistently arranged in alphabetical order. The category “‘subspecies” should have been employed more frequently. I have long believed that use of that category would solve some of the most difficult taxonomic problems among southeastern woody plants—for example, in such closely related pairs as Juniperus virginiana—J. silicicola, Taxodium distichum—T. ascendens, Myrica cerifera-—M. pusilla, Nyssa sylvatica-N. biflora, Toxicodendron radicans-T. toxicarium, Cyrilla racemiflora—C. parvifolia, Tilia © President and Fellows of Harvard College Journal of the Arnold Arboretum 71: 389, 390. ait 1990. 390 JOURNAL OF THE ARNOLD ARBORETUM americana-T. heterophylla, Clethra alnifolia-C. tomentosa, Halesia diptera var. diptera and var. magniflora, and probably many more. The subspecific treatment of the southern sugar maples is a good example of what can thereby be achieved. The original descriptions and keys are accurate, thorough, well done, and reflective of the author’s field acumen. The keys particularly are readily usable for identification and for differentiation of taxa because of the emphasis on vegetative characteristics. Excellent full-page drawings are supplied for almost all species, both indigenous and naturalized, and add much to the value of the book. Statements on habitats and distribution again reflect the broad field knowledge of the author and his careful study of floristic literature. A useful glossary follows the introduction. Several pages of pertinent refer- ences precede the indexes to common and scientific names (I would have preferred that the two indexes be combined). Proofreading must have been painstaking, for I found very few errors. On page 132 A. iIncanna and on page 133 4. incarna are used for Asimina incana (Bartram) Exell. Apparently God- frey prefers to use Bartram’s original misprint A. /ncarna for this species, despite Exell’s selection of 4. incana as the intended correct name. Job Kuyt might be surprised at the misspelling of his name on page 4 This book is a ““must” addition to the bookshelf for anyone interested in identification of woody plants of the American Southeast or in the southeastern flora and woody plants, ecology, and phytogeography in general. As book prices go nowadays, $50 is reasonable for such an informative, well-illustrated, well- printed, and well-bound tome.— RosBert F. THORNE, Rancho Santa Ana Bo- tanic Garden, Claremont, California 91711. Now Available and Complete FLORA OF THE LESSER ANTILLES by Richard A. Howard and Collaborators The Flora of the Lesser Antilles, a long-term project of Dr. Richard A. Howard, former director of the Arnold Arboretum, has been brought to a conclusion with the publication of volumes 5 and 6 during 1989. Other volumes in the series are still available, either individually or as part of a complete set: VOLUME |. Orchidaceae $20 VOLUME 2. Pteridophyta 25 VOLUME 3. Monocotyledoneae 35 VoLuME 4. Dicotyledoneae, Part | 73 VoLuME 5. Dicotyledoneae, Part 2 85 VoLuME 6. Dicotyledoneae, Part 3 85 Special price for the complete 6-volume set $260* Special price for volumes 4-6 $203" Also Available ARNOLD ARBORETUM PLANT INVENTORY A comprehensive plant inventory of the living collections of the Arnold Arboretum, together with a statement of the Arboretum’s accession policy and a fee schedule for obtaining propagating material from the collections. 21" Orders with payment in U. S. funds, including a $2 ($4 foreign) shipping and handling charge per book, should be addressed to the attention of Frances Maguire, Arnold Arboretum, 125 The Arborway, Jamaica Plain, Massachusetts 02130, U. S. A. Checks should be made payable to the Arnold Arboretum. *Prices marked with an asterisk include postage and handling. Journal of the Arnold Arboretum July, 1990 CONTENTS OF VOLUME 71, NUMBER 3 The Genera of Cupressaceae (Including Taxodiaceae) in the South- eastern United States. JEFFREY A. HART AND ROBERT A. PRICE ............0..00000- The Phylogenetic Significance of Stomata and Trichomes in the Labiatae and Verbenaceae. PHILIP DCANTING: 446 bosek cob asad 4 6tdetdsbecedidetati ht! On the Genus Burretiodendron Sensu Lato (Tiliaceae). REN AHOGE: nv ane taca ret be re oe Be oe Snares nore ates Desmodium schubertiae (Leguminosae), A New Species from Cam- bodia and Vietnam. HimOVOSET ORASH .s.osccuevaaceledw bed kaueeibeescereiwwss A Note on the Chilean Endemic Drabha thlaspiformis (Brassicaceae). IHSAN A. AL-SHEHBAZ .........0 00000 c cee eee teen teens Book Review | c645 24 S044 be hk oe og 2 Kw hE Eas a ees 275-322 323-370 371-380 381-384 385-387 389, 390 Volume 71, Number 2, including pages 145-273, was issued 5 April 1990. JOURNAL OF THE ARNOLD ARBORETUM HARVARD UNIVERSITY VOLUME 71 NUMBER 4 ISSN 0004-2625 Journal of the Arnold Arboretum The Journal of the Arnold Arboretum (ISSN 0004-2625) is published quarterly in January, April, July, and October for $70.00 per year, plus $10.00 or $15.00 postage for addresses outside of the United States, by the Arnold Arboretum of Harvard University. It is printed and distributed by Allen Press, Inc., 1041 New Hampshire Street, Lawrence, Kansas 66044. Second-class postage paid at Lawrence, Kansas. POSTMASTER: send address changes to Journal of the Arnold Arboretum, % Allen Press, Inc., P.O. Box 368, Lawrence, Kansas 66044. Because publication of the Journal of the Arnold Arboretum is being suspended after Volume 71 (1990) has been completed, subscriptions can only be accepted for the present volume. Remittances for Volume 71 should be sent to Journal of the Arnold Arboretum, 1041 New Hampshire Street, Lawrence, Kansas 66044, U.S.A. Claims will not be accepted after six months from the date of issue. Volumes 1-51, reprinted, and some back numbers of volumes 52-56 are available from the Kraus Reprint Corporation, Route 100, Millwood, New York 10546, Loa EDITORIAL COMMITTEE S. A. Spongberg, Editor E. B. Schmidt, Managing Editor P. F. Stevens, Book Review Editor P. S. Ashton K. S. Bawa C. E. Wood, Jr. Printed at Allen Press, Inc., Lawrence, Kansas COVER: The stylized design appearing on the Journal and the offprints was drawn by Karen Stoutsenberger. THIS PUBLICATION IS PRINTED ON ACID-FREE PAPER. JOURNAL OF THE ARNOLD ARBORETUM VOLUME 71 OcToBER 1990 NUMBER 4 THE GENERA OF CARDUEAE (COMPOSITAE; ASTERACEAE) IN THE SOUTHEASTERN UNITED STATES! RANDALL W. ScotTT? Tribe Cardueae Cassini, Jour. Phys. Chim. Hist. Nat. Arts 88: 155. 1819, “Carduineae.” Perennial to annual herbs [rarely shrubby or tree- like]. Leaves alternate, sin- ile) and partly radiating marginal flowers. Involucral bracts ovate to lanceolate [bristles], imbricate in several series, often with an apical spine, bristle, or membranaceous appendage. Receptacle setaceous, bristly, hairy or honey- combed [naked or with connate bracts partially split at the upper margin]. f+} 1 epared for the Ger S United States, a long-term project made possible by er from the National Science Foundation and at this writing supported by BSR-8415769 (Carroll E. Wood, Jr., principal investigator), under which this account was prepared, and BSR- 8415637 (Norton G. Miller, tie ene This treatment, the 134th in the series, follows the format established in the first paper (Jour. Arnold Arb. 39: 296-346. 1958) and continued to the present. The area covered by the Generic Flora includes North and South Carolina, Georgia, Florida, Tennessee, Alabama, Mississippi, Arkansas, and Louisiana. The descriptions are based primarily on the plants of this area, with information about extraregional members oe family, sublamnys § aoe subtribe, or soraee in brackets [ ]. Those references that I have not verifi I want to thank Carroll Wood for his extensive and constructive comments, both editorial and eh pasos thanks go to Tina Ayers without whose support this project might never have been completed. [Ihsan Al-Shehbaz and Barbara Callahan were very helpful with bibliographic search- es, while I was at Harvard. For assistance manifested in a variety of ways, I want to thank David Boufford, Barbara Nimblett, Cathy Paris, John Pruski, Stephen Spongberg, and Emily Wood. The illustration was iS raw bode an earlier NSF grant by Rachel A. Wheeler supervised by George K. Brizicky from ted by C. E. Wood, Jr. (a-e, j) and J. K. Small (8/22, GH) in Florida, an ne 2 R. B. Channell and H. F. L. Rock in southeastern North Carol old Arboretum of Harvard University, 22 Divinity Avenue, ane cee a. - 138. Present address: Department of Biolosieal Sciences, Northern Arizona University, P. O. B Flagstaff, Arizona 8601 © President and Fellows of Harvard College, 1990. Journal of the Arnold Arboretum 71: 391-451. Renee 1990. 392 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Pappus of simple or plumose bristles or paleaceous, uniseriate (Echinops) or multiseriate, free or basally fused, of nearly equal length (Carduinae), or the inner longer (Centaureinae). Corollas tubular (Echinops), actinomorphic, di- vided into tube and limb, the outer florets occasionally zygomorphic (falsely radiate) (Cirsium, Centaurea), lobes narrow or linear [triangular] and, in many cases, elongated. Anthers with sagittate basal appendages, these entire or den- ticulate [laciniate]; filaments hairy, papillose or glabrous; pollen variable, but basic type spherical, spiny, tricolporate (see Wagenitz, 1955; Stepa). Styles very long, united and only diverging at the tips (see Solbrig, 1963, fig. 2, /) or widely divergent or very short and divergent, often with a distinct collar of hairs at or below the base of the flat, elliptical or semi-cylindrical, acute to acuminate, often papillose branches, the stigmatic papillae on the inner surface. Achenes mostly obovate, elongate-conoidal or fusiform, commonly somewhat bossed on the upper, abaxial side, occasionally laterally compressed, glabrous or sparsely covered with one-celled hairs or shortly forked twin hairs to densely villous (Echinops);, often with a distinctly marked apical crown or margin; detachment area of achenes straight and basal (Echinops), straight and lateral- adaxially, lateral-abaxially, rarely basally oriented (Carduinae) or concave and lateral-adaxially oriented (Centaureinae); seed often with a basal or lateral (Centaurea and related genera) hilum. (Carlineae Cass., Jour. Phys. 88: 152. 1819; Centaurieae Cass., Ibid. 154; Echinopseae Cass., Ibid. 157; Cynareae Less., Linnaea §: 128. 1830, nom. illegit. [includes Carduus L., type of Cardueae Cass.].) Type GENUS: Carduus L. Nearly 80 genera (ca. 2500 spp.; Dittrich, 1977) traditionally placed in four subtribes, but more recently (Dittrich, 1977) segregated into three tribes: Echi- nopseae Cass., Carlineae Cass., Cardueae Cass., with the last divided into sub- tribes Carduinae Dumort. and Centaureinae Dumort. Genera known from the Southeast belong to the Cardueae (Dittrich, 1977). Arctium L., Carduus L., Cirsium Miller, Onopordum L., and Silybum Adanson were placed in the Carduinae and Centaurea L. and Cnicus L. in the Centaureinae. The Echinopseae Cass. may be represented in the Southeast by Echinops L., but there is no solid evidence that any of the cultivated species has escaped from cultivation there. Centers of diversity are in Europe and western Asia. About 35 species in eight genera occur in the southeastern United States. The great majority of the North American representatives of this tribe have been introduced from Europe. Native taxa from the Southeast occur only in Cirsium (all but two of the nearly fifteen species are indigenous) and Centaurea (one of eight species 1s native). Considerable diversity of opinion has been expressed in the nomenclatural history of the Cardueae. In a serial treatment that extended over a period of nearly fifteen years, Cassini (1816-1830) described numerous taxa in the tribes Cardueae Cass., Centaurieae Cass., Carlineae Cass., and Echinopseae Cass. These tribes were first described under their French names (Cassini, 1817), but were later given Latin names by Cassini (1819). Two years after Cassini finished his serial treatment, Lessing presented a more inclusive treatment of the family wherein he placed most of six of Cassini’s tribes under the illegitimate name Cynareae Less. Bentham (1873a, b) included four of Cassini’s tribes in the 1990] SCOTT, GENERA OF CARDUEAE 393 Cardueae [““Cynaroideae”’] in what came to be the traditionally accepted view of the tribe. Most recently, Dittrich Se recognized three tribes in lieu of Bentham’s single broadly circumscribed o As indicated by Dittrich (1977), the ee name Cynareae, coined by Lessing, has been used by many authors (e.g., Boissier; Bentham; Hoffmann) and still persists in contemporary works (e.g., Carlquist, 1965; Cronquist, 1955, 1977; Dittrich et a/., 1979, 1980; Clapham et al/.; Jeffrey, 1968; Johnson 1974, 1975). The following genera known from the southeastern United States were in- cluded in Cassini’s Cardueae: Arctium (as Lappa Scop.), Carduus, Cirsium, Onopordum, and Silybum. He placed Cnicus and Centaurea in the Centaurieae. Several taxa treated by Cassini as subtribe Xeranthemées of the tribe Carlineae have been placed in the Cardueae by subsequent workers. In a treatment published a year before Cassini finished his extended serial treatment of the Compositae, Dumortier retained the tribes Echinopseae (as Echinopsideae) and Cardueae (as Carduaceae), but reduced Cassini’s tribes Carlineae and Centaurieae to subtribes of the Cardueae. Lessing incorporated all or parts of five of Cassini’s tribes (Arctotideae, Calenduleae, Cardueae, Centaurieae, Echinopseae) and the subtribe Xeran- theminae Cass. as subtribes of the tribe Cynareae. Many members of Cassini’s tribe Carlineae were treated under the subtribe Carduinae, while two new subtribes, Cardopateae Less. and the Othonninae Less., were proposed for other representatives of this group. With few exceptions, A. P. de Candolle, among others, accepted Lessing’s broad view of the Cardueae. De Candolle reduced subtribe Othonninae to a division of the Calenduleae, elevated three of Lessing’s divisions (Carlineae, Serratuleae, Silybeae) to subtribal status and extracted a twelfth subtribe (Car- thaminae) from Lessing’s subtribe Centaurieae. As treated by Bentham (1873a), the Cardueae (““Cynaroideae”’) contained nearly 36 genera in the four subtribes Carduineae, Carlineae, Centaurieae, and Echinopsideae. Lessing’s subtribes Cardopatiinae and Xerantheminae (sensu De Candolle) were incorporated into the subtribe Carlineae, while De Candolle’s subtribes Carthaminae and Silybinae were included in the Centau- rieae and Carduineae, respectively. Similarly, the subtribe Serratulinae (Less.) DC. was divided between the Carduineae and Centaurieae. Cassini’s tribes Arctotideae and Calenduleae were resurrected from the subtribal status ac- corded them by Lessing and De Candolle. Cassini’s system was considered to be a “natural” one by Dittrich (1977), who divided the Cardueae into tribes Echinopeae [sic] (2 genera), Carlineae (11 genera), and Cardueae (66 genera). These he grouped together under the subfamily name Cynaroideae, a name used by Bentham at the tribal level, but without any prior usage at the subfamilial level known to the present author. Dittrich noted that no single character unites the three tribes as he delimited them, not even the traditional stylar character, which he found to be limited to the Cardueae sensu stricto, and that between these three groups no connecting forms are known. The Echinopseae are distinguished by their globose, sec- ondary heads comprised of numerous uniflorous heads and by their corollas 394 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 with long, slender lobes. The Carlineae have capitula with numerous florets, characteristically short corolla lobes, and receptacles with connate paleae. The Cardueae have long, slender corolla lobes, bristly receptacles, flat and elliptical style branches (in contrast to the half-cylindric or cylindric, usually acuminate branches of the other two tribes), papillose to hairy anther filaments, and laterally compressed achenes. Dittrich further divided the Cardueae into the Carduinae and Centaureinae on the basis of characteristics of the pappus (bris- tles of almost equal length and form vs. bristles or scales elongated from outside to inside; the internal pappus elements often morphologically different from the outer, forming a ‘‘double pappus’’) and the detachment areas of the achenes (straight, lateral-adaxially, rarely basally oriented vs. always concave and of lateral-adaxial orientation). Whereas Bentham recognized 10 genera of the subtribe Centaureinae, Dittrich (1977) recognized 27, many of which are seg- regates from Centaurea. In a chemical review of the Cardueae sensu lato, H. Wagner (1977) preferred to view the group as comprising two tribes, the Carlineae and the Cardueae. The Echinopseae are apparently treated as a subtribe of the Carlineae, while the Cardueae are divided into two subtribes, the Carduinae and the Centau- reinae. The taxa of both tribes are inclusively referred to as the Cynareae throughout his paper. As delimited by Bentham (1873a), the Cardueae comprise the largest tribe of Compositae in the North Temperate region of the Old World. Bentham thought that the tribes with the closest connections to the Cardueae (Cynareae) were the principally South American Mutisieae (subtribe Gochnatinae) and the South African Arctotideae (subtribe Gorterinae). The latter tribe he considered to be a possible link between the northern tribes Cardueae and Anthemideae. On the basis of seed anatomy, Lavialle linked the Cardueae (Cynarées) to the Mutisieae (Mutisées) through Yeranthemum L. Cronquist (1955, 1977) reaffirmed the close relationship between Mutisieae, Vernonieae, Arctotideae, and Cardueae, but disagreed with Carlquist (1961, 1976), who grouped them with the Lactuceae in subfamily Cichorioideae. While the close relationship of the Arctotideae and Cardueae is not disputed, the position of the Arctotideae was considered problematic by Cronquist (1955, 1977), since he considered them a bridge between the radiate and discoid tribes. Carlquist (1976) proposed two subfamilies for the Compositae (Asteraceae) and aligned the Cardueae with tribes Arctoteae, Lactuceae (“Cichorieae’’), Eupatorieae, Vernonieae and Mutisieae in the subfamily Cichorioideae (“the mutisioid line” vs. the subfamily Asteroideae or ‘“‘helianthoid line”). Both Carlquist (1976) and Jeffrey (1977) have suggested that the ligulate flowers of the Cichorioideae are modified bilabiate flowers distinctly different and of a separate origin from the ray flowers found in the Asteroideae. Carlquist con- sidered the expanded flowers found in the Arctoteae and Cichorioideae to have originated in the same manner and included the Arctoteae in this subfamily. Cronquist (1977) disputed Carlquist’s conclusions about the Arctotideae on the grounds that the external morphology of the ray flowers found in many representatives of the Arctoteae is much like that of the Asteroideae and that Carlquist rested his opinions concerning the alignment of the Arctoteae on 1990] SCOTT, GENERA OF CARDUEAE 393 Didelta tomentosum Less., the ray flowers of which have three, four, or five teeth. Carlquist (1976) thought that the Cardueae originated in the Old World (Northern Hemisphere) and that the Arctoteae are the result of “explosive evolution of a stock related to a cichoriead-carduaead complex in southern Africa.’ While warning that wood anatomy is not a particularly good character on which to establish tribal relationships, Carlquist (1965) noted that the wood anatomy of the Cardueae would not contradict Bentham’s (1873b) proposed relationships between the tribes Cardueae, Mutisieae, and Vernonieae. Jeffrey (1978) recognized 17 tribes and the two subfamiles Lactucoideae (correctly, Cichorioideae Kitamura) and Asteroideae in his treatment of the Asteraceae. He placed the tribes Cardueae and Arctoteae in subfamily Lac- tucoideae, along with tribes Lactuceae, Arctoteae, Eremothamneae (a mono- typic tribe from South Africa), Vernonieae, Liabeae, and Eupatorieae. The Cardueae were said by Jeffrey to have three subtribes that we may presume to be Carduinae, Centaureinae, and Echinopsinae, since he did not follow Dit- trich’s recognition of the Echinopseae at the tribal level. In a study of particular interest, Bremmer tacitly supported Dittrich’s di- vision of the Cardueae into three tribes. Bremmer contended that the phylo- genetic relationships between the Arctoteae and Dittrich’s three tribes are un- clear and that the Echinopseae, Carlineae, and Cardueae are conceivably monophyletic taxa. If Dittrich’s three tribes are treated as a single tribe from which the Arctoteae are excluded, Bremmer claimed that the Cardueae sensu lato may not represent a monophyletic group. This is illustrated in the clado- gram presented by Bremmer in which the four groups are shown as independent lineages arising from a common ancestor. The Cardueae sensu stricto were characterized by three unique characters: 1) papillose to hairy anther filaments, 2) testa epidermal cells in a lignified palisade layer, becoming hard and dark brown to black, and 3) nuclear endosperm. The Echinopseae were distinguished by their secondary heads composed of numerous uniflorous heads and by echinopsine alkaloids, but they were noted as lacking the lactiferous tissue typical of other members of the subfamily Cichorioideae. Neither the Carlineae nor the Arctotideae are delimited by unique characters. The Carlineae are set off as having paleaceous receptacles and shortly lobed disc corollas, while the Arctotideae are said to have true apically three-lobed ray florets and to lack the caudate anthers found in the other three tribes. Haslett ef a/. and Boulter et a/. used the analysis of plastocyanin amino acids extracted from species representing eight tribes of the Compositae in a phy- logenetic study of the family. Their data support the view that the Compositae are a natural group and indicate a close relationship between the Lactuceae and Cardueae. Indicative of the Cardueae’s possibly primitive position among the Compositae was the basal position occupied by Centaurea in the phylogeny generated from the data of Boulter et a/. Bremmer noted that the Cardueae and Lactuceae are the only ‘“‘cichorioid”’ tribes included in these studies and that the positions of Senecio L. and Taraxacum Wigg. indicate a lack of suf- ficient data from which to draw an unequivocal conclusion. Support for the primitive position of the Cardueae can be inferred from Bolick’s studies on the pollen of Compositae. Bolick described two basic pollen 396 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 types in the Compositae: a caveate Helianthoid pattern and a non-caveate Anthemidoid pattern, which Bolick claimed to be less specialized in its structure than the Helianthoid type. The Cardueae were found to have Anthemidoid- type pollen similar to that of the Mutisieae and Vernonieae. Helianthoid pollen is characteristic of the remaining tribes, except for the Anthemideae, Arctoteae, and Lactueae, tribes in which both types of pollen were found. Recent analyses of chloroplast DNA (Jansen & Palmer) support the hypoth- esis that the Mutisieae (subtribe Barnadesiinae), not the Cardueae, may be the most primitive subtribe of the family, since members of this tribe lack in their chloroplast genome an inversion characteristic of the other members of the family. The phylogenetic interpretations of Jansen & Palmer have been ques- tioned by Robinson, who showed, in figure form, a hypothetical phylogeny based on Jansen & Palmer’s proposal that closely aligned the Cardueae with the Mutisieae (save for the subtribe Barnadesiinae). Wagner characterized the chemistry of the group as being for the most part similar to the rest of the family. He reported volatile and non-volatile terpe- noids, acetylenes, phenolic compounds, and on hydroxycinnamic acid derivatives and flavonoids as ompounds of the tribe. In addition to a variety of common monoterpenes aa triterpenoids of the beta-oleanane and lupane type, Wagner noted that sesquiterpene lactones of the germacrane type are characteristic of the Carduinae and Centaureinae. Further discussion of the structural similarities of these compounds can be found in Hermout & Sorm, who noted that the germacranolides isolated from the Cardueae are strikingly uniform in their structure. The guaianolides cynaropicrin and cos- tuslactone have been found in Cynara L. and Saussurea DC., while eleman- olides have been found in Centaurea species. Seaman suggested that the pres- ence of certain C-6 trans-lactonized guaianolides and germacrolides and occasionally elemanolides found in 11 genera of the Cardueae may be useful in determining tribal limits. Wagner (1977) noted that the subtribe Centau- reinae seemed to be the best differentiated in that in this group are found nearly all types of acetylenes found in the Cynaroideae, a variety of highly methox- ylated flavonoids (including flavanones and lignans) and sesquiterpenoids. Whereas the Carduinae contain special acetylenes (e.g., C,, acetylenes, acetylene glycosides), sesquiterpenoids of the guaianolide type, monomethoxylated fla- vonoids, and simple cinnamic acids and derivatives, the Centaureinae have acetylene aldehydes, acetylene chlorohydrins and acetates, sesquiterpene lac- tones of the germacranolide type, highly methoxylated flavonoids, and typical fully methoxylated lignans. Nowak and coworkers have indicated that cyna- ropicrin may prove to be a chemical marker for subtribe Centaureinae. Acetylenes appear to be well differentiated among the subtribes of the Car- dueae (Bohlmann et a/.). Thiophenes in considerable variety were said to be typical of Echinops and Xeranthemum, furan derivatives were found among members of the Carlininae (Carlina L., Atractylis L.), and ene-tetrayenes are common constituents in members of the Carduinae (Arctium, Carduus, Cir- sium, Silybum, Onopordum). While the Centaureinae are not well differentiated chemically from the other subtribes, they do have unique chlorohydrines and -acetates (Centaurea, Carthamus L.) not found in the other subtribes, as well as unsaturated aldehydes in the aerial parts and ene-tetrayene-ene in the roots. 1990] SCOTT, GENERA OF CARDUEAE 397 Flavonoids occur predominantly as flavones and flavonols (flavanones and flavanonols are restricted to Carthamus, Centaurea and Silybum). Silybum is unique in the tribe in that it produces flavonolignans. Highly methoxylated flavonols and flavones are commonly found in Centaurea and Cirsium. Wagner further reported that flavonoids are normally found in the glycosidic form with 3-O-glycosides and 4’-O-glycosides as the most common forms. Other phenolic compounds reported by Wagner for the Centaureinae and Carduinae are cinnamic acids and their derivatives. Tannins are generally absent or occur sporadically. Species of Arctium, Carthamus, and Centaurea have the dimeric phenylpropane derivatives arctiin and hydroxyarctiin. These compounds were reported to be of physiological importance in the seeds of these taxa and were thought to be restricted to the Carduinae and Centaureinae (Wagner). Alkaloids are widespread in the Cardueae, but the structures of only a few have been elucidated (Wagner). Amines and cyanogens appear to occur spo- radically with only tyramine and histamine reported from Si/ybum and choline from Onopordum, Saussurea, and Centaurea (Wagner). Moore & Frankton (1962) proposed a base chromosome number of 17 for the tribe. They suggested that lower numbers have been derived through re- duction brought about by translocations with a possible loss of non-essential chromatin and inactivation of excess centromeres. They further noted that this process has occurred in many genera and may have contributed to speciation. They thought that genera with the same chromosome number are not neces- sarily closely related, nor are those with different chromosome numbers dis- tantly related, since different chromosome numbers do not prevent ready hy- bridization between taxa. They stressed the importance of evaluating both chromosome morphology and chromosome number when assessing generic relationships. Achene characters have proved to be the most reliable ones in the delimi- tation of genera within the Cardueae and have been the subject of several in- depth studies (Dittrich, 1968a & b, 1970; Isley; Lavialle; Singh & Pandey). The four subtribes of the Cardueae are readily distinguished on the basis of their achenes. According to Lavialle and Dittrich, achene characters clearly segregate the Echinopsinae and Carlininae from the closely aligned Carduinae and Centaureinae. Pappus characters and those of the apical region of the achene allowed Dittrich (1970) to distinguish three groups of genera within the Carduinae: 1) Cirsium, Carduus, and Silybum, 2) Cynara, Notobasis Cass., Ptilostemon Cass., Arctium, and Onopordum, 3) Galactites Moench, Picnomon Adanson, 7yrimnus Cass., and Jurinea Cass. Dittrich’s study (1968a & b) of the Centaureinae informally grouped taxa of this subtribe on the basis of wheth- er the hilum was basal (e.g., Serratula L., Rhaponticum Hill, Leuzea DC.), lateral (Carduncellus Adanson, Carthamus, sections of Centaurea), or caudate (Cnicus, sections of Centaurea). The Cardueae have economic significance in several areas, perhaps, most notably, as a source of noxious weeds that have overgrown much agricultural land throughout the world. Species of Centaurea, Cirsium, Echinops, and other genera of the tribe are cultivated as ornamentals. Cynara Cardunculus L., the cardoon, has long been cultivated for the petioles of its leaves that are eaten 398 JOURNAL OF THE ARNOLD ARBORETUM [voL. 7] like celery. The bracts and the receptacle of immature heads of C. Scolymus L., which is probably derived from C. Cardunculus, are commonly eaten as the globe artichoke. Species of Centaurea have been noted as important sources of nectar for honey bees (Goltz). Echinops, a genus of about 120 species (Mabberley) distributed in southern Europe, North Africa, temperate and subtropical Asia, northward to Japan, and naturalized in parts of North America, has been included by Cronquist (1980) asa member of the flora of the southeastern United States. Three species of the genus (/. sphaerocephalus L., E. exaltatus Schrader, and E. bannaticus Rochel ex Schrader) are often grown as ornamentals (FE. bannaticus frequently as I. Ritro L., see Karlsson). Cronquist (1980) noted that F. sphaerocephalus has escaped from cultivation and become casually established in parts of the United States and southern Canada. Echinops sphaerocephalus is the only species of Echinops treated in floras of the northeastern United States (e.g., Fernald; Gleason & Cronquist), but it has not been included in any of the floras for the Southeast (e.g., Chapman; Long & Lakela; Radford ef a/.; Rickett; Small, 1933, among others), nor have I seen specimens that would indicate that E. sphaerocephalus has become established in the Southeast. Since the three species mentioned above are commonly cultivated, it is reasonable to expect that one or more of them may eventually become established in some areas of the Southeast. Echinops is included here in brackets in the key to genera, and various selected references are among those that follow. REFERENCES: ArANo, H. The karyotype analysis and its karyotaxonomic considerations in some genera of subtribe Carduinae. Jap. Jour. Genetics 32: 323-332. 1957 BARKLEY, T. M. A manual of the flowering of Kansas. 402 pp. Kansas State University Endowment Association, Manhattan, Kansas. 1968. [Key to Cynareae, 373; Arctium, Cirsium, Carduus, oe Cana. Centaurea, 374-376. See also GREAT PLAINS FLORA ASSOCIATION. | BENTHAM, G. Compositae. 7/17 G. BENTHAM & J. D. Hooker. Gen. PI. 2: 163-533. 1873 nth ] f history and g 2 phi l dictmbhiiti n of Compositae. Jour. Linn. Soc. Bot. 13: 335-582. 1873b. Bosrov, E. G., & S. K. CZEREPANOV, eds. Cynareae. (In Russian.) Fl. URSS (SSSR). 28. xx + 655 pp. Moscow, Leningrad. 1963. [Thirty-nine genera by 13 authors. Only genera ee to the southeastern United States are cited here: Carduus, 4-39, by S.G MSHIAN; Cirsium, 51-125, 111 species, by A. L. KHARADZE; Cynara, 225, 226, by S. G. TAMAMSH en Silybum, 227, 228, by S. G. TAMAMSHIAN; Ono- pordum, 228-240, by S. G. TAMAMSHIAN; Centaurea, 370-578, 178 species, 22 subgenera by M. V. KLoxkov, Si I. SosNovsky, N. N. TSVELEv, & S. K. CZEREPANOV; ae 581-587, by S. A. oo Cnicus, 587, 588, by M. M. ILan; Echinops, 27: 1-53. 1962, by E. G. B BOHLMANN, es T. BURKHARDT, & C. ae er occurring acetylenes. 547 pp. London & New York. 1973. [Various taxa and their constituents listed, 442-461; Arctium, aa Carduus, 447, Cirsium, 447, 448.] BoissiER, E. Flora orientalis. Vol. 3. 1033 pp. Geneva. 1875. [Subtribe Carduinae, 457- 711; pei (as Lappa Tourn.) 4 spp., 457, 458; Cirsium, 74 spp., 523-553; Sil- yburn, I sp. Bae Centaurea, 183 spp., 614— 695: Cnicus, | sp., 705, 706.] Botick, M. R. Tai nomic, evolutionary and functional considerations of Compositae pollen neeeeran and sculpture. Pl. Syst. Evol. 130: 209-218. 1978. 1990] SCOTT, GENERA OF CARDUEAE 399 BORNMULLER, J. Revisionsergebnisse einiger orientalischer und zentralasiatischer Arten der Gattung Echinops. Beih. Bot. Centralbl. 36: 200-228. 1918. [Section Oligolepis revised and new sect. Pleiacme and ten new species described.] Bouter, D., J. T. GLeaves, B. G. HAsLett, D. PEAcock, & U. JENSEN. The relationships of 8 tribes of the Compositae as . by plastocyanin amino acid sequence data. Phytochemistry 17: 1585-1589. BREMMER, K. Tribal interrelationships of re Asteraceae. Cladistics 3: 210-253. 1987. Carduus, 9 spp., 55-94; Centaurea 17 spp., 94-215; Arctium, 4 spp., 264-278.] Brizicky, G. K. Subgeneric and sectional names: their starting points and early sources. Taxon 18: 643-660. 1969. Bunce, A. von. Uber die Gattung Echinops. Bull. Acad. Sci. St. Pétersbourg 6: 390- 411. 1846 Burtt, B. L. Compositae and the study of functional evolution. Trans. Proc. Bot. Soc. Edinburgh 39: 216-232. CANDOLLE, A. P. DE. Cynareae. Prodr. Syst. Nat. 6: 449-678. 1837. [Centaurea 565- 605; Cnicus 606-609; Silybum 616: Carduus 621-633; Cirsium 634-657; Arctium (as Lappa Tourn.) 661-662. CARLQUIST, S. Comparative plant anatomy. New York. 146 pp. 1961. —. Wood anatomy of Cynareae (Compositae). Aliso 6: 13-24. 1965. —. Tribal interrelationships and phylogeny of the Asteraceae. Aliso 8: 395-492. 1976 Cassini, H. Compositae. Jn: G. Cuvier, ed., Dictionnaire des sciences naturelles 1-60. oe - Sais 1816-1830. [A description of the tribe ‘““Carduacées”’ appears in vol. in vol. 7, p. 376; ‘““Echinopsées” in vol. 14, p. 200 ela appeared in vol. 60, p. 568. See also, Kinc, R. M., & H. D. DAWSGN: Cassini on Compositae. Oriole editions, New York. 1975. 3 vols. A valuable fac- simile compilation of Cassini’s works collected from the original Dictionnaire ar- ticles.] Sixiéme mémoire sur l’ordre des Synanthérées, contenant les caractéres des tribus. Jour. Phys. Chim. Hist. Nat. Arts 88: 150-163. 1819. [In this article, Cassini gave both Latin names and his tribes p described in French; see citation above for Cassini 1816- 1830. Sixiéme mémoire sur l’ordre des Synanthérées. Jn; Opuscules phytologiques. Vol. 1. Ixvii + 426 pp. Paris. 1826. CHAPMAN, A. W. Flora of the southern United States. ed. 3. xxx1x + 641 pp. Cambridge, Massachusetts. 1897. [Cynareae, 268, 269. CLAPHAM, A. R., T. G. Tutin, & D. M. Moore. Flora of the British Isles. ed. 3. xxix + 688 pp. Cambridge, England. 1987. [Echinops, Carlina, Arctium, Carduus, Cir- sium, Silybum, Onopordum, Cnicus, Centaurea, 479-490.] CORRELL ii : ,& M.C. JoHNsTon. Manual of the vascular plants of Texas. xv + 1881 er, Texas. 1970. [Centaurea, Cirsium, Carduus, Silvbum, Carthamus, On- ars es 1713-1720.] CRONQUIST, Phylogeny and taxonomy of the Compositae. Am. Midl. Nat. 53: 478- 1. 195 . The Cees revisited. Brittonia 29: 137-153. | —. Asteraceae. Vascular flora of the southeastern United on Vol. |. xv + 261 p. Chapel Hill, North Carolina. 1980. [Key to Cynareae, 11, 12; Arctium, Carduus, Ca Silybum, Onopordum, Echinops, Centaurea, Cnicus, 218-228. An integrated system of classification of flowering plants. xxvii + 1262 pp. New York. 1981. [Order Asterales, 1020-1028. . History of generic concepts in the Compositae. Hy 34: 6-10. 1985. Davis, P. H., & A. J. C. Grierson, eds. Compositae. Jn: P. H. Davis, ed., Flora of Turkey. Vol. 5. 890 pp. Edinburgh. 1975. [Arctium, cae by F. K. Kupicua; Cynara, Silybum by F. K. KupicHa; Carduus by P. H. Davis, Centaurea by G. 400 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 WAGENITZ; Cirsium by P. H. Davis & B. S. Parris; Echinops by I. C. HEDGE; Onopordum by A. DANIN.] DittricH, M. Karpologische Untersuchungen zur Systematik von Centaurea und ver- wandten Gattungen. Bot. Jahrb. 88: 70-162. 1968a. Morphologische Untersuchungen an den Friichten der Subtribus Cardueae- Centanreinae (Compositae). Willdenowia 5: 67-107. 1968b. Morphologische und anatomische Untersuchungen an Friichten der Carduinae (Compositae). I. Morphologischer Teil. Candollea 25: 45-67. 1970. Cynareae—a systematic review. Chapter 36, pp. 999-1016 in V. H. HEYwoop et al., eds., Biology and chemistry of the Compositae. Academic Press, London, New York, and San peer 1977. , F. PETRAK, K. H. RECHINGER, & G. WAGENITZ. Cynareae. Jn: K. H. RECHINGER, ed. Fl. Iran. 139a (October, 1979) & 139b (April, 1980). 2 vols. [Arctium, Carduus, Cnicus, Cynara, Echinops, Onopordum, Silybum by K. H. RECHINGER; Cirsium by . PETRAK; Centaurea by M. DitrricH & K. H. RECHINGER.] Dominauez, X. A., A. GONZALEZ, M. A. ZAMUDIO, & A. GARZA. aT te from Stevia Berlandieri and Cirsium texanum. Phytochemistry 13: 2001. 1974. Don, G. The Linnean arrangement and its supplement. /n: J. C. pees Loudon Hortus Britannicus. London. 1830. [Source of early sections for the genus Cirsium, 330, 331, and possibly Carduus, 329, 330; Onopordum, 331. Other genera treated: Centaurea, 360-362; Cynara, 332, Carthamus, 332; Silybum, 330. A general history of the dichlamydeous plants. 4 vols. London. 1831-1838. DUMORTIER, B.C. Florula belgica, operis majoris prodromus. Tornaci Nerviorum. 172 pp. 1827. [Source of the subtribes Carlineae, Centaureinae, Carduinae; see BRizicKy Duncan, W. H., & J. T. KARTEsz. Vascular flora of Georgia. An annotated checklist. Athens, Georgia. 1981. Euuiort, S. A sketch of the botany of South-Carolina and Georgia. 2 vols. Charleston, South Carolina. 1821-1824. [Cirsium, 7 spp., included in synonymy under Cnicus, pp. 268-272.] Faun, A. Secretory tissues in plants. x + 302 pp. London, New York, San Francisco. 1979. [Carthamus, Centaurea, sim 64, 65. FERNALD, M. L. Gray’s manual of botany. ed. 8. Ixiv + 1632 pp. New York. 1950. [Key to Cynareae, 1537; Arctium, pase Carduus, Cirsium, Onopordum, Cen- taurea, Cnicus, 1537-1545. ForRCELLA, F., & H. Woop. Sequential flowering of thistles ane Asteraceae) in southern Australia. Austral. Jour. Bot. 34: 455-462. 1986 FRANKTON, C., & R. J. Moore. Cytotaxonomy, phylogeny, and Canadian distribution of Cirsium undulatum and Cirsium Flodmanii. Canad. Jour. Bot. 39: 21-33. 1961. GAERTNER, J. De fructibus et seminibus plantarum. Stuttgart/Tiibingen. 1788-1791. [Source of original description of Si/vyhum Marianum, 2: 378: source of accounts of various taxa in the Cardueae, including, among others, Carduus, Centaurea, Cirsium, Cnicus, Echinops.] Gisss, R. D. Chemotaxonomy of flowering plants. 4 vols. xxii + 2372 pp. Montreal & London. 1974. [Cirsium, Centaurea, Carduus, ae Silybum, Arctium, 1: 94-96; Centaurea, Cirsium, Silybum, 2: 1195, 1197. — H. A., & A. Crongquist. Manual of a. plants of northeastern United es and adjacent Canada. li + 810 pp. Boston. 1963. [4rctium, Carduus, Cirsium, Sibun. Onopordum, Echinops, Centaurea, Cnicus, 747-754.] GoLpBLATT, P. Index to plant chromosome numbers 1975-1978. Monographs in sys- tematic botany. Vol. 5. 553 pp. Missouri Botanical Garden. 1981. [Asteraceae, 70- 14 l. —. Index ae chromosome numbers 1979-1981. Monographs in systematic botany, _ 8. 427 pp. Missouri Botanical Garden. 1984. [Asteraceae, 52-109.] 1990] SCOTT, GENERA OF CARDUEAE 401 ndex to plant chromosome numbers 1982-1983. Monographs in systematic botany. Vol. 13. 224 pp. Missouri Botanical Garden. 1985. [Asteraceae, 31-59.] Gottz, L. Honey and pollen plants. Part II. The thistles. Am. Bee Jour. 126: 667-669. 1986. [Cirsium, Silybum, Centaurea, Echinops, Onopordum, Carduus, discussed as nectar sources. ] Gray, A. Contributions to North American botany. I. Characters of new Compositae, . Synoptical flora of North America. Vol. 1. Pt. 2. New York. 1884b. [Cynaro- ideae, including Arctium, Carduus, Cirsium (as Cnicus), Onopordum, Silybum, Cen- 07. taurea, 396-4 GREAT ana FLorA ASSOCIATION. Flora of aay Plains. vii + 1392 pp. Lawrence, Kansas. 1986. [Asteraceae, 838-1021, by T. M. BARKLEY; key to Cynareae, 844, 845; tien 908-914, by R. E. Brooks; pone 864, 865; Carduus, 895-897; Sees 897, 898; Centaurea, 898-901; Onopordum, 979, 980, by R. L. McGreEGor.] GupTA, R. "D. & Y.S.Murta. Studies on seed storage and viability in some weed seeds of family Asteraceae. Seed Res. New Delhi 14: 115-125. 1986.* Gutyas, S., & J. Pestr. Angaben zur Anatomie der Nektarien der Centaureae. Acta Biol. Szeged. He 17-24. 1966. HASLETT, B. G., GLEAVES, & D. BouLtrer. N-terminal amino acid sequences of plastocyanins ce various members of the Compositae. Phytochemistry 16: 363- 365. 1977. Hea, G. Compositae. Illustrierte Flora von Mittel-Europa 6(2): 394-1367. Taf. 269- 273. 1928-1929. [Genera of Cynareae, 807-992: includes Echinops, Carlina, Arc- tium, Saussurea, Jurinea, Carduus, Cirsium, Silybum, Onopordum, Cynara, Ser- ratula, Crupina, Rhaponticum, Centaurea, Carthamus, Cnicus.] Compositae. II. Matricaria-Hieracium. G. WAGENITz in Illustrierte Flora von 1413; includes genera noted above. Text of ed. | essentially unchanged; many black & white illustrations changed; new data largely in the Nachtrége, 1353-1452.] Herser, C. B., & T. W. WHITTAKER. Chromosome number, polyploidy and growth habit in California weeds. Am. Jour. Bot. 35: 179-186. 1948. [First reports for Centaurea Calcitrapa, 2n = 20; C. solstitialis, 2n = 16; C. picris, 2n = 26; and Silybum Mar- ianum, 2n = 34. Counts for Cirsium arvense and C. vulgare, 2n = 34, and 2n = 68, respectively, confirmed earlier reports for these taxa.] HERMOUT, V., & F. SorM. eluent eg orl of the sesquiterpenoids of the Compositae. In: J. HARBORNE & T. Swain, eds., Perspectives in phytochemistry. xi + 235 pp. London and New York. 969 [Various members of Cardueae, 139-165 7 HesLop-HArRIsOoN, Y., & K. R. SHIVANNA. The receptive surface of the angiosperm stigma. Ann. Bot. II. 41: Sreey hy 1977. [Centaurea, 1240, listed as having a papillate stigmatic surface with unicellular papillae. ] HoFFMANN, O. Compositae. /n; A. ENGLER & K. PRANTL, Nat. Pflanzenfam. IV. 87- 391, 1894. [Tubifloreae—Cynareae, 312-333; Cirsium, Carduus, Centaurea, o. opordum, Cnicus, published Sept., 1893.] Howe .t, J. T. Distributional data on weedy thistles in western North America. Leafl. West Bot. 9: 17-29. 1959. [Early records noted for Arctium, Carduus, Carthamus, Centaurea, Cirsium, Cnicus, Cynara, Echinops, Onopordum, Silybum.] IsLtey, D. Seed characteristics of weedy thistles. Assoc. Off. Seed Anal. Proc. 37: 93-98. 1947. [Cirsium, Cnicus, Onopor don (sic.), Silymbium (sic.), Carduus keyed by seed characteristics. ] JANSEN, R. K., & J. D. PALMER. A chloroplast DNA inversion marks an ancient evo- 402 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 lutionary split in the sunflower family (Asteraceae). Proc. Natl. Acad. Sci. U.S. A. 84: 5818-5822. 1987. JEFFREY, C. Notes on the Compositae: II]. The Cynareae in East Tropical Africa. Kew Bull. 22: 107-140. 1968 . Corolla forms in Compositae—some evolutionary and taxonomic speculations. Chap. 6. Pp. 111-118 in V. H. HEywoon et a/., eds., The biology and chemistry of the Compositae. London. 1977. . Compositae. Pp. 263-268 in V. H. HEywoop, consultant ed., Flowering plants of the world. New York. 1978. JENSEN, S. R., B. J. NIELSEN, & R. DAHLGREN. Iridoid compounds, their cre and systematic importance in the angiosperms. Bot. Not. 128: 148-180. 1975. Jounson, M. F. Cynareae Ciena in Virginia: Cirsium, a Onopordum. Virginia Jour. Sci. 25: 152- 974. . Cynareae oe in Viren: Arctium, Centaurea, Cnicus. Castanea 40: 63-73. 1975. — &H.H. Ixris. Preliminary reports on the flora of Wisconsin. No. 48. Compositae = family |. Trans. Wisconsin Acad. Sci. 52: 255-342. 1963. Kartesz, J. T., & R. KARTESZ. A synonymized checklist of the vascular flora of the United a Canada and Greenland. Chapel Hill, North Carolina. 1980. [Arctium, 51; Carduus, Carthamus, Carlina, Centaurea, 62; Cirsium 65-67; Echinops, 71; Onopordum, 94; Silybum, 102. KITAMURA, S. Compositae Japonicae. Pars prima. Mem. Coll. Sci. Kyoto Univ. B. 13. 1937. [Source of name of subfamily Carduoideae Kitamura, 5; tribe Cynareae, 5- 212; Echinops, 14-18; eRe 32, 33; Cirsium, 33-134. LAMARCK, J. B. DE, & A. P. DE CANDOLLE. Flore Francaise. ed. 3. Paris. 1805. [Com- positae. Vol. 4. pt. LARSEN, K. Chromosome numbers of some European flowering plants. Bot. Tidsskr. 50: 163-174. 1954. [Centaurea Cineraria L. reported as 2n = 18; Silybum Marianum as 2n = 34; confirmations of earlier counts. ] Chromosome studies in some Mediterranean and South European flowering plants. Bot. Not. 109: 293-307. 1956. LAVIALLE, P. Recherches sur le pas ee, de l’ovaire en fruit chez les Composées. Ann. Sci. Nat. Bot. LX. 15: 39- 1912. LEONHARDT, R. Phylogenetisch- a cae Betrachtungen. I. Betrachtungen zur Sys- tematik der Compositen. Osterr. Bot. Zeitschr. 96: 293-324. 1949. [Cynareae as most ar tribe; Campanulaceae most closely related to the Asteraceae. ] LessInG, C. F. Synopsis generum Compositarum. xi + 473 pp. Berlin. 1832. Love, A., & D. Léve. Cytotaxonomical studies on boreal plants. Arkiv Bot. 31(A): I- 22. 1944 . Chromosome numbers of Central and Northwest European plant species. Opera Bot. Lund. 5: 1-581. 1961. [Compilation of chromosome numbers and sources for the Asteraceae, 339-362; Carduus, 339; Cirsium, 340; Cynara, Silybum, Onopor- dum, Arctium, 341; es 342; Carlina, Echinops, Cnicus, 344.] MABBERLEY, D. J. The plant-book. xii + 706 pp. Cambridge, England. 1987. MacRoserts, D. T. The vascular ae of Louisian na. Daa ann notated checklist and bibliography of the vascular plants reported to grow Itivation in Louisiana. Bull. Mus. Life Sci. Louisiana State Univ. oe 6: 1-165. 1984. se minus, 61; Carduus see 62; Cirstum ed C. carolinianum, C. discolor C. horridulum, C. Lecontei, C. muticum, C. Nuttallii, C. terrae-nigrae, C. eee tum, C virginianum, C. fiat 62; Centaurea americana, C. Cyanus, C. maculosa, 62; Silybum Marianum, 67.] McVauau, R. Flora Novo-Galiciana. Vol. 12. Compositae. 1157 pp. Ann Arbor, Mich- igan MEHRA, P. N., B. S. Git, J. K. MEHTA, & S. S. Srpuu. Cytological investigations on 1990] SCOTT, GENERA OF CARDUEAE 403 the Indian Compositae. I. North-Indian taxa. Caryologia 18: 35-68. 1965. einer mosome numbers reported for certain taxa; Cynareae, 43; Arctium Lappa, n = | Carduus nutans, 2n = 40; Cirsium arvense (as Cnicus arvensis Hoffmann), n = 7 Centaurea Ca n= 12; Carthamus tinctorius, n = 12 & P. REMANANDAN. Cytological investigations on n Indian Compositae V. Tribes Arctotidae, Cynareae, Calenduleae, and Mutisieae. Nucleus 19: 8-12. 1976. Two species of Echinops reported as n = id revious counts of n = 18 confirmed, meiosis was said to be normal and pollen fertility reported as 95 percent; Carduus nutans counted as n = 20, while ‘laggards and bridges” were said to disturb meiosis; Cirsium arvensis (as Cnicus arvensis) was reported as n = 17, in confirmation of earlier reports. ] Moore, D. M., ed. Cardueae. Jn: T. G. TuTIn et al., eds., Fl. Europaea 4: 208-304. 1976. [Arctium by F. H. Perrinc; Carduus (Key by M _L. RocHA AFONSO), Cnicus, Onopordum, & Silybum by J. do Same FRANCO; Centaurea by J. DOSTAL; Cirsium K. WERNER; Echinops by S. KoZUHAROV Moore, R. J., ed. Index to plant ae auibem 1967-1971. Reg. Veg. vol. 90. Utrecht. 1973. C. FRANKTON. aaron studies in the tribe Cynareae (Compositae). Canad. Jour. Bot. 40: 281-293. 1962. . The thistles a ete Monograph No. 10. Research Branch, Canada Dep. Agr. 1974. [Includes treatments of Echinops, ae Cirsium, Carduus, Saus- surea, Silybum, Onopordum, Centaurea, Cnicus with discussions of economic uses. } Morton, J. K. Chromosome numbers in Compositae a Canada and the U. S. A. Bot. (om Linn. Soc. 82: 357-368. 1981. [Cirsium arvense, 2n = 34; C. discolor, 2n = 20; and Cnicus benedictus, 2n = 22, reported.] Mutter, J. Fossil pollen records of extant angiosperms. Bot. Rev. 47: 1-142. 1981. [Cirsium-type pollen confirmed from the Pliocene, 1 NAKAI, T. Ordines familiae. Imperial University, Tokyo, Japan. 1943. Nowak, G., B. Drozpz, W. Kroszczynskl, & M. HoLus. Sesquiterpene lactones XXX. Cynaropicrin in species of the subtribe Centaureinae. (In English; Polish summary.) Acta Soc. Bot. Poland. 55(1): 17-22. 1986. NutTTALL, T. The genera of North American plants. 2 vols. Philadelphia. 1818. Parra, O. Morfologia de los granos de polen de los Compuestas Cynareas Chilenas. Bol. Soc. Biol. Concepcién 42: 89-96, 1970.* Persoon, C. H. Synopsis plantarum. Pars secunda. 657 pp. Paris. 1807. PHILIPSON, W.R. The relationships of the Compositae, particularly as illustrated by the morphology of the inflorescence in the Rubiales and the Campanulatae. Phyto- morphology 3: 391-404. 1953 . Ovular morphology and major classifications of the dicotyledons. Bot. Jour. Linn. Soc. 68: 89-108. 1974 PODDUBNAJA-ARNOLDI, W. Ein Versuch der Anwendung der embryologischen Methode bei der Lésung einiger systematischer Fragen. I. Vergleichende embryologischen- zytologische Untersuchungen iiber die Gruppe Cynareae, Fam. Compositae. Beth. Bot. Zentralbl., Abt. 2, 48: 141-237. 1931. RapForD, A. E., H. E. AHLES, & C. R. Bett. Manual of the vascular flora of the Carolinas. Ixi + 1183 pp. Chapel Hill, North Carolina. 1968. [Compositae by H. E. Centaurea, 1039-1041; Carduus (including Cirsium), 1041-1044; Arctium, 1045.] Rickett, H. W. Wildflowers of the United ey Vol. 2. The Southeastern States. Part 2. Pp. 606-612. 1966. [Cirsium pp. 606-6 arduus, Cnicus, Onopordum, Arctium pp. 610-612. Color photos of Cirsium fates C. arvense, C. discolor, C. hor- ridulum, C. Lecontei, C. muticum, C. Nuttallii, C. pumilum, C. repandum, C. vir- ginianum;, Carduus acanthoides, C. nutans, Arctium ae mis Acan- thium.| 404 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Rosinson, H. Some suggestions regarding the cy aaa of chloroplast DNA variation in the Asteraceae. Phytologia 63: 316-324. 1987. RypBERG, P. A. Flora of the Rocky aon and adjacent plains. Published by the author. New York. 1917. [Correct date is 1918, since the shipment of books was held up in a railroad strike was not distributed until January, 1918, according to A. CRONQUIST (pers. comm.).] —. Flora of the prairies ee plains of central North America. New York. 1932. Scopo.l, J. A. Methodus Plantarum. 28 pp. Vienna. 1754 Flora Carniolica. xxii + 607 pp. Vienna. 1760. . Flora Carniolica. ed. 2. 2 vols. (Vol. 1, Ixxi1 + 448 pp.; vol. 2, 496 pp.) Vienna. 1772. SEAMAN, F. C. Sesquiterpene lactones as taxonomic characters in the Asteraceae. Bot. Rev. 48: 121-595. 1982. SHISHKIN, B. K., & E. G. Bosrov, eds. Cynareae. (22 genera by 9 authors; in Russian.) F1. URSS (SSSR). 27: xxi1 + 760 pp. Moscow, Leningrad. 1962. [Echinops, 2-53, 58 species, by E. G. Bosrov; Arctium, 93-107, 8 species, by S. V. YUZEPCHUK. See Bosrov & CZEREPANOV, eds., vol. 28, for other genera.] SINGH, R. P., & A. K. PAnpey. Development and structure of seeds and fruits in Compositae-Cynareae. Phytomorphology 34: 1-10. 1984. SMALL, J. K. Manual of the southeastern flora. xxii + 1554 pp. New York. 1933. SMALL, J.S. The origin and development of the Compositae. New Phytologist 18(1&2): 1-35; (3&4): ee 89; (5&6): 129-176; (7): 201-234. London. 1919, Smitu, E. B. An atlas and annotated list of the vascular plants of Arkansas, ed. 2. iv 489 pp. Fayetteville, Arkansas. 1988. on Arctium minus; pei CrISPUS, a nutans, Centaurea (3 spp.); Cirsium (7 spp. Silybum Marianum.] SoLsric, O. The tribes of Compositae in the southeastern United ie. Jour. Arnold Arb. 44: 436-461. 1963. —. Subfamilial nomenclature of Compositae. Taxon 12: 229-235. 1963. SpPAcH, E. Histoire naturelle des végétaux. 10: 10-15, 65-72. 1841. ne of several subgeneric names in Centaurea. ] SPRAGUE, T. A. XXX¥V. Additions to the Index Kewensis. Bull. Misc. Inf. Kew 1925: 312-314. 1925. [A list of species taken from Wallich’s Catalogue and given descrip- tions by G. Don Stepa, I. S. Ad cognitionem pollinis Se encas generum nonnullerum tribus Cy- nareae familiae Compositae. Not. Syst. Geogr. Inst. Bot. Thbilissi. 20: 54-62. 1959. . Morfologija pyl’cy roda Cirsium Mill. i blizkih rodov triby Cynareae (Com- positae). Trudy Tbilis. Bot. Inst. 21: 81-126. 1960. STEUDEL, E. G. Nomenclator botanicus. ed. |. Stuttgart, Tiibingen. 1821-1824. [Later edition published in 1840. TAKHTAJAN, A. L. Outline of the classification of flowering plants (Magnoliophyta). Bot. Rev. 46: 225-359. 1980. Tucci, G. F., & F. MAGGIni. Ribosomal RNA genes in species of the Cynareae Tribe (Compositae) I. irene 132: 76-84. 1986. UNITED STATES DEPARTMENT OF AGRICULTURE. Common weeds of the United States. (Reprint.) Dover Siblitioas New York. 1971. WAGENITz, G. Compositae. Pp. 484-497 in H. iy ais ed., A. Engler’s Syllabus der Pflanzenfamilien. ed. 12. I] Band. Berlin. 4. . Systematics and phylogeny of the ene (Asteraceae). Pl. Syst. Evol. 125: 29-46. 1976. WaGnerR, H. Cynareae—a chemical review. Chapter 37 Hf ue H. HEYwoop e7 al., eds., Biology and chemistry of the Compositae. London. WALLicH, N. Numerical list of dried plants in the Eas : a Company’s museum. London. 1831 [1828-1849]. [Often cited as ““Wallich’s Catalogue”; years of publi- cation vary with number. Most of the “‘new species” listed in this catalogue are 1990] SCOTT, GENERA OF CARDUEAE 405 nomina nuda, since they lack descriptions. A number were later published by G. Don, 1831-1838. A list of these species can be found in SprAGugE, 1925.] WIEGAND, K. M., & A. F. EAMEs. The flora of the Cayuga Lake Basin. Cornell Univ. Agr. Exper. Sta. Mem. 92. 491 pp. Ithaca, New York. 1926. — J. C. A dictionary of the seats. plants and ferns. ed. 7. Revised by H. K. ry SHAW. Cambridge, England. 1 ce R. P. Guide to the vascular plants of central Florida. Univ. Presses Florida. 1982. Youna, D. A., & D. S. SEIGLER, eds. Phytochemistry and angiosperm phylogeny. New York. | Youna, D. J., & L. Watson. The classification of dicotyledons: a study of the upper levels of the hierarchy. Austral. Jour. Bot. 18: 387-433. 1970. ZOHARY, M. Evolutionary trends in the fruiting head of Compositae. Evolution 4: 103- 109. 1950. ZWOFLER, H. Preliminary list of phytophagous insects attacking wild Cynareae (Com- ore ee in Europe. Commonwealth Inst. Biol. Control Tech. Bull. 6: 81- 154. ec eee und Ergebnisse der Co-Evolution von phytophagen und ento- mophagen Insecten und hdheren Pflazen. Sonderbande Naturwiss. Ver. Hamburg. 2: 7-50. 1978. KEY TO THE GENERA OF CARDUEAE IN THE SOUTHEASTERN UNITED STATES A. Achenes attached to the receptacle by the base; flowers all alike, marginal ones not enlarged. B. Heads 1-flowered, united into a spherical, secondary head. ....... [Echinops. ] B. Heads with more or less numerous flowers, not united into a secondary head. C. Receptacle fleshy, deeply alveolate, naked or only very shortly bristly; plants with strongly spiny-winged stem and tomentose herbage. . 4. Onopordum. C. Receptacle with scales or setae or densely bristly; plants without both stem strongly spiny winged and herbage tomentose throughout. D. Staminal filaments united below; leaves white mottled. ...3. Silybum. D. Staminal filaments separate; leaves not white veined or variegated. E. Involucral bracts hooked at tip; leaves not bristly or spiny, the lower broadly ovate and mostly cordate; style branches partly distinct. ... Eh a ii tava te sea CO Ate res Wace Bide weeutaeny Gurnee nates 5. Arctium. E. Involucral bracts not hooked; leaves bristly or spiny on the margin, lanceolate or narrow; style branches coherent, commonly with a pu- bescent ring at bas F. Pappus beers plumose; stem not conspicuously spiny winged (except in C. vulgare). oo... eee 1. Cirsium F. Pappus bristles barbellate, not plumose; stem conspicuously spiny WAN RODE cease Spices eo he Wea Wee Mises Saal a arduus. A. Achenes obliquely or laterally attached to the receptacle; marginal eon often enlarged and neutral, appearing ray-like, occasionally undifferentiated from inner ee G. Achenes with 10 horny teeth at the summit, and with a biseriate pappus of 10 long awns alternating with 10 shorter inner ones; leaves prickly saci te flowers NOW. cis aie eee ce tae eae eon ae cas aac aaa us deb ohn ote . Cnicus. . Achenes without horny apical teeth, and with a pappus of several series of short (seldom elongate) scales or bristles, or the pappus wanting: leaves not prickly; flowers in most species anthocyanic (or white), seldom yellow. .. 6. Centaurea. ey) 406 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Subtribe CARDUINAE Dumortier, Fl. Belg. Prodr. 72. 1827. 1. Cirsium Miller, Gard. Dict. Abridg. ed. 4. 1: ord. alph. 1754, emend. Scopoh, Fl. Carniol. ed. 2. 123. 1772. Biennial or perennial [annual] herbs. Stems to 2.5 m tall, branched or simple, arachnoid-tomentose to glabrous, unarmed or spiny winged, wings triangular with a short apical spine. Leaves sessile to long petiolate; blades lanceolate to oblanceolate, the upper surface glabrescent or setaceous [densely pubescent], glandular, or glandular-pubescent, lower surface glabrescent or pubescent along the primary veins to white-tomentose throughout, glands capitate or punctate, hairs crispate or sinuate, multicellular; margins entire to more often incised or laciniate, weakly to strongly spiny. Capitula (heads) homogamous or hetero- gamous, discoid, ovoid to globose, terminal or axillary, solitary or in clusters of 3 to several heads on short to somewhat elongate peduncles [sessile]. In- volucre ovoid, to ca. 7 cm high and ca. 9 cm wide, bracts imbricate, multiseriate, linear-lanceolate to lanceolate, acute to acuminate, often spine tipped, usually with a simple apical spine, adpressed to reflexed, glabrous to sparsely pubescent [arachnoid-pubescent], often with a glutinous dorsal ridge (vitta), entire or ciliate [lacerate]. Receptacle flat to subconic, scales numerous, setaceous. Flow- ers perfect (rarely imperfect and then the heads functionally so and the plants partly or wholly dioecious). Corollas tubular, actinomorphic, white, cream- yellow, lavender, purple, or garnet red [red], glabrous or sparsely glandular, tube narrow, elongate, throat short, undivided, lobes linear to lanceolate, elon- gate. Staminal filaments glabrous to more often papillose-hairy, anthers white or stramineous, saggitate, apical appendage acute, small. Styles smooth below a distinct collar of hairs, branches papillate, scarcely divergent, narrowly trun- cate to acute. Achenes 3-6 mm, basifixed or nearly so, glabrous, smooth, oblong, compressed, gibbous, with a distinct, apical margin surrounding a subconical central projection. Pappus of several rows of distinctly plumose setae, basally fused, deciduous as a ring [or persistent]; the inner setae somewhat longer than the outer and simple, lanceolate, occasionally flattened and ciliate towards apex; pappus of outermost florets often with fewer, minutely barbellate or scabrous setae. Lecrorype species: Cirsium heterophyllum (L.) Hill. (Carduus hetero- phyllus L.) = Cirsium helenioides (L.) Hill; see Britton & Brown, Illus. Fl. No. United States, Canada. ed. 2. 3: 548. 1913: Britton & Millspaugh, Bahama Flora, 458. 1920; R. McVaugh, Flora Novo-Galiciana, 229. 1982. Werner treated Cirsium heterophyllum as a synonym of C. helenioides (L.) Hill in Flora Europea. (Name Latin, from cirsion, thistle, used by Dioscorides, from kirsos or cirsos, a swollen vein, which it was said to cure; see Loudon; Fernald; Correll & Correll; Munz.)— THISTLE, PLUMED THISTLE. Tenuous morphological distinctions and a misunderstood nomenclatural history have led authors to treat members of Cirsium as representatives of both Carduus (Elliott; Nuttall; Radford ef al.) and Cnicus (Gray, 1874, Pammel, 1901). Linnaeus did not recognize Cirsium, but described the species later treated as Cirsium as representatives of Carduus. The year following the pub- lication of Species Plantarum, Miller proposed Cirsium on the basis of char- 1990] SCOTT, GENERA OF CARDUEAE 407 acteristics of the spines on the leaves and the lack of spines on the “cup of the flower” (1.e., involucre). Miller used polynomials under his generic descriptions until the 1768 edition of The Gardener’s Dictionary (see Druce for commen- tary), yet in this edition Miller merged Cirsium with Carduus. Because of this, it was Scopoli (1772) who provided the first binomials in Cirsium and who noted the plumose pappus as a distinguishing character for the genus. Adanson was wrongly credited as the author of Cirsium for his realignment of Cnicus Tournefort (Cnicus L. is now a conserved name). Primarily on the basis of characters of the achene, Cassini, Lessing, and De Candolle considered Cirsium to be closely related to Carduus, but distant from Cnicus. Bentham and Asa Gray (1874) both included Cirsium in Cnicus, but Hoffmann maintained its generic status. Dittrich (1977) assigned Cirsium and Carduus to the tribe Car- dueae (subtribe Carduinae Dumort.) and aligned them with Modestia Charadze & Tamamshian (a genus of three species of Central Asia) and Si/ybum on the basis of achene characters (apical plate, pappus, and detachment area). While the division between Cirsium and Carduus has long been noted as somewhat arbitrary (cf. Scopoli, 1772; McVaugh) and various authors (e.g., Correll & Johnston; Willis) have suggested that Cirsium should be included within Carduus. Ahles in Radford et a/. (1968) is the only recent author who has adopted this position. Recent floras of Europe (e.g., Flora Europaea, Flora of Turkey, Flora Iranica) and the Orient have maintained Cirsium and Carduus as separate genera. A number of sectional names have been proposed for Cirsium since Du- mortier first subdivided the genus into sections Erio/epis (Cass.) Dumort. (in- volucres woolly), Onotrophe (Cass.) Dumort. (involucres spiny), and Cirsioty- pus Dumort. (involucres unarmed). Dumortier was followed shortly by Duby, who borrowed on De Candolle’s unpublished material and described the sect. Epitrachys DC. ex Duby (leaves with upper surface spiny; flowers purple). In a treatment that has gone largely unrecognized, George Don (in Loudon) divided Cirsium into seven sections: Solitaria (leaves decurrent; flowers red, heads subsolitary), Polyanthema (leaves decurrent; flowers red, heads aggre- gated), Leucanthum (leaves decurrent; flowers whitish), Eriophora (leaves ses- sile; flowers red, heads subsolitary), 4ggregata (leaves sessile; flowers red, ag- gregated), Chrysantha (leaves sessile; flowers yellow or white), and Dubia. Don’s sect. Eriophora contained several species found in the southeastern United States (C. discolor, C. altissimum, C. muticum Michaux, C. virginianum (L.) Michaux, C. arvense (L.) Scop.), as well as the species later chosen as lectotype of the genus. Don’s treatment of other families had closely followed De Can- dolle’s accounts published in his Prodromus, but his treatment of the Com- positae preceded De Candolle’s by several years. De Candolle recognized six sections: Lophiolepis (Cass.) DC., Eriolepis, Or- thocentron (Cass.) DC., Corynotrichum DC., Cephalonoplos DC., and Ono- trophe DC. The last section contained several species (C. repandum Michaux, C. horridulum Michaux, C. muticum, C. virginianum) from the southeastern United States, and, along with Don’s sect. Eriophora, is currently placed in synonymy under sect. CirstuM. Koch, who claimed to have based his treatment on a manuscript of De Candolle’s treatment of Compositae for the Prodromus, 408 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 \\ \ Sh 3 ~S WY = \ \\\\ Se f => add LY ED IE: oe ig EEE 7 yaa Yee 1990] SCOTT, GENERA OF CARDUEAE 409 attributed two sections to De Candolle (Epitrachys, Chamaeleon) that De Can- dolle did not recognize, and erected yet two others (Breea (Less.) Koch and Picnomon (Cass.) Koch), each of which contained a single species (C. arvense and C. acarna DC., respectively). Koch’s sect. Chamaeleon was later placed in synonymy under sect. Crrstum, and his sect. Breea is considered to be a synonym of sect. CEPHALONOPLOS, a section set aside for dioecious species. Although Bentham (1873a) did not produce a sectional treatment of Cirsium, Hoffmann, whose treatment of the Compositae bore many similarities to Ben- tham’s, recognized seven sections of Cirsium: Notobasis (Cass.) Hoffm., Ce- phalonoplos DC., Epitrachys DC. ex Duby, Chamaeleon DC. ex Koch, Ery- throlaena Don in Sweet, Ptilostemon (Cass.) Hoffm., Lamyra (Cass.) Hoffm., and Ancathia (DC.) Hoffm Petrak (1917), the most recent monographer of Cirsium, considered the American species to be members of subgenus Fucirsium (= Cirsium) in which he recognized six sections for the North American taxa: Cirsiopsis Petrak, Mastigophyllum Petrak, Dermatolepis Petrak, Echenais (Cass.) Petrak, Ono- trophe and Erythrolaena (see Ownbey & Hsi). The nearly 15 indigenous species known from the Southeast (excluding two European introductions) were placed in three subsections of sect. Onotrophe (Petrak, 1917; Frankton & Moore, 1969: Ownbey & Olson), a section later considered to be a synonym of sect. CiRsIUM (Petrak, 1979). he two naturalized European species (Cirisum arvense (L.) Scop. and C. vulgare (Savi) Tenore) are not considered to be closely related to the native species of the Southeast. Werner included C. vulgare in sect. Eriolepis, but Petrak (1979) placed this species in sect. Epitrachys subsect. Lanceolata Petrak. While the name Epitrachys may more closely characterize the group of taxa included under this name, Petrak’s description of Epitrachys and Werner’s description of sect. Eriolepis are not easily distinguished. As treated by Werner, the sect. Crrsrum includes C. helenioides (L.) Hill under which he synonymized C. heterophyllum, the taxon previously designated as the lectotype species - the genus (see above). Cirsium arvense was retained in sect. CEPHALONOPI DC. by both Petrak (1979) and Werner. Of the nearly 75 species of Cirsium recognized in the continental United States, only 12 to 15 are native to the Southeast. Many are biennials or mono- Figure |. Cirsium. a-i, C. vittatum: a, small flowering plant, base of plant to right, upper part with flowering head to left, x ‘4; b, flower, showing protruding stamens and style, x 2; c, three of the five anthers from within, the central one in more detail, x 4: d, base of an anther from within to show appendages, filament removed, x 10; e, detail of style, showing stylar brush at middle and lowermost part of united style arms with stigmatic line, x 30; f, achene with pappus, = 1; g, h, basal, central, and distal portions of a single pappus filament, x 4; 1, achene from which pappus has become detached, x 5. j, C. horridulum: flowering head, showing ea false involucre, x ‘2. k, C. Leconte: flowering head, the cobwebby hairs omitted, 410 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 carpic taxa that need three to five years to come to flowering and fruition. Petrak (1917) treated the southeastern taxa in three subsections (Odorata, Campanulata, and Acanthophylla) of sect. ONOTROPHE DC. Taxa of subsections Odorata and Campanulata tend to inhabit the Atlantic and Gulf Coastal Plains and are often associated with salt or fresh marshes, bogs, savannas and pine barrens. The subsection Acanthophylla includes taxa often found in woodlands, thickets, and pasture areas of the Piedmont of Georgia and the mountains of Kentucky, Tennessee, and Arkansas. Cytological studies (Ownbey & Hsi; Own- bey & Olson; Moore & Frankton, 1969) have indicated that species of subsect. Odorata have chromosome numbers of 2” = 30 and above, while those of subsections ee and Acanthophylla have lower chromosome num- bers (2n = 18, 2 Recent treatments of the species of Cirsium in the eastern and southeastern United States have followed Petrak’s taxonomic system, except for a realign- ment of two species at the subsectional level (Ownbey & Olson). The transfer of C. Lecontei Torrey & Gray and C. repandum Michaux from subsect. Cam- panulata to subsect. Odorata was made on the basis of their apparent cytological homogeneity with the other species (C. horridulum Michaux, C. pumilum (Nutt.) Sprengel, C. vittatum Small) of the latter subsection. The transfer of C. repan- dum and C. Lecontei left C. Nuttallii DC. the lone member of subsect. Cam- panulata, a group initially characterized by oblong-cylindrical or campanulate capitula. Cirsium repandum, C. Nuttallii, and C. Lecontei are morphologically similar species often found growing in sandy soils of pine barrens and savannas of the Coastal Plain. Cirsium Lecontei appears to inhabit somewhat wetter hab- itats (edges of swamps, peat bogs, low, moist meadows) than the other two species, which occur in open pine woodlands, on sand ridges, and along road- sides. All three species flower primarily from mid-June to mid-September (Cronquist, 1980; Radford ef a/., 1968). No occurrence of hybridization among these species has been reported. Moore & Frankton (1969) considered Cirsium repandum and C. Lecontei to be uniform species that show characteristics of other species groups, but remain strikingly distinct. They suggested that C. repandum may have developed from hybridization between the similar more northern species C. pumilum and the C. horridulum complex (Moore & Frank- 1969), a group that Petrak (1917) and Moore & Frankton believed to be unrelated in origin to other North American species. Several taxa (e.g., C. horridulum, C. Smallii Britton, C. vittatum), recognized as series Horridula of subsect. Odorata, form a distinct group easily recognized in the field by the presence ofa large false involucre (Moore & Frankton, 1969). Elements of this complex, distinguished from one another by one or two subtle characters (Cronquist, 1980), can be found along the Atlantic and Gulf Coastal Plains from Maine to Mexico in the sandy soil of low pine woods, savannas, or along the edges of fresh or salt marshes. Chromosome numbers of 2” = 30 were reported for C. pumilum and C. repandum, 2n = 32 for both C. Lecontei and C. horridulum, and 2n = 32, 34 (33, 35) for C. vittatum (Ownbey & Olson). Moore & Frankton (1969) reported 2n = 28 for C. Lecontei and suggested that this species arose from a past 1990] SCOTT, GENERA OF CARDUEAE 411 hybridization between stock of the series Horridula and C. Nuttallti (subsect. Campanulata). Two chromosome numbers (2” = 24 and 2n = 28) were re- ported for C. Nutta/lii, and probable accessory chromosomes were observed in association with 12 bivalents in a plant from Florida, yet normal meiosis was noted among 14 bivalents in plants from Louisiana. Ownbey & Olson investigated plants of C. Nuttallii from populations in Florida, Georgia, and South Carolina and found the chromosome number to be consistently 2” = 24, except for one plant from Georgia with a chromosome number of 2n Included in a third subsect., Acanthophylla, were four species from the South- east (C. altissimum (L.) Sprengel, C. discolor (Muhl.) Sprengel, C. miuticum Michaux, C. virginianum (L.) Michaux). This subsection was differentiated primarily by ovate to ovate-globose capitula and weakly spiny involucres (Pe- trak, 1917). Later, C. carolinianum (Walter) Fern. & Schub. and C. tferrae- nigrae Shinners (= C. Engelmannii Rydb.; see Shinners, 1964) were added to this subsection (Ownbey & Olson). Many of the taxa comprising the subsection inhabit inland woods, thickets, and pasture areas of higher elevations and, in some cases, have become ready colonizers of disturbed areas. Cirsium muticum is found in swamps and moist woods from central Canada to New Brunswick, south to the Gulf Coast, and west to eastern Texas. Cirsium discolor occurs in drier habitats, mostly thickets and grasslands, from New England south along the Atlantic Coast to North Carolina, west to Tennessee, and north to Min- nesota. The closely related C. a/tissimum (L.) Spreng. occupies slightly more shaded, but otherwise similar habitats to those of C. discolor (see, Ownbey, 1964). The distributional range of C. altissimum, from southern New York to northern Florida, west to the Dakotas and Texas, is more southerly than that of C. discolor. According to Frankton & Moore (1963), this southern extension of C. altissimum may be attributed to the reduction of the diploid chromosome number from 20 to 18 and the coincident plasticity that produced the southern phenotypes of C. a/tissimum, which are sometimes recognized as species. Subsection Acanthophylla appears to be morphologically and cytologically homogeneous. At least three species (Cirsium carolinianum, C. muticum, C. discolor) have chromosome numbers of 2” = 20. Extra chromosomes arising from meiotic irregularities and triploid populations have been reported for C. muticum (Frankton & Moore, 1963; Ownbey & Olson), which is known to hybridize with C. discolor (Ownbey, 1951; Bloom). In turn, C. discolor has been shown to hybridize with C. al/tissimum, 2n = 18 (Ownbey, 1964), the two not easily distinguished in parts of their ranges (Davidson) and considered to be conspecific by some authors (e.g., Gray, 1884b; Wiegand & Eames). More than 98 percent of both pollen and achenes produced by the progeny of the cross between C. discolor and C. altissimum aborted before maturation, yet rare instances of introgression were reported (Ownbey, 1964). A similar pattern was found in the hybrid progeny of the C. muticum x C. discolor crosses. In this case, the hybrids produced less than three percent fully formed pollen (vs. over 90 percent for the parental species), and less than five percent of the flowers in hybrids produced mature achenes; yet some evidence of introgression of C. discolor into C. muticum was reported (Ownbey, 1951). 412 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 On the basis of concepts set forth in an earlier paper (Moore & Frankton, 1962b), Frankton & Moore (1963) considered Cirsium discolor and C. altis- simum to be slightly more advanced morphologically than C. muticum, because of the number of pappus setae and the length of the corolla tube relative to the rest of the corolla. Petrak placed C. discolor and C. altissimum in the series Altissima and C. muticum alone in the series Mutica, an arrangement with which Frankton & Moore (1963) concurred. More recently, these three species, along with C. Engelmannii Rydb. (C. terrae-nigrae Shinners) and C. caroli- nianum, were grouped in the series A/tissima (Ownbey & Olson). Cirsium virginianum (L.) Michaux, a species of the Atlantic Coastal Plain that inhabits savannas, bogs, and wetlands from New Jersey to northern Flor- ida, 1s morphologically similar to C. carolinianum, a species that occurs in open woods on dry, sandy soils from southern Ohio to the mountains of North Carolina, northwestern Georgia, Alabama, Arkansas, Louisiana, and east- ern Texas. Cirsium virginianum has been reported to have chromosome num- bers of 2n = 28 (Ownbey & Olson; Moore & Frankton, 1969) and n = 11 (Jones, 1968), while C. carolinianum has been reported as 2n = 20 (Ownbe & Olson) and nv = 11 (Jones, 1970). Ownbey & Olson included C. carolinianum in series A/tissima, but left C. virginianum in series Virginiana where it was initially placed by Petrak. Cirsium virginianum was not considered by them to be closely allied to other species of the Southeast, and its relationships to other North American species were considered to be obscure (Ownbey & OI- son). Cirsium arvense (L.) Scop. is a widespread, pernicious weed that occurs in fields and wasteplaces across the northern half of Europe and the United States. It extends south into Iowa (Pammel; Hayden) and Kansas (Barkley) and west to California, but it only skirts the northern part of the Southeast: northwestern North Carolina, northern Tennessee, and northwesternmost Arkansas (E. B. Smith). This species causes considerable losses in agricultural yields (Hayden; Hodgson, 1958, 1968; U. S. Dep. Agr.), and its control has been the subject of studies too numerous to be accounted for in this treatment (for partial reviews, see Haggar ef a/.; Hodgson, 1968). The separation of “sexes” in Cirsium arvense was first noted by T. Smith (1822), who described in detail the morphological differences between what he considered to be separate “‘male” and “female” plants. Correns (1916) more accurately described the imperfectly dioecious nature of this species and noted staminate plants bearing seeds of sexual origin. Correns’s findings were con- firmed by Delannay (1977), who showed that the stigmas of staminate plants are narrower and more elongated and the stigmatic papillae less well developed than on carpellate plants. According to Delannay, the result of these differences is less retention and germination of pollen on the staminate plants. Both De- lannay (1977, 1979) and Lloyd & Myall upheld Correns’s proposal that dioecy in Cirsium evolved from gynodioecy, a condition common to many European species of Cirsium (e.g., C. palustre, Delannay, 1979). While early studies (Bakker; Correns; Groh; Lloyd & Myall) reported very few occurrences of ‘“hermaphrodites” or strict dioecy among populations of C. arvense, Kay took into account that populations of this species differentiate into clones with pollen-bearing flowers and clones with strictly carpellate ones and found that 1990] SCOTT, GENERA OF CARDUEAE 413 15 percent of the pollen-bearing clones among three populations in southern Britain were perfect-flowered plants that produced from ten to 65 seeds per capitulum. Another 11 percent were “‘subhermaphrodites” that produced two to ten seeds per capitulum. Both “hermaphrodites” and ‘‘subhermaphrodites”’ had high seed viability, but experiments 1n self-pollination resulted in low seed set among the perfect-flowered plants. Williams studied the germination of achenes of five Cirsium species and reported that of those studied C. arvense was the most variable, with the percent of germination ranging from 13.4 to 24 in four separate studies. Williams also noted that the achenes of many thistles are commonly parasitized by beetle larvae, fruit flles (Urophora sp.), and rusts (Puccinia sp.). Guyot & Guillemat recorded a considerable decrease in viability of C. arvense seeds over a period of five years (less than ten percent germination after five years). Kumar & Irvine reported an increase in germination rate in Cirsium arvense achenes in response to prechilling, photoperiod (eight hours of light), and higher temperatures (30°C). Similar results were reported by Amor & Harris in populations found in Vic- toria, Australia. Allelopathic properties have been attributed to C. arvense by several authors (Amor & Harris; Bendall; Helgeson & Konzak; Stachion & Zimdahl). Cirsium vulgare, a wind-dispersed, monocarpic, tetraploid (2n = 68) weed introduced from Eurasia, is widely established in disturbed areas in much of North America. Although primarily an insect-pollinated outcrossing plant, self- pollination or apomixis has been shown to produce viable, if somewhat fewer, seeds. Achenes produced by either self-pollinated flowers or apomixis were significantly heavier than achenes produced by cross-pollinated flowers. In addition, achenes produced by cross-pollination germinate more slowly, and fewer seedlings survive their first year, than achenes produced by apomixis or self-pollination (Van Leeuwen, 1981b). This led Van Leeuwen to propose that achenes produced by cross-pollination are better dispersed and perhaps better equipped genetically (i.e., genetically more differentiated) to colonize areas apart from the parental plants than are achenes produced otherwise. Predis- persal seed predation by birds, mice, and insects (particularly Cheilosia grossa and Epiblema scutulana) resulted in an average loss of seven percent of the achenes in populations of C. vulgare. Stem predation by rabbits and stem- boring insects brought about an average 38 percent loss of flowering heads on these plants (Van Leeuwen, 1981la, 1983). De Jong & Klinkhammer proposed a different hypothesis for the effects of microorganisms on the germination of seeds of C. vulgare. The range of chromosome numbers from 27 = 34 to 2m = 18 among species of Cirsium has been proposed to be an aneuploid reduction series brought about by translocations between non-homologous chromosomes (Frankton & Moore, 1961; Moore & Frankton, 1962a; Ownbey & Hsi; Bloom). Intraspecific variation in chromosome numbers and accessory chromosomes are considered to be common (Frankton & Moore, 1963; Moore & Frankton, 1962a; Ownbey et al.; Pinkava & Keil). American species have been consistently reported to be diploids, while species introduced from Europe (C. arvense, C. vulgare) are known to be polyploids (Bloom). Triploid, tetraploid, and hexaploid species (all with multiples of 17 chromosomes) have been noted among Japanese 414 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 species (Aishima, Arano). Taxa from the putative center of origin in Eurasia have been reported as diploids (2” = 34) or polyploids (Moore & Frankton, 1962a; Ownbey, Raven, & Kyhos; Bloom). Moore & Frankton (1962a) sug- gested that the European species of Cirsium are on the base level of cytological evolution, since they share a base chromosome number of 17 Few investigations have been made of the phenolic compounds of Cirsium species. The flavonoids tricin-5-O-glucoside, quercetin-3-O-digalactoside and quercetin-3-O-rhamnoglucoside were reported as major constituents in C. ar- vense (Wallace). Shelyuto and colleagues (1970, 1972b) isolated apigenin, teolin, apigenin-7-O-glucoside, apigenin-7-O-rhamnoglucoside and 3-O-me Ikaempferol from C. arvense (see Wagner for review). The en cirsimaritin, first found in C. brevistylum, was determined to be a minor con- stituent of C. arvense, but was not found in C. edule, C. undulatum, or C. vulgare (Wallace & Bohm). The flavonoids kaempferol-3-O-glucoside, quer- cetin-3-O-glucoside, quercetin-3-O-galactoside, apigenin-7-O-diglucoside, and genkwanin-4'-O-glucoside were reported for C. lanceolatum (L.) Hill (Mc- Gowan & Wallace’). The phenolic acids p-coumaric, caffeic, ferulic, p-hy- droxybenzoic, protocatechuic and vanillic were also reported for C. /anceo- latum (McGowan & Wallace). Taraxasterol was found in C. texanum, but tests for alkaloids and sesquiterpene lactones were negative for this species (Dom- inguez et a/.). Taraxasterol and triterpenes are known from C. arvense (Dutta et al.), and stigmasterol has been reported from C. o/eraceum (L.) Scop. (Piatak & Eichmeier). Alkaloids of unknown structure have been found in C. arvense and in 12 other species of Cirsium (Willaman & Schubert; Willaman & Li). REFERENCES: Under tribal references see BARKLEY; BENTHAM, 1873a; BoBRov & CZEREPANOV; BrizicKy; DE CANDOLLE; CASSINI; CORRELL & JOHNSTON; CRONQUIST, 1980; DittTRIcH, 1977, DominGuez ef al.; Don; DUMORTIER; ELLIOTT; GOLDBLATT, 1981, 1984, 1985; GRAY, 1884a, 1884b; HeEG1, 1928-1929, 1987 HOFFMANN: IsLeEy; LESSING; MACROBERTS; McVauGu; Morton; MuLLER; NUTTALL; PERSOON; RADFORD ef al., 1968; RICKETT; SMALL, 1933: E. B. SMITH: STEPA, 1959, 1960; UNITED STATES DEP. AGR.; WAGNER; and WILLIS. AHu_es, H. E., C. R. BELL, & A. E. eryeee Species new to the flora of North or South Carolina. Rhodora 60: 10-32. 8. [Cirsium arvense. Airy SHAW, H. K. On eos penne Hill and C. /anceolatum (L.) Scop. Not. Syst. Paris 13: 55-58. AISHIMA, T. Chromosome eee in the genus Cirsium. I. Bot. Mag. Tokyo 48: 150- 3McGowan & Wallace cited [Ahles in] Radford et al. as their source for this specific name; both Cirsium ne eolatum and Cirsium vulgare (Savi) Tenore were listed at the end of Ahles’s treatment of Carduus lanceolatus L. Cirsium vulgare is probably the correct name of the species McGowan & Wallace studied, since Cirsium lanceolatum is a European species unknown in America, and the vlant material used by McGowan & Wallace was collected in Jackson County, North Carolina. The name Cirsium lanceolatum (L.) Hill was misapplied by Small (1933) to es vulgare (Savi) Tenore and, apparently for this reason, was cited by Ahles (in Radford et a/.) in his account of this species. (For commentary on this nomenclatural problem see Airy Shaw; Arenas, oe ) 1990] SCOTT, GENERA OF CARDUEAE 415 151. 1934. [Chromosome numbers for thirty species in 8 series were 17 or multiples thereof. Author suggests 17 as base number for the genus Amor, R. L., & R. V. Harris. Distribution and seed production of Cirsium arvense (L.) Scop. in Victoria, Australia. Weed Res. 14: 317-323. 1974, Arenas, J. A propos de Cirsium loneeciani Hill et de Cirsium lanceolatum Scop. Not. Syst. Paris 13: 59-60. 1947. Les Composées-Cynarocephales de Belgique. Bull. Jard. Bot. Bruxelles 24(4): 955. 274, 1954. ARrENO, H. The karyotype analysis and its karyo-taxonomic considerations in some genera of subtribe Carduinae. Jap. Jour. Genet. 32: 323-332. 1957. [All members of Cirsium classified into 3 types from the viewpoint of the karyotype, and Cirsium type is subdivided into 3 subtypes that agree with Kitamura’s taxonomic systems.] BAKKER, D. A comparative life-history study of Cirsium arvense (L.) Scop. and Tussilago Farfara L., the most troublesome weeds in the newly reclaimed polders of the former Zuiderzee. Brit. Ecol. Soc. Symp. 1: 205-222. 1960 BEAMAN, J. H., & B. L. TURNER. Chromosome numbers in Mexican and Guatemalan Compositae. Rhodora 64: 271-276. 1962. [Cirsium Skutchii from Guatemala re- ported as n = | BENDALL, G. M. The marae ae of California thistle (Cirsium arvense) in Tasmania. Weed Res. 15: . 1975 BLoom, W. L. Chromosomal ae tes between Cirsium discolor and C. muticum and the origin of supernumerary chromosomes. Syst. Bot. 2: 1-13. 1977. [Analyses of F, hybrids between these taxa indicate that the chromosome arrangements of these species differ by two or three translocations and one paracentric inversion.] BOHLMANN, F., & W. R. ABRAHAM. Aplotacene epoxide from Cirsium hypoleucum. Phytochemistry 20(4): 855, 856. 1981. Britton, N. L., & C. F. MittspauGH. The Bahama flora. Published by the authors. New York. 1920. [Cirsium Smallii Britton described; C. pinetorum Small placed in synonymy. ] CHARADZE, A. L. Cirsium. og Russian.) /n: B. K. SHISHKIN & E. B. Bosrov, eds., Flora URSS. 27: 51-215. 19 CLuTe, W. N. The te of plant names. XLII. The thistles. Am. Bot. 35: 72-76. 1930. CorreELL, D. S., & H. B. CorreELL. Flora of the Bahama Archipelago. (50) + 1692 pp. frontisp. Vaduz. 1982. [Cirsium vittatum (illustrated) and C. horridulum, 1466-1468; fresh stems of latter species chopped up for salads or “‘soaked in acetic acid in the manner of pickles.”’] CoRRENS, C. Untersuchungen uber oo bei Distelarten. Sitz-ber. Acad. Wiss. Berlin 20: 448-477. 1916. CzaPik, R. Karyological studies in xine of Cirsium Mill. em. Scop. occurring in Poland. (In Polish.) Acta Soc. Bot. Polon. 27: 483-489. 1958. DABYDEEN, S. Natural eee 2 the genus Cirsium: C. Flodmanii x C. undu- latum. Rhodora 89: 369-373. DANDENO, J. B. The parachute se : thistle down. Science II. 22: 568-572. 1905. Davipson, R. A. Initial biometric survey of morphological variation in the Cirsium altissimum-C. discolor complex. Brittonia 115: 222-241. Davis, D. E. The thistle epons horridulum.| No. 1 pasture weed? Highlights Agr. Res. Ala. Sta. 8(1): 1 Davis, P. H., & B. S. sere Cie: In: P. H. Davis, ed., Fl. Turkey 5: 370-412. 1975. DELANNAY, X. ee study of dioecy in Cirstum arvense. Phytomorphology 27: 419-425. 1977 . La gynodioecie dans le genre Cirsium Miller. Bull. Soc. Bot. Belg. 111: 10-18. 1978. 416 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 —. Evolution of male sterility mechanisms in gynodioecious and dioecious species of Cirsium (Cynareae-Compositae). Pl. Syst. Evol. 132: 327-332. 1979. DerSCHEID, L. A., & R. E. SHuttrz. Achene development of Canada thistle ide arvense) and perennial sowthistle (Sonchus arvensis). Weeds 8: 55-62. 1960 DetMers, F. Canada thistle, Cirsium arvense Tourn. Ohio Agr. Exp. Sta. Bull. 414: l- 45. 1927 Dewey L. H. Canada thistle Gece arvensis (L.) Robs.). U. S. Dep. Agr. Bot. Circ. 27: 1-14. Revised ed. June, 1901. DEWoLrF, G. P. The stemless es Cirsium acaule. Baileya 5(3): 118, 119. 1957. DittricH, M. Anatomishe Untersuchungen an den Friichten von Carthamus L. und page sae Candollea 24: 263-277. 1969. Druce, G. C. e abridgement of Miller’s Gardeners Dictionary of 1754. Rep. Bot. Exchange aa Brit. Isl. 3: 426-436. 1914 Dusy, J. E. Aug. Pyrami de Candolle Botanicon gallicum. ed. 2. 1828. Paris. [Source of sect. Epitrachys DC.] Dutta, C. P., L. P. K. Ray, & D. N. Roy. oo & its derivatives from Cirsium arvense. ‘Phytochemistry 11: 2267-2269. 1972. FERNALD, M. L. Minor transfers and forms in ose Sarai 45: 353-354. 1943a. . Eastern extension of Cirsium Flodmani. Ibid. 356. 1943b. ForsyTH, S. F., & A. K. WATson. Predispersal seed predation of Canada thistle. Canad. Entomol. 117: 1075-1081. 1985. FRANKTON, C., & R. J. Moore. Cytotaxonomy, phylogeny and Canadian distribution of Cirsium undulatum oe Cirsium Flodmani. Canad. Jour. Bot. 39: 21-33. 1961. [Includes distribution & Cyt eps of Cirsium muticum, Cirsium discolor, and Cirsium altissimum. Ibid 41: 73-84. 1963. Guiyzin, V. L, V. L. SHELvuTO, A. V. Patupin, & N. T. BuBon. Mater S’ezda. Farm. B. SSR. 3: 153-156. 1977. I. F. Urvantsev, ed., Minsk. Gos. Med. Inst. USSR.* [Of 62 Cirsium spp. almost all contained apigenin, which was found only in inflores- cences of some spp. Cynaroside found in infl. of 17 spp., pectolinarin found in 19 Goprrey, R. K. Studies in the Compositae of North Carolina I. An enumeration of noteworthy distribution records. Jour. Elisha Mitchell Sci. Soc. 66: 186-194. 1950. [C. carolinianum, C. Smallii. —— & J. W. ane: Aquatic and wetland plants of southeastern United States. Dicotyledons. x + 933 pp. Athens, Georgia. 1979. [Cirsium, 877-881; includes C. Lecontei, C. Nuttall G: ene. C. muticum, and an excellent account of the C. horridulum complex Gray, A. A synopsis of the North American thistles. Proc. Am. Acad. 10: 39-48. 1874. [A treatment of native Cirsium species as Cnicus.] Grou, H. The inflorescence of the Canada thistle (Cnicus arvensis). Ontario Nat. Sci. Bull. 3: 41-42. 1907. [Author reported on imperfectly dioecious nature of this species and morphological changes thus brought about, while noting that some plants of this taxon do not produce seed.] Guyor, L., & J. GUILLEMAT. Semences et plantules des principales mauvaises herbes. Haacacar, R. J., A. K. OswaLp, & W. G. RICHARDSON. A review of the impact and oa = creeping thistle (Cirsium arvense L.) in grassland. Crop Protect. 5(1): 73- 6. ees ie Distribution and reproduction of Canada thistle in Iowa. Am. Jour. Bot. 21: 355-373. 1934. HELGEson, E. A., & R. Konzak. Phytotoxic effects of ts of field bindweed and of Canada thistle. A preliminary report. N. Dakota Agr. ee Sta. Bull. 12: 71- 76. 1950.* 1990] SCOTT, GENERA OF CARDUEAE 417 Hit, E. J. The pasture-thistles, east and west. Rhodora 12: 211-214. 191 Hopason, J. M. Canada thistle (Cirsium arvense Scop.) control with cultivation; crop- pings = chemical sprays. Weeds 6: I-11 58. —— iation in ecotypes of Canada thistle. Ibid. 12: 167-171. 1964. ena ecology and control of Canada thistle. U. S. Dep. Agr. Tech. Bull. 1386. : pp. 1968. Ho.vs, J. Brief aes on the 4th volume of Flora Europaea. Preslia 49(4): 311- 327. 1977. [Numerous nomenclatural notes on a variety of Compositae, Cirsium and Centaurea in particular; none of the former of concern for the southeastern United States.] Howe i, J. T. Sertulum Greeneanum. III. Studies in Cirsium. Am. Midl. Nat. 30: 29- 39, 1943. . Studies in Cirsium. I. Leafl. West. Bot. 9(1): 9-15. — [Nomenclature and concepts of several species from the western United Stat . Distributional data on weedy thistles in western North America. Ibid. 9(2): 17- 29. 1959b. [Species from 9 genera keyed and representative specimens noted, along with earliest record known to the author from the western United States.] Hs1, Y. T. Taxonomy, distribution and relationships of the species of Cirsium belonging to the series Undulata. Ph.D. thesis, Univ. Minnesota. Univ. Microfilms, Ann Arbor, Michigan.* [Also, Diss. Abstr. 21(8): 2085-2086. 1961.*] Hunter, J. H. Integrated approach to Canada thistle control. Pp. 8-12 m 1984 Manitoba Weed lee se 1984. sat yoe and Winnipeg, Manitoba. p.* o, & G.I. McIntyre. Some effects of a on the growth and eee a pepe arvense. Bot. Gaz. 146: 524-530. 1985. JEFFREY, C. Notes on Compositae: III. The Cynareae in eee Africa. Kew Bull. 22: 107-140. 1968. Jones, S. B., Jk. Chromosome numbers in southeastern United States Compositae. I. Bull. Torrey Bot. Club 95: 393-395. 1968. [Cirsium virginianum (as Carduus vir- ginianus L.) reported as n = 11.] romosome a in the Compositae. Ibid. 97: 168-174. 1970. [Cirsium carolinianum, n= JonG, T. J. pe, & P. G. ne The negative effects of litter of parent plants of Cirsium ee on their offspring; autotoxicity or immobilization. Oecologia 65: influence plant growth. This hypothesis differs from those that attribute allelopathic activity to C. arvense; see STACHEN & ZIMDAHL. | Kay, Q. O. N. Hermaphrodites and subhermaphrodites in a reputedly dioecious plant, Cirsium arvense (L.) Scop. New Phytol. 100: 457-472. 1985. KITAMURA, S. Compositae Japonicae. I. Mem. Coll. Sci. Kyoto Imp. Univ. B. 2: 87- 387. 1937. [Subtribe Carduinae, 31-212; Ci, 33-134, ca. 90 species in 3 sections ea Sas Kitamura, Pseudo- cas Nakai, Qnotrophe ).] Kocu, W. D. Synopsis Flora Germanicae et Helveticae. 1x + 844 pp. Francofurti ad a ieee 1837. [Early treatment of sections of Cirsium. ] KRASNIKOV, A. A. Chromosome numbers in some representatives of the family Aster- aceae from Siberia USSR. Bot. Zhur. ae 70(12): 1702-1703. 1985. Kumar, V., & D. E. G. IRvine. Germination of seeds of Cirsium arvense (L.) Scop. Weed Res.11: 200-203. 1971. [Germination is increased with exposure to light, higher temperatures, and pre-chilling. } LapycIna, E. YA. Anatomical and histochemical study of the leaves of Cirsium setosum (Willd.) Mb. (in Russian.) Farmatsiya Moscow 29(6): 17-23. 1980. [Flavonoids mainly located in vacuoles of upper epidermal cells. LALONDE, R. G., & J. D. SHORTHOUSE. Growth and development of larvae and galls of 418 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Urophora cardui (Diptera, Tephritidae) on Cirsium arvense (Compositae). Oecologia 65: 161-165. 1985. Lamp, W. O.,& M. K. McCarty. Predispersal seed predation ofa native thistle, Cirsium oe ee Environ. Entomol. 11: 847-851. 1982. LEEUWEN, B. vAN. Influences of micro-organisms on the germination of the mono- carpic Cirsium vulgare in relation to disturbance. Oecologia 48: 112-115. 1981a. The role of pollination in the population biology of the monocarpic species Cirsium ie and Cirsium vulgare. Ibid. 51: 28-32. 1981b. The consequences of predation in the biology oat monocarpic species Cirsium palustre and Cirsium vulgare. Ibid. 58: 178-187. 1983. Lewis, W. H., H. L. StripLinc, & R. G. Ross. Chromosome numbers for angiosperms ofthe southern United States and Mexico. Rhodora 64: 147-161. 1962. [C. horri- dulum, 156, | = 34] Lioyp, D. G., & A. J. MYALL. Sexual dimorphism in Cirsium arvense (L.) Scop. Ann. Bot. Il. 40: 115-123. 1976 Lona, R. W. Additions and nomenclatural changes in the flora of southern Florida— I. Rhodora 72: 17-46. 1970. [Cirsium horridulum var. vittatum (Small) Long, 45.] & O. LAKELA. A flora of tropical Florida. xvii + 962 pp. Coral Gables, Florida. 1971. [Cirsium horridulum var. horridulum and var. vittatum, C. Nuttallii, the only representatives of Cynareae Love, A., & D. Léve. Jn: IOPB chromosome number reports LXXV. Taxon 31: 344- ae 1982. [Cirsium altissimum, 2n = 20; C. Drummonadii, 2n = 34; C. Flodmanii, 1 = 22: C. muticum, 2n = 20; C. undulatum, 2n = 26.) Lovins, M. D., & J. L. HAmrick. Genetic organization and ee history in o North American species of Cirsium. Evolution 42: 254-265. 8 eG. R.S., & L. C. HADERLIE. Seasonal variations in Aaa thistle (Cirsium arvense) root bud growth and root carbohydrate reserves. Weed Sci. 33(1): 44-49. 1985. [Root-bud growth highest during late fall and winter months following death of aerial shoots; no obvious seasonal patterns in presence of root buds or their ability to elongate at different times of the year. McCarty, M. K., C. J. Scirres, & L. R. Rosinson. A descriptive guide for major Nebraska thistles. Nebraska Agr. Exp. Sta. Bull. 493. 24 pp. 1967. [Chiefly Carduus and ae McGowan, S. G., & J. W. WALLAcE. Flavonoids and phenolic acids from Cirsium ae Phytochemistry 11: 1503, 1504. 1972. MEEHAN, T. Cirsium discolor. Meehan’s Monthly 6: 161. Meta, K. Der An clegtantieed Bau des Stengels bei den Compositae Cynareae. Diss. ]. Fak. Univ. GGttingen. 1907. ey W.A. Cirsium Small Britton. Quart. Jour. Fla. Acad. Sci. 12: 65, 66. 1949. Moore, R. J., & C. FRANKTON. ean: studies in the tribe Cynareae (Com- positae). Canad. Jour. Bot. 40: 281-293. 1962a & y erred and Sa as distribution of Cirsium edule and Cir- sium brevistylum. Ibid. 40: 1187-1196. maps. 1962b. & Cytotaxonomy of Cirsium Hookerianum and related species. Ibid. 43: 597-613. 1965. —— & . An evaluation of the status of Cirsium pumilum and Cirsium Hillii. Ibid. 44: 581-595. 1966. & Cytotaxonomy of foliose thistles Se spp. aff. C. foliosum) of western North America. Ibid. 45: 1733-1749. 19 Cytotaxonomy of some Cirsium ae of the eastern United States with a key to eastern species. Ibid. 47: 1257-1275. 9. Morton, J. K. A cytological study of the Compositae uae Hieracium and Ta- raxacum) of the British Isles. Watsonia 11: 211-223. 1977. Netson, A., & J. B. MACBripe. A teratological thistle. Am. Bot. 20: 136, 137. 1914. {Description OFA fasciated specimen of C. Drummondii.] 1990] SCOTT, GENERA OF CARDUEAE 419 Ownsey, G. B. Natural aac in the genus Cirsium—lI. C. discolor (Muhl. ex Willd.) Spreng. x C. m m Michx. Bull. Torrey Bot. Club 78: 233-253. 1951. —. Natural hbridization in the genus Cirsium—II. C. altissimum x C. discolor. Michigan Bot. 1: 87-97. 4. Nuttall’s Great se species of Cirsium: C. undulatum and C. canescens. Rhodora 54; 29.-35. 1952. . T. Hst. Chromosome numbers in some North American species of the genus Cirsium. Il. Western United States. Madrono 20: 225-228. 1970. & W. A. OLson. Cytotaxonomic notes on the species of Cirsium native to the southeastern United States. Rhodora 71: 285-296. 1969. ,P. H. RAven, & D. W. KyHos. Chromosome numbers in some North American species of the genus Cirsium. III. Western United States, Mexico and Guatemala. Brittonia 27: 297-304. 1975. PAMMEL, L. H. The thistles of Iowa, with notes on a few other species. Proc. lowa Acad. Sci. 8: 214-239. 1901. Pavone, P., C. M. Terrasi, & A. Z1zzA. In: Chromosome number reports LX XII. Taxon 30: 695, 696. 1981. PETRAK, F. Die mexikanischen und saree amerikanischen Arten der Gattung Cirsium. Beih. Bot. Centralbl. 27: 207-255. 1910. —. Beitrage zur Se oe es und zentral-amerikanischen Cirsien. Bot. Tidsskr. 31: 57-72. Die pordamertanischen Arten der Gattung Cirsium. Beih. Bot. Centralbl. 35: 932 567. 1917. . Uhr einige Arten und Bastarde der Gattung Cirsium. Mitt. Thur. Bot. Ges. 2: 13-41. : pea In: K. H. Recuincer, ed. Fl. Iran. 139a: 231-280. 1979. PiaTAK, D. M., & L. S. EICHMEIER. Plant investigations, we Hexane extract of Cirsium arvense. Trans. Illinois State Acad. Sci. 64: 300. 1971. Pinkava, D. J., & D. Ker. Chromosome counts of ee from the United States and Mexico. Am. Jour. Bot. 64: 680-686. 1977. [Four species of Cirsium reported: C. arizonicum, 2n = 15u, C. ea 2n = 17; C. Flodmanii, 2n = 12m; C. undulatum, 2n = 120 and 2n = | PODDUBNAJA-ARNOLDI, W. Ein Versuch der Anwendung der embryologischen Methode bei der Lésung einiger systematischer Fragen. I. Vergleichende embroyologisch- zytologische Untersuchungen iiber die Gruppe Cynareae, Fam. Compositae. Beih. Bot. Centralbl. 48: 141-237. 1931. PuTNAM, B. L. A white form of Carduus arvensis. Asa Gray Bull. 7: 37. 1899. [Cirsium ense.| Rosrnson, B. L. Notes on the genus Cirsium. Rhodora 13: 238-240. 1911. RostovtseEvA, T. S. Chromosome numbers of some species of the family Asteraceae. Biol. Zhur. Leningrad 64: 582-589. 1979. [Cirsium incanum (S. G. Gmel.) Fisch. ex Bieb., 2n = 28; Carduus crispus, 2n = 18.] . Chromosome numbers of some species of the family Asteraceae 2. [bid. 68: 660-664. 1983. [Arctium Lappa L., 2n = 36; Cirsium heterophyllum (L.) Hill, n = 17.] RypsBerG, P. A. Studies on the Rocky Mountain flora— XXIV. Bull. Torrey Bot. Club 37: 541-557. 1910. [Notes on numerous Cirsium taxa (as Carduus), particularly hybrids; four new species of Cirsium described as Carduus. Flora of the Rocky Mts. and Adjacent Plains. xii + 1110 pp. Published by the author, New York. 1917 Flora of the Prairies and Plains of Central North America. vil + 969 pp Published by the New York Botanical Garden, New York. 1932. [Cynareae, 879- 886; Cirsium, 879-883, 18 spp. treated.] Scopout, J. A. Methodus Plantarum. 28 pp. Vienna. 1754. [Cirsium, 19, listed with Carduus and Arctium (as Lappa) under Classis Flosculosae.] 420 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 ra Carniolica. xxii + 607 pp. Vienna. 1760. [Cirsium treated with description; seven nsec listed without binomials. ora Carniolica. ed. 2. Ixxii + 448 pp. Vienna. 1772. [Cirsium, 123; 15 species Fie and sae binomials.] ScruGul, A., E. Boccieri, & C. DEL PReTeE. Chromosome numbers for the Italian flora, part 331 -365. Inf aa Ital. 9(2): 116-140. 1977. [Cirsium vulgare ssp. sylvaticum (Tausch) Dostal, 2” = 68; tetraploid of x = 17.] oo V.L., V. I. Gryzin, & A. I. BAN’Kovsku. Flavonoids of Cirsium arvense. m. Nat. Compounds 6(3): 366. 1970.* me , & N. T. Buson. Flavonoids of Cirsium oleraceum flowers. Khim. Prir, Soedin. 7: 371 ff. 1971.* ———, ———, & N. T. Buon. Flavonoids of Cirsium palustre. Ibid. 2: 241 ff. 1972a.* ——, ——,& . 3-O-Methylkaempherol from Cirsium arvense. Ibid. 7: 371 72b.* SuiH, C. Notulae de plantis tribus Cynarearum familiae Compositarum Sinicae. II. (In Chinese; English summary.) Acta Phytotax. Sinica 22: 445-455. 1984. [Forty-nine species of Cirsium reported for China (9 newly described, 4 new records reported) divided among 8 sections (3 newly described). ] SHINNERS, L. H. Notes on Texas Compositae—I. Field & Lab. 17: 23-30. 1949. [Cirsium terrae- ae Shinners described. ] names and records for Texas Compositae. Sida 1: 373-379. 1964. [Cirsium eee Shinners placed in synonymy under C. Engelmannii.] Stepa, I. S. Morphology of pollen of the genus Cirsium Mill. and of close genera of the tribe Cynareae (Compositae). (In Russian.) Trudy Tbilis. Bot. Inst. Acad. Nauk SSSR. 21: 81-126. 1961. SKALINSKA, M., J. MALECKA, R. IZMAILow, et a/. Further studies in chromosome num- bers of Polish angiosperms. XII. Acta Biol. Cracov. Bot. 19: 107-148. 1974. SMALL, J. K. Additions to the flora of subtropical Florida. Bull. New York Bot. Gard. 3: 419-440. 1905. [Description of Carduus vit ll (= Cirsium vittatum (Small) Small).] of Miami. xii + 206 pp. Published by the author. New York. 1913. to Cirsium vittatum (Small) Small, C. pinetorum (Small) Small, and C. horridulum, 199 SmitH, T. On certain species of Carduus and Cnicus which appear to be dioecious. Trans. Linn. Soc. London 13: 592-603. 1822. STACHEN, W. J., & R. L. ZimpauH. fae are activity of Canada thistle (Cirsium arvense) in ees Weed oe a 83-86. 1980. STOUTAMIRE, W. P., & J. H. BEA Chromosome studies of Mexican alpine rie Brittonia 12: oe 230. 1960. Casun cernuum and C. subcoriaceum, both 17. Strip, A., & R. FRANZEN. - Chromosome number reports LX XII. Taxon 30: 829- 842. 1981. jae Cirsium species from Greece reported to be 2n = 34; C. vulgare confirmed : 2n = 68.] THIERET, J. W. Twenty-five ae of vascular plants new to Louisiana. Proc. Louisiana Acad. Sci. a 78-82. 9. [C. terrae-nigrae Shinners new to Louisiana: Caddo Parish, 2 mi. SE of reer Thieret 25999 det. G. B. Ownbey, who was reported to be revising N. American Cirsium. Cf. SHINNERS, 1964.] Tonyan, T. R. Caryological data on species of Cirsium Mill. growing in Armenia. (In Russian.) Biol. Zhur. Armenii. SSR. 34(7): 769-772. 1981. [C. arvense, C. incanum (S. G. Gmel.) Fisch. ex Bieb., and C. echinus (Bieb.) Hand.-Mazz. reported as 2” = 34] 1990] SCOTT, GENERA OF CARDUEAE 421 —. New chromosome numbers of the ans of Cirsium in Armenia. (In Russian.) Ucen. Zap. Erevan Univ. 3: 115-120. 2% Toumey, J. W. Fasciation in Cnicus pains Bot. Gaz. 16: 236. 1891. [C. vulgare.] TURNER, B. L., W. L. ELLIson, & R. M. KING. Chromosome numbers in the Compositae. IV North American species, aes phleti interpretations. Am. Jour. Bot. 48: 216- 223. 1961. [C. horridulum, n TURNER, S. K., P. K. Fay, E. L. oe p, & D. C. SANDs. Resistance of Canada thistle, Cirsium arvense, ecotypes toa on pathogen, Puccinia obtegens. Weed Sci. 29(6): l. VAN Loon, J. C., & H. DE Jonc. /n: IOPB chromosome number reports LIX. Taxon 27: 53-61. 1978. [C. arvense, 2n = 34; C. palustre, 2n = 34.] WALLACE, J. W. Tricin-5-O-glucoside and other flavonoids of Cirsium arvense. Phy- tochemistry 13: 2320, 2321. 1974. [Quercetin-3-O-digalactoside and quercetin-3- O-rhamnoglucoside found to be major constituents; cirsimaritin-4’-O-rutinoside reported as a minor constituent. ] & Boum. Cirsimaritin-4’-O-rutinoside, a new flavone glucoside from Cirsium brevistylum. Phytochemistry 10: 452-454. 1971. [Compound not found in C. le, C. undulatum, C. vulgare, or C. arvense;, bibliography includes citations for early work on several Japanese taxa. Watson, A. K., R. BEAUREGARD, & W. KEOGH. Biological control of Canada thistle in Quebec. Canada Thistle Symp. Agr. Canada, March, 1980, Regina, Saskatchewan. WELLS, H. Hy bridization and genetic recombination of Cirsium californicum and Cir- sium occidentale (Asteraceae: Sap ee 30: 12-30. 1983. WERNER, K. Cirsium. In: T. G. TuTin et al., , Fl. Europaea 4: 232-242. 1976. WILLAMAN, J. J.. & B. G. SCHUBERT. oe eee plants and their contained al- kaloids. U.S. Dep. Agr. Tech. Bull. 1234. 287 pp. 1961. [Unnamed alkaloids reported in Cirsium arvense and C. setigerum Ledeb., .L. Lt. Alkaloid- ieee? plants and their contained alkaloids. Lloydia 33 (Supplement): vii + 286 pp. WILLIAMS, J. T. Variation in the es of several Cirsium species. Trop. Ecol. 7: 1966. [Five species, including C. arvense and C. vulgare, studied. Germination rate of C. arvense achenes found to be quite variable, with a range of 34.3 to 66.3 percent. Range for C. vu/gare was between 26.2 and 41.6 percent. Common par- asitization of Cirsium achenes by beetle larvae, fruit flies, and rusts also noted.] Witson, M. F., P. K. ANDERSON, & P. A. THomas. Bracteal exudales)s in two Cirsium species as possible deterrents to insect rene of seeds, Cirsium discolor , Cirsium Flodmani. Am. Midl. Nat. 110: 212-214. WILson, R. G., & M. K. McCarry. a and seedling and rosette develop t of Flodman thistle (Cirstum Flodmanil). Weed Sci. 32: 768-773. 1984. [Light im- portant in seed germ a greatest seedling emergence occurred when seeds were planted on soil surfac Wu, T. S., C. S. Kuon, & S. I. Jen. Constituents of Formosan folk medicine. VIII. Flavonoids of the leaves of Cirsium albescens, C. japonicum var. australe, C. Ka- wakamui f. variegatum and C. Hosogawa. T’ai-wan Yao Tsa Chih. 32(2): 88-90. 1981. [Linarin, a flavone glycoside, was isolated from leaves of all; C. Hosogawa leaves also yielded luteolin, luteolin-7-glucoside, and linarin.] Wynn-WILiiAMs, M. Observations pate a seeds and germination of some thistle species. Int. Seed Test. Assoc. 6: 4 ZILKE, S., & L. A. DERSCHEID. oe sae germination of Canada thistle and perennial sowthistle seeds. N. Cent. Weed Control Conf. Proc. 14: 42, 43. 1957. 422 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 2. Carduus Linnaeus, Sp. Pl. 2: 820. 1753; Gen. Pl. ed. 5. 358. 1754. Biennial [annual or perennial] herbs. Stems to ca. 2 m tall, fibrous, glabrous to arachnoid-tomentose, spiny winged, wings triangular to palmate, occasion- ally with an apical spine 1-5 mm long. Leaves sessile to more often petiolate; blades lanceolate to oblanceolate, deeply lobed to 2-pinnatisect [subentire], to ca. 25cm x 10cm, the upper surface glabrescent to sparsely [densely] pubes- cent, occasionally punctate glandular, the abaxial surface glabrous to densely arachnoid-villous and often glandular, hairs crispate or sinuate, multicellular, glands capitate or punctate; margins often with an apical spine to 5 mm [ex- ceeding 10 mm]. Capitula homogamous, discoid, subglobose [globose], 1.5- 2.5 [8] cm wide, sessile or in clusters of 2-4 [10 or more] on narrowly winged peduncles. Involucre ovoid, bracts imbricate in 5-8 [10] series, linear-lanceolate to oblanceolate, acute to acuminate, spine tipped or sharply mucronate, spread- ing or reflexed [adpressed], glabrous to densely pubescent, occasionally mi- nutely glandular; margins entire, scabrous or ciliate; outer bracts long-acu- minate to acute, inner bracts longer than middle, acute [obtuse], shortly contracted to a spinule, obscurely 1-5 veined, occasionally contracted slightly below middle. Receptacle densely setose. Flowers perfect, actinomorphic or zygomorphic. Corolla purple, lavender, [pink or white], tube narrow, throat short and unequally divided by sinuses of corolla lobes with 2 sinuses deeper than the others, lobes linear, elongate (ca. 4-5 mm long), glabrous or sparsely glandular. Staminal filaments sparsely to densely pubescent, anthers white, sagittate, with slender, entire or lacerate basal appendages, apical appendage acute, small. Styles smooth below a distinct collar of hairs at base of the relatively short, papillose, scarcely divergent, narrowly truncate to acute branches. Achenes basifixed, 3-5 mm long, somewhat compressed, glabrous, smooth [rugulose when dried] with 5-10 ribs and a distinct annular margin at apex. Pappus bristles in several rows, connate at base, unequal, inner setae longer than outer, white, not plumose, smooth to barbellate. LECTOTYPE SPECIES: Carduus nutans L.; see Britton & Brown, Illus. Fl. No. U. S. Canad. ed. 2. 3: 554. 1913. (Name the ancient Latin one.)— PLUMELESS THISTLE. A genus of 91 species, according to Kazmi (1963, 1964), distributed in Eurasia and northern Africa. Two introduced and naturalized species occur in the southeastern United States. As established by Linnaeus, Carduus was composed of 23 species, several of which were later placed in Cirsium by Scopoli (1772) (see discussion under Cirsium). While the strength of the characters used to distinguish these genera (pappus of smooth or barbellate setae vs. pappus of plumose setae) has been questioned (Correll & Johnston; Willis) most workers have treated Cirsium and Carduus separately. Recently, though, Ahles in Radford et a/. included in Carduus 13 taxa that are considered in this paper to be members of Cirsium. The 91 species recognized by Kazmi (1963, 1964) were placed in two sub- genera (CARDUUS, AFROCARDUUS Kazmi) of Eurasian and African distribution. He later (1964) proposed a third subgenus (ALFREDIA (Cass.) Kazm1) based on characters of the corolla lobes, staminal filaments, and pappus. Subgenus AFRO- CARDUUS was delimited by hooked corolla lobe apices, glabrous or sometimes 1990] SCOTT, GENERA OF CARDUEAE 423 tuberculate filaments, and opaque achenes. AFROCARDUUS is comprised of ca. 21 African species that are placed among sections ACAULON Kazmi (7 spp.), AFROCARDUUS Kazmi (8 spp.), and PINNATISQUAMA Kazmi (6 spp.). For subg. ALFREDIA, represented by species from Afghanistan, western Pakistan, eastern Russia, Mongolia, Tibet, and China, Kazmi (1964) proposed sections ALFRE- DIA, PTEROCAULON, and APTERON but did not give a treatment of species. Moore & Frankton (1962) reported that sect. ALFREDIA contains six species distin- guished by their yellow flowers. Subgenus CarpDuus (70 spp.) (Kazmi, 1964) is distinguished by smooth corolla lobes, villous staminal filaments, and pappus usually of whitish or, rarely, stramineous bristles. The distribution of this group extends from the Mediterranean region to northern Europe with several widespread species that are now encountered in many areas of the United States. The two species Carduus acanthoides L. and C. nutans L., musk-thistle, known to occur in the Southeast are members of subg. CARDuus. Both belong to the type section, one of two (CARDUUS, LEPTOCEPHALI Reichenb. f.) recog- nized in the group (Kazmi, 1964; Franco, 1976). Carduus crispus L., a species sparingly introduced in the United States, and also a member of sect. CARDUUS, has not been reported in the Southeast, but eventually may be encountered in the northern parts of the area along roadsides or in disturbed areas (Johnson, 1974), Both Carduus nutans and C. thoides are widely established in the United States along roadsides and in pastures and disturbed areas. Hensley observed that both species thrive in areas of high-calcium limestone deposited at very shallow depths. Carduus acanthoides does not extend south of a few localized occurrences in western North Carolina (Radford et a/.) and, perhaps, Tennessee, yet C. nutans has been collected as far south as Georgia, Mississippi, and Louisiana. Both species are monocarpic, outcrossing (Smyth & Hamrick) bien- nials that can be differentiated by capitulum and involucre size and the ori- entation of the capitula (heads nodding in C. nutans vs. upright or erect in C. acanthoides). Moore & Frankton (1974) recognized three subspecies of C. nutans in Canada, and Kazmi (1964) added a fourth from northern Europe. Two subspecies of C. acanthoides were described by Kazmi (1964). Cytological studies have reported chromosome numbers of 2n = 16 (n = 8) for C. nutans (Poddubnaja-Arnoldi; Gorecka; Léve & Love, 1944; Moore & Frankton, 1962; Devesa; Van Loon & De Jong; Van Loon & Kieft; Van Loon & Snelders) and 2n = 22 for C. acanthoides (Gorecka; Love & Love, 1944; Moore & Frankton, 1962; Poddubnaja-Arnoldi; Van Loon & Kieft). Mehra and colleagues and Mehra & Remanandan have reported chromosome numbers of 2n = 40 (n = 20) for C. nutans with laggards and bridges mildly disturbing meiosis. The chromosome number for C. crispus has consistently been reported as 2n = 18 (Arano; Goérecka; Léve & Live, 1944; Poddubnaja-Arnoldi; Pro- batova & Sokolovskaya; Rostovtseva). Moore & Frankton (1962) suggested a base chromosome number of 8 or 11 for the genus. The lower numbers were thought to result from reduction brought about by fragmentation and fusion. Polyploids appear to be rare in Carduus, with 2n = 54 first reported for C. tenuiflorus Curt. and C. pycnocephalus L. by 424 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Moore & Frankton (1962). They speculated that these species are hexaploids of the base number 9 formed by hybridization between a diploid and a tet- raploid species followed by doubling of the chromosome number of the hybrid. Hybridization is commonly reported in Carduus (Arénes, 1949; Briquet, 1931; Devesa; Kazmi, 1964; McCarty; Moore & Mulligan, 1956). Kazmi (1964) listed crosses between C. acanthoides and 14 other species and between C. nutans and four others. Moore & Mulligan’s (1959) study of hybridization between C. nutans and C. acanthoides (= C. x orthocephalus Wallr.) showed that several morpholog- correlation between morphology and chromosome number (1.e., the more closely a hybrid approached the morphology of one of the parental types the more likely it was to have a chromosome number either close to or the same as that parent). A few hybrid plants had chromosome numbers of 2” = 16, but the majority of the hybrids had chromosome numbers of 2” = 22, which led Moore & Mulligan to hypothesize that backcrossing to both parents had occurred to produce the chromosome numbers of the parental stock. Of the carpellate parents with a known chromosome number of 2” = 22, 23 of the 24 hybrid progeny examined had 2n = 22. Wagner reported the alkaloids acanthoine and acanthoidine for Carduus ac anthoides and referred readers to Bohlmann et ai. for a list of polyacetylenes. colleagues’ report of flavonoids from C. pycnocephalus appears to be one of the few reports of flavonoids isolated from Carduus. There are no economically important members of this group. REFERENCES: Under tribal references see ARANO; BOBROV & CZEREPANOV; BOHLMANN et al.; BRIQUET, 1931; CORRELL & JOHNSTON; CRONQUIST, 1980; GoLDBLATT, 1981, 1984, 1985; Hea, 1928-29, 1987; JoHNsoN, 1974; Love & L6ve, 1944; MABBERLEY; MACROBERTS; MEHRA et al, MEHRA & REMANANDAN: D. M. Moore; Moore & FRANKTON, 1962, 1974; PODDUBNAJA-ARNOLDI, RADFORD ef a/.; RICKETT; SMALL, 1933; Scopout, 1772; E. B. SMITH; WAGNER; WILLAMAN & SCHUBERT; WILLAMAN & Lr; and WILLIS ABDEL-SALAM, N. A., I. Morel, & S. CATALANO. An unusual triterpene ester in a compositaceous plant, Ts getulus. Tat Jour. Crude Drug Res. 21: 79-80. 1983.* Amer, M. M.A., O. M. SALAMA, & A. A. OMAR. Flavonoids of Carduus pycnocephalus. Fitoterapia 1985(1): 61. 1985. ARENES, J. ro eo 1 étude du genre Carduus. Mém. Mus. Natl. Hist. Nat. Paris. Il, 24: 183-255. 1949. —. Surla sstémation de quelques Carduus. Not. Syst. Paris 15(4): 390-410. 1959. BATRA, S. W. T. Insects and fungi associated with Carduus thistles oe U.S. Dep. Agr. Sci. & Educat. Admin. 100 pp. Washington, D. C. Cory, V. L. Six thistles introduced into Texas. Madrofio 5: coon. 1940. [C. pyc- nocephalus, C. nutans. | Davis, P. H. Carduus. Notes Bot. Gard. Edinburgh 33: 413. 1975. . Carduus. In: P. H. Davis, ed. Fl. Turkey 5: 420-438. 1975. — = 1990] SCOTT, GENERA OF CARDUEAE 425 Devesa, J. A. Contribuci6én al estudio cariologico del genero Carduus en la peninsula Iberica. Lagascalia 10: 65-81. 1981. [Chromosome numbers of 22 taxa of Carduus; base numbers of 8, 9, 10, & 11 reported; C. nutans reported as n = 8, 2n = 16; C. tenuiflorus reported as 2n = 54, presumably an allopolyploid from x = 8 and x = 11 ancestors Franco, J. do AMARAL. Carduus. In: - G. TuTIn et al., eds., Fl. Europaea 4: 220-232. 1976. (Key by M. L. RocHa AFo Fries, R. E. Revision der tropisch- ache Carduus-Arten. Acta Horti Berg. 8: 11-38. 1923. [Twenty-two species of Carduus of which 14 are newly described.] Gorecka, A. Cytological studies in three species of Carduus L. (In Polish; Enelish summary.) Acta ae Bot. Polon. 25: 719-731. GrREMAuUD, M. An attempt at experimental taxonomy in the defloratus group of the genus Carduus Goose II. Morphological variation, numerical taxonomy. (In ae Rev. Cytol. Biol. Veg. Bot. 4: 111-171. 1981. [Includes illustrations and maps. Hamrick, J. L., & J. M. Lee. Effect of soil surface topography and litter cover on the germination, survival, and growth of musk thistle (Carduus nutans). Am. Jour. Bot. 74: 451-4 84. HARRIMAN, N. . In: IOPB oo number reports XLVIII. Taxon 24: 367-372. 1975. [C. acanthoides, 2n = 22.] Hens.ey, M. S. Distribution of | an and Carduus thistles in Rockingham County, Virginia. [Abstr.] Virginia Jour. Sci. 24: 140. 1973. Hitt, E. J. Carduus Hillii perennial. Pl. World 2: 127. 1899. HowELt, J. a Carduus in California. Leafl. West. Bot. 2: 212, 213. 1939. Kazmt, 8S. M.A. Revision der Gattung Carduus (Compositae), I. Mitt. Bot. Staatssamml. Miinchen 5: 139-198. 1963; II. Zbid. 279-550. 1964. KITAMURA, S. Compositae Japonicae. Mem. Coll. Sci. Kyoto Univ. B. 13: 1-421. 1937. [One species of Carduus, C. crispus L., treated, 32, 33.] Kok, L. T. Status of two European weevils for the biological control of Carduus thistles in the U.S.A. Eanes conicus, Ceuthorhynchidius horridus.] Acta Phytopathol. 16: 139-142. 1981.* Kuzmanov, B., & S. GeorcievA. /n: IOPB chromosome number reports LIII. Taxon 25: 483-500. 1976. [C. acanthoides, 2n = 22.] Lee, J. M., & J. L. Hamrick. Demography of two natural populations of musk thistle (Carduus nutans). Jour. Ecol. 71: 923-936. 198 McCarty, M. K. A nursery study of large- flowered taxa of Carduus. Weed Sci. 33: 664-668. 1985. C. J. Scirres, & L. R. Ropinson. A descriptive guide for major Nebraska thistles. Nebraska Agr. Exp. Sta. SB 493. 24 pp. 1967. (U.S. Dep. Agr. cooperating.) [Chiefly Carduus and Cirsium.] Mestre, J. C. Composées-Cynarées; développement de l’embryon chez le Carduus nutans L. a. Rend. Acad. Sci. Paris 245: 355-358. 1957. Moore, R. J., & G. A. MULLIGAN. Natural pepaeae between Carduus acanthoides and eee nutans in Ontario. Canad. Jou t. 34: 71-85. 1956 A study of natural ssa A es ae Carduus acanthoides and Carduus nutans. Proc. 10th Int. Congr. Genetics, Montreal, 1958. 2: 193. 1959 Morton, J. K. A cytological study of the Compositae (excluding Hieracium and Ta- raxacum) of the British Isles. Watsonia 11: 211-223. 1977. MULLIGAN, G. A., & C. FRANKTON. The plumeless thistles (Carduus spp.) in Canada. Canad. Field Nat. 68: 31-36. 1954 R. J. Moore. Natural selection among hybrids between Carduus acanthoides and C. nutans in Ontario. Canad. Jour. Bot. 39: 269-279. map. 1961. Muri, A., & M. VACHOVA. “In: Index of chromosome numbers of Slovakian flora. Part 6. Acta Fac. Rerum Nat. Univ. Comenianae Bot. 26: 1-42. 1978. 426 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 OuivierRI, 1. Comparative electrophoretic studies of Carduus pycnocephalus L., Carduus tenuiflorus Curt. (Asteraceae), and their hybrids. Am. Jour. Bot. 72: 715-718. 1985. , M. Swan, & P. H. Gouyon. Reproductive system and colonizing strategy of two species of ’ arduus (Compositae). staal 60: 114-117. oo RECHINGER, K. ‘arduus. In: K. H. RECHINGER, ed., FI. Iran. 139a: 218-230. 1979. Rees, N. E. Ta species ud milk thistle (C “arduus spp.) as hosts of feiss si CONICUS. Weed Sci. 34: 241-242. 1986. — _S. nes mosome numbers of some species of the family Asteraceae ort. (In Russian.) Bot. Zhur. Leningrad 64: 582-589. 1979. [Carduus crispus, i = i i SmiTH, L. M., Il, & L. T. Kox. Dispersal of musk thistle (Carduus nutans) seeds. Weed Sci. 32: Sos 1984. SmyTuH, C. A., & J. L. HAmrick. Variation in estimates of i in musk thistle, Carduus nutans, populations. Jour. Hered. 75: 303-307. VacHova, M. fn: Index of chromosome numbers of na oo Part 5. Acta Fac. Rerum Nat. Univ. Comenianae Bot. 25: 1-18. 197 VAN Loon, J. C., & H. DE JonG. Jn: IOPB chromosome number reports LIX. Taxo 27: 53-61. 1978. [Carduus nutans, 2n = 16; Centaurea maculosa, 2n = 18, Onan, arvense, 2n = 34, - Chromosome number reports LX VII. Taxon 29: 538-542. 1980. [Carduus Bi age = 16; C. acanthoides, 2n = 22. — & H.C. M. Snevpers. /n; IOPB chromosome number reports LXV. Taxon 28: 632-634. 1979. [C. nutans, oe = 16.] 3. Silybum Adanson, Fam. Pl. 2: 116, 605. 1763, nom. cons. Annual or biennial herbs. Stems 20-150 cm tall, simple or sparingly branched, robust, glabrous or slightly tomentose, lacking wings. Leaves near base of stem petiolate, upper cauline leaves sessile and auriculate-clasping; blades to 80 cm long and nearly 40 cm wide, broadly lanceolate to oblanceolate, lower leaves deeply but irregularly lobed, upper leaves entire or shallowly lobed, white veined or variegated, glabrescent to puberulent; margins spiny. Capitula 3-6 cm wide, homogamous, discoid, solitary at the ends of branches. Involucre hemispheric to ovoid; bracts imbricate, multiseriate, the outer ones with a broad, firm, spinulose-ciliate base and a coriaceous, subfoliose, strongly re- flexed, spine-tipped and basally spine-margined appendage, inner bracts lan- ceolate, erect and unappendaged. Receptacle flat to slightly convex, densely bristly. Flowers perfect. Corolla slightly zygomorphic, purplish; tube narrow, elongate, glabrous to puberulent, throat short, abruptly expanded, unequally divided by the sinuses of the long, linear, acute lobes. Staminal filaments glabrous, connate below; anthers shortly tailed at base, apical appendage firm, triangular, glabrous. Style smooth below a distinct collar of hairs at base of the elongate, papillose, connate branches. Achenes basifixed, obovoid-oblong, 6- 7 mm long, somewhat compressed, glabrous, black or brown with lighter streaks, with a distinct annular rim subtending the pappus. Pappus in several series, outer setae white, minutely barbellate, innermost bristles considerably shorter, capillaceous, basally connate and deciduous as unit. Base chromosome number 17. eee sPEcIES: Carduus Marianus L. = Silybum Marianum (L.) Gaertner, type s. (Name the Latin one for the milk-thistle, from the Greek name sillybos ¢ or silybon.)— MILK-THISTLE. 1990] SCOTT, GENERA OF CARDUEAE 427 A genus of two species, Si/vbum Marianum (Mariana Mariana (L.) Hill) and S. eburneum Coss. & Dur., both native to the Mediterranean area, the former now widespread in North and South America. Si/ybum Marianum, milk-thistle, 2” = 34, is naturalized locally in many parts of the United States, particularly in drier waste areas, along roadsides, and in disturbed pastures. In the Southeast it is rare and may be little more than a waif locally escaped from cultivation. Small (1933) reported it south to Alabama; Cory listed it west to Texas; yet it was not treated by either Correll & Johnston or Barkley). Johnson (1978) recorded it from Goochland County, Virginia; Thieret (1968) reported it as new to Louisiana, and E. B. Smith (1988) noted it as a local escape from cultivation in three widely separated localities in Arkansas. This species has become a common and troublesome weed in the pampas of Argentina (Mab- berly) and in areas with a Mediterranean climate along the Pacific coast of North America (Howell, Munz), where a seed weevil, Rhinocyllus conicus, (Hawkes et al., 1972) has been released to control it. Silvbum Marianum was treated as a species of Carduus by Linnaeus and was initially given generic status as VWariana by Hill. Shortly afterward Adanson proposed the now conserved name Si/ybum for the same taxon. Cassini grouped Silybum with Alfredia Cass. (= Carduus subgenus Alfredia (Cass.) Kazmi) and Echenais Cass. (treated as Carduus by Lessing, Cnicus by Bentham) as an unspecified division of the Carduinae. Lessing placed Si/ybum with Tyrimnus Cass. and Galactites Moench in an unranked division, Silybeae, of the subtribe Carduinae, on the basis of their monadelphous filaments, multiseriate pappus, and compressed, glabrous achenes. The same genera were recognized as subtribe Silybeae by De Candolle. A detailed study of seed anatomy and morphology led Dittrich (1970, 1977) to group Si/ybum informally with Carduus, Cirsium, and Modestia Charadze & Tamamshiam, while Galactites and Tyrimnus were aligned with Picnomon Adanson and Jurinea Cass. Silybum is distinguished from Galactites, Cirsium, and Modestia by its simple or barbellate (not plumose) pappus bristles. It differs from Tyrimnus and Car- duus in its uniseriate pappus and mottled cauline leaves that are markedly different from the reduced leaves that subtend individual heads. In addition, Silybum is distinguished from other genera of Cardueae of the Southeast by its connate staminal filaments. Chromosome counts of 2 = 34 have been reported for Sil/ybum Marianum (Goldblatt, 1981; Heiser & Whittaker; Larsen). Moore & Frankton (1962) found that the chromosomes of Si/ybum are somewhat shorter than those of Cirsium but are similar in width and appearance. They also noted that the thickness of the chromosomes and the total amount of chromatin in the nucleus are similar in species of Si/ybum, Carduus, Cirsium, Cnicus, and Onopordum. Wagner reported that flavanones and flavonols were restricted to Si/ybum, Centaurea, and Carthamus. Both species of Si/ybum were noted as exceptional for their production of flavonolignans, compounds known from only one other genus, Hydnocarpus (Flacourtiaceae). Flavonoids isolated from Si/ybum in- clude silybin, silydianin, silychristin, and silymarin (Wagner ef al., 1971; Wag- ner, 1977; Cappelletti & Caniato, 1983, 1984; Belikov, 1985; Hassan, 1985). The antihepatotoxic activity of the flavonolignans has been of considerable 428 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 interest for medicinal purposes (Hahn et a/.; Sharma & Singh; Schrall & Becker; Fiebig & Wagner; Hikino ef a/.; Wagner, 1986). Mabberley noted that the antihepatotoxic qualities of S. Marianum have been known since the time of Dioscorides and that they are now known to work as effective antidotes for Amanita poisoning, since they are able to displace phalloidin from membrane receptors. In addition, the alkaloids tyramine and histamine have been isolated from Si/yhum (see Wagner, 1977, for references). REFERENCES: Under tribal references see BARKLEY, 1986; Boprov & CZEREPANOV; CORRELL & JOHNSTON; CRONQUIST, 1980; DE Capone DittrRicH, 1970, 1977; GAERTNER; GOLDBLATT, 1981; HeGi, 1928-29, 1987; HEISER & WHITTAKER; HOWELL; LARSEN; LAVI- ALLE; LESSING; MABBERLEY; MACROBERTS; MOORE & FRANKTON, 1962, 1974; Poddub- naja-ARNOLDI; SINGH & PANDEY; SMALL, 1933; E. B. SMITH; WAGNER, 1977; and ZWGFLER, 1965 AustIN, M. P., R. H. Groves, L. M. F. Fresco, & P. E. KAve. Relative growth of 6 thistle species along a nutrient gradient with multispecies competition. Jour. Ecol. 73: 667-684. 1985. Betikov, V. V. Determination of the content of flavanol derivatives in the fruits of Silybum Marianum. (In Russian.) Rastit. Resur. 21: 350-358. 1985. Botpt, P. E., & L. T. Kok. Bibliography of Phinocyllus conicus Froel. (Coleoptera: Curculionidae), an introduced weevil for the biological control of Carduus and Silybum thistles. Bull. Entomol. Soc. Am. 28: 355-358. 1982.* CApPPELLETTI, E. M., & R. CANIATO. Chemical characterization of wild Italian popula- tions of Silvbum Marianum (L.) Gaertn. Plant-Med-Phytother. 17: 209-214. 1983. [Flavonoid content and the ratios of flavolignan isomers silybin, silydianin and silychristin investigated from populations in central and southern Italy. & . Silymarin localization in the fruit and seed of Si/yhum Marianum. Herba Hung. 23: 53-66. 1984. [Silymarin (2,4-dinitrophenyl-hydrazine) localized in pericarp, seed integument, and endosperm; 2, 4- DNPH localized in subepidermal and membraniform layers of pericarp. All layers of seed integument contain fla- volignans; 2,4 DNPH and flavolignans were not found in cotyledons as suggested by LANGHAMMER. Flavolignans may play an ecological role as allelopathic agents through flavolignan-induced [AA oxydase inhibition in other plants; also report on a subepidermic layer of thin-walled cells, some devoid of and others a with “‘a dark brown substance” responsible for the mottled appearance of achen Cory, V.L. Six thistles introduced into Texas. Madrofio 5: 200-202. 1940. (First report of Silybum Marianum in Texas.] Fiesic, M., Neue Sie wirksame Flavonolignane aus einer weissbliihenden Silybum-Varietat. Pl. Med. 51: 310-313. 1984. FRANCO, J. DO AMARAL. Si/ybum. In: T. G. saa et al., eds., Fl. Europaea 4: 126. 1976 tH > 7 Haun, G., H. D. LEHMANN, M. Kuerten, H. UEBEL, & G. VoGEL. Zur Pharmakologie und Toxikologie von Silymarin, des antihepatotoxischen Wirkprinzips aus Si/ybum Marianum (L.) Gaertner. Arzneim. Forsch. 18: 698. 1968.* Hassan, N. M., F. M. HAMMoupDé~, A. M. Rizk, H. Rimpcer, & S. I. Ismait. Isolation of silymarin from Silybum Marianum growing in Egypt. 33rd Annual Congress including the Business Section of the Society for Medicinal Plant Research, Re- gensburg, West Germany, Sept. 23-28. Acta Agron. Acad. Sci. Hung. 34(Suppl.): 107. 1985. 1990] SCOTT, GENERA OF CARDUEAE 429 Hawkes, R. B., L. A. ANpRes, & P. H. DUNN. Seed weevil released to control milk ctle. Calif, Agr. ee i“ 1972.* [Rhinocyllus conicus. | Hixtino, H., Y. Kiso, H. WAGNER, & M. Fiesic. Antihepatotoxic actions of flavonolig- nans from Silybum Vem fruits. Pl. Med. 50: 248-250. 1984 Jounson, M. F. Studies in the Virginia flora: species new to Virginia. (Abstr. ) ASB Bull. 25: 75. 1978. [Silyvbum Marianum in acne County, Virgini Kou, A. K., A. K. WAKHLU, & J. L. KArRIHALOO. Chromosome numbers of some flowering plants of Jammu (Western Himalayas) II. Chromosome Inf. Serv. 20: 32, 6 Koztowskl, J., & M. Hotynskxa. Effect of mineral fertilization on the crop yield of Silybum Marianum fruit and on the content and yield of silymarin. (In Polish.) Herba Polon. 31: 51-60. 1986. [Rate of growth tied to both atmospheric conditions and to NPK (nitrogen, phosphorus, potassium) fertilization, which were found to be independent of each Kupicua, F. K. Silybum. In: P. H. Davis, ed. Fl. Turkey 5: 369. LANGHAMMER, L. Anatomie und Histochemie der Friichte von aan Pl. Med. 17: 268-275. 1969. [Chemical composition of cell walls and storage substances in S. Marianum.] Lotter, H., & H. WAGNER. The stereochemistry of silybin (Si/ybum Marianum). Zeit. Naturforsch. C. Biosci. 38: 339-341. 3: Mericu, A. H. Si/ybum Marianum in Turkey. Istanbul Univ. Eczacilik Fak. Mecm. Pook, E. W. The effect of shade on the growth of variegated thistle (Si/vbum Marianum L.) and cotton thistle (Onopordum sp.). Weed Res. 23: 11-17. 1983. SCHRALL, R., & H. Becker. Callus und Suspensionckulturen von Silybum Marianum. II. Mitteilung: Umsetzung von Flavonoiden mit Coniferyalkol zu Flavonolignanen. (English abstr.) Pl. Med. 32: 27-32. 1977. [Isolation of silybin from tissue culture.] SHARMA, B. M., & P. SINGH. Pharmacognostic study of seeds of Silybum Marianum. Pl. Sci. Lucknow 6: 34-37. 1974. SINGH, J. P., B. K. Kapaui, & Y. K. Sarin. Ecology of Si/ybum Marianum Gaertn. Jour. ae Taxon. Bot. 3: 665-668. 1982. Spirova, I. Germination dynamics and rate of germination of milk thistle (Si/ybum Marianum (L.) Gaertn.) seeds. UVTIZ (Ustav Vedeckotech Inf Zahradnictivi) Praha 11: 322-329. 1984. THIERET, J. W. Additions to the vascular flora of ey Proc. Louisiana Acad. Sci. 31: 91-97. 1968. [Sijvbum Marianum, among other Wacner, H. Antihepatotoxic flavonoids. /n: V. aa E. Mipp.eton, Jr., & J. B. HARBORNE, eds., Plant flavonoids in biology and medicine: biochemical, pharma- cological and structure-activity relationships: proceedings. Buffalo, New York, July 2-26, 1985. New York. 1986. —.,, O. SELIGMANN, L. H6RHAMMER, M. Seitz, & J. SONNENBICHLER. Zur Struktur von Silychristin, einem zweiten Silymarin- eee aus Silybum Marianum. Tet- rahedron Lett. 22: 1895. 1971. 4. Onopordum Linnaeus, Sp. Pl. 2: 827. 1753; Gen. Pl. ed. 5. 359. 1754. Biennial herbs. Stems stout [very short], fistulose, much branched, spiny winged, densely tomentose, rarely glandular. Leaves near base of stem petiolate, middle to upper cauline leaves sessile, semi-auriculate, clasping; blades lan- ceolate to oblanceolate, toothed to pinnatifid [pinnatisect], glabrescent to arach- noid-tomentose; margins spiny. Capitula homogamous, discoid, terminal or axillary, solitary to corymbosely arranged, peduncles usually prickly winged. 430 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Involucre hemispherical to globose [glabrescent or] arachnoid-tomentose; bracts imbricate in several series, narrowly lanceolate [oblanceolate], coriaceous, spine tipped, the margins scabrous; outer bracts commonly triangular, erect to re- curved, adpressed to lax; inner bracts narrowly lanceolate, glabrescent. Recep- tacle flat to slightly convex, glabrous, fleshy, alveolate with short, basally fused scales along the margins of the pits, but never densely bristly. Flowers perfect, homogamous, actinomorphic to slightly zygomorphic. Corollas lilac to pur- plish, glabrous or with scattered sessile glands; tube slender, elongate, throat short, slightly expanded, actinomorphic or saccate and unequally divided by the sinuses of the linear, elongate lobes. Staminal filaments glabrous, attached mi appendages often extending beyond corolla lobes. Style smooth or rugose below a distinct collar of hairs at base of the papillate, connate to slightly divergent branches. Achenes basifixed, 4-5 mm long, tetragonal, commonly laterally compressed, brown, rugulose, glabrous. Pappus bristles multiseriate, unequal, capillary or barbellate [plumose], often flattened, basally connate, deciduous as unit. Base chromosome number 17. LECTOTYPE SPECIES: Onopordum Acan- thium L.; see Britton & Brown, Illus. Fl. No. U. S. Canada 3: 556. 1913; also, Hitchcock & Green, Nom. Prop. Brit. Bot. 179. 1929; Jackson, Index Linn. Herb. 109. 1912. (Name Latinized from the ancient Greek name of the plant, Onopordon, from onos, donkey, and, porde, flatulence. Pliny stated that it caused flatulence in donkeys.) —SCOTCH THISTLE, COTTON THISTLE. A genus of about 40 species of Mediterranean Europe and Western Asia. Onopordum Acanthium is naturalized in various parts of the world. Onopordum is distinguished from other thistles by the absence of bristles or paleae on the receptacle. The genus appears to be closely related to the large genus Cousinia Cass. and more remotely to Jurinea Cass. and Saussurea DC. Onopordum Acanthiumis often found escaped from cultivation, established along roadsides, in dry, waste areas, and in disturbed pastures. It is the only species of this genus reported from the Southeast (Small, 1933; Cronquist, 1980), but it does not appear in the Flora of the Carolinas (Radford, et al.). McGregor reported O. Acanthium as a serious weed of infrequent, localized occurrence and often found in feedlots; Johnson (1974) noted only two collections from Virginia. Among Onopordum species, it can be distinguished by its large seeds and numerous slender, reflexed, involucral bracts (+ 2 mm wide at the base). Nearly 100 years have passed since the last revisionary treatment of the genus was published by Rouy, who recognized 24 species in three sections (Acaulia Rouy, Erecta Rouy, Reflexa Rouy) based on characters of habit, involucral bracts, and pappus. Rouy placed Onopordum Acanthium in sect. Reflexa (now ONOPORDUM), which was characterized by erect, winged stems, recurved involucral bracts, and scabrous or semi-plumose pappus bristles. Franco established the monotypic subgenus ACAULON and recognized four sections (ONOPORDUM, ERECTA, ECHINATA Franco, RECURVATA Franco) under subg. ONOPORDUM. The monotypic sect. ONOPORDUM was differentiated by its lanate or tomentose pubescence of unicellular hairs; lack of reticulate veins on 1990] SCOTT, GENERA OF CARDUEAE 431 the lower side of the leaves; linear-subulate, erecto-patent involucral bracts shorter than the flowers; eglandular corolla lobes; and scabrid pappus hairs. A base chromosome number of 17 was indicated by Moore & Frankton’s (1962) report of the chromosome number 2n = 34 for O. Acanthium (see also Poddubnaja-Arnoldi and Morton, among others), O. illyricum L., and O. Sib- thorpianum Boiss. & Heldr. All subsequent chromosome counts for species of Onopordum have been 2n = 34 or n = 17 (see Moore, 1973; Goldblatt, 1981, 1984, 1985, for references). Moore & Frankton (1962) noted that the chro- mosome number and morphology of Onopordum are indistinguishable from those of Cirsium. Relatively few phytochemical studies have been undertaken on Onopordum. The phenolic compound cynarin was reported from O. i//yricum, and the al- kaloids choline and stachydrine were reported from O. Acanthium (see Wagner for references). The sesquiterpene lactone onopordopicrin has been isolated from O. Acanthium, O. leptolepis DC., O. tauricum Willd., and two costunolide and melitensin sesquiterpene lactones were reported from QO. /eptolepis DC. (see Seaman for references). Rustaiyan et al. have isolated onopordopicrin and two related esters, eudesmanolides, and eudesmane derivatives from O. car- manicum (Bornm.) Bornm. Rustaiyan ef al. reported that the 15-hydroxyl germacranolide with an 8-alpha-acyloxy group isolated from O. carmanicum is known from Jurinea (Hermout & Sorm) and that similar sesquiterpene lactones have been reported from Centaurea, Arctium, and Cnicus species, as well as from Dicoma Cass., of the Mutisieae. REFERENCES: Under tribal references see Boprov & CZEREPANOV; BOISSIER; CRONQUIST, 1980; Dittricu, 1970, 1977; FERNALD; GLEASON & CRONQUIST; GOLDBLATT, 1981, 1984, 1985; Go.tz: Heat, 1928-29, 1987; HerMout & SorM; HOWELL; JOHNSON, 1974; LAVIALLE; MABBERLEY; MCGREGOR; Moore; Moore & FRANKTON, 1962; MORTON; PODDUBNA- JA-ARNOLDI; RICKETT; SEAMAN; SINGH & PANDEY; SMALL, 1933; WAGNER; and ZWGFLER, 965 Borsster, E. Onopordon. Fl. Orient. 3: 558-564. 1875. Cory, V. L. Six thistles introduced into Texas. Madrono 5: 200-202. 1940. [First report of O. Acanthium in Danin, A. Onopordum. In: P. H. Davis, ed., Fl. Turkey 5: 356-369. 1975. Dress, W. J. Notes on the cultivated Cae 9: Onopordum. Baileya 14: 75-86. 1966. c, A. Revision of the Onopordon species of Palestine, Syria and adjacent countries. Palest. Jour. Bot. Jerusalem Ser. 2(4): 185-199, 1942.* ForceLLa, F., & J. T. Woop. Colonization potentials of alien weeds are related to their native distributions: Implications for plant quarantine. Jour. Austral. Inst. Agr. Sci. 50: 35-41. 1984. [Comparative study of distributions of weedy species of Carduus, Centaurea, and Onopordon in Australia and in their native habitats. ] FRANCO, J. DO AMARAL. Onopordum. In: T. G. Tutin et al., eds., Fl. Europea 4: 244— 248. 1976. GONZALEZ-COLLADO, I., F. A. MActAs, G. M. MASSANET, J. M. Oxtva, F. L. RODRIGUEZ, & C. VERGARA. Chemical constituents of Onopordum nervosum eke (In Spanish; English summary.) Ann. Quim. C. Quim. Org. Boiquim. 80: 100, 101. 1984.* 432 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 MARTICORENA, C., & M. QUEZADA. oa eae anew record toh COmposiiae (Cy eile in Chile. Bol. Soc. Biol. Con : 301, 302. 1977 RECHINGER, K. H. Onopordum. In: K. H. ss saan ed., Fl. Iran. 139a: 156-164. 1979. Rouy, G. Revision du genre Onopordon. Bull. Soc. Bot. France 43: 577-599. 1896. Rustalyvan, A., B. AHMADI, J. JAKUPovic, & F. BOHLMANN. Sesquiterpene lactones and Laue etic from Onopordon carmanicum. Phytochemistry 25: 1659- 1662. 5. Arctium Linnaeus, Sp. Pl. 2: 816. 1753; Gen. Pl. ed. 5. 357. 1754. Biennial herbs. Stems erect, striate, branching, glabrous to less often arach- noid-hairy; taproots stout, elongate. Leaves petiolate, petioles solid or fistulose; blades ovate to ovate-oblong, obtuse to cordate based, the upper surface with scattered sessile glands, the lower surface commonly arachnoid-hairy and glan- dular, the glands numerous, sessile; blade margins entire to serrate or rarely laciniate, often remotely dentate. Capitula homogamous, discoid, ca. 2-5 cm wide, borne in several- to many-headed racemes or cymes. Involucre subglo- bose, green [rarely purple], glabrous [or with arachnoid indumentum], bracts multiseriate, imbricate, narrowly lanceolate, subequal, coriaceous, basally ap- pressed and slightly keeled, apices subulate, inwardly hooked, and on the outer bracts strongly reflexed. Receptacle flat, with numerous, white, subulate, sca- brous scales. Flowers perfect, actinomorphic, corollas lavender to purple (rarely white), glabrous, tube narrow, equal to or slightly longer than the slightly expanded throat and acute to acuminate, elongate, glabrous lobes. Staminal filaments glabrous, anthers stramineous, sagittate, apical appendages with mu- cronate tips. Style slightly swollen at base, smooth below an indistinct collar of hairs found at the base of papillate, cuneate branches. Achenes basally attached, 4-7 mm long, oblong, somewhat compressed, apically truncate, com- monly rugose, glabrous, many nerved, light to dark brown, often mottled. Pappus bristles numerous, unequal, relatively short (ca. 2 mm long), barbellate, individually and readily deciduous. LEcroTyPeE SPECIES: Arctium Lappa L.; see Britton & Brown, Illus. Fl. No. U. S. Canada 3: 547. 1913; also see Hitchcock & Green, Nom. Prop. Brit. Bot. 179. 1929. (Name from Greek, arction, a plant name from arctos, a bear, because of the rough involucre; cf. Munz. URDOCK. A Eurasian genus of about ten species, some of which are widely naturalized in southern Europe and Asia Minor, China(?), Japan, and North America. Three introduced species, 4. Lappa L., great burdock; 4. minus (Hill) Bernh., common burdock; and 4. tomentosum Miller, woolly burdock, are known from the southeastern United Perring recognized five species in Europe and, in the following quote, clearly stated the difficulties encountered in the taxonomy of the genus. “Specific limits within this genus cannot be clearly defined, each species showing great variation in hairiness of leaves and capitula, length of peduncles and colour of capitula and florets. All taxa are interfertile and although they are normally autogamous, outbreeding sometimes occurs. This has cies in innumerable intermediates which are fully fertile and breed true from s Four species (4rctium Lappa, A. nemorosum a tecime & Courtois, 4. fo- 1990] SCOTT, GENERA OF CARDUEAE 433 mentosum, and A. minus) based on characteristics of the achenes, corollas, involucral bracts, leaves and arrangement of the inflorescence were recognized by Fernald & Wiegand, but they were uncertain of the role hybridization had played among the taxa found in America. Arénes (1950) recognized four species ppa, A. minus, A. tomentosum, and A. Chabertii) and two sections, Felandulosa (i.e., ARCTIUM including 4. Lappa and A. minus) and GLANDULOSA. Each species was subdivided into various taxa. In addition, Arénes (1950) recognized nine interspecific hybrids involving A. Lappa or A. minus, two of which (4. x mixtum Nym., A. x nothum (Ruhm.) Weiss) were said to occur in North America (Moore & Frankton, 1974; Gross et a/.). The division between glandular and eglandular taxa set forth by Arénes was considered to be a seemingly natural separation by Moore & Frankton. Cronquist (1980) notes that A. nemorosum may also be of hybrid origin. The three species in the Southeast are found in waste places, along roadsides and streambanks, and in overgrazed pastures. As Gross et a/. noted, the biennial habit of these species confines them to areas not cultivated or otherwise dis- turbed on an annual basis. Wiegand & Eames found that 4A. minus and A. Lappa are common in both rich soil and clay. Rollo et al. reported a strong overlap in distribution between 4. minus and A. Lappa in their study area. All three are common in the Northeast, but only 4. minus occurs with any regularity south of the Carolinas. Throughout its range in the Southeast it is more likely to be found in the mountains than in the Piedmont. Mulligan & Kevan described Arctium minus as an autogamous biennial weed that is frequently visited by insects, despite the lack of any noticeable odor. These findings corroborated the work of Mulligan & Findlay, who noted that A. minus set viable seed even when covered with a “pollination bag.” Gross & Werner reported that 4. minus, which reproduces only by seed, will not flower in the second year if conditions are poor. Gross & Werner also deter- mined that the probabilities of a rosette’s flowering in a given year are linked to rosette size, but that rosette size is not linked to age. Total achene (hence seed) production was shown to be correlated with either rosette size or the height of the flowering stem, yet differences in total fruit production were shown to be a function of the number of capitula on a plant. No relationship was found between rosette size and the number of fruits in a capitulum (Gross & Werner). Hawthorn & Hayne discovered that the mean number of fruits per plant in A. minus (13,400) was significantly greater than in A. Lappa (8200), as were the mean numbers of heads per plant (338 versus 112, respectively). Predis- persal seed predation by the microlepidopteran Metzneria lappella accounted for the loss of 28 to 71 percent of the achenes of A. minus but only 15.5 percent of those of A. Lappa. Hawthorn & Hayne also found that only about one fifth of the heads of 4. minus were shed during the winter compared to 62 percent of the heads of 4. Lappa. This pattern of herbivory, along with the observation that very few achenes are dispersed from heads of A. minus because of the contraction of the involucre led Hawthorn & Hayne to conclude that A. Lappa is more efficient than A. minus in seed dispersal. Rollo et al. observed four biotypes of Arctium present in their study popu- 434 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 lations. These biotypes, including the two parental species (4. minus and A. Lappa), were differentiated on the basis of size of capitula, length of peduncles, stem color, involucral characters, degree of insect herbivory, and allometric measurements. The relative abundance of these biotypes changed yearly. A “scenario” involving differential insect predation, hybrid inferiority, and dif- ferential dispersal was suggested to explain the dynamics of burdock popula- tions. In accord with the study by Hawthorn & Hayne, Rollo et a/. noted that the capitula of A. minus remain tightly closed allowing few achenes to fall out, while the capitula of 4. Lappa open widely, allowing many achenes to fall beneath the parent plant. In addition, Rollo et a/. observed that A. minus is more likely to colonize new, disturbed habitats than is 4. Lappa, suggesting that this was due to the tightly closed capitula of 4. minus that are not so readily deciduous as those of 4. Lappa. They concluded that the capitula of A. minus are more suited to dispersal by animals to which they are readily attached by their hooked involucral bracts. Noting common burdock’s ex- tremely high seed mortality from the moth Metzneria lappella, Rollo et al. characterized A. minus as a fugitive species that depends on sis distance dispersal by animals to escape moth predation in both time and sp On the basis of counts of nm = 18 (Nakajima), 2n = 36 (Tarnavschi) and 2n = 32 (Sugiura) for A. Lappa; 2n = 36 (Léve & Live, 1944; Tarnavschi) for A. nemorosum, 2n = 36 (Tarnavschi) and 2” = 32 (Wulff) for A. minus; and 2n = 36 for A. tomentosum (Poddubnaja-Arnoldi; Tarnavschi), Love & Love (1961) considered Arctium to have a basic chromosome number of nine. Mul- ligan has since reported 2 = 36 for A. minus. Despite many claims of hybridization between species of Arctium (Arénes, 1950; Moore & Frankton, 1974; Weigand & Eames), only one study has directly addressed this subject. Rollo et a/. studied the allometric and electrophoretic variation between four biotypes representing two parental types (4. minus, A. Lappa) and two putative hybrids. They concluded that three of the four biotypes could be clearly distinguished by allometric measurements. Electrophoretic patterns showed unique isozymes in both parental biotypes which also differed from one another to the greatest extent allometrically. The putative hybrid biotypes were found to share various isozymes from the two parental types, thus supporting the hypothesis that these biotypes were hybrids. The sesquiterpene lactone arctiopicrin was found in Arctium minus, A. Lap- pa, A. nemorosum, and A. tomentosum and onopordopicrin was isolated from A. minus (see Seaman for sources). In addition, the triterpenes 6-cudesmol, arctiol, eremophilene petsitolone, and fucinanolide were reported for A. Lappa (see Wagner, 1977, for sources). Wagner (1977) reported the flavonoids iso- quercitin, astragalin, quercetin 3-O-arabinoside, quercimeritrin, rutin, and kaempferol 3-O-rhamnoglucoside from A. minus. Alkaloids of unknown struc- ture have also been reported for 4. Lappa and three other unnamed species of Arctium (Wagner, 1977). Some of the many economic uses of Arctium that have been well documented by Moore & Frankton (1974) include root extracts used as laxatives, diuretics, sudorifics, blood purifiers, and hair restorers, as well as constituents in ““wood tea” and as an emergency coffee substitute (Hegi, 1929; Youngken, 1948). 1990] SCOTT, GENERA OF CARDUEAE 435 Moore & Frankton also cited Rousseau & Raymond’s report that the young leaves and dried roots were eaten by the American Indians until fairly recently, and Yanovsky’s record that A. Lappa was eaten as greens and the roots were cooked for soup and dried and stored for winter food. Arctium Lappa is cul- tivated for its edible roots (gobo) in Japan, and the young leaves are used in salads in both Japan and Scandinavia (Mabberley). The hooked involucral bracts of Arctium were the inspiration for the now widely used “Velcro” fabric closures REFERENCES: Under tribal references see CRONQuIST, 1980; Dirrricu, 1970, 1977; GAERTNER; GOLDBLATT, 1981, 1984, 1985; Heai, 1928-29, 1987; JoHNSON, 1975; LAVIALLE; LESSING; LINNAEUS; LOvE & LOve, 1944, 1961; MABBERLEY; Moore & FRANKTON, 1974; PODDUBNAJA-ARNOLDI; RADFORD ef ai.; Rickert: SEAMAN; SHISHKIN & BOBROV; SMALL, 1933; WAGNER, 1977; and WEIGAND & EAMES. ARENES, J. Monographie du ane es : ie Jie Bot. Bruxelles 20: 67-156. 1950. [Four species in two sections (4. a & A. minus in sect. Eglandulosa; A, oe us & A. Chabertii in ae pera all divided into various subspecific DE ee J. E. Les bardanes (genre Arctium L.) de Belgique et des régions voisines. Nat. Belges 47: 21-29. 1966. FERNALD, M. L., & K. M. WIEGAND. A synopsis of the species of Arctium in North America. Rhodora 12: 43-47. 1910. [4. Lappa, A. tomentosum, A. nemorosum, and A. minus, representative specimens cited.] Gross, R. S., & P. A. WERNER. Probabilities of survival and reproduction relative to rosette size in the common burdock (Arctium minus, Compositae). Am. Midl. Nat. 109: 184-193. 1983. . WERNER, & W.R. HAWTHORN. The biology of Canadian weeds 38. Arctium minus and Arctium Lappa. Canad. Jour. Pl. Sci. 60: 621-634. 1980. HawtTuHorn, W. R., & P. D. Hayne. Seed production and predispersal seed predation in the biennial Composite species, Arctium minus (Hill) Bernh. and A. Lappa L. Oecologia 34: 285-295. 1978. KupicHa, F. K. Arctium. In: P. H. Davis, ed., Fl. Turkey 5: 354-356, 1975. Morita, K., Y. NIsHIIMA, & J. KADA. Chemical nature of a desmutagenic factor from burdock (4rctium Lappa). Agr. Biol. Chem. 49; 225 932. 1985. MULLIGAN, G. A., & J. N. FINDLAY. d colonization in Canadian weeds. Canad. Jour. Bot. 48: 859- 860. 1970. & P. G. . Color, brightness, and other floral Seas attracting insects to the blossoms of some Canadian weeds. /bid. 51: 1939-1952. NAKAJIMA, G. Chromosome numbers in some Heaps and wild angiosperms. aie Jour. Genet. 12: 211-218. 1936. [A. Lappa, n = 18.] Naya, K., K. Tsuyz, & U. Haku. Conviexk of Arctium Lappa. Chem. Lett. 3: 235 ff. 1972.* ParrisH, M. D., & J. M. Barirp. Predispersal seed predation of common burdock (Arctium minus) by Abruchid coleoptera. Bull. New Jersey Acad. Sci. 27: 26. 1982. PERRING, F. H. Arctium. In: T. G. Tutin et al., eds., Fl. Europaea 4: 215. 1976. RECHINGER, K. H. Arctium. In: K. H. RECHINGER, ed., Fl. Iran. 139a: 105-107. 1979. Reep, F. C., & S. N. STEPHENSON. Factors affecting seed number and size in burdock, Arctium minus. Proc. Michigan Acad. Sci. 5: 449-455. 1973.* Roserts, H. A., & J. E. Nettson. Seed survival and periodicity of seedling emergence in 12 weedy species of Compositae. Ann. Appl. Biol. 97: 325-334. 1981. 436 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Rotio, C. D., J. D. sy peespaeie es B. 5. SMITH. Electrophoretic and allometric vari- ation in burdock (Arctium s its ecological implications. Canad. Jour. Bot. 63: 1255-1261. an Romincer, J. M., J. M. Ricketson, & W. HopGson. Distribution of common burdock, Arctium minus Bernhardi (Compositae), in Arizona. Jour. Arizona Nevada Acad. Sci. 20: 71. 1985.* Rousseau, J., & M. RAYMOND. Etudes ethnobotaniques québécoises. Contr. Inst. Bot. Univ. Montréal 55. 1945. RostTovtsevA, T. S. Chromosome numbers of some species of the family Asteraceae 2. Biol. Zhur. Leningrad 68: 660-664. 1983. [Arctium Lappa L., 2n = 36. SALEH, N. A. M., & B. A. Boum. Flavonoids of Arctium minus (Compositae). Experi- mentia 27: 1494. 1971. SuciuRA, T. A list of chromosome numbers in angiospermous plants. Bot. Mag. Tokyo 45: 353-355. 1931. . Studies on the chromosome numbers in higher plants, with special references o cytokinesis. I. Cytologia 7: 544-595. 1936. i. I. T. Die Chromosomenzahlen der Anthophyton-Flora von Ruménien mit einem Ausblick auf das Polyploidie-Problem. Bull. Jard. Mus. Bot. Univ. Cluj 28(Suppl.): 1-130. 1948. WASHINO, T., M. YOSHIKURA, & S. OBATA. New sulfur from Arctium Lappa. Agr. Biol. Chem. Tokyo 50: 263- 269. 1985.* Wu trr, H. D. Chromosomenstudien an der schleswig-holsteinischen Angiospermen- flora. I. Ber. Deutsch. Bot. Ges. 55: 262-269. 1937.* YANOVSKY, E. Food yo of the North American Indians. U.S. Dep. Agr. Misc. Publ. 237. 84 pp. 1936.* YOUNGKEN, H. W. A textbook of pharmacognosy. Ed. 6. 1063 pp. Philadelphia. 1948. tu] ‘ d Subtribe CENTAUREINAE Dumortier, Fl. Belg. Prodr. 72. 1827. 6. Centaurea Linnaeus, Sp. Pl. 2: 909. 1753; Gen. Pl. ed. 5. 389. 1754. Annual to perennial herbs [rarely dwarf shrubs with spiny branches or larger evergreen shrubs]. Stems tomentose or scabrous to hirsute with multicellular hairs, rarely glabrous, commonly with sessile glands. Leaves sessile to (less often) petiolate, alternate or all basal; blades narrowly lanceolate to ovate, entire to pinnatifid; margins scarcely or not at all prickly. Capitula heterogamous, discoid or falsely radiate, hemispherical to subglobose, terminal or axillary, commonly solitary or in loose cymose clusters, subsessile to long pedunculate. Involucre ovoid, subglobose [nearly cylindrical, oblong or fusiform], commonly contracted apically, bracts imbricate in several series, often striate, glabrous to tomentose, either shortly spine tipped or more commonly with an orbicular, lanceolate or triangular, scarious or hyaline, erose to pectinate appendage, or occasionally the appendage absent and the apical margin merely pectinate or lacerate. Receptacle flat to slightly convex, with numerous smooth, white bris- tles. Florets perfect or sterile. Corollas lavender, blue, yellow, or white; marginal florets sterile [perfect], occasionally with an irregularly expanded, falsely radiate corolla; inner florets perfect, discoid, actinomorphic, tube often elongate, slen- der, throat expanded, lobes long, linear to lanceolate. Staminal filaments gla- brous; anthers stramineous or commonly the same color as corolla, tailed at base, though often shortly, the apical appendage narrow, acute. Styles smooth 1990] SCOTT, GENERA OF CARDUEAE 437 below a distinct collar of hairs at base of the short, slightly divergent, narrowly truncate to acute, papillate branches. Achenes obliquely or laterally attached to receptacle, oblong or ovoid, compressed or obtusely 4-angled, seldom evi- dently nerved, glabrous at maturity. Pappus of several series of unequal sca- brous or barbellate [plumose] bristles or narrow scales, often much reduced, or absent. LECTOTYPE SPECIES: Centaurea Centaurium L.; see Britton & Brown, Illus. Fl. No. U. S. Canada, ed. 2. 3: 556. 1913; also, R. McVaugh, Fl. Novo- Galiciana, 209. 1982. (Name from Greek, of the centaurs, who were said to have used it in healing.)— KNAPWEED, STAR-THISTLE. A genus first proposed by Linnaeus for 50 species, Centaurea currently is comprised of some 450 (Mabberley) to 600 species (Holub, 1972a; Wagenitz, 1985a). Various members of this genus have become widespread throughout the world, but species diversity is concentrated in the Mediterranean and Near Eastern regions. Dostal (1976) included 221 species in his treatment of Cen- taurea for the Flora Europaea, and Wagenitz recognized 172 in the Flora of Turkey (1975) and 89 in Flora Iranica (1980). Wagenitz (198 5a) indicated that 260 species in 38 sections are known from Southwest Asia, with the greatest numbers in southern and eastern Turkey and adjacent Iran and Iraq. In this area, there are several widespread species, as well as 201 endemics, many of which occupy restricted areas. The large number of endemics and the coherent distribution of most groups (often delimited as subgenera, sections or, in some cases, genera) led Wagenitz (1985a) to conclude that speciation has occurred in relatively recent times and /n situ. Further, Wagenitz believed that the diversity of the group can only be understood by taking into account the capitulum which, according to both Wagenitz and Burtt, functions as a single flower and its fruit might in other families. To illustrate this point, Wagenitz noted the wide range in the number of flowers per capitulum in different species (4 or 5 in C. virgata Lam. to over 200 in C. macrocephala (Muss.) Puschk. ex Willd.), as well as several forms of dispersal: a well-developed pappus as a means of wind dispersal (anemochory),; small, deciduous capitula with hooked appendages suited for animal dispersal of whole capitula (epizoochory); and an elaisome that mediates dispersal by ants (myrmecochory). (See also Dittrich, 1968a, b.) Only two species are indigenous in the United States: Centaurea americana Nutt., which enters the Southeast in western and southwestern Arkansas (E. B. Smith, 1988) and western Louisiana, and C. Rothrockii Greenman, a species of the Great Plains. However, nearly twenty other species have become estab- lished in parts of the United States, some as pernicious weeds. Species of Centaurea in the southeastern United States are most often found in disturbed fields, waste places, and along roadsides. Cronquist (1980) recog- nized eight species in the Southeast, but several taxa (C. dubia Suter., C. nigra L., C. Jacea L., C. Calcitrapa L.) that he treated occur, ifat all, only sporadically along the northern boundaries of this region. Three of these species (C. dubia, C. Jacea, C. nigra) are perennials that are widespread in the Northeast. They are distinguished from other taxa of Centaurea in the Southeast and from each other by leaf characters and subtle distinctions of the involucral bract append- ages (Cronquist, 1980, among others). 438 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 The dubious occurrence of various taxa is reflected in Johnson’s inclusion (see distribution maps in Harvill et al.) of C. Calcitrapa, C. dubia, C. nigrescens Willd., and C. x pratensis Thuill. (C. Jacea x C. nigra), which are not treated by Ahles in Radford et a/. Small (1933) considered only C. solstitialis L., C. melitensis L., and C. Calcitrapa to be naturalized and C. Cyanus L. to have escaped from cultivation in the Southeast. Centaurea Cyanus, bachelor’s-button or cornflower, is a widely cultivated winter annual that is periodically found to be well established in fields and waste places and along roads. It is distinguished by its narrow, nearly linear leaves and heads with an outer series of elongate, zygomorphic flowers, the corollas of which are usually bright blue (or occasionally pink or white). Centaurea maculosa Lam., spotted knapweed, a short-lived perennial that has become one of the most serious weeds of western rangelands (Boggs & Story), is commonly encountered in old fields and pastures and along roadsides in the Southeast. Although similar to C. Cyanus in having an outer series of elongate, expanded florets, C. maculosa differs in its pinnatifid leaves and lavender flowers As noted by Cronquist (1980), C. melitensis, a yellow-flowered European species similar in many respects to C. solstitialis (also yellow flowered), is found only occasionally in South Carolina and Georgia. Centaurea solstitialis, a wide- spread and pernicious weed in several western states, is not well established in the Southeast. Centaurea Calcitrapa, another uncommon introduction to this region, can be differentiated from C. solstitialis by its purple florets, wingless stems, and epappose achenes. Both of these species are readily differentiated from others of Centaurea by their strongly spiny involucral bracts. Centaurea americana, an annual plant common in grasslands in the Great Plains and most readily distinguished by its large, unarmed heads and well- developed pappus, enters the Southeast along the eastern border of its range which extends from Missouri southward Se western Arkansas and the border of Louisiana with Texas, as noted abov From the time of Cassini to the present, the question of generic boundaries in the subtribe Centaureinae and related sul within Centaurea sensu lato has generated considerable controversy. The ‘rather lengthy account that follows outlines a part of the difficulty and complexity of the group Early in his serial publications on the Compositae in the Dictionnaiie des Sciences Naturelles, Cassini (1817) recognized the tribes Cardueae (as “‘Car- duacées’’) and Centaureeae (as ‘‘Centauriées’’), both names invalidly published, to which he attributed 17 and 11 genera, respectively. Later (1819), he validated these and other tribes with Latin names. By 1830, when he completed his treatment for publication, Cassini had come to recognize in the Centaureeae 41 genera, nearly 30 of which he described himself. He aligned the Centaureeae with the Cardueae and Carlineae, but considered the possibility of recognizing the Cardueae and Centaureeae as subtribes of a single tribe, Cardueae. Initially, Cassini had differentiated Centaureeae from Cardueae by characteristics of the anther appendages (apically fused and curved, not straight as in the Cardueae), achene, and pappus. He noted that while both the Centaureeae and Cardueae have laterally compressed, obovate achenes with four more or less pronounced 1990] SCOTT, GENERA OF CARDUEAE 439 sides or ribs, the basal areole (“‘l’aréole basilaire’’) or point of attachment of the achene of the Cardueae is sessile, broad (“‘large’’), smooth (“‘plane’’), curved (‘‘arrondie’’), slightly oblique (‘un peu oblique-antérieure”’), and lacking a rim or margin. The Centaureeae, on the other hand, were reported to have sessile, very oblique basal areoles strongly adherent to the receptacle (“la substance du clinanthe”’) and situated in a broad, diamond-shaped notch with curvilinear sides (“‘une large échancrure en losange, a bords curvilignes’’). As Cassini observed, the pappus of the Centaureeae is considered to be double or two ranked. The outer rank is comprised of several series of imbricate squamellae, the outermost of which are extremely short, flattened, obtuse, and barbellate. There is a gradual transition from the outer squamellae to an inner series of longer, cylindrical, bristle-like setae. A single series of very short, membranaceous, truncate squamellae constitutes the innermost rank of the double pappus. The pappus of the Cardueae is comprised of numerous series of unequal barbellate setae; the outermost series is composed of very short, filiform bristles, while the inner series is of barbellate setae flattened at the later descriptions, Cassini (1830) noted the outwardly curved corollas and outer series of sterile flowers in the Centaureeae (absent in the Cardueae) and the unique zygomorphic nature of the corollas in the Cardueae as differentiating characters. These distinctions, which have been the foundation for many sub- sequent classifications, were used by Dittrich (1977). Lessing placed in synonymy under Centaurea nearly 30 genera that Cassini had described. Until recently, Lessing’s concept of the tribe and of Centaurea has been followed by most workers. Spach’s tribal arrangement, patterned on Cassini’s, included many of the characters Cassini used to delimit the tribes. Spach’s proposed classification of Centaurea included 27 subgenera, six of which (Chartolepis Cass., Jacea Miller, Cyanus (Miller) Pers., Lopholoma Cass., Hymenocentron Cass., Calcitrapa (Heist ex Fabr.) Pers.) he treated in detail. Although he utilized suites of char- acters from the flowering capitula in his descriptions, differences in the invo- lucral bracts were stressed in his segregation of these six taxa. Bentham followed Lessing’s tribal treatment in many respects but returned the Arctoteae and Calenduleae to tribal status. The subtribe Centaureinae comprised ten genera in Bentham’s treatment. Centaurea was divided into 22 series (““vix Sectiones’’) most of which were based on Cassini’s genera. The first infrageneric treatment for Centaurea was put forth by A. P. de Candolle in the third edition of Lamarck & De Candolle’s Flore Francaise (1805). Six sections were proposed by De Candolle (see Brizicky) as reductions of Jussieu’s generic concepts: Centaurea (involucral bracts entire, foliaceous, not spinose), Rhaponticum (Ludwig) DC. (involucral scales scarious, not ciliate or spinose), Cyanus (Miller) DC. (involucral scales ciliate, not spinose), Seridia (Juss.) DC. (involucral scales terminated with numerous prickles disposed like fingers of the hand), Calcitrapa (Heist. ex Fabr.) DC. (involucral scales ter- minated by one spine that is laterally divided towards the base), and Crocodylium (Hill) DC. (involucral scales terminated by a simple spine) Shortly afterward, Persoon (1807) proposed ten subgenera for Centaurea 440 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 (Amberboa Pers., Phrygia Pers., Cyanus (Miller) Pers., Jacea (Miller) Pers., Stoebe (L.) Pers., Calcitrapa (Heist. ex Fabr.) Pers., Seridia (Juss.) Pers., Cro- codylium (Hill) Pers., Verutum Pers., and Crupina Pers.) based primarily on involucral bract characters similar to those used by De Candolle 1n his sectional treatment. In the year that Cassini finished his series of articles for Cuvier’s Dictionnaire des Sciences Naturelles, George Don (1830) put forth a classification with six sections, including five of those proposed by De Candolle (Centaurea, Cyanus, Calcitrapa, Seridia, Crocodylium). Don described the section Crupina (invo- lucral scales entire), as well as sul for most of the other sections included in his treatment. De Candolle’s treatment in his Prodromus (1837) divided Centaurea into five series and 31 sections. Article 33.4 of the International Code of Botanical Nomenclature rules that De Candolle’s sections and series were invalidly pub- lished because of their inverted order (i.e., sections as subdivisions of series). As with his earlier sectional treatment (see above), De Candolle relied on characters of the involucre to differentiate his series, but expanded his delim- itation of sections to include characteristics of the involucre, flowers, achenes, and pappus. The majority of the sections recognized by De Candolle were originally described by Cassini as genera. De Candolle’s inclusive treatment was followed by many subsequent workers, including Bentham (see above) and Hoffmann. Utilizing characters of the in- volucre and pappus, Hoffmann recognized in Centaurea 41 sections, many of which were based on Cassini’s genera. Several of De Candolle’s sections were treated as genera by Boissier (e.g., Aetheopappus Cass., Psephellus Cass., Phaeo- pappus (DC.) Boiss.), who also recognized as genera groups treated as infra- generic categories by other workers (e.g., Stizolophus Cass., Acroptilon Cass., Melanoloma Cass.). Boissier distinguished 16 sections in Centaurea. Hayek’s (1901) classification comprised nine subgenera (Centaurium, Mi- crolophus (DC.) Hayek, Calcitrapa, Cyanus, Jacea, Odontolophus (Cass.) Hay- ek, Crocodylium, Cheirolophus (Cass.) Hayek, Plectocephalus Hayek) and 18 sections (mostly in the subgenera Calcitrapa, Cyanus, and Jacea) in a well- documented treatment of Centaurea in Austria-Hungary. Hayek proposed five new sections (Fucalcitrapa (subg. Calcitrapa), Eucyanus and Pannophyllum (subg. Cyanus), Eujacea (subg. Jacea), Eucheirolophus (subg. Cheirolophus)) differentiated, for the most part, by characteristics of the involucral bract ap- pendages.* More recently, still other systems have been presented in anatomical, paly- nological, cytological, and floristic studies. Wagenitz’s (1955) work on the pollen morphology and systematics of Centaurea has been used by subsequent workers as a basis for various systematic schemes (e.g., Dostal, 1973; Dittrich, 1977: Guinochet; Guinochet & Foissac). Wagenitz recognized eight pollen types (Ser- ratula, Centaurium, Crupina, Scabiosa, Dealbata, Jacea, Montana, Cyanus) in Centaurea sensu lato based on differences in size and shape of the grains and the sculpturing and structure of the exine. His data came from about 26 ‘The use of the prefix ““Eu-” is specifically ruled against in Article 21.3 of the International Code. 1990] SCOTT, GENERA OF CARDUEAE 441 sections of Centaurea and from 16 closely related genera. Wagenitz’s obser- vations have been elaborated on by Stepa (1959, 1960) and Velari. Velari noted, in confirmation of Wagenitz, that Centaurea presented solitary, isopolar and radiosymmetric pollen grains, the shape of which varied from suboblate to euprolate. In addition, the grains were reported to be tricolporate with operculate apertures, and the sculpturing was described as echinate, echinulate, or verrucate. Dittrich (1968a, b) studied in considerable detail the morphology and anat- omy of achenes from eleven genera and in Centaurea from nearly 25 sections. He concluded that the form of the hilum (basal, lateral, or caudate), the mech- anisms for fruit detachment (the elaiosomes or simple parenchymatic tissues), achene pubescence, and the consistency of the pericarp are important for the determination of generic boundaries in the Centaureinae. While considerable variation can be noted in pappus characters of taxa in the Centaureinae, Dittrich proposed a new definition of single and double pappus in this group and concluded that this character is of limited value for defining genera in this tribe. Dittrich (1977) recognized 27 genera in the subtribe Centaureinae. In a manner very similar to that of Cassini, he characterized the group by the concave and lateral-adaxial orientation of the detachment area of their fruits and their “double pappus” comprised of an outer rank of bristles or scales elongated from outside to inside and an internal rank often different from the outer. Utilizing Wagenitz’s (1955) seven basic pollen types as the groundwork for his “inner taxonomy,” Dostal (1973) resurrected numerous old generic names, 50 segregates from Centaurea sensu lato. Combinations in several of the genera recognized by Dostal were made by both Holub (1972b) (Acosta Adanson, Colymbada Hill emend. Holub, Cyanus Miller, Jacea Miller) and Sojak (Cy- anus, Jacea, Calcitrapa Heister ex Fabr., Xanthopsis (DC.) Koch, Heterolophus Cass., Acosta; lectotypes were provided for Behen Hill, Psora Hill, and Sagmen Hill). The lack of detailed knowledge of the western Asiatic and Mediterranean groups led Dostal (1975) to reconsider making the numerous nomenclatural combinations necessary if these segregates are recognized as genera as he had earlier proposed (Dostal, 1973). Instead, his account for the Flora Europaea (Dostal, 1976) treated Centaurea as comprising |2 subgenera and 26 sections.° Jeffrey (1968) listed 59 genera of the Centaureeae from the most ““Mutisioid” ‘The combination subgenus Microl/ophus (Cass.) Dostal appears to be based on a a of the earlier combination subgenus Microlophus (DC.) Hayek. The rejection of Hayek’s name appears to e due to Hayek’s use of De Candolle’s illegitimate sectional name (section roth (Cass.) DC.) (see Article 33.4 in the International Code of Botanical Nomenclature; also see discussion above referring to De Candolle’s s eries and sectional eee ae Dostal’s pee should be rejected 1980) have tacitly, if not explicitly, rejected De Candolle’s series nomenclature by their gectelance of his sectional nomenclature 442 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 71 genera to those that he considered most specialized. Jeffrey’s list begins with Warionia Benth. & Coss., a genus assigned to the Mutisieae by Dittrich. The next nine taxa listed by Jeffrey were placed by Dittrich (1977) in the tribe Carlineae, and these were followed by Echinops and Acantholepis Less., of the tribe Echinopseae. The remaining taxa of Jeffrey’s list were assigned to the tribe Cardueae by Dittrich, but Jeffrey’s list intermixed taxa placed in the subtribes Centaureinae and Carduinae by Dittrich (1977). Jeffrey did not intend to in- dicate acceptance of taxa at the generic level by their inclusion in his list, except for the 11 genera (Centaurea, Cirsium, Cnicus, Cyanus, Cynara, Echinops, Goniocaulon, Plectocephalus, Rhaponticum, Silybum, Volutaria) included in his key to genera of East Tropical Africa. Jeffrey included Cnicus, Goniocaulon, Plectocephalus, and Rhaponticum in his key, even though they had not been recorded for this area. A more extensive key was presented by Jeffrey for 17 genera that he considered worthy of recognition apart from Centaurea on the basis of differences in filament and style-branch pubescence, characters Jeffrey found to correlate well with the pollen types established by Wagenitz (1955). He further noted that subgenus CENTAUREA Is a Small group more distinct from other members of Centaurea sensu lato than many groups that have been segregated as genera by earlier workers. For a consistent treatment, Jeffrey held that Centaurea would have to be restricted to subg. CENTAUREA. If such a system were adopted, however, Jeffrey believed that the numerous nomencla- tural recombinations that would have to be made would result in a “‘clearly undesirable” situation. Jeffrey recognized nine subgenera of Centaurea based on filament and achene pubescence and the pollen types of Wagenitz. The eight species of Centaurea (C. americana, C. Calcitrapa, C. Cyanus, C. dubia, C. Jacea, C. maculosa, C. nigra, C. solstitialis) found in the Southeast have, as noticed above, been placed in different subgeneric categories and, in some cases, different genera by European workers. Centaurea americana 1s the type species of Plectocephalus D. Don (Dostal, 1973), but within Centaurea it has been placed in sect. Plectocephalus (D. Don) DC. (Wagenitz, 1955; Dittrich, 1968a). Dostal (1976) incorrectly synonymized C. dubia under the later name C. transalpina Schleicher ex DC., a species that he included in subgenus Jacea sect. Nigrescentes (Hayek) Dostal. Hayek (1901), however, treated C. transal- pina as a synonym of C. dubia in (subgenus Jacea) sect. Eujacea Hayek. Centaurea Jacea, the type species of Jacea Miller (Dostal, 1973), and C. nigra were both placed in sect. Jacea (Miller) DC. (Wagenitz 1955, 1975; Dittrich, 1968a) and in subg. Jacea (Miller) Hayek (Clapham ef a/.). Dostal (1976) retained C. Jacea in subg. Jacea, sect. Jacea, but placed C. nigra in sect. Lepteranthus (DC.) Dumort. Since these two species frequently hybridize (see below), certain workers (Briquet, 1931; Wiegand & Eames) have considered them to be conspecific. Centaurea Cyanus (2n = 24; Love & Léve, 1961, as Cyanus arvensis Moench), the type species of Cyanus Miller (Dostal, 1973), was placed in subgen. Odon- tolophus (Cass.) Hayek by Dostal (1976) and in sect. Cyanus (Miller) DC. by Wagenitz (1975 Centaurea maculosa(2n = 18, 36; Love & Love, 1961, as Acrolophus maculo- sus (Lam.) Cass.) has been placed variously in the genus Acosta Adanson 1990] SCOTT, GENERA OF CARDUEAE 443 (Holub, 1972b), subg. Acro/ophus sect. Maculosae (Hayek) Dostal (Dostal, 1976), and sect. Acrolophus (Cass.) DC. (Wagenitz, 1955; Hayek, 1901), while C. Calcitrapa (2n = 20; Love & Love, 1961, as Calcitrapa stellata Lam.), the type species of Calcitrapa Heister ex Fabr. (Dostal, 1973), has been placed in sect. Calcitrapa (Wagenitz, 1955, 1975) and subg. Calcitrapa (Dostal, 1976). Centaurea solstitialis (2n = 16) has been placed in the genus Leucantha S. F. Gray (Live & Love, 1961), subg. So/stitiaria (Hill) Dobrocz. (Dostal, 1976; Clapham et a/.), and sect. Mesocentron (Cass.) DC. (Wagenitz, 1955). The only comprehensive treatment of Centaurea for the United States and adjacent Canada is that by Moore (1972), who treated 28 species. Within his key to species, taxa were grouped according to sectional characteristics. Moore & Frankton (1954) put forth cytological and morphological justification for segregating C. repens L., a widespread weed of the Western States, as Acroptilon repens (L.) DC. (see also Dostal, 1973). Sectional classification based on pollen data (Wagenitz, 1955) was considered to be congruent with classifications based on involucral characters and chro- mosome number by Guinochet, who tentatively proposed that groups with high base chromosome numbers, nearly entire involucral bracts, relatively complex pollen structure (Serratula type), generally large capitula, large, entire leaves, and perennial habit gave rise to groups with lower chromosome num- bers, complex involucral bracts, relatively simple pollen structure (Jacea type), normally small capitula, often deeply divided leaves, and annual or biennial habit. Guinochet postulated base chromosome numbers of 8, 9, 10, 11, 12, 13, and 15 for Centaurea. His study of the karyotypes showed a correspondence be- tween the size and the number of chromosomes. Guinochet hypothesized that the lower chromosome base numbers evolved from the higher ones by way of translocations and loss of chromosomes. These chromosomal mutations were considered by him to have played a primary role in the differentiation of species, subspecies, and varieties, and further that polyploidy had affected speciation in only a secondary manner. Hypotheses of hybridization and introgression have been used to explain plants morphologically intermediate between taxa otherwise considered dis- tinct (Arénes, 1948, 1949, 1957; Chiapella; Georgiadis; Marsden-Jones & Tur- rill; Roy). Other studies (Ockendon et a/.; Elkington & Middlefell) have found little or no evidence of hybridization and claim that certain Centaurea species (e.g., C. nigra L.) are highly variable. In particular, the taxa included in sect. JAceA (including C. Jacea, C. nigra, C. dubia, and C. nigrescens of the South- eastern species) have been studied extensively, and the number of species recognized in this group has fluctuated from one (Briquet, 1902, 1931) or a few variable species (Guinochet) to as many as seven distinct ones (Arénes, 1957; Marsden-Jones & Turrill). All the species of this section have chromo- some numbers of 2n = 22 or 2n = 44 (Guinochet; Guinochet & Foissac; Roy; Love & Live, 1961: Gervais; Moore, 1968; Marsden-Jones & Turrill). Extensive and well-documented crossing experiments (Marsden-Jones & Turrill) indicate that C. nigra and C. Jacea produce fertile offspring when selfed and will, in general, breed true for characters of the involucre and flower color. 444 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Hybrids between C. Jacea and C. nigra were readily evident, and backcrossing between hybrid offspring and the parental lines produced a wide range of flower and involucral characteristics. In some cases, interspecific crossing seemed to lead to abnormalities in the formation of the androecium. Marsden-Jones & Turrill reported that in some extensive hybrid swarms, pure forms of C. Jacea have disappeared The distribution of sesquiterpene lactones has been used in attempts to clarify taxonomic boundaries within both Centaurea and the subtribe Centaureinae (Geppert et al.; Nowak et al.; see above under tribal discussion), but no clas- sification based on phytochemistry has been proposed to date. The numerous sesquiterpene lactones known from Centaurea have been reviewed by several authors (Wagner, Seaman, Nowak ef a/., Geppert et al.) and will not be reit- erated here. Compounds isolated from species known from the Southeast include cy- naropicrin from C. americana (Ohno et al.), cnicin from C. Calcitrapa (Jakupo- vic et al.), scabiolide from C. Calcitrapa and C. solstitialis, and chlorohysso- pifolin A from C. nigra and C. solstitialis. (See Seaman for citations and sesquiterpenes known from Centaurea up to 1982.) wide variety of flavonoids, including three flavones (centaurein, jacein, Jaceoside) from C. Jacea and two (swertisin, isoswertisin) from C. Cyanus (see Wagner for citations and structural components), has been reported from Cen- taurea. Stevens (1982) and Stevens & Merrill reviewed the wide range of biological activity (cytotoxicity, phytotoxicity, actinioplasticity, and allergenicity, among others) attributed to sesquiterpene lactones and suggested that these compounds may play a role in the allelopathy attributed to C. repens and C. solstitialis. They concluded that these compounds affect root growth in a similar, yet more pronounced, manner to auxins, but do not affect seed germination. REFERENCES: Under tribal references see BENTHAM, 1873a & b; BoBROV & CZEREPANOV; BOISSIER: Bouick; BOULTER ef a/.; BRIQUET, 1931; BRizicKy; BURTT; CASSINI; CLAPHAM et al.: CRONQUIST, 1980; DittricH, 1968a, 1968b, 1977; Don, 1830; Heci, 1928-29, 1987: HOFFMANN; JEFFREY, 1968; LAMARCK & DE CANDOLLE, LESSING; LOVE & LOVE; PERSOON: SEAMAN; E. B. SMITH: SPACH; STEPA, 1959, 1960; WAGNER, 1977: WIEGAND & EAMES; ZowARy: ZWOFLER, 1965, 1978 APPENDINO, G., P. GARIBOLDI, & F. BELLIARDO. Sesquiterpene lactones from Centaurea oe subsp. nervosa. Phytochemistry 25: 2163-2165. 1986. ARENES, J. Les Centaurea pias de Herbier du Muséum d’Histoire Naturelle. Bull. Soc. Fr. Ech. Pl. Vasc. 2: 27-30. 1948: 3: 34, 35. 1949. Le groupe spécifique du Centaurea paniculata L. sensu haan Mém. Mus. Hist. Nat. Paris Bot. 2: 175-266. 1951. [Subgen. Acroloph . Les Centaurées de la sous-section Jacea. Systématique, ae cies et phylogénie. Bull. Jard. Bot. Bruxelles 27: 143-157. 1957. Sur les Centaurées de la sous- -section Phrygiae Boiss. Not. Syst. Paris 15: 376- 390, 1959. [Five species and a number of subspecific taxa keyed and described, and their phylogeny discussed; putative hybrids between these and other species listed. ] ArmirTaGE, A. M. Spotlight on centaureas. Am. Nurseryman. Pp. 60-65, January, 1988. 1990] SCOTT, GENERA OF CARDUEAE 445 BLANCA Lopez, G. ere . género Centaurea L. sect. Willkommia G. Blanca. Lagascalia 10: 131-205. BLARINGHEM, L. Etudes sur le cues floral. 11. Variabilité, sexualité et fecondité du Centaurea pratensis Thuill. Bull. Soc. Bot. France 67: 311-318. 1920. [C. pratensis treated as a distinct species (in contrast to Gadeceau’s earlier treatment) with con- siderable variation in flower, pollen, and achene morphology, as well as variation in the distribution of flower types in the capitula.] Bocas, K. W., & J. M. Story. The population age structure of spotted knapweed (Centaurea maculosa) in Montana. Weed Sci. 35: 194-198. 1987. BOHLMANN, F., & R. K. Gupta. Guaianolides from Centaurea canariensis. Phytochem- istry 20: 2773, 2774. 1981. BRIQUET, J. Monographie des Centaurées des Alpes Maritimes. Bale et Genéve. 1902.* Britton, C. E. Note on Centaurea. Jour. Bot. 57: 340-342. 1919. CARLSTROM, A. New taxa and notes from the SE Aegean area and SE Turkey. Willdenow- ia 15: 73-78. 1986. CHIAPELLA, L. F. Biosistematica di alcune centauree della sezione Acrocentron Cass. Giorn. Bot. Ital. 113: 237-252. 1979. CHICOINE, T. K., & P. K. Fay. The longevity of viability of spotted knapweed seeds in Montana soils. Proc. West. Soc. Weed Sci. 37: 204-207. 1984.* [C. maculosa.] Cory, V. L. Six igre introduced into Texas. Madrofio 5: 200-202. 1940. [Centaurea melitensis, C. Pict CuSMA VELARI, T. oe electron microscopic studies of Centaurea L. s.s. pollen. Giorn. Bot. Ital. 116: 201-210. 1984 VuERICH. Pollen and systematic morphology in Centaurea section Eujacea. Meeting of the Societa Botanica Italiana Congresso Sociale (Social Congress of the Italian cae Society), Prato, Italy, Oct. 12-14. Giorn. Bot. Ital. 116(Suppl. 1.): 76, 77 ea S. Commentationes de systematica Centaureinarum. I, II. (In Russian.) Not. Syst. Leningrad 19: 442-457. 1959; 20: 457-489. 1960. DosrocHayeEva, D. M., & I. M. Kotov. Taxonomy of the genus Centaurea L. s.l. (English summary.) Ukrayin. Bot. Zhur. 19: 39-44. 1962. [Includes new species.] two mutually exclusive forms: prismatic and curvilinear crystals. Prismatic crystals found in 81 of 112 species. This character was shown not congruent with Wagenitz’s (1955) classification based on pollen morphology.] DosTAL, J. Etude préliminaire de l’évolution et de la systématique des types du genre Centaurea sect. Cyanus subsect. Montanae. Preslia 10: 57-69. 1931. axonomick-studie o celedich Lycopodiaceae a Asteraceae a o rodech Lyco- podium a Centaurea. Acta Univ. Palackianae Olomucensis, Fac. Rer. Nat. 31: 5- 1969.* —. Preliminary notes on the subtribe Centaureinae. Acta Bot. Acad. Sci. Hung. 19: 73-79, 1973. New nomenclatural combinations and taxa of the Compositae subtribe Cen- taureinae in Europe. [n; V. H. HEywoop, ed., Fl. Europaea. Notulae systematicae ad floram Europaeam spectantes. No. 18. Bot. Jour. Linn. Soc. 71: 191-210. 1975. . Centaurea. In: T. G. TuTin et al., eds., Fl. Europaea 5: 254-301. 1976. Drozpz, B., & J. Plotrowski. Lactones of Carduinae subtribe. Pol. Jour. Pharmacol. Pharm. 25: 91-94. 1973.* Ear_e, F. R., I. A. Woirr, & Q. Jones. Search for new industrial oils. III. Oils from the Compositae. Jour. Am. Oil Chem. Soc. 37: 254-256. 1960. [Centaurea Cyanus reported to have high concentrations of dienoic acids suitable for applications de- pendent on reactions of linoleic acid.] 446 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 ELKINGTON, T. T., & L. C. MIDDLEFELL. es variation within Centaurea nigra L. in the Sheffield region. Watsonia 9: 109-116. 1972. FELDMAN, L. J., & E. G. Cutrrer. Regulation of leaf a in Centaurea solstitialis L. 1. Leaf development of whole plants in sterile culture. Bot. Gaz. 131: 31-39. 1970. II. The development potentialities of excised leaf primordia in sterile culture. bid. 39- 49. [Heteroblastic leaf development regulated in part by endogenous levels of gib- berellins. Suggested that in transition to flowering levels of endogenous gibberellins rise, resulting in production of small, entire leaves below capitula, whereas lower levels of gibberellins had brought about production of deeply lobed leaves on lower portions of stem. Plants grown on media with high concentrations of gibberellic acid produce only entire, unlobed leaves, while plants grown on media with a gibberellic acid inhibitor produce lobed leaves. ] FERNANDEZ Casas, J. Numericos aus de plantas serves III. Lagascalia 6: : C. antennata, 2n = 22; C. Haenseleri, 2n 4. FLETCHER, R. A. & A. J. RENNEY. A growth inhibitor founds in Centaurea spp. Canad. Jour. Pl. Sci. 43: 475-481. 1963. GADECEAU, E. Observations critiques sur le groupe de Centaurea Jacea L. Bull. Soc Bot. France 67: 55-63. 1920. [Discussion of taxa closely related to C. Jacea, including C. nigra and C. sertees as varieties. Garbou, C. Sur la signification de la pectination des saan des écailles chez Centaurea Jacea L. s.1. Aik ce Bot. France 112: 153-163. Analyse du mode de reproduction chez Coma Jacea : sensu lato. Compt. ae anes Sci. Paris, Sci. Nat. 265: 912-914. 1967. r quelques tentatives d’ ice aed expérimentale intraspecifique chez Cen- is "lacea L. s.l. Ibid. 1195- . Caryosystématique aes de la section Acrocentron Cass. Bull. Soc. Bot. France 116: 29-38. 1969. Recherches cas eae sur la section Jacea Cass. et quelques sections voisins du genre Centaurea L. en France et dans les régions limitrophes. Feddes Repert. 83: 311-472. 1972. GEORGIADIS, T. Problémes de différenciation et d’introgression dans Centaurea subg. Acrolophus Compositae en Gréce. Bot. Jahrb. Syst. 102: 321-337. 1981. Pollen de Centaurea en Gréce. (English summary. ) Pollen Spores 27: 29-51. 1985. [Differentiation in pollen morph Soci eographical orientation and cytological characters; a elation betuecn pollen ana floral morphology noted.] GEPPERT, B., B. DRozpz, M. KEILCZEwSsKI, & M. HoLus. Sesquiterpene lactones XXIII. Isolation of eee p aen lactones from Centaurea L. species. Acta Soc. Bot. Polon. 52: 23-34. 1983. [18 taxa analyzed; lactones reported were salonitonolide, cnicin, cynaropicrin, an repin, and janerin. Study did not involve species that occur in the southeastern U Gervais, C. Annotated list of chromosome numbers of the northeastern American vascular flora. Nat. Canad. 106: 451-462. 1979. GiLLoT, X. Sur la gynodioecie de la Centaurea Jacea L. Ann. Soc. Bot. Lyon 21: 67- 1896. Ne} = Ne) ON \O ~ ~ — Goncare: A. G., J. BERMEJo, & G. M. MAssaANneT. Aportacion al estudio quimiotax- onomico ae genero Centaurea. Rev. Latinoamer. Quim. 8: 176-181. 1977. , J. BERMEJO-BARRERA, T. ZARGOZA-GARCIA, & F. ESTEVEZ-Rosas. Sesquiterpene lactones from Centaurea species. Phytoc hemistry 23: 2071, 2072. 1984. GuGLeER, W. Die Centaureen des ungarischen National-Museums. Ann. Hist. Nat. Mus. Hung. 6: 15-297. 1907. ae rai M. Contribution a l'étude caryologique du genre Centaurea L. sens. lat. ull. Soc. Hist. Nat. Afr. Nord. 48: 282-300. 1957. [C. ruthenica Lam., n = 15.] J. Forss Sur les caryotypes de quelques espéces du genre Centaure Let leur signification taxonomique. Rev. Cytol. Biol. Veg. 25: 373-389. 1990] SCOTT, GENERA OF CARDUEAE 447 Guy as, S., & J. Pestr. Data on the anatomy of nectaries of Centaurea species. Acta Biol. Szeged. * 17-24. 1966. Harvitt, A. M., Jr., T. R. BRADLEY, & C. E. Stevens. Atlas of the Virginia flora. Part II. cep. 147 pp. Farmville, Virginia. | HAYEK, A. von. Die Centaurea-Arten Osterreich- -Ungams. Denkschr. Acad. Wiss. Wien, — -Nat. 70: 585-773. 1901. —— rodromus florae peninsulae Balcanicae. Fedde Repert. Spec. Nov. Beith. 30. x 735- 795. 1931. [C. nigra, C. Jacea, C. dubia, C. solstitialis, C. maculosa treated as separate species; C. dubia Suter. with three subspecific divisions corre- sponding to C. nigrescens Willd., C. vochinensis Bernh., and C. smolinensis Hayek.] —— a oe Y., & K. R. SHIVANNA. The receptive surface of the angiosperm a. Ann. Bot. II. 41: 1233-1258. 1977. [Centaurea said to have dry, papillate a : Hous, J. On correct generic names of Acrocentron Cass. and Acrolophus Cass. (Cen- taurea L. s.1.). Preslia 44: 215-218. 1972a. New nomenclatural combinations in Centaureinae. Acta Bot. Acad. Sci. Hung. 19: 73-79. 1972b. Brief comments on the 4th volume of Flora Europaea. Preslia 49: 311-327. L, E. D. ee maculosa in Indiana. Rhodora 48: 391. 1946. (pero J., Y. Jia, V. P. PATHAK, F. BOHLMANN, & R. M. Kina. Bisabolone deriv- atives and sesquiterpene lactones from Centaurea species. Pl. Med. 5: 399-401. 1986. Jones, G. N. Centaurea maculosa in Illinois. Rhodora 49: 84. KAMANzI, K., J. RAYNAUD, & B. Vorrin. Flavonoid C- seat from the flowers of Centaurea eins I (Compositae). Pl. Med. Phytother. 17: 47-51. Lack, A. J. Competition for pollinators in the ecology of Centaurea Scabiosa and Centaurea nigra. I. Variation in flowering time. New Phytol. 91: 297-308. 1982a. II. Observations on nectar production. /bid. 309-320. 1982b. III. Insect visits and the number of successful pollinations. bid. 321-339. 1982c. Lopez, G. B. Notas cariosistematicas en el género Centaurea section Wilkommia G. Blanca. I. are Anal. Jard. Bot. Madrid 38: 109-126. 1981. Mappox, D. MAYFIELD, & N. H. Porirz. Distribution of yellow starthistle (Centaurea se one and Russian knapweed (Centaurea repens). Weed Sci. 33: 315-327. 1985. [Light infestations (<450 ha.) of C. solstitialis reported Davidson ounty, Tennessee, and Halifax County, N. Carolina. None of C. repens reported a states of the southeastern U. S., but heavy infestations reported in large areas of Virginia. ] ee tes E. M., & W. B. Turritt. British knapweeds. 201 pp. Dorking, En- gland. Printed for the Society of London. i [A thorough treatment with many valuable notes, particularly on subg. Jace MassioT, G., A. M. Morraux, L. LE MEN Cae J. Bouquant, A. Mapaci, A. MAHAMOUD, M. CuHopova, & P. AcLInou. Guailanolides from the leaves oe an. taurea incana. cc 25: 258-261. 1986 McGrecor, R. L. Centaurea. Pp. 898-901 in Great Plains Flora Association, Flora of the Great Plains. Lawrence, Kansas. MERRILL, G. B., & K. L. STEVENS. Sesquiterpene lactones from Centaurea solstitialis. Phytochemistry 24: 2013-2018. 1985. Moore, R. J. Jn A. Léve, IOPB Chromosome number reports. X VIII. Taxon 17: 419- 422. ie ribution of native and introduced knapweeds (Centaurea) in Canada and the United States. Rhodora 74: 331-346. 1972. & C. FRANKTON. Cytotaxonomy of three species of Centaurea adventive in Canada. Canad. Jour. Bot. 32: 182-186. 1954. [Support segregation of C. repens as 448 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Acroptilon Cass., distinguishing characters: all flowers perfect and similar; achenes smooth, slenderly obovoid; scar basal, not oblique; pappus 9 mm long, caducous; hyaline margin of “tegules” (phyllaries) entire; » = 13, unknown in other sections Nitsson, A. The flora along the railways in the Landskona area southernmost Sweden. Svensk. Bot. Tidskr. 78: 293-307. 1984. OcKENDON, D. J., S. M. WALTERS, & T. P. WHIFFEN. Variation within Centaurea nigra L. Proc. Bot. Soc. Brit. Is. 7: 549-552. 1969. Ouno, N., H. Hirar, H. YosHioka, X. A. DOMINGUEZ, & T. J. MABRY. Cynaropicrin, a sesquiterpene lactone from Centaurea americana. Phytochemistry 12: 221, 222. 1973. Oxsuz, S., & H. Avyi_piz. Sesquiterpene lactones from Centaurea coronopifolia. Phy- tochemistry 25: 535-537. 1986. ,& C. JoHANsson. 6-Methoxylated . 16 glycosyl flavonoids from Cen- laurea species. Jour. Nat. Prod. 47: 902, 903. Pavietic, Z. A study of taxonomical relations ae the species Centaurea rupestris and Centaurea Fritschit and their spontaneous hybrid ra sordida. (Astera- ceae-section Acrocentron). Acta Bot. Croat. 42: 137-14 983. Puitieson, B. A. An enumeration of the Le ae ann epecies of Centaurea. Jour. Bot. 77: 227-234. 1939. [Provides a key to 2 a Provan, J. Centaureele Romaniei (Centaureae Romaniae), Monographie. (Subgen. Cy- anus). (In Hungarian & German.) Bul. Acad. ole Studii Agron. Cluj Mem. 1. 256 pp., 63 pls. 1930.* [Monograph of Centaurea in Romania.] Rocue, B. F., Jr., G. L. Piper, & C. J. TALBotr. Knapweeds of ee Exten. Bull. Wash. State Univ. Coop. Exten. Serv. 1393. 41 pp. Nov Roster, H., A. E. Star, & T. J. wire a 6- Methoxyflavonols from Centaurea Jacea. Phytochemistry 10: 450, 451. Roy, B. Chromosome numbers in some mcs and hybrids of Centaurea. Jour. Genet. 35: 89-93. 1937. SAARISALO-TAUBERT, A. A study of hy bridization 1 in Centaurea section Jacea in eastern Fennoscandia. Ann. Bot. Fenn. 3: 86-95. 1966.* StncH, R. P., & A. K. PANDEY. Development and structure of seeds and fruits in Compositae- Cynareae. Phytomorphology 34: 1-10. 1984. SoJAK, J. Nomenklatorické poznamky (Phanerogamae). Cas. Nar. Mus., Odd. Prir. 140: 127-134. 1972. Sosnovsky, D. I. es zwei Sektion der Gattung aoe, s.str. (In Russian; German summary.) Monit. Jard. Bot. Tiflis. I]. 5: 21-34. 1931. Spacn, E. Histoire aie des végétaux. 10: 10-15, 65-72. 1841. Paris. [Source of several subgeneric names in Centaurea. ] STEFANOV, B., & T. GEoRGIEV. Beitrag zur Begrenzung der Arten der Gattung Centaurea L. von der Sekt. Cyanus DC. Spis. Balg. Akad. Nauk. 44: 133-193. 1931.* STEVENS, K. L. Sesquiterpene lactones from Centaurea repens. Phytochemistry 21: 462, 463. 1982. . Allelopathic polyacetylenes from Centaurea repens (Russian knapweed). Jour. Chem. Ecol. 12: 1205-1211. 1986. G. B. MERRILL. jae lactones and allelochemicals from Centaurea species. Pp. 83-98 in ACS. Symp. Ser. Am. Chem. Soc. 268. 1985. [Based on “Symposium on the chemistry of allelopathy, biochemical interactions among plants,” April 1984, St. Louis, Missouri.] & . Wona. Structure of chlororepdiolide, a new sesquiterpene lactone from Centaurea repens. Jour. Nat. Prod. 49: 5 837. 1986. SULYOK, G., & A. LASLo-BENcsIk. Cyanidin 3-(6-succinyl glucoside)-5-glucoside from flowers of seven Centaurea species. Phytochemistry 24: 1121, 19 TAKEDA, K., & S. TominaGa. The anthocyanin in blue flowers of Ceniaurea Cyanus. Bot. Mag. Tokyo 96: 359-363. 1983. 1990] SCOTT, GENERA OF CARDUEAE 449 Tamas, M., & S. Srotertu. Chromatographic identification of anthocyanidins and an- thocyanosides in some flowers and fruits of indigenous plants. Stud. Cercet. Biochim. 19(1): 113-120. 1976.* TAMURA, H., T. Konpo, Y. Kato, & T. Goro. Structures of a succinyl anthocyanin and a malonyl] flavone. Two constituents of the complex blue pigment of cornflower, Centaurea Cyanus. Tetrahedron Lett. 24: 5749-5752. 1983.* TsANkKoVA, E., & I. OGNYANOvV. New sesquiterpene lactones from Centaurea phrygia. Pl. Med. 5 465, 466. 1985. TZVELEV, N. Not. Syst. Leningrad 19: 409-441. 9: Upapuyaya, M. K. Induction ieee by gibberellic acid in rosettes of diffuse (Cen- taurea diffusa) and spotted (Centaurea maculosa) knapweed. Canad. Jour. Bot. 64: 2428-2432. 1986. [Bolting induced by gibberellic acid, a significant correlation between rosette age and gibberellic acid treatment noted; gibberellic acid did not induce bolting in plants less than 40 days o VALDES-BERMEJO, E., & M. P. A. MATA. Estudios cariologicos en especies ibéricas del género Centaurea L. (Compositae). I. Anales Jardin Bot. Madrid 40: 119-142. 1983. VAN Loon, J. C., & H. DE Jonc. Jn: IOPB chromosome number reports LIX. Taxon 27: 53-61. 1978. [Centaurea diffusa, 2n = 36; C. maculosa ssp. maculosa, 2n = 18.] WAGENITZ, G. Pollenmorphologie und Systematik der Gattung Centaurea 142: 213-279. 1955. [The definitive study on which many subsequent claseineations have been based. . Uber einige Arten der Gattung Centaurea aus der Tiirkei. Willdenowia 2: 456- 468. 1960a. ——. Centaurea L. sect. Cynaropsis, eine neue Sektion der Gattung aus Vorderasien. Ibid. 469-494. 1960b. Die Eingleiderung der Phaecopappus-Arten in das System von Centaurea. Bot. Jahrb. 82: 137-215. 1963. . Centaurea. In: Materials for a Fl. Turkey XXX: Compositae, I. (Compiled by P. H. Davis.) Notes Bot. Gard. Edinburgh 33: 217-231. 1974. . Parallele Evolution von Merkmalen in der Gattung Centaurea. Phyton Austria 16: 301-312. 1975. . Centaurea. In: P. H. Davis, ed., Fl. Turkey 5: 465-585. 1975. —. Centaurea. In: K. H. Recutncer, ed., Fl. Iran. 139b: 313-420. 1 ‘ . Centaurea and the Index Kewensis. Taxon 32: 107-109. 1983. [Of the 2070 entries examined by the author, nearly 25 percent (488) were found to be incorrect ace in addition, “‘a considerable number” of names were reported by the author e been omitted from the Index e flora of Arabia 7. Centaurea i in the Arabian Peninsula. Notes Bot. Gard. Edinburgh 41: 457-466. 1984. A new species of Centaurea (Compositae) from Iraq. Kew Bull. 40: 793, 794. 1985. 1; Centaureae L. (In Russian.) o Centaurea in Southwest Asia: patterns of distribution and diversity. Second MAL-ELDIN. Zur Kenntnis der griechischen Centaurea-Arten der Sektion Paso: ia (English summary.) Bot. Jahrb. 107: 95-127. 1985. [Revision of the 25 species of section Acrocentron known from Greece.] 7. Cnicus Linnaeus, Sp. Pl. 2: 826. 1753; Gen. Pl. ed. 5. 358. 1754, nom. cons. Erect annual herbs, the stems branching, arachnoid-villous. Leaves of the lower stem commonly short-petiolate, upper leaves sessile; blades lanceolate, up to 20 cm x 5 cm, subcoriaceous, sinuate-dentate or pinnatifid, scarcely or 450 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 not at all decurrent, lower surface conspicuously veined and densely glandular, margins weakly spiny or ciliate. Capitula heterogamous, discoid, solitary at the end of branches, closely subtended by ovate or lanceolate cauline leaves. In- volucre campanulate to ovoid, often obscured by cauline leaves; bracts im- bricate, multiseriate, ovate-lanceolate, the outer mucronate-subulate, the inner longer, with a pectinate to spine-like apical appendage. Receptacle flat; scales numerous, setaceous. Outer flowers inconspicuous, sterile, with a very slender 2- or 3-lobed corolla; inner flowers perfect, their corollas slightly zygomorphic, yellow, glabrous or minutely glandular, with the tube narrow, elongate, ca. 8- 10 mm long, the throat expanded, slightly saccate, unequally divided by the sinuses of the lanceolate, glabrous corolla lobes with two sinuses considerably deeper than the other three. Staminal filaments pubescent, connected at base of throat; anthers stramineous with a firm, narrow, apical appendage, shortly saggitate tailed at base. Style smooth below a distinct collar of hairs at base of the short, scarcely divergent, papillate branches. Achenes obliquely attached to receptacle, ca. 8 mm long, subterete, strongly 20 ribbed, glabrous, with a firm 10-toothed crown. Pappus biseriate, the outer of 10 long, minutely scabrid, stramineous awns, alternating with 10, much shorter, white, minutely hairy and sparsely pectinate inner bristles. (Carbenia Adanson.) TYPE SPECIES: C. benedictus L., typ. cons. (See Hitchcock & Green, Nomencl. Prop. Brit. Bot. 179. 1929, for discussion.) (A Latin name of the safflower (Carthamus L.), from Greek cnecos.)— BLESSED THISTLE. A single species of the Mediterranean region widely naturalized in many parts of the Northern Hemisphere. Cronquist (1980) characterized Cnicus benedictus in the Southeast as spar- ingly established in fields and waste places. This species is found scattered throughout the Piedmont and along the Coastal Plain from Virginia, south through Georgia and Alabama, but apparently it neither extends as far west as Texas (Correll & Johnston), nor is 1t found in the Great Plains flora (Barkley, 1986). Linnaeus included five species in his treatment of Cnicus. Adanson recog- nized the genus Carbenia for two species, one of which was C. benedictus. Apparently unaware of Adanson’s Carbenia, Gaertner redefined Cnicus as a monotypic genus based on the unique pappus characters of C. benedictus. ersoon incorporated Cnicus as a subgenus of Carduus, but treated Cnicus benedictus as a member of Centaurea subgenus Calcitrapa. While Don rec- ognized Cirsium and placed therein many taxa treated by earlier authors under Cnicus, he, too, treated C. benedictus under Centaurea (sect. Calcitrapa). Les- sing’s recognition of Cnicus for C. benedictus was followed by De Candolle, who provided an extensive list of excluded species. Both Lessing and De Can- dolle placed Cnicus in the subtribe Centaureinae. Bentham recognized Carbenia Adans. for Cnicus benedictus and broadly circumscribed Cnicus to include many taxa currently placed in Cirsium. Neither Hoffmann nor Spach followed Bentham but treated Cnicus as it was delineated by Gaertner. Asa Gray (1884a, b, 1874), on the other hand, did follow Bentham’s circumscription (see dis- cussion under Cirsium) and included Cnicus benedictus in Centaurea. Such 1990] SCOTT, GENERA OF CARDUEAE 451 nomenclatural confusion led to the conservation of Cnicus L. emend. Gaertner in 1935 (Hitchcock & Green). Contrary to most floristic treatments of the past fifty years, Ahles in Radford ef a/. placed Cnicus in synonymy under Centaurea from which it is readily distinguished by its pappus structure and fruit characters (Dittrich, 1970). Chromosome numbers reported for Cnicus benedictus are 2n = 20 and 2n = 17 (see Mehra et a/.; Goldblatt, 1981, for references). Medicinal qualities have been attributed to Cnicus (Franco) particularly for the glucoside cnicin for its antimicrobial and antibiotic activity (Wagner; Schneider & Lachner; Vanhaelen-Fastré) and for treatment for gout (Mabber- ley). In addition to cnicin, another sesquiterpene lactone, salonitenolide, and monoterpenes have been isolated from C. benedictus (Seaman). REFERENCES: Under tribal references see ADANSON; BARKLEY, 1986; BENTHAM, 1873a; BoBROV & MEuRA et al.: MORTON; MuNz; PERSOON; RADFORD et al.; SEAMAN; SPACH; and WAGNER. ARYVAND, A. Jn: IOPB chromosome reports LVIII. Taxon 26: 557-565. 1977 FRANCO, J. DO AMARAL. Cnicus. In: T. G. Tutin et al, eds., Fl. Europaea 4: 301, 302. Korte, F., & G. BECKMANN. Uber die Bitterstoffe aus Cnicus benedictus. Naturwissen- schaften 45: 390. 1958. KupicHa, F. K. Cnicus. In: P. H. Davis, ed., Fl. Turkey 5: 588-590. 1975. RECHINGER, K. H. Cnicus. In: K. H. RECHI HNGER: ed., Fl. Iran 139b: 428-430. 1980. SCHNEIDER, G., & I. LACHNER. Beitrag zur Analytik und Wirkung von Cnicin. Pl. Med. 53: 247-251. 1987. VANHAELEN-FASTRE, R. Antibiotic and cytotoxic activities of cnicin, isolated from Cni- cus benedictus. Jour. Pharm. Belg. 27: 683. 1972.* MILLER, MELIACEAE 453 THE GENERA OF MELIACEAE IN THE SOUTHEASTERN UNITED STATES! NorRTON G. MILLER? MELIACEAE A. L. de Jussieu, Gen. 263. 1789, ‘Meliae’, nom. cons. (MAHOGANY FAMILY) Small to large trees [shrubs, or rarely suffrutescent herbs]; new growth from terminal buds or from axillary buds, if axillary, then the branch apices dying back at end of growing season; bud scales imbricate [or absent], pubescent or not, deciduous; pith homogeneous [or with clusters of fibers]; new growth near leaf insertions or without extrafloral nectaries. Leaves alternate, exstipu- late, once or twice odd- or even-pinnate [trifoliolate, or simple]; spirally ar- ranged [rarely ene trichomes of young leaves simple and hooked, glan- ‘Prepared for t! ic Fl ft ted States, a long-term project made possible through the support of National Science eee ia ee BSR-8716834 (N. G. Miller, ae ipal investigator), under which this account was largely prepared, and ice 8717333 (C. E. Wood, Jr., principal investigator). The 135th in the series, this paper follows the format established in ae first one (Jour. Arnold Arb. 39: 296-346. 1958) and continued to the present. The area covered by the Several botanical colleagues helped in various ways in the preparation of this paper, nga GB, Wood, Vey Weds Grits, GC. Tucker, Js. enenbat: Gy i ROges » Te J Rosa , B. G. Schubert, and B. T. Styles, all of and/or comments on a draft of the paeucce Dh I also thank Stephen A. Spongberg for his editorial review of the manuscript. The staff of the New York State Library, and especially Alta Beach, Senior Librarian, and of the Libraries of the Arnold Arboretum and Gray Herbarium, Harvard University, gave prompt assistance with the sometimes challenging task of See my pleas for literature that retum a Ta or liquid-preserved (a—g) material collected by D. Sturrock in West Palm ee Florida; a-g and n are by Rachel Wheeler (with dissections by Wood), h—-m are by Sue Sar This paper is published as Contribution Number 654 of the New York ete Science Servic *Biological Survey, New York State Museum, The State Education Department, Albany, New ‘York 12230. © President and Fellows of Harvard College, 1990. Journal of the Arnold Arboretum 71: 453-486. October, 1990. 454 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 dular, dendritic, or stellate [or peltate scales]; leaflets symmetrical or asymmetrical, serrate or entire, deciduous with rachis or not [rachis rarely winged, sometimes with an intermittently growing terminal “bud”’]. Plants polygamous or monoecious [or rarely dioecious]; inflorescences axillary, large or small bracteate thyrses [panicles, or rarely spikes, sometimes cauliflorous or ramiflorous, or flowers epiphyllous], cymules 3-flowered, the terminal (first- opening) flower perfect or carpellate, lateral flowers staminate; staminate flow- ers deciduous after anthesis. Flowers regular (actinomorphic); in Melia perfect and staminate flowers similar at anthesis, in Swietenia staminate and carpellate flowers dimorphic. Sepals [2—4 or]5(6)[or 7], separate or fused basally and calyx then 5-(rarely 4- or 6-)lobed [or calyx circumscissile]. Petals [3, 4]5(rarely 6)[8 or 14], free [or sometimes fused below to staminal tube], in 1 [or rarely 2] whorls, glabrous or pubescent abaxially, alternate with the sepals, the aesti- vation convolute, imbricate [or contorted or valvate]. Stamens united into a cylindrical or urn-shaped [cyathiform] tube [or filaments free], tube fringed [or not] with [8 or 9]10(or 12) teeth, individual teeth deeply cleft or not; anthers [3-]10(1 2)[-23], [rarely septate], in one [rarely 2] whorl[s], sessile on inside [or top] of tube [or with short filaments from top of tube], basifixed or dorsifixed, glabrous [or pubescent], alternating with the teeth or seemingly opposite two narrow teeth, dehiscence introrse, connective short [or sometimes greatly pro- longed and filiform]; pollen [3- or] 4-colporate. Gynoecium syncarpous, the ovary superior [very rarely inferior], [3-](4)5 or 6[-15]-locular, each locule with 2 superposed ovules or with numerous ovules in 2 rows [ovules sometimes few and collateral], the placentation axile [rarely parietal], ovules anatropous, nectariferous disc annular, entirely below the ovary or extending slightly up- ward, obscurely lobed and free from the ovary [or cyathiform, tubular, rarely a stipe supporting the gynoecium, or absent]. Ovary in perfect flowers pyriform and tapered into a long style, stigma rounded, scarcely wider than style; in carpellate flowers ovary globose, style short, stigma discoidal and nearly as wide as mouth of staminal tube [or obconical, globose-capitate, or 3—6-lobed], anthers withered [staminodia rarely present]; in staminate flowers ovary nar- rowly pyriform, style abruptly differentiated and longer than in carpellate flow- ers, stigma discoidal. Fruit a drupe, endocarp 5- or 6-locular, keeled, usually one seed per locule [or endocarps separate, thin or thick walled] or a 5-locular capsule, septicidally [loculicidally] dehiscent from base [or apex], columella persistent, [or fruit a fleshy or leathery berry, or very rarely a nut]. Seeds retained within the endocarp or winged and dispersed after capsule dehiscence [com- monly with a brightly colored arillode or sarcotesta]; cotyledons collateral, longer than broad or broader than long, embryo short, straight [or curved], plumules minute or absent; endosperm fleshy, oily or more commonly thin and appearing absent; germination phanerocotylar or cryptocotylar. (Including Cedrelaceae R. Brown in Flinders, Voy. Terr. Austral. 2: 595, 1814, ““Cedre- leae’’; tribe Aitonieae Harvey in Harvey & Sonder, Fl. Capensis 1: 243. 1860 [Sapindaceae].) Type GENUS: Melia L. A mainly tropical family of moderate size (50 genera, 550 species, according to recent monographic study; Pennington & Styles), with 14 genera (eight native, 1990] MILLER, MELIACEAE 455 six introduced) represented in the New World. Only Cabralea A. Juss. (one sp.), Ruagea Karsten (ca. five spp.), Cedrela P. Browne (seven spp.), Schmar- daea Karsten (one sp.), and Swietenia Jacq. (three spp., one in our area) are endemic to the Western Hemisphere; Trichilia P. Browne (85 spp.; largely lowland tropical America, some species 1n Africa and a few in the Indo-Malayan region), Guarea Allamand ex L. (35 spp. in tropical America, five in tropical Africa), and Carapa Aublet (two spp.; tropical America and Africa) are dis- junctly distributed among portions of the New and Old World tropics. Of the introduced genera only one species of Melia L., M. Azedarach L., is widely established in tropical and warm temperate parts of North and South America, including the southeastern United States. Over the past 20 years a great deal of new and important information has been discovered about the taxonomy of the Meliaceae, mainly by botanists associated with the Commonwealth Forestry Institute, University of Oxford, England. Their studies have been wide ranging; those pertaining to generic concepts are summarized by Pennington & Styles in a lengthy paper that con- tains many original observations and new analyses and interpretations. In addition the family has been monographed for the Flora Neotropica series (Pennington, 1981), and many problems involving difficult species complexes have been clarified, in part aided by the recent availability of new and more adequate collections and a better understanding of the biology—especially the floral biology—of the family. The Meliaceae, excluding the Ptaeroxylaceae J. F. Leroy, are a reasonably coherent group of monoecious, dioecious, or polygamous woody plants (gen- erally trees), mostly with alternate, pinnate leaves and regular pentamerous flowers containing a staminal tube and a hypogynous nectariferous disc. Four subfamilies are recognized (Pennington & Styles), two of which are represented in our area. Subfamily Melioideae [Harms] (plants polygamous or dioecious, ovules | or 2, superposed or collateral, fruit a drupe, berry, or loculicidal capsule, seeds not winged, rays of wood usually | or 2 seriate) contains Melia Azedarach, a naturalized tree in our area, in tribe Melieae [DC.]. Six additional tribes represented in the neotropics, paleotropics, or both. Subfamily Swietenioideae Harms (plants monoecious, ovules usually many in two rows, fruit a septicidal capsule with a central columella, seeds winged, rarely otherwise, rays of wood generally 3-6 cells wide) includes Swietenia and eight other genera in tribe Swietenieae (A. Juss.) Spach,’ plus two additional tribes that include neotropical and/or paleotropical genera not represented in our area. In general the flowers of members of the Swietenioideae are small or 3Adrien de Jussieu in his ‘Mémoire sur le Groupe des Méliacées” recognized two families, Meliaceae and Cedrelaceae, which he divided into groups of genera, providing a name for each group and Latin diagnoses but no designation of rank. Spach (Hist. Nat. Vég. Phan. 3: 161-205. 1834) used the rank tribe for De Jussieu’s groups. The De Jussieu “Mémoire” was published in 1832 (1830 is the year on the title page of volume 19 of Mém. Mus. Hist. Nat. Paris in which the monograph was published, but the volume was issued in 1832; see Pennington, 1981, p. 4). The names and diagnoses of the new species and the subfamilial groups proposed by De Jussieu were, however, published in 1830 (Bull. Univ. Sci. Industr. Sect. 2 (Bull. Sci. Nat. Géol.) 23: 234-241) and also prior to 1832 in Linnaea 6(Lit.): 107-115. 1831. 456 JOURNAL OF THE ARNOLD ARBORETUM [vVoL. 71 sometimes minute, in contrast to the larger, more showy flowers of members of the Melioideae. Two monotypic subfamilies, Quivisianthoideae Pennington & Styles and Capuronianthoideae Pennington & Styles, accommodate genera endemic to the Malagasy Republic. The poorly known Quivisianthus Baillon in Grandidier has flowers similar to certain members of the Melioideae but differs substan- tially from them and other genera in the subfamily in having a loculicidal capsule and dry winged seeds. Capuronianthus J. F. Leroy is like members of the Swietenioideae in having a septicidal capsule, but it has naked buds, de- cussate leaves, and two superposed ovules (with two others aborting). Thus these subfamilies have certain characteristics of either the Melioideae or Swie- tenioideae, as well as some unique features. Much is known about the chromosome cytology of the Meliaceae. About 100 species (of the ca. 550 species in the family) have been studied (Styles & Khosla), and this work has revealed greater variation in chromosome numbers (2n = 12 to ca. 360) than has been found in other woody, mainly tropical angiosperm families. Most counts are of mitotic figures from root tips (Styles & Khosla), although some counts are based on anther squashes. In many species the chromosomes are minute (0.5 to 3.5 wm), even in cells of the root tip. Accurate determination of the higher numbers has been difficult because of staining problems (Datta & Samanta) and chromosome size. Most chromo- somes have submedian to median centromeres. Polyploid series are present in some genera (e.g., Chisocheton Blume and Dysoxylum Blume of the Indo- Malayan region south to Australia and/or New Zealand). Within species vari- ation in chromosome number eae aneuploidy) is also known, for ex- ample, in Swietenia, Toona (Endl.) M. J. Roemer, and other genera. There is considerable disagreement about ieee base numbers in the Meliaceae. Some authors have suggested x = 7 (Mehra et a/.). Others cite evidence favoring two base numbers, x = 6, x = 7 (Khosla & Styles) or multiple base numbers, x = 9, 10, 11, 12, 13, 14 (Datta & Samanta). The most frequent haploid chromosome number in the family is 25. Chromosome numbers and karyo- types do not generally provide independent substantiation for subfamilies and tribes defined on the basis of morphology. Flower morphology is extremely diversified in the Meliaceae. Characteristics of the androecium are particularly useful taxonomically at the generic level. The filaments are generally connate into a staminal tube (the shape of which may differ considerably among genera) or rarely are free. The anthers are inserted in the throat of the tube or at its summit, and they either are sessile or have short extensions of the filaments. Teeth occur along the distal edge of the tube, and the shape of these differs in taxonomically significant ways. The shape of the nectariferous disc, which is always located below the gynoecium, is also variable, as is the shape of the stigma. The patterns these structures present may be consistent within or among genera. It has been suggested that the variability in floral structure may reflect ad- aptations for specific insect pollinators (White, in Pennington & Styles). How- ever, the pollination biology of the family has been incompletely investigated. The flowers of some species are reported to be fragrant. This suggests insect 1990] MILLER, MELIACEAE 457 pollination, as does the uniform presence of floral nectaries. Flowers of Guarea rhopalocarpa Radlk. open at night and are probably moth pollinated (Bullock et al.), Moths were found to pollinate various species of Guarea and Cedrela in Costa Rica (Bawa et al., 1985), and hymenoptera have been seen to effect pollen transfer in Trichilia havanensis Jacq. (White, in Pennington & Styles). It has only recently been realized that genera characterized by perfect flowers are unusual in the Meliaceae (Lee, 1967; Styles, who mentioned that this condition was restricted to Turraea L. and a few related genera; suspected also in some species of Guarea (Pennington, 1981)). What numerous authors in the past have described as perfect flowers are in reality either functionally staminate or functionally carpellate. In some cases (e.g., Swietenia) the carpellate and staminate flowers are conspicuously dimorphic, in others (e.g., Melia) the sta- minate and perfect flowers are superficially similar but can be told apart by inspection of the ovules, which appear aborted even in young staminate flowers. Poor staining quality of pollen or withered anther sacs characterize the func- tionally carpellate flowers, which only rarely have obvious staminodia. It may be necessary to study plants in the field to ascertain whether a species is Monoecious, dioecious, or polygamous. The reason for this, as Pennington (1981) points out, is that the staminate flowers are deciduous soon after an- thesis, and they may not be present in herbarium specimens. In some dioecious species (e.g., Trichilia Poeppigii C. DC., of northwestern South America) not only are the flowers dimorphic, but the inflorescences are as well. A Costa Rican population of the dioecious Guarea rhopalocarpa, carefully followed for two years (Bullock ef a/.), showed complex patterns of flowering and fruiting in discrete episodes at irregular intervals during the study period, with two or three episodes per year per tree. Flowering occurred discontinuously over nine months of the year, and certain individuals flowered more or less at the same time. Trees with staminate or carpellate flowers were in about even proportion, but the number of staminate inflorescences during the census period was nearly always much greater Mechanisms of fruit and seed dispersal are also varied in the family. In many members of subfam. Melioideae the outer integument of the ovule becomes elaborated into a small or large sarcotesta, which can be rich in oils. The details of development of this structure are poorly known in most genera, but in some species 1t appears to originate from a specific part of the ovule. The meliaceous sarcotesta does not seem to be a proliferation of the funiculus, and therefore it is not exactly equivalent to an aril. In species with dehiscent fruits the sarcotesta is red or orange, and it contrasts with the black or brown unmodified seed coat. Such bright colors attract bird or mammal disperal agents. Seed disperal in Guarea glabra Vahl, an understory tree producing abundant fruit, has been studied in Panama (Howe & De Steven). Seventy percent of the visits and 60 percent of the seeds removed involved four North American migrant birds (great crested flycatcher, Swainson’s thrush, red-eyed vireo, and Tennessee warbler). Fruiting and the northward migration of these birds were synchronized. Of ten species of Ag/aia Lour. studied by Pannell & Koziot in Malaysia and Indonesia three had dehiscent fruits revealing seeds with a red sarcotesta (bird-dispersed), and seven had indehiscent fruits and seeds with a 458 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 yellow, white, or translucent sarcotesta (five of these were primate-dispersed). More lipids were present in sarcotestas associated with bird dispersal; the coverings of the primate-dispersed seeds were gelatinous, low in lipid content, and high in sugars. Bats are likely to disperse the fleshy fruits (and endocarps) of Azadirachta indica A. Juss. in West Africa (Ayensu) and perhaps elsewhere in the tropics. Fruits of Melia Azedarach are eaten and dispersed by birds in the United States and by birds and fruit bats in South Africa (White, 1986). Seeds of Carapa guianensis Aublet are eaten by rodents, monkeys, and wild pigs, which may be agents of dispersal (White, 1983). The buoyant seeds of this species, which prefers swamp forests in at least part of its range, are transported by water (ibid.). In Amazonia several kinds of fish have been observed to eat the seeds of C. guianensis, but the seeds appear to be destroyed in the process (Gotts- berger). Wind is the presumed dispersal agent for those species (mainly mem- bers of subfam. Swietenioideae) with dry, winged seeds Published morphological and anatomical studies of the Meliaceae have dealt mainly with wood anatomy and aspects of vegetative structure that are unusual in seed plants. The early investigations of Kribs and of Panshin into the secondary xylem of representatives of the family yielded character sets that for many years were thought to be diagnostic for genera. However, wood of about one-half of the known species of Meliaceae has now been examined (Pennington & Styles), and some of the conclusions drawn by Kribs and Panshin are no longer tenable. Few genera of the Meliaceae can be distinguished on the basis of wood anatomy alone, but anatomical characters sometimes correlate with other morphological ones in taxonomically significant ways. Wood provides characters (e.g., fibers septate, terminal bands of apotracheal parenchyma absent vs. fibers nonseptate, apotracheal parenchyma present) that are helpful in delimiting subfamilies, and within the Melioideae in placing genera in tribes. Leaves (as well as the bark and secondary xylem) of many Meliaceae have secretory cells. In leaves they are located in the mesophyll and with back lighting are visible as translucent dots. (Secretory cells are evidently lacking in Melia Azedarach and Swietenia Mahagoni, however.) The pinnate leaves of species of Guarea are unusual (Skutch) because they exhibit intermittent, indeterminate growth from a crozier-like “bud” at the rachis apex, which remains meriste- matic. On the basis of anatomy and development such “‘leaves”’ are leaf homo- logs, although in their continuous growth (including increases in secondary xylem thickness) they are analogous to branches (Steingraeber & Fisher). Ex- periments performed by Fisher showed that leaflets in G. Guidonia (L.) Sleumer exhibited either a “sun” or a “shade”’ morphology and that within a given leaf the expression of one form or the other was plastic and related to whether the leaflets were initiated in the shade or in the sun. Leaves of G. rhopalocarpa are estimated to be 7-11 years old at abscission (Skutch). Such indeterminate leaves are also found in species of Chisocheton, a few species of which also have few- flowered, epiphyllous inflorescences. Vascular bundles supplying such inflo- rescences arise from the stele of the rachis with no evidence that the bundles are adnate to the rachis vasculature (Mabberley, 1979). 1990] MILLER, MELIACEAE 459 Basal leaflets in some Meliaceae are modified into stipule-like structures that are appressed to the leafbase-stem junction. The point of attachment is, how- ever, the leaf rachis. Pseudostipules of varying form occur in species of Trichilia (Pennington, 1981), and they are known also in some other families of the Rutales (e.g., Sapindaceae; Weberling & Leenhouts). Sac domatia occur on the abaxial surfaces of leaves of Dysoxylum Fraseranum Benth. of Australia (Met- calfe & Chalk, 1979). The palynology of the Meliaceae has been summarized by Pennington & Styles on the basis of studies of about two-fifths of the species in the family, including representatives of all genera. The family is stenopalynous, with little variation from the basic pattern (pollen subprolate or prolate-spheroidal, 3- or 4-colporate, psilate, sometimes verrucose) among genera or even subfamilies. Pollen has been helpful in the placement of certain genera once included in the Meliaceae, for example Flindersia R. Brown in Flinders to the Rutaceae and Ptaeroxylon Ecklon & Zeyher (which has neither meliaceous or rutaceous pol- len) to the Ptaeroxylaceae (with Cedrelopsis Baillon). Turraea and allied genera (pollen generally oblate-spheroidal, 3-colporate, exine scabrous to verrucose) are the most disparate elements in the family palynologically. Pollen is of limited value in defining genera. Chemotaxonomic studies of the Meliaceae have focused largely on the dis- tribution and systematic significance of limonoids, a group of oxidized triter- penes otherwise known to occur in the Rutaceae and Cneoraceae. These sec- ondary metabolites impart a bitter taste to the plant tissues in which they occur. A mixture of limonoids is present in most species, and different limonoids may be present in different parts of a plant (Taylor, 1983). Sometimes limonoids occur in only one plant organ. Only the most highly oxidized limonoids appear to be significant taxonomically. Various kinds of limonoids are partitioned 1n mostly nonoverlapping patterns among the Meliaceae, Rutaceae, and Cneo- raceae, and within the Meliaceae between subfamilies Melioideae and Swie- tenioideae, but less clearly among the tribes recognized by Pennington & Styles. Melia and Azadirachta A. Juss. (both in tribe Melieae) have many limonoids in common, although Azadirachta has some that are lacking in Melia (Taylor, 1983). Alkaloids are reported from only five members of the Meliaceae (Mester), and coumarins are known in relatively few genera (Gray). Flavonoid chemistry has not been much used as a chemotaxonomic tool in the family (Harborne), and it is unclear how much potential it has. The paleobotanical record of the Meliaceae consists of pollen, leaves, seeds, and fruits, mainly of Tertiary age. A few Cretaceous fossils have been attributed to the family (e.g., Graham, 1962). Structurally preserved wood from the Ter- tiary of Europe and North Africa is similar to that of the extant genera Carapa, Entandrophragma C. DC., and Lovoa Harms (Louvet, 1973, 1975; Madel; Selmeir, 1983, 1987), and these occurrences are cited as examples of tropical or subtropical elements in pre-Quaternary paleofloras of the Mediterranean Basin region. Cedre/a is represented by leaves, fruits, and pollen in the Eocene and Miocene floras of the western United States (MacGinitie; MacGinitie ef al.). Dispersed pollen of Cedrela and/or Guarea has been recovered from Oli- gocene and Miocene sediments in the Caribbean Basin (Graham & Jarzen; 460 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Graham, 1977). These occurrences suggest possible migrational tracks and temporal limits for the migration of neotropical plant elements occurring in the Tertiary paleofloras of the southeastern United States. The absence of palynological diversification and distinctive pollen types in the Meliaceae may limit how much information can be obtained from studies of dispersed fossil pollen. Most members of the Meliaceae are forest trees, usually reaching the canopy or subcanopy, but sometimes only the understory. In the Neotropics the family is especially well represented in nonflooded lowland rain forest (terra firma) and in seasonally flooded lowland forest (varzea). Various genera are also represented in montane forests and sometimes in the cloud forests of central and northern South America. Certain species grow on drier sites, especially in the West Indies where species of Trichilia, Guarea, and Cedrela occur on dry (sometimes mesic) soils over limestone hills. Xy/ocarpus granatum Koenig and X. mekongensis Pierre, are mangroves, occurring in parts of the area from East Africa to tropical Australia and Polynesia (Tomlinson). The family is of considerable commercial importance, primarily as a source of valuable timbers that are used to make high quality furniture. Khaya A. Juss., African mahogany, species of Swietenia, the true mahoganies, and species of Entandrophragma of tropical Africa yield perhaps the most valuable hard- wood lumbers. A limonoid, azadirachtin, extracted from the seeds of Azadi- rachta indica A. Juss., neem tree, has elicited much interest as a growth inhibitor and an antifeeding agent against insects. Azadirachtin is presently under study for possible use in the control of insects that cause damage to food and other crop plants (Schmutterer & Ascher). Oil is extracted from seeds of species in a number of genera (Vaughan). Fruits of Lansium parasiticum (Osbeck) Sahni & Bennet (L. domesticum Jack), langsat, and Sandoricum Koetjape (Burm. f.) Merrill, santol, are eaten in Southeast Asia. The former is considered to be one of the best fruits of the Malayan region (Popenoe). The Meliaceae are placed in the Rutales (Danie et al., Takhtajan, Thorne) or Sapindales (Cronquist), both of which are variously circumscribed. Ther is general agreement, however, that these orders are allied and form a distinct evolutionary line, linked directly to magnolioid ancestors (Meeuse). Phyto- chemical markers (triterpenes), in addition to morphological criteria, indicate close evolutionary relationships among the Meliaceae, Simaroubaceae, and Rutaceae. Limonoids are known from the Meliaceae and Rutaceae, whereas quassinoids, which are biochemically derived from them, are restricted to the Simaroubaceae (Seigler), suggesting that the Simaroubaceae are advanced, at least in this character. REFERENCES: ABDULLA, P. Meliaceae. Jn: E. NAsir & S. I. Au, eds., Flora of West Pakistan 17. 8 pp. 1972. [Swietenia epee and S. macrophylla (cultivated); Mf. Azedarach, ‘“‘wild in W. Himalaya, up to 1700 m. — and naturalized in parts of Iran, China, Burma, Turkey, ae & W. Pakistan.’ ApESIDA, G. A., E. K. ADESOGAN, D. A Oxon, D. A. H. TayLor, & B. T. STYLEs. The limonoid chemistry of the genus Khaya (Meliaceae). Phytochemistry 10: 1845- 1990] MILLER, MELIACEAE 461 1853. 1971. [Six species defined on morphological criteria had distinct limonoid profiles AyeENsu, E. §. Plant and bat interactions in West Africa. Ann. Missouri Bot. Gard. 61: 702-727. 1974. [Frugivorous bats implicated in the rapid colonization of the Accra Plains (Ghana) by Azadirachta indica, which was introduced into West Africa in the early 1900’s.] BAHADUR, K. N. Monograph on the genus Joona (Meliaceae). 251 pp. Dehra Dun. 1988. [Toona (seven spp.) and Cedrela kept as aa genera. Bartey, L. H., E. Z. BAmey, & L. H. BArtey Hortorium Starr. Hortus third. xiv + 1290 pp. New York & London. 1976. [Aelia, 124, incl. cultivar ‘Floribunda’ (bushy, very floriferous) and cultivar ‘Umbraculifera’, Texas umbrella tree (foliage drooping, branches radiating); Swietenia, 1086.] BAILLON, H. Méliacées. Hist. Pl. 5: 470-508. 1874. [Melia Azedarach, 470, 471, 493; Swietenia, 478-480, 504; English transl., The natural history of plants 5: 470-508. 1878. BALAOGuN, A. M., & B. L. FeruGa. Fatty acid composition of seed oils of some members of the Meliaceae and Combretaceae families. Jour. Am. Oil Chem. Soc. 62: 529- 531. 1985. [Species in six genera of Meliaceae studied.] BARREIROS, H. S. pe. Cedrela (Meliaceae): Formas de crescimento. Taxonomia—lI. (English Abstr.) Arq. Jard. Bot. Rio de Janeiro 21: 135-139. 1977; IL. (English Abstr.) Rodriguésia 30: 253-277. 1978. [Features of tree architecture related to taxonomy; Bawa, K.S.,S.H. Buttock, D. R. Perry, R. E. CoviLie, & M. H. GRAYUM. pean biology of tropical lowland rain eee trees. II. Pollination systems. Am. Jour. Bot. 72: 346-356. 1985. [Moths pollinate Cedrela (1 sp., monoecious) and ue (3 Spp., eae and “small anal insects” pollinate Trichilia (1 sp., dioecious). ] , D. R. Per & J. H. Beach. Reproductive biology of oe lowland rain forest trees. I. Sexual oan and incompatibility mechanisms. Am. Jour. Bot. 72: 331-345. 1985. [Monoecy in Carapa (1 sp.) and Cedrela (1 sp.) and om in Guarea (6 ae ) and Trichilia (1 sp.).] ER. Spatial relationships between staminate and pistillate plants is pee tropical forest trees. Evolution 31: 64-68. 1977. [Guarea Luxii C. DC. 1 J. D. Sm. (= G. glabra Vahl); random distribution of staminate and ae ener in natural forest Hae in Costa Rica. BeEcKER, P., & M. Wong. Seed dispersal, seed predation, and juvenile mortality of Aglaia sp. (Meliaceae) in eee dipterocarp rainforest. Biotropica 17: 230-237. 1985. [Seeds disgorged after sarcotesta is detached in crop of hornbills, seeds remain viable after regurgitation; dispersal also by squirrels.] BENTHAM, G., & J. D. Hooker. Meliaceae. Gen. Pl. 1: 327-340. 1862. oon by Hooker; 37 genera in four tribes; Addenda and Corrigenda, 994, 995.] BERNARDO, F. A., C. C. JESENA, JR., & D. C. RAmireEz. Parthenocarpy and apomixis in Lansium Es Correa. Philip. Agr. 44: 415-421. 1961. [= L. parasiticum.] BOESEWINKEL, F. D. Development of the seed of Trichilia grandiflora Oliv. (Meliaceae). Acta Bot. Neerl. a 459-464. 1981. [Large seed size due to pachychalazy; sarcotesta derived from the outer integument and the chalaz BorcHERT, R. Phenology and control of Le in aa trees. Biotropica 15: 81- 89. 1983. [Incl. Cedrela mexicana M.J.R er (= C. odorata L., fide Pennington).] BREWBAKER, J. L. The distribution and Bae significance of binucleate and trinucleate pollen grains in the eer Am. Jour. Bot. 54: 1069-1083. 1967. [Binucleate and trinucleate types in Meliace Bucuincer, M., & R. FALCONE. Las Meliaceas tae Rev. Invest. Forest. 1: 9-58. 7 pls. 1957. aan Cabralea, a Trichilia BULLock, S. H., J. H. BEAcH, & K. AWA. Episodic flowering and sexual dimorphism in Guarea rhopalocarpa ina ee Rican rain forest. Ecology 64: 851-861. 1983. [Analysis of phenological observations made over two years.] 462 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 CANDOLLE, A. P. DE. Meliaceae. DC. Prodr. 1: 619-626. 1824. [Sixteen ee: in three tribes (Melieae, Trichilieae, Cedreleae); ee (7 spp.), Swietenia (3 sp CANDOLLE, C. DE. Méliacées. Monogr. Phan. 1: 399-752, 756-758. ie 4 9. 1878. [Thirty-five genera in four tribes (Melieae, Fein Swietenieae, Cedreleae); Me- lia Azedarach, pl. 6 (fig. 9); Swietenia neha ae ni, pl. 9 (fig. CARREIRA, L. M. M., & R. SEcco S. DE. orfologia polinica de plantas cultivadas no Parque do Museu Goeldi— IL eee ceae. Bol. Mus. Paraense Emilio Goeldi Bot. 1: 5-22. 1984. [LM and SEM of Swietenia Mahagoni, S. macrophylla, also Carapa, Cedrela, Guarea.] CHAKRABORTY, D. P. Family Rutaceae: a biochemical systematic viewpoint. Bull. Bot Soc. Bengal 18: 103-118. 1964. [Bitter constituents ally Rutaceae, Meliaceae, and Simaroubaceae. | CHAKRABORTY, T., & P. C. Darra. Chemical and botanical characters as aids to the taxonomy of Meliaceae. Bot. Soc. Bengal Sen Mem. Vol.: 437-454. 1969. Review and extensive bibliography. ] CHANG, C. The Meliaceae of Taiwan: its taxonomy and floristic relationships. Korean Jour. Pl. Tax. 18: 1-7. 1988. [Melia, Aphanamixis Blume, Aglaia, Dysoxylum, Chi- socheton. | CHANG, K. T., & F. H. WANG. Morphology of pollen grains of Meliaceae. (In Chinese; English a ) Acta Bot. Sinica 5: 253-265. 1956. [Incl. Melia and Swietenia, plus 10 other genera.] Co.ey, P. D. In ee variation in herbivory on two tropical tree eae Ecology 64: 426-433. 1983. [Trichilia Cipo A. Juss. and Cecropia insignis (Moraceae).] Corner, E. J. H. The seeds of dicotyledons. Vol. |. xii + 311 pp. Vol. 2. viii + 522 pp. Cambridge, London, New York, and Melbourne. 1976. [Meliaceae, Vol. 1, 185—- 193; Vol. 2, 316-331. Melia Azedarach, M. dubia, Vol. 1, 190, 191; Vol. 2, 327, 328. Swietenia, Vol. 1, 191, 192; Vol. 2, 329.] Cronquist, A. An integrated system of classification of flowering plants. Frontisp. + xvii + 1262 pp. New York. 1981. [Meliaceae, 813-815; Melia Azedarach, detailed illustration, 814; Meliaceae, Staphyleaceae, Sapindaceae, Hippocastanaceae, Acer- aceae, Burseraceae, Anacardiaceae, Simaroubaceae, Rutaceae, Zygophyllaceae, et al. in Sapindales DAHLGREN, R. M. T., S. ROSENDAL-JENSEN, & B. J. NIELSEN. A revised classification of the angiosperms with comments on apie as iced chemical and other char- acters. Pp. 149-204 in D. A. YounG & D. S. SEIGLER, eds., Phytochemistry and angiosperm eee New York. 1981. [References to Meliaceae throughout. ] Datta, P. C., . SAMANTA. Cytotaxonomy of Meliaceae. Cytologia 42: 197-208. 1977. (Oviginal data - seven spp. in six res Sane Melia Azedarach, 2n = 28, Swietenia Mahagoni, = 54, S. macrophylla, = 54; idiogram ms.] Davis, G. L. Systematic pees ap oo x + 528 pp. New York, London, & Sydney. 1966. [Meliaceae, 173, 174.] Doria, J. J. Neem: the tree insects ey Garden 5(4): 8-11. 1981. Duke, J. A. Keys for the identification of seedlings of some prominent woody species in eight forest types in Puerto Ri ico. Ann. Missouri Bot. ie 52: 314-350. 1965. [Melia Azedarach. 324; Swietenia Mahagoni: cryptocotylar, cophylls alternate, B17, fig. 73:8. eae ca cryptocotylar, eophylls opposite, 320, fig. 72; also Guarea and Trichilia On tropical tree seedlings. I. Seeds, edie. systems, and systematics. Ibid. 56: 125-161. 1969. [Meliaceae, incl. Carapa, Cedrela, Melia, Swietenia, Trichilia.] EIcHLer, A. W. Blithendiagramme construirt und erlaiitert. Vol. 2. Leipzig. 1878. [Meliaceae, 327, 328; incl. Melia Azedarach.] ErDTMAN, G. Pollen morphology and plant taxonomy. Angiosperms. (Corrected reprint + new addendum.) Frontisp. + xiv + 553 pp. New York. 1966. [Meliaceae, pollen of species in 14 genera, incl. Melia Azedarach (illus.) and Swietenia Mahagoni, described briefly, 268, 269.] 1990] MILLER, MELIACEAE 463 FisHER, J. B. Sun and shade effects on the leaf of Guarea (Meliaceae): plasticity of a branch analogue. Bot. Gaz. 147: 84-89. 1986. [G. Guidonia; growth of leaves under different light regimes yields plastic response in the development of sun or shade forms of sequentially produced leaflets GARUDAMMA, G. K. Studies in the Meliaceae I. Development of the embryo in Azadi- rachta indica A. Juss. Jour. Indian Bot. Soc. 35: 222-225. 1956. Il. Gametogenesis in Melia Azadirachta Linn. Ibid. 36: 227-231. 1957. [= Azadirachta indica.] GERSHENZON, J., & T. J. MABRY. Secondary metabolites and the higher classification of angiosperms. Nordic Jour. asta 3: 5-34. 1983. [Limonoids, the most useful tri- terpenoids in angiosperm taxonomy, unique to the Meliaceae, Rutaceae, and Cneo- raceae; the biochemically allied quassinoids only in the Simar a e.] as P. K., & S. K. Roy. Chisochetonoxylon bengalensis gen. et ., a new fossil ood of Meliaceae from the Tertiary beds of Birbhum District, West eae India. Curr Sci. Bangalore 48: 737-739. 1979. Gipss, R. D. Chemotaxonomy of flowering plants. Vol. 3. Pp. 1275-1980. Montreal and London. 1974. [Meliaceae, 1674, 1675, 1679-1685; chemistry summarized in Table 71; similar seed-fats in Burseraceae, Meliaceae, and Rutaceae. Grrarpi, A. M. M. [Grrarpi-Deiro, A. M.] Contribuga6 ao estudo de sage e anatomia foliar das Meliaceae do Rio Grande do Sul: I. Guarea Lessoniana A. Jus (camboata). Iheringia Bot. 18: 34-47. 1973.* Il. Trichilia elegans Juss. (p nw ervilha). (English abstr.) Bol. Soc. Argent. Bot. 16: 183-196. 1975. III. Trichilia Catigua A. Juss. (catigua). [heringia Bot. 20: 91-104. 1975.* IV. Trichilia Schu- manniana Harms, Trichilia Casaretti C. DC. (catigua-branco), Trichilia Hieronymi Griseb. (catigua-vermelho) e ane columnata A. M. Girardi (arco-de-paneira). (English Abstr.) /bid. 21: 81-101. 1975. GOTTSBERGER, G. Seed dispersal by a in the inundated ce of Humaita, Amazonia. Biotropica 10: 170-183. 1978 [Carapa guianensis, 174, GRAHAM, A. Ficus Ceratops Knowlton and its affinities nn — living genus Guarea. Jour. Paleontol. 36: 521-523. pl. 90. 1962. [Upper Cretaceous, Wyoming; three- dimensional casts, with two-layered pericarp visible in some fossils. udies in neotropical paleobotany. II. The Miocene communities of Veracruz, Mexico. Ann. Missouri Bot. Gard. 63: 787-842. 1977[1976]. [Upper Miocene pollen of Cedrela (fig. 151) and Guarea (figs. 152, 153).] _M. JarzEN. Studies in neotropical paleobotany. I. The Oligocene com- munities of Puerto Rico. Ibid. 56: 308-357. 1969. [Guarea pollen, 328.] Gray, A. I. Structural diversity and distribution of coumarins and chromones in the Rutales. Pp. 97-146 in P. G. WATERMAN & M. F. GRuNDON, eds., Chemistry and chemical taxonomy of the Rutales. London & New York. 1983. [Coumarins in Melia Azedarach and species in three other genera; no chromones reported from the Meliaceae. ] GrupMa, P., & B. T. StyLes. Bibliografia selectiva sobre Meliaceas. Centro Interam. Doc. Inf. Agr. L.1.C.A. Bibliog. 14. 143 pp. 1973. [Lengthy bibliography; all subjects included; indexed by genus and species. ] Groom, P. Excretory systems in the seas, xylem of Meliaceae. Ann. Bot. 40: 631- 649. pl. 20. 1926. [Produced by cambiu HaARBOoRNE, J. B. The flavonoids of the atice Pp. 147-173 in P. G. WATERMAN & . GRUNDON, eds., Chemistry and chemical taxonomy of the Rutales. London & ‘New York. 1983. [Common flavonol glycosides listed for five species ~ a genera) of Meliaceae; species in six other genera have flavonoids in one or a - bination of the following classes of compounds: methylated flavonols, cao uied flavones, flavanones.] Harms, H. Meliaceae. Jn: ENGLER & PRANTL, Nat. Pflanzenfam. III. 4: 258-308. 1896. Addenda in, Nachtrag und Register zu Teil II-IV, 208, 209. 1897; Erganzungsheft I, 36, 37. 1900; Erganzungsheft II, 188-190. 1906; Ergénzungsheft III, 161-163. 1914. [Forty-four genera in three subfamilies (Cedreloideae, Swietenioideae, Me- 464 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 lioideae), plus one genus of uncertain placement; ve i 274, 275 (S. Mahagoni illustrated); Melia, 286-288 (M. Azedarach illustrated).] . Meliaceae. Jn; ENGLER & PRANTL, Nat. spear ed. 2. 19bI: 1-172. 1940. [Fifty genera in three subfamilies; Swietenia, 70-74 (S. Mahagoni illustrated); Melia, 99-102 (M. Azedarach illustrated). ] Hermscu, C., Jk. Comparative anatomy of the secondary xylem in the Gruinales and Terebinthales of Wettstein, with reference to taxonomic grouping. Lilloa 8: 83-198. Howarp, R. A. Flora of the Lesser Antilles, Leeward and Windward Islands. Vol. 4. Dicotyledoneae—Part |. 673 pp. Jamaica Plain, Massachusetts. 1988. [Meliaceae, 581-596; incl. Azadirachta, Carapa, Cedrela, Guarea, Melia, Swietenia, Trichilia.] Howe, H. F., & D. De STEVEN. Fruit production, migrant bird visitation, and seed dispersal of Guarea glabra in Panama. Oecologia 39: 185-196. 1979. [Seed dispersal by resident and migratory birds. ] Jussieu, A. pE. Mémoire sur le groupe des Méliacées. Mém. Mus. Hist. Nat. Paris 19: 153-304. pls. 12-23. 1832 [1830]. [Meliaceae (28 genera in two groups, rank not given); Pee (8 genera in two groups, rank not given); new names published in Bull. Univ. Sci. Industr. Sect. 2 (Bull. Sci. Nat. Géol.) 23: 234-241. 1830.] KHOosLA, P. K.., & B. T. Sryies. Karyological studies and ee evolution in Meliaceae, Silvae Genet. 24: 73-83. 1975. [Two series based on x and x = 7 established for the sare pein chromosome races in Soieieni i KOENIGUER, J. C., & P. Lou Sur la présence d’un bois de Méliacées ae le Tertiaire u Fezzen oriental: ie Boureaui Louvet. Palaeobotanist 17: 33-35. ] pl. 1968. [Permineralized; Eo-Oligocene of Libya.] KOSTERMANS, A. J.G. H. A monograph of Agiaia, sect. Lansium Kosterm. (Meliaceae). Reinwardtia 7: 221-282. 1966. [Subfam. Mende. 15 spp. of the Indo-Malayan region.] Kriss, D. A. C 1 t of the woods of Meliaceae. Am. Jour. Bot. 17: 724— 738. 1930. [Key to 33 genera based o on wood structure; suggests recognizing Swie- teniaceae for genera of subfam. Swietenioideae on the basis of uniformity of ana- tomical and morphological characters. ] Lee, H. Y. Study on the thyrse, a mixed inflorescence. Taiwania 13: 131- 145. 1967. [Inflorescences of Melia A which the cymules are three-flowered dichasia, terminal flower perfect i in Melia, carpellate in Swietenia. | Leroy, J.-F. Contributions 4 l’étude des foréts de esa ean Jour. Agr. Trop. Bot. Appl. 7: 455, 456. 1960. [Ptaeroxylaceae J. F. Leroy, fam. . Essais de taxonomie syncrétique 1. Etude sur les Mielnceae de Madagascar. (English abstr.) Adansonia II. 16: 167-203. 1976. [Khaya, Neobeguea J. F. Leroy, Capuronianthus, Xylocarpus Koenig, eee Neomangenotia J. F. Leroy; mor- phology, habit development, ecology, phylogeny.] Lersten, N. R., & R. W. POoHLt. oo ates in Cipadessa (Meliaceae). Ann. Bot. I. 56: 363-366. 1985. [On leav Litt.e, E. L., Jr. Checklist of United ate trees (native and naturalized). U. S. Dep. Agr. Forest Service Agr. Handb. 541. iv + 375 pp. 1979. [Melia Azedarach, 172; Swietenia Mahagoni, 280, 281.] & . WaApsworTH. Common trees of Puerto Rico and the Virgin Islands. Vol. |. U. S. Dep. Agr. Handb. 249. x + 548 pp. 1964. [Meliaceae, including Melia Azedarach, Swietenia Mahagoni, S. macrophylla, 242-255. Louver, P. Sur les affinités des flores tropicales ligneuses africaines Tertiaire et actuelle. Bull. Soc. Bot. France 120: 385-395. 1973. [Paleoecological inferences (e.g., distri- bution of tropical rain forest and savannas) based on fossil woods; incl. several Meliaceae in the form-genera Entandrophragmoxylon and Lovoaxylon.] 1990] MILLER, MELIACEAE 465 Sur trois eee oe du Tertiaire de Libye. Ibid. 121: 269-280. 1975 [1974]. [Petrified wood na j Magnieri (Oligocene) similar to wood of aot il elie Candollei Harms] Lussock, J. A contribution to our knowledge of seedlings. 2 vols. New York. 1892. [Meliaceae, 1: 334-337; seedling of Melia Azedarach illustrated. MAaBBERLEY, D. J. Meliaceae. Pp. rae 202 in V. H. HEyYwoop, consultant ed., Flowering plants of the world. New York. . The species of Chisocheton ee) Bull. Brit. Mus. Bot. 6: 301-386. 1979. [Fifty-one spp. of the Indo-Malayan region; many general notes of biological interest; incl. Megaphyllaea Hemsley] MacainitTig, H. D. The Kilgore flora, a late Miocene flora from northern Nebraska. Univ. Calif. Publ. Geol. Sci. 35: 67-158. 16 pls. 1962. [Cedrela Trainii Arnold, leaflets and winged fruits, 114, p/s. 3, 6, 7.] , E. B. Leopotp, & W. L. RoHReER. An sre mire Eocene flora from the Yellowstone Absaroka Volcanic Province, northwestern Wind River Basin, Wyo- ming. [bid. 108. 103 pp. 45 pls. 1974. [Cedrela Saas (Lesquereux) MacGinitie, 74, pl. 16, impression fossils of leaflets; also pollen identified as Cedrela cf. mexi- cana. MApe., E. MahagonihGlzer der Gattung Carapoxylon n. g. (Meliaceae) aus dem euro- paischen Tertiar. Senckenberg. Lethaea 41: 393-421. 1960. [Based on structurally preserved wood from the Upper Miocene of southwestern Germany. MANGENCT, S., & G. MANGENOT. Nombres chromosomiques nouveaux chez diverses dicotylédones et monocotylédones d’ Aftique occidentale. Bull. Jard. Bot. Bruxelles 27: 639-654. 1957. [Fourteen species in seven genera; endopolyploidy in Entan- drophragma angolense C. DC.] Martin, A.C. The comparative internal morphology of seeds. Am. aa fen 36: 513- 660. 1946. Ai ees 618, 619, 646; incl. Melia Azedarach, Swietenia.] Matupba, E. Meliaceas de Chiapas. Anal. Instit. Biol. (México) 19: 407-425. 1948. [Melia, nee Trichilia, Guarea, Cedrela.] Meeusg, A. D. J. The concept of the Rutales. Pp. 1-8 in P. G. WATERMAN & M. R. GRUNDON, eds., Chemistry and chemical taxonomy of the Rutales. London & New York. 1983. [Rutales and Sapindales distinct on the basis of feeding behavior of swallowtail oe larvae (superfamily Papilionidae). Meura, P. N., T. S. SAREEN, & P. K. KHosLA. Cytological studies on Himalayan Me- liaceae. Jour Amold Arb. 53: 558-568. 1972. [Melia Azedarach, n= 14 (two sources); M. composita, n = 14, M. Toosendan, n = 14; counts in 15 other species in nine ] Mester, I. Structural diversity and distribution of alkaloids in the Rutales. Pp. 31-96 in P. G. WATERMAN & M. F. GRuNDON, eds., Chemistry and chemical taxonomy of the Rutales. London & New York. 1983. [Alkaloids reported from five members of the Meliaceae.] METCALFE, C. R., & L. CHALK. Meliaceae. Anat. Dicot. 1: 349-358. 1950. [Leaves, axes, bark, wood, roots; extensive bibliography. ] & Anatomy of the dicotyledons. ed. 2. Vol. 1. Oxford. 1979. [Epirachial flowers and inflorescences, sac domatia, interxylary cork.] Minrray, E. Contribution a l'étude caryo-taxinomique des Méliacées. Bull. Soc. Bot. France 110: 180-192. 1963a. [Ten species in eight genera, incl. Melia and Swietenia. } e noyau et les chromosomes somatiques de deux Méliacées. Bull. Mus. Hist. Nat. Paris IL 35: 527-531. 1963b. [Neobeguea, Carapa.] Mirra, C. R. Neem. [v] + 190 pp. /7 pls. Hyderabad. 1963. [Azadirachta indica; medicinal uses and chemistry. ] Morton, J. F. Atlas of medicinal plants of Middle America, Bahamas to Yucatan. XXvill + 1420 pp. Springfield, Illinois. 1981. [Melia Azedarach, Swietenia Mahagoni, S. macrophylla, 403-407.] 466 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 MuL ter, J. Fossil pollen records of extant angiosperms. Bot. Rev. 47: 1-142. 1981. [Meliaceae, 69, 70 Murty, Y. S., & S. Gupta. Morphological studies in Meliaceae. I]. A reinvestigation of floral anatomy of members of Swietenieae and Trichilieae. Proc. Indian Acad. Sci. B. 87: 55-64. 1978. [Swietenia Mahagoni, also oe of Soymida A. Juss., Chukrasia A. Juss., Dysoxylum, eeeaeehees and 7ri Nair, N. C. Early endosperm development in Meliaceae. Sci ‘Culture 22: 34, 35. 1956. [Melia Azedarach; Azadirachta, Ci Oe pees a.] Studies on Meliaceae I. Floral morphology and embryology of Naregamia alata W. & A. Jour. Indian Bot. Soc. 38: 353-366. 1959. [Flowers, micro- and megaspo- ae apes Napa etes seeds.] I. Floral morphology and embryology Melia Azedarach Linn.—a reinvestigation. [bid.: 367-378. 1959. [Formation of aici (2-5) eta opie in some ovules; polyembryony; triple fusion ob- served; numerous other well documented details.] III. Floral morphology and em- bryology of Sandoricum indicum Cav. Phyton Argentina 10: 145-151. 1958. V. Morphology and anatomy of the flower of the tribes Melieae, Trichilieae and Swie- tenieae. Jour. Indian Bot. Soc. 41: 226-242. 1962. [Incl. Melia and Swietenia.] V1. Morphology and anatomy of the flower of the tribe Saeed and discussion on the floral anatomy of the family. Ibid. 42: 177-189. 3. & K. Kanta. Studies in Meliaceae IV. Floral eereee and embryology of Azadirachta indica A. Juss.—a reinvestigation. [bid. 40: 382-396. 1961. NarRAYANA, L. L. Floral anatomy and embryology of Cipadessa baccifera Bia Jour. Indian Bot. Soc. 37: 147-154. 1958a. [Incl. micro- and megasporogene . Floral anatomy of the Meliaceae. I. Ibid. 37: 365-374. 1958b. (Melia, een Swietenia eae 365-369; also Cedrela, Walsura Roxb., Aglaia.] Il. Ibid. 38: 288-295. 1959. [Turraea, Soymida, Heynea Roxb. ex Sims (= Trichilia); Chlorox- ylon DC. cane NETOLITZKY, F. Anatomie der Angiospermen-Samen. Handb. Pflanzenanat. IJ. Arche- 182. PageTow, W. Embryologische Untersuchungen an Taccaceen, Meliaceen und Dilleni- aceen. Planta 14: 441-470. 1931. De ramiflorum Miq.] PANDEY, Y. N. Studies on the cuticular characters of some Meliaceae. Bull. Bot. Surv. India 11: 377-380. 1972 [1969]. [4zadirachta indica, Melia Azedarach, M. Bir- manica Kurz, Swietenia Mahagoni, S. macrophylla, Soymida febrifuga A. Juss., Cedrela Toona Roxb.; M. Azedarach vs. M. Birmanica & S. Mahagoni vs. S. macro- phylla distinguished on the basis of epidermal characters. ] PANNELL, C. M., & M. J. Kozior. Ecological and phytochemical diversity of arillate seeds in Aglaia (Meliaceae): a study of vertebrate dispersal in tropical trees. Philos. Trans. Roy. Soc. London B. 316: 303-333. 1987. [Ten spp.; dispersal by birds, primates, and civet.] PANSHIN, A. J. Comparative anatomy of the woods of the Meliaceae, sub-family Swie- tenioideae. Am. Jour. Bot. 20: 638-668. pls. 37-40. 1933. PENNINGTON, T. D. Materials for a monograph of the Meliaceae I. A revision of the genus Vavaea. Blumea 17: 351-366. 1969. [Subfamily Melioideae; four spp. in two sections; ee wood anatomy; Malayan region (Sumatra to Fiji ceae. Fl. Neotrop. Monogr. 28: 1-449, 462-470. 1981. "(Swietenioideae by B. T. Cee 359-418; Melia Azedarach, 24, 25; Swietenia, 3 spp., plus putative hybrids, 391-406.]} RUKHAN. Manual para la identificacion de campo de los principales arbores tropicales de México. vii + 413 pp. Instituto Nacional de Investigaciones Forestales. 1968. [Cedrela odorata L., Ce glabra, Melia Azedarach, Swietenia macrophylla, Trichilia havanensis, all illustrated, 238-247.] & B. T. eneric monograph of the Meliaceae. Blumea 22: 419-540. 1975. [Introduction by F. Wuite, 419-422: 51 genera in four subfamilies (two 1990] MILLER, MELIACEAE 467 monotypic); comprehensive literature survey; much new information about wood anatomy, floral morphology, and palynology; Melia (Melioideae, Melieae), 463, Swietenia (Swietenioideae, Sctenioad) 521, 523.] Poprenoge, W. Manual of tropical and subtropical fruits. xv + 474 pp. New York. 1920. [Sandoricum Koetjape, Lansium domesticum (= L. parasiticum), 426-428. RECORD, S. J. Mahogany and some of its substitutes, a descriptive key based on gross and lens characters. Jour. Forestry 17: 1-8. 1919. [Thirteen families and 27 genera, including 11 genera in the Meliaceac.] ———. American timbers of the ican family. Trop. Woods 66: 7-33. 1941. [Seven ee Wide incl. notes on harvesting mahogany in South America, 19-31.] s on tropical timbers. /hid. 80: 1-6. 1944. [Wood of Swietenia and Cedrela dsenestned complete ring of parenchyma forms annually in S. macrophylla, there- fore age determinations are possible.] Ripiey, H. N. The dispersal of plants throughout the world. xx + 744 pp. + /6 pis. Ashford, Kent. 1930. [A4zadirachta, 348; Melia, 477, 482, 487; Swietenia, 120.] Rortn, I. Estructura anatomica de la corteza de algunas ‘especies arboreas venezolanas de Meliaceae. Acta Bot. Venezuela 6: 239-259. 1972. [Carapa, Cedrela, Trichilia.] SAHNI, K. C., & S. S. R. BENNET. Correct name of ‘langsat.’ Indian Forester 100: 202. 1974. [Lansium parasiticum (Osbeck) Sahni & Bennet.] SCHMUTTERER, H., & K. R. S. AscHer, eds. Natural pesticides from the neem tree (Azadirachta indica A. Juss.) and other tropical plants. Proc. 2nd Internatl. Neem Conf., Deutsche Ges. fiir Technische Zusammenarbeit. 587 pp. ScHoiz, H. Meliaceae. Jn: H. MELCHIOR, A. ENGLER’s Syllabus der Pflanzenfamilien. ed. 12. 2: 270-272. 1964. [Three subfamilies recognized.] SEIGLER, D. S. Terpenes and plant phylogeny. Pp. 117-148 in D. A. Younc & D.S SEIGLER, eds., Phytochemistry and angiosperm phylogeny. New York. 1981. [Bio- genesis of limonoids and derivative compounds indicates that the Meliaceae and Simaroubaceae arose from Rutaceae-like ancestors. SELMEIR, A. Carapoxylon ortenburgense n. sp. (Meliaceae) aus dem untermiozanen Ortenberger Schotter von Rauscher6d (Niederbayern). Mitt. Bayer. Staatssam. Pa- laontol. Hist. Geol. 23: 95-118. 1983. panes details 1 in petrified wood com- parable to those eal and FE 2 t Meliaceae); lower Miocene of southern Germany. ] Cana n. gen. (Meli Evy se sekundarer Lagerstatte von Seibersdorf am Inn (Bayern). /bid. 27: 123-144. 7. [Anatomical details in petrified wood comparable to those of Cedrela; spice ] Sttva, M. F. DAs G. F. pa, & O. R. GoTTLies. Evolution of quassinoids and limonoids in the Rutales. Biochem. Syst. Ecol. 15: 85-103. 1987. eigen eet peepee among and within Simaroubaceae, Rutaceae, Cneoraceae, and Meliaceae on t basis of indices of skeletal specialization and oxidation state of limonoid and quas- sinoid sas (triterpenoids). ] ——— . L. Dreyer. Evolution of limonoids in the Meliaceae. /bid. 1 299- 310. tee [Limonoid chemistry correlates with the subfamily sae scheme of Pennington & Styles; pathways of limonoid synthesis outlined. Skutcu, A. F. A compound leaf with annual increments of growth. Bull. Torrey Bot. Club 73: 542-546. 1946. [Guarea rhopalocarpa; rachises of compound leaves ter- minated by “resting buds;”’ 2 or 3 leaflet pairs produced each year.] Smit, C. E., Jr. A revision of Cedrela (Meliaceae). Fieldiana Bot. 9: 295-341. pls. 7- 14. 1960. [Six species, plus one of uncertain status, in Caribbean region, Central and South America; summary of economic uses, paleobotany, morphology, history of the genus. ] . Flora of Panama. Part VI. Family 92. Meliaceae. Ann. Missouri Bot. Gard. 52: 55-79. 1965. [Melia Azedarach (cult.), Swietenia macrophylla (“thoroughly har- vested”’); Cedrela, Carapa, Trichilia, Guarea.] 468 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 STEINGRAEBER, D. A., & J. B. FisHeR. Indeterminate growth of leaves in Guarea (Me- liaceae): a twig analogue. Am. Jour. Bot. 73: 852-862. 1986. [Anatomical evidence that pine of G. Guidonia and G. glabra, while analogous to branches, are leaf hom oo. STYLES, “ T. he flower biology of the Meliaceae and its bearing on tree breeding. Silvae ae 21: 175-181. 1972. [Review;, dichogamous staminate and carpellate flowers in Swietenia spp.; perfect and staminate flowers in Melia Azedarach, few data available on pollinators of any Meliaceae.] . KHosLa. Cytology and aes biology of Meliaceae. Pp. 61-67 in J. Burey & B. T. Sry.es, eds., Tropical pee variation, breeding and conservation. inn. Soc. Symp. Ser. 2. xv + 243 pp. 1976. [Mentions program to improve Melia Azedarach, hybrids ioe species of cee intraspecific chromosome races in Swietenia spp., considers bees and moths the main pollen vectors. ] &C Chromosome numbers in the Meliaceae. Taxon 20: 485-499. 1971. [Fifty-eight spp. in 30 genera, incl. reports for Melia (2 spp.), Swietenia (3 Spp.). TAKHTAJAN, A. Systema magnoliophytorum. (In Russian.) 439 pp. Leningrad. 1987. [Meliaceae, Rutaceae Birseraceae Px) TaAyYLor, D. A. H. Chemotaxonomy, the occurrence of limonoids in the Meliaceae. FI. Neotrop. Monogr. 28: 450-459. 1981. [Extensive bibliography. ] . Biogenesis, distribution, and systematic significance of limonoids in the Meli- aceae, Cneoraceae, and allied taxa. Pp. 353-375 in P. G. WATERMAN & M GRuNDON, eds., Chemistry and chemical taxonomy of the Rutales. London & New York. 1983. The chemistry of the limonoids from Meliaceae. Fortschr. Chem. Org. Natur- stoffe 45: 1-102. 1984. THORNE, R. F. Phytochemistry and angiosperm phylogeny, a summary statement. Pp. 233-295 in D. A. YounGc & D. S. SEIGLER, eds., Phytochemistry and angiosperm phylogeny. New York. 1981. [Meliaceae, Rutaceae, Simaroubaceae, Burseraceae, Anacardiaceae, Leitneriaceae ef a/. in suborder Rutineae, with suborder Sapindineae et al. comprising Rutales, superorder Rutiflorae TOMLINSON, o B. The botany of mangroves. xil + 413 pp. Cambridge. 1986. [Xylo- carpus, 2 spp., incl. original observations of vegetative structure and biology of X. granatum Koenig; pneumatophores. Umapevt, I., M. DANIEL, & S. D. SABNIS. Chemosystematics of some Indian members of the family Meliaceae. Feddes Repert. 99: 195-197. 1988. [Melia Azedarach, M. composita (= M. Azedarach, fide Mabberley, 1984 (under Melia references)), Swie- tenia Mahagoni, S. macrophylla, plus species in | 1 other genera (in all four subtribes) surveyed; most contained flavonols and proanthocyanins. | VAUGHAN, J. G. The ae and utilization of oil seeds. xv + 279 pp. London. 1970. [Melia Azedarach, 156, 157, 159; Azadirachta, Carapa, Trichilia, Amoora Roxb. (= Aglaia WATERMAN, P.G. Alkaloids of the Rutaceae: their distribution and systematic signifi- cance. Biochem. Syst. Ecol. 3: 149-180. 1975. [Relationships of Meliaceae, Sima- roubaceae, and Rubiaceae as indicated by limonoid, coumarin, alkaloid and chro- mone chemistry, 174-176.] WEBERLING, F., & P. W. LEENHOUTS. Systematisch-morphologische Studien an Tere- binthaies-Familien (Burseraceae, Simaroubaceae, Meliaceae, Anacardiaceae, Sap- indaceae). (English summary.) Abh. Akad. Wiss. Mainz Math.-Naturw. 10: 495- 584. 1966 [1965]. [Pseudostipules (modified basal leaflets) in Trichilia and Toona.] White, F. Long distance dispersal, overland migration and extinction in the shaping of tropical African floras. Bothalia 14: 395-403. 1983. [Carapa, Xylocarpus, 400, 401.] The taxonomy, chorology and reproductive biology of southern African Meli- aceae and Ptaeroxylaceae. /bid. 16: 143-168. 1986. [Melia Azedarach, 154, 155.] 1990] MILLER, MELIACEAE 469 B. T. Styves. Meliaceae. Fl. Zambesiaca 2: 285-319. 1963. [Ten genera; Melia 4zedarach, 315 Wiis on, P. Meliaceae. N. Am. FI. 25: 263-296. 1924. [Melia Azedarach, Swietenia (four spp.), plus five other genera.] KEY TO THE GENERA OF MELIACEAE IN THE SOUTHEASTERN UNITED STATES General characteristics: Small to large trees; leaves alternate, exstipulate, once or twice odd-pinnate or even- ee plants monoecious or polygamous; i axillary thyrses,; flowers perfect, staminate, or carpellate (if ee staminate and c Lisa then dimorphic), regular, een epals free or fused basally; petals re stamens hy- pogynous, united into a fringed ie or tube Cie in hoe teeth, anthers oe wider than style; ovules two, superposed, or many in 2 rows; fruit a drupe or a septicidal capsule splitting from base to apex; seeds remaining in endocarp or seeds free and winged. A. Leaves once to twice odd-pinnate, leaflets serrate; flowers large, showy, perfect and staminate flowers isomorphic; sepals separate; staminal tube fringed with narrow teeth; stigma rounded, as wide as the elongate style; nectariferous disc inconspicuous; TRU APCIUDS xuenneciesoe de camacenseeakes Nees Hata Ieee 1. Melia. A. Leaves even-pinnate, leaflets entire; flowers small, whitish, staminate and carpellate flowers dimorphic; sepals fused basally, staminal tube terminating in a ring of deltoid teeth; stigma discoidal, style narrow, short or long, nectariferous disc conspicuous, orange; fruit a septicidal capsule ......00 2.000. eee 2. Swietenia. Subfam. MELIOIDEAE [Harms in Engler & Prantl, Nat. Pflanzenfam. III. 4: 267. 1896.] 1. Melia Linnaeus, Sp. Pl. 1: 384. 1753; Gen. Pl. ed. 5. 182. 1754. Small [to large] trees [or shrubs]; branch apices dying back, new chee initiated from axillary buds; bud scales stellate pubescent. Leaves deciduou once to twice, rarely thrice, odd-pinnate, a pair of glands (extrafloral nectaries? located on new branches near leaf insertions; young leaves with a mixture of dendritic (or stellate) hairs and simple, hooked hairs; leaflets petiolulate, mostly symmetrical at base, acuminate, serrate. Plants polygamous; inflorescences borne in the axils of the early leaves, terminal flower of cymule perfect, lateral flowers staminate; perfect and staminate flowers similar at anthesis. Sepals 5(6?), mostly free. Petals 5(6?), weakly [or strongly] pubescent abaxially, alter- nate with the sepals. Staminal tube cylindrical, outer surface smooth or with linear appendages, inner surface with long hairs, tube 20(24?)-toothed; anthers 10(12?), sessile, basifixed, inserted inside the tube opposite pairs of teeth, as long as the teeth and bent inward over the stigma at anthesis, connective slightly prolonged; pollen + prolate, exine smooth to slightly scabrate. Ovary 5- or 6- locular, each locule with 2 superposed, anatropous ovules, style long-cylindri- see for lobes erect]; nectariferous disc annular, obscure, entirely below the vary. Fruit a + globose [or ovoid] drupe, endocarp spheroidal [or narrowly ellipsoidal], keeled, 5- or 6-locular, or locules fewer by abortion; one, rarely 470 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 TSS or vob ss a”. of ADC Ficure 1. Melia. a-g, M. Azedarach: a, inflorescence and leaf, x 4; b, flower, x 3; c, staminate flower in partial vertical section (petals and two sepals removed), showing gynoecium, nectariferous disc, staminal tube, ae and anthers, x 4; d, fruit (drupe), showing scars of staminate flowers on pedicel, x 1; e, fruit in diagrammatic cross section, showing 6-locular stone, 2 locules without seeds, endosperm stippled, embryos with 2 or 3 cotyledons, x 2; f, stone, from side, x 2; g, stone, from above, x 2 two seeds per locule. Seed coats thin, brown; embryo ellipsoid, cotyledons flat, longer than broad, plumule minute; endosperm conspicuous, fleshy, oily. Lecrotype species: M. Azedarach L.; see N. L. Britton, N. Am. Trees 593. 1908. (Ancient Greek name for manna ash [fraxinus Ornus L.].)—CHINABERRY. A small genus indigenous to temperate, subtropical, and tropical regions of Asia and Africa, with one species, Melia Azedarach, chinaberry, China tree, pride of India, Carolina mahogany, 2” = 28, introduced into the Americas where it is now widely naturalized. The original range of M. Azedarach may be impossible to ascertain because the species occurs throughout a large part of warm-temperate and tropical Asia. Hiern (in Hooker) noted that MM. Azeda- rach was “wild in the sub-Himalayan tract, alt. 2-3000 ft.”’; Rechinger reported that M. Azedarach was spontaneous in the western Himalayas. Other authors (e.g., Coode & Cullen) concluded that it is native to India and China. Mabberley has recently contended that its native range encompasses portions of the area from Nepal, India, Burma, and southern China, through parts of the Malay Archipelago to New Guinea, the Solomon Islands, and tropical Australia (see discussion that follows). The introduction of Melia Azedarach into the United States is credited to André Michaux, who is said to have grown it in his garden near Charleston, 1990] MILLER, MELIACEAE 471 South Carolina, in the late eighteenth century. At the close of the second decade of the 1800’s, F. A. Michaux reported that M. Azedarach had become abundant in coastal areas of the southern United States, and, about the same time, Elliott, in reference to South Carolina and Georgia, wrote that it was “‘perfectly nat- uralized” and ‘“‘springing from seed in cultivated land and around enclosures.” It is now widely grown and self-seeding throughout our area except in the mountains. Mabberley suggested that plants in North America seemed to have at least two distinct origins, - from Indian plants via the Middle East and from Chinese plants via Japa The cultivar ‘Umbraculifers’, Texas umbrella tree, 27 = 28, with a dense, flattened crown of foliage and the main branches radiating from the trunk like the supports of an umbrella, was first observed in Texas (where it may have originated). Most of the naturalized trees are of a much more open form. The Texas umbrella tree is widely planted in the southeastern United States. Propagation of Melia Azedarach for horticultural purposes is from seeds or cuttings. The seeds germinate while they are still enclosed in endocarps, and one fruit may produce up to four seedlings. Germination percentages are usually high. Precocious flowering (sometimes even in the seedling stage) has occa- sionally been observed (van Steenis). The circumscription of Melia is uncertain. Most authors state that the genus consists of 15 or fewer species, but it may actually contain about five species (Pennington & Styles, family references), or probably even a smaller number. Jacobs emphasized several important differences between Melia and Azadi- rachta A. Juss., which are sometimes treated as congeneric. The latter consists of two species, A. indica A. Juss. (M. Azadirachta L.), neem, a tree cultivated throughout India and held sacred by the Hindus, and A. excelsa (Jack) Jacobs of the Indo-Malayan region. Some of the characters that distinguish Me/ia and Azadirachta are: leaves twice- or thrice-pinnate vs. once-pinnate; large extra- floral nectaries(?) near petiole bases, one pair (both circular) vs. two pairs (one pair circular, the other linear); ovary 4—8-locular vs. 3-locular; style broad, stigma 4—6-lobed vs. slender and 3-lobed; and ovules superposed vs. collateral. However, Corner (family references) questioned whether Azadirachta, Melia, and Cipadessa Blume have been distinguished satisfactorily. He also mentioned that seeds of species of Melia and Cipadessa are similar. Flowers and fruits of Melia Azedarach have been illustrated and described repeatedly. Less well-known species of Melia are included in various standard floras and other works that treat tropical or subtropical Asia or Africa. While numerous names exist for the considerable morphological diversity presented by the Asian plants, Mabberley concluded that only one polymorphic species, M. Azedarach, exists in that region. He proposed an informal infraspecific classification for M. Azedarach that consists of three categories, wild plants (incl. M. dubia Cav., M. composita Willd., M. australasica A. Juss., and other synonyms), Chinese cultivars (M. Toosenden Sieb. & Zucc. and other syn- onyms), and Indian cultivars (numerous synonyms). Many of these names have been used in floras or to document phytochemical or other investigations into the biology of Melia. The two groups of cultivars originated in different parts of Asia through selection for desirable horticultural qualities. 472 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 Wild plants of Melia Azedarach occur in forests in an area including India and southern China, southward through the Indochinese Peninsula, parts of Malesia to New Guinea and tropical Australia, where they sometimes are large trees (to 40 m). Wild plants, which have larger leaflets (to 6 cm long) and smaller flowers, are evidently not hardy in cool-temperate regions. In contrast, plants naturalized or cultivated in temperate regions have smaller, serrate or lobed, usually glabrous leaflets and large bluish, pink, or white flowers. The chinaberry of the southeastern United States is presumably partly derived from plants introduced into Europe from the Middle East, which in turn are believed to have come at least as early as the 9th century B.C. from plants of Indian origin (Mabberley). Cultivars selected in Japan (from putative Chinese stocks) have been a second source of introductions into European and North American horticulture. A name often applied to Mabberley’s concept of wild plants of Melia Azeda- rach is M. dubia Cav.; flowering and fruiting specimens so-named in the her- barium of the Arnold Arboretum are markedly distinct from Melia Azedarach as it is represented in the Southeast. Mabberley has added M. dubia and several other names to the synonymy of M/. Azedarach, citing his inability to correlate differences in flower color and leaf characters with geography. It would be interesting to know if the “‘wild’’ plants are reproductively isolated from those in cultivation, in view of the fact that hybrids are known among at least some of the cultivars. Melia Azedarach is widely naturalized in Africa, and one or perhaps two other species appear to be indigenous to that continent. Melia Bombolo Welw. is reported from West Africa (Sénégal, Gabon, Democratic Republic of the Congo, and Angola; Staner & Gilbert) and MM. Volkensii Guerke from tropical East Africa (Uganda; Giirke). The relationship between these species and M. Azedarach s. |. is unclear. The genus obviously needs monographic study. The sole representative of the genus in the Southeast, Melia Azedarach, is reported to be polygamous on the basis of observations of trees growing in Taiwan (Lee, Styles (1972), family references). Individual cymules in an inflo- rescence are mostly three-flowered dichasia in which the terminal flower is perfect (and the first to open) and the two lateral flowers are staminate (and caducous following anthesis). Sometimes all three flowers are staminate. Perfect and staminate flowers are indistinguishable at anthesis, but the fate of indi- vidual flowers can be followed by observing the pattern of fruit set. My ex- amination of herbarium specimens collected in the Southeast revealed that fruits are usually at the ends of pedicels that bear opposite scars (which represent the places where the staminate flowers were attached). Polygamy should be confirmed in our area by observations of plants in flower. Plants of Melia Azedarach have many uses, although the species is not of great commercial importance. The wood has been used in cabinets, furniture, and cigar boxes, and in the manufacture of fiberboard. Pulp from the wood has been made into various kinds of paper in India. Styles & Khosla (family references) consider 4. Azedarach to be a “species of enormous forestry po- tential,’ and report that a program to improve the species genetically is under way in Argentina. 1990] MILLER, MELIACEAE 473 A decoction of bark from the roots is reputed to be an effective vermifuge, probably reflecting the presence of vanillic acid (Chiang & Chang). Other parts of the plant are also used occasionally in folk medicine. The seeds contain about 40 percent oil. The endocarp beads, for example, in rosaries. A triterpenoid, azadirachtin, originally isolated from the fruits of Azadirachta indica, but also found in the fruit of M. Azedarach (Morgan & Thornton), inhibits feeding in desert locusts (Schistocera gregaria). Other kinds of insects are repelled by extracts of the plant. Fermented fruits of M. Azedarach were used as a source of alcohol during the American Civil War (Mabberley). Fruits and leaves of Melia Azedarach are reported to be poisonous to humans and certain domestic animals (Carratala, Kwatra ef al.). However, the poison- ous principle has been elusive. A toxic alkaloid, tazetine, has been found in the bark and fruit (Morton, family references), and the presence of an alkaloid, azedarine, is mentioned by Carratala. The work of Morrison indicated that the toxicity of the fruit derives from an unidentified alkaloid, which acts in concert with a resin. However, Schulte and coworkers did not detect toxic substances in the fruits of 4. Azedarach. Oelrichs and colleagues isolated four limonoids (meliatoxins) from the flesh of fruits from trees in Queensland. These proved toxic to pigs and mice in clinical trials. Birds and fruit bats play a role in the dispersal of Melia Azedarach (White, 1986). In North America the robin (7urdus migratorius migratorius) 1s reported to eat quantities of the fruit (Beal, Elliott). REFERENCES: Under family references see ABDULLA; BAILEY et al.; BAILLON; BENTHAM & HOOKER; A. DE CANDOLLE; C. DE CANDOLLE; CORNER; CRONQUIST; DATTA & SAMANTA; DUKE, 1965, 1969; EICHTER: ERDTMAN; GRAY; GRIJPMA & STYLES; HARMS, 1896, 1940; HOowARD; Nair, 1959b, 1962; NARAYANA, 1958b; PANDEY; PENNINGTON, 1981; PENNINGTON & SARUKHAN; PENNINGTON & STYLES; RIDLEY; SCHOLZ; SMITH, 1965; STYLes; STYLES & KHOSLA; STYLES & VosA; TAYLOR, 1981, oo 1984; UMADEVI et al.; VAUGHAN; WHITE, 1983, 1986; WHITE & STYLES; and WILsSO ALEXANDER, E. J. Melia Azedarach. Addisonia 12: 17, 18. pl. 393. 1927. ANONYMOUS [C. S. SARGENT?]. The pride of China tree. Garden Forest 7: 92, 95. 1894. [M. Azedarach, “. . . introduced into the United States about a hundred years ago by the French botanist Michaux .. .”; umbrella tree (var. umbraculifera) “supposed ANONYMOUS. [INSTITUTUM BOTANICUM, ACADEMIA SINICA, ed.] Iconographia Cormo- phytorum Sinicorum. Vol. 2. Papaveraceae-Cornaceae. iv + 1312 pp. Peking. 1972. [Melia, 3 spp., 566, 567. AusTIN, D. R. Exotic plants and their effects in southeastern eas Environ. Conserv. 5: 25-34. 1978. [M. Azedarach naturalized in three countie Backer, C. A., & R. C. BAKHUIZEN VAN DEN BRINK, JR. Flora af faee Vol. 2. 1v + [l- 72] + 641 pp. + foldout map & table. Cea 1965. [Melia, 3 spp., 120.] BALL, O. M. Formation of adventitious roots in the umbrella China tree. Bot. Gaz. 46: 303, 304. 1908. [Internal roots in hollow trunks of old trees.] BALozeT, L. Note sur une variété de Melia Azedarach L. originaire d’Argentine. Revue Int. Bot. Appl. Agr. Trop. 33: 461-463. 1953. [Lowermost leaflets deeply incised.] 474 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 EAL, F. E. L. Food of the robins and bluebirds of the United States. U. S. Dep a Bull 171. 31 pp. 1915. [Fruits and/or endocarps he ern in oe os robins, ceeds disgorged after pulp is digested, 13, BENAYOUN, J., & T. SAcus. Unusual xylem differentiation ae mature leaves of Melia. Israel Jour. Bot. 25: 184-194. 1976. [M. Azedarach; in partially defoliated branches xylem consists of small diameter vessels, no fibers, and large vascular rays.] Bonner, F. T., & C. X. GRANO. Melia Azedarach L., Chinaberry. Pp. 535, 536 in Seeds of woody plants in the United States. U. S. Dep. Agr. Handb. 450. viii + 883 pp. 1974. [Seed biology.] BowDEN, W. are A ae a numbers in higher plants. I. Acanthaceae to Myrt t. 32: 81-92. 1945. [M. Azedarach, 2n = 28; M. Azedarach f. wmbrac mie ae ee (M. Azedarach var. umbraculiformis), 2n = 28.] CARRATALA, R. E. Intoxicacion mortal por frutos de Melia Azedarach L. (Paraiso ve- getal). Estudio toxicologico. Revista Asoc. Med. Argent. 53: 338-340. 1939. [Ex- periments on rabbits and frogs. CHANG, F. C., & C. K. CHIANG. Tetracyclic triterpenoids from Melia Azedarach, L. 11 2-oxa-trans-bicyclo[3,3,o]-octanones. Tetrahedron Lett. 1969: 891-894. 1969. [From bark. ] CHaAuvIN, R. Sur la substance qui, dans les feuilles de Melia Azedarach, repousse les criquets. Compt. Rend. Acad. Sci. Paris 222: 412-414. 1946. [Crickets repelled by extract of leaves. ] CHIANG, C. K., & F. C. CHANG. Tetracyclic triterpenoids from Melia Azedarach, L.— Ill. Tetrahedron 29: 1911-1929. 1973. Cooper, M. J. E., & J. CULLEN. Melia L. in P. H. Davis, ed. Flora of Turkey 2: 520, 521. 1967. [M4. Azedarach, “native of India and China.’ De Sitva, L. B., W. STOCKLIN, & T. A. GEISSMAN. The isolation of salannin from Melia dubia. Phytochemistry 8: 1817-1819. 1969. [From fruits collected in Ceylon Exona, D. E. U., O. FAKUNLE, A. K. Fasina, & J. I. OkoGuUN. The meliacins (limonoids). Nimbolin A and B, two new meliacin cinnamates from Azadirachta indica L. and Melia Azedarach L. Jour. Chem. Soc. Chem. Commun. 1969: 1166, 1167. 1969. ELLIOTT, - A sketch of the botany of South-Carolina and Georgia. Vol. 1. 606 pp. 6 pls. 1821. [M. Azedarach, 475, 476; fruit a favorite food of the American robin.] EXELL, A. - & F. A. MENDONCA. Meliaceae. Conspectus Florae Angolensis 1: 305- 320. pls. 13-15. 1951. [Melia Azedarach, M. dubia (M. Bombolo), 317, 318.] Grrarpt, A. M. M. Flora ilustrada do Rio Grande do Sul. Fasciculo 10. Meliaceae. Bol. Inst. Biocién. Univ. Fed. Rio Grande do Sul 33 [Bot. 3]. 61 pp. 1975. [Melia Azedarach, grown as an ornamental in southernmost Brazil, 14, 15.] Gourke, [R. L. A. M.] Meliaceae in A. ENGLER, ed., Die Pflanzenwelt ae, und der Nachbargebiete, Theil C: 230-232. 1895. [M. Volkensii Guerke, Harpin, J. W., & J. M. ARENA. Human poisoning from native and vultiveicd Pane ed. 2. xii + 194 pp. Durham, North Carolina. [Melia Azedarach, fruits and tea made of leaves poisonous; toxic principle probably a “‘resinoid.”’] Heir, C. E. Germination studies and testing methods for chinaberry and Chinese parasol tree. Newslett. Assoc. Official Seed Anal. 48(4): 25, 26. 1974. [Maximum germination at 20-30°C; with prior soaking of endocarps in water (two days). ] Hooker, J. D. The flora of British India. Part 1. viii + xl + 740 pp. London. 1872. [Meliaceae, 540-569, _ W. P. Hiern: Melia, five spp., including original obser- vations on M. dubia C Jaconss, M. The generic ean for Melia excelsa Jack. Gard. Bull. Singapore 18: 71- 75. 1971. [Melia and Azadirachta distinguished.] Kaptan, E. R., & N. SAPEIKA. Chemical composition of the fruit of Melia Azedarach L. South Afr. Jour. Med. Sci. 36: 83, 84. 1971. [Isolation and identification of fatty acids from fruit.] 1990] MILLER, MELIACEAE 475 NG, G. Materials for a flora of the Malayan Peninsula. No. 7. Jour. Asiatic Soc. Bengal II. Nat. Hist. 64: 16-137. 1895 [1896.] [Melia, two spp., including notes on M. composita Willd., which is kept separate sa M. dubia, 17-21. KwatTra, M.5S., B. Sincu, D.S. Horui, & P. N. DHINGRA. Poison ning by Melia Azedarach in pigs. Veterin. Record 95: 421. 1974. [Death after ingestion of fruits and leaves.] Lavie, D., M. K. Jain, & I. Kirson. Terpenoids—V. Melianone from fruit of Melia Azedarach L. Tetrahedron Lett. 1966: 2049-2052. 1966. MasBeERLEY, D. J. A monograph of Melia in Asia and the Pacific: the history of white cedar and Persian lilac. Gard. Bull. Singapore 37: 49-64. 1984. [One variable species recognized, its native range and the distribution of the two main cultivars specified; a bibliographic synthesis, analysis of morphological or other characters not included; M. Azedarach typifie ee E. Melia Bombolo essence congolaise a croissance trés rapide. Bull. Re- cherches Agron. Gembloux II. 1: 576-601. 1966. [= M. dubia, fide Hiern (see HOoKER).] Micuaux, F. A. The North American sylva. Vol. 3. 285 pp. pls. 10/-156. Philadelphia. 1819. [M. Azedarach, 4-6, pl. 102, “*...so abundant and so easily multiplied in the maritime parts of the Southern States, as to be ranked among their natural produc- tions.’’] Moraan, E. D., & M. D. THorNToNn. Azadirachtin in the fruit of Melia Azedarach. Phytochemistry 12: 391, 392. 1973. [A triterpenoid that under test conditions in- hibited the feeding of the desert locust (Schistocera gregaria).] MoraGan, P. W., & J. I. DurHAM. Ethylene production and leat abscission in Melia Azedarach L. Pl. Physiol. 66: 88-92. 1980. [‘. . . C,H, in concert with those hor- mones which govern sensitivity to C,H,, regulate aaa leaf fall . Morrison, F. R. A contribution to the chemistry of the fruit obtained from the white cedar tree (Melia Azedarach, L. var. australasica, C. D.C.; syn. M. australasica, A. Juss.) ani in New South Wales, with notes on its reputed toxicity. Jour. Proc. . New S. Wales 65: 153-177. 1932. [Mixture of an alkaloid and a resin pe ‘from fruit poisoned guinea pigs.] Murty, Y.S., &S. Gupta. Morphological studies in the Meliaceae. II. A reinvestigation of floral anatomy of Azadirachta and Melia. Jour. Indian Bot. Soc. 57: 195-204. 1978. [M. Azedarach, M. Birmanica Kurz, M. composita aie Ocui, M., H. Korsuxi, K. Hirotsu, & T. TokKoroyAMA. Sendanin, a new limonoid from Melia Azedarach Linn. var. japonica Makino. Ot aea Lett. 1976: 2877- 2880. 1976. Oexricus, P. B., M. W. Hitt, P. J. VALLELY, J. K. MACLEop, & T. F. Mouinski. Toxic tetranortriterpenes of a fruit of Melia Azedarach. Phytochemistry 22: 531-534. 983. [Chemical characterization of four limonoids (meliatoxins); interpopulational variation in limonoid presence; toxicity demonstrated through clinical trials.] Ocim1, C. Studies on bacterial gall of chinaberry (Melia Azedarach Linn.) caused by Pseudomonas meliae n. sp. (In Japanese; English summary.) Sci. Bull. Fac. Agr. Univ. Ryukyus 24: 497-556. 1977. Oxocun, J. L, C. O. FAKUNLE, D. E. U. Exona, & J. D. CoNNOLLy. Chemistry of the meliacins ‘(limonoids). The structure of meliacin A, a new protomeliacin from Melia Azedarach. Jour. Chem. Soc. Perkin Trans. I. 1975: 1352-1356. 1975. [From wood.] PELLEGRIN, F. Méliacées. Jn: H. HumBert & F. GAGNEPAIN, eds., Suppl. Fl. Gén. Indo- chine. Fasc. 5: 683-700. 1946. Fasc. 6: 701-728. 1948. [Melia, 3 spp., 684, 685.] Perkins, K. D., & W. W. Payne. Guide to the poisonous and irritant plants of Florida. Florida Coop. Ext. Ser. Circ. Univ. Florida 441. 1-7 + [79] pp. 1978. [Melia Aze- darach, No. 299, fruit eee to humans and livestock.] PiccoLo, A. L. G., L. I. THomazini, & O. Cecar. Melia Azedarach L.: oa mies vegetative. Revista Agr. Piracicaba 47: 71-74. 1972. [Offshoots from ro & . GREGOLIM. Fenologia de Melia Azedarach L. no sul do Brasil. ade 30: 107- 109. 1980. [Observations over 12 months.] 476 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 PURUSHOTHAMAN, K. K., K. DURAISWAMY, & J. D. CONNOLLY. Tetranortriterpenoids from Melia dubia. Phytochemistry 23: 135-137. 1984. [Limonoids, compositin, and compositolide, from leaves and seeds; M. dubia = M. Azedarach, fide Mabberley Reba ieee, K. H. Meliaceae. FI. cia 133. 3 pp. 1987. [M. Azedarach, ‘In montibus Himalaya occidentalibus sponta Ripvey, H. N. The flora of the Malay Peninsula. Vol. 1. xxxv + 918 pp. London. 1922. [Melia, 3 spp., including M. Azadirachta, 383, 384.] SAUNDERS, E. R. Floral morphology, a new outlook with special sees to the inter- pretation of the gynaeceum. Vol. |. Cambridge. 1937. [M/. Azedarach, 89. Scuutte, K. E., G. RUcker, & H. U. Matern. Uber einige Tnhaltsstoffe oa on und Wurzel von Melia Azedarach L. Planta Med. 35: 76-83. 1979. [Many com- ounds isolated and identified; toxic substances not detected in fruits.] Smus, J. Melia Azedarach. Common bead-tree. Bot. Mag. 27: t. 1066. 1807. Smncu, M. M., S. K. PURKAYASTHA, P. P. BHOLA, & A. K. Gupta. A reappraisal of the suitability of Melia Azedarach as a paper making raw material. Indian Forester 103: 641-650. 1977. [May be used for wrapping, writing, and printing papers if tree has grown slowly. ] STANER, P., & G. GILBERT. Meliaceae. Fl. Congo Belge Ruanda-Urundi. Spermatophytes 7: 147-213. 1958. [Melia, 2 spp., 172-175, including description and illustration of M. Bombolo.]} STEENIS, C. G. G. J. VAN. General amen Fl. Malesiana I. 4: xii—Ixix. 1948. [Precocious flowering in M. Azedarach, xxi. Tutin, T. G. Melia L. Fl. Europaea 2: 31. 1968. [M. Azedarach, “widely planted in S. Europe for ornament and shade and locally naturalized.”’] UmapEvI, I., M. DANIEL, & S. D. SABNIS. Sapwood-heartwood conversion in Melia Azedarach Linn.—a chemical study. Jour. Econ. Tax. Bot. 10: 411-415. 1987. [Lip- ids, hemicelluloses, lignin, alkaloids, organic acids, ethanol, and water soluble frac- tions summarized in a table; most compounds identified, some not. ZERPA, D. M. DE. Los cromosomas de Melia Azedarach. Agron. Trop. 2: 257. 1953. [n = 14] Subfam. SWIETENIOIDEAE Harms in Engler & Prantl, Nat. Pflanzenfam. III. 4: 267. 1896. 2. Swietenia Jacquin, Enum. Syst. Pl. Ins. Carib. 4, 20. 1760. Small to large trees; bark dark brown, shallowly fissured; scales of terminal buds glabrous. Leaves even-pinnate (rarely odd-pinnate), apex of rachis abort- ing; young leaves with numerous scattered glandular hairs and a few long, simple hairs; leaflets petiolulate [or sessile], oblique at base, lower part of lamina of each narrower than upper, short [to long] acuminate, entire; leaflets usually deciduous before rachis. Plants monoecious; inflorescences borne on new growth; terminal flower of cymules carpellate, lateral flowers staminate. Flowers im- perfect; perianth of staminate and carpellate flowers similar, calyx of 5 (rarely 4 or 6) nonoverlapping lobes, glabrous, margin uneven or ciliate; petals 5 (rarely 4 or 6), convolute in bud, entire or ciliate, otherwise glabrous. In staminate flowers the stamen tube urn-shaped, 10 (rarely 12)-toothed; anthers 10 (or 12), dorsifixed, inserted at a position below the stigma, alternating with the teeth; pollen spherical, + psilate, margins of colpi thickened; ovary narrowly pyr- iform, (4)5(6)-locular, ovules rudimentary, style long, stigma narrower than 1990] MILLER, MELIACEAE 477 mouth of tube; nectariferous disc annular, obscurely lobed, extending slightly above the insertion of the ovary. In carpellate flowers stamen tube urn-shaped, anthers small and withered, inserted above the stigma, anthers and teeth of stamen tube + flexed over stigma at anthesis, ovary globose, (4)5(6)-locular, each locule with numerous anatropous ovules in 2 rows; placentation axile; style short, stigma as broad as mouth of stamen tube, discoid, indistinctly 5-rayed, stigmatic surface on lower side; nectariferous disc annular, obscurely lobed, extending slightly above the insertion of the ovary. Fruit a 5-locular capsule, dehiscing septicidally from the base [or apex and base or from the middle to the ends], pericarp 2-layered, the woody outer layer separating first, 5-ridged columella persistent. Seeds large, in 2 rows, winged, attached by the funiculus near the apex of the axis; wing mostly elaborated from outer integ- ument, weakly [or strongly] aerenchymatous at base and around seed; inner seed coat thin; embryo transversely elliptic, located at the bottom of the seed below the wing, radicle oriented perpendicularly to the long axis of the seed and to the position of the micropyle, cotyledons broader than long, plumule inconspicuous; endosperm very thin, inconspicuous, oily. Type SPECIES: S. Mahagoni (L.) Jacq., the only one included in the genus when it was established by Jacquin. (Named in honor of Gerard von Swieten, 1700-1772, Dutch phy- sician and botanist, who worked in Vienna during the last third of his life and who was instrumental in the establishment of the botanic garden at Sch6nbrunn and the University of Vienna.]— MAHOGANY, CAOBA. Three species of tropical and subtropical America; one, Swietenia Mahagoni, West Indian mahogany, native in our area only at the southern end of Florida (Monroe and Dade counties, including the Florida Keys), but hardy northward in Florida. Otherwise, it is indigenous to the islands of the western and northern Caribbean region (Greater Antilles, Bahamas) but evidently was introduced in Puerto Rico, the Virgin Islands, and the Lesser Antilles (Little; Little & Wads- worth, family references). However, the exact extent of its native range is not known because this species was planted extensively, probably beginning in the 1700’s, and some populations represent escapes from cultivation. Indeed, this important tree has been introduced throughout the tropics as a source of timber and as a shade tree. It seeds freely, and isolated mature trees are sometimes surrounded by numerous seedlings, as, for example, in the Lesser Antilles (Howard). Swietenia Mahagoni is allopatric with S. macrophylla King, Honduran ma- hogany, known from Mexico and Central America (Veracruz and Chiapas, Mexico, and Belize south to Panama) and South America (Colombia and Venezuela and disjunct to Peru, Bolivia, and Brazil), and S. humilis Zucc., restricted to the Pacific slope of Mexico (Sinaloa southward), Guatemala, Hon- duras, El Salvador, Nicaragua, and Costa Rica. Swietenia macrophylla and S. humilis are sympatric in parts of Mexico, Guatemala, and Costa Rica (Lamb, 1966; Styles in Pennington, 1981, family references), but detailed studies of co-occurring populations evidently have not been undertaken. Putative, spon- taneous hybrids between open-pollinated trees of S. macrophylla and S. Ma- hagoni, S. humilis and S. macrophylla, and S. humilis and S. Mahagoni have 478 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 FiGure 2. Swietenia. a—n, S. Mahagoni: a, inflorescence, x 1; b, staminate flower, and placentation, x 12; g, inflorescence with very young fruit, and showing scars of staminate flowers, x 1; h, leafy branch with mature fruit, x 4; i, j, stages in opening of fruit, x '4; k, axis of fruit (columella) with seeds after fall of woody valves, x '; 1, ridged columella with seeds removed, note two rows of scars, (cf. f), x 2; m, seed, x 1; n, embryo, oriented as in seed, radicle at left, cotyledon wider than long and extending above and below the position of radicle, cotyledons strongly coherent (second cotyledon not visible), x 2. 1990] MILLER, MELIACEAE 479 been recognized either in plantations of the parental species or in areas of sympatry (Whitmore & Hinojosa). The hybrids present various morphological traits that are intermediate between the parents, but detailed studies of progeny resulting from controlled crosses have not been undertaken to confirm these observations. Only in Taiwan have crosses between a tree of S. macrophylla (pollen parent) and S. Mahagoni been performed (Lee, 1968). The stomatal length to width ratio and leaflet size in the hybrid seedlings were in general intermediate between measurements of these features in seedlings from seeds produced through the self-pollination of flowers on the parental trees. Chro- mosome numbers of the plants used in this study were not determined. The three species of Swietenia differ in a combination of traits, including vegetative, floral, and fruit characters. Leaflets in S. Mahagoni are generally smaller (mostly 4-6 cm long, 1.5—2.5 cm wide), petiolulate (especially the lower ones), and acute and sometimes aciculate (vs. mostly 7-9 cm long, 2-3 cm wide, subsessile, and long acuminate in S. humilis and mostly 9-13 cm long, 3-5 cm wide, petiolulate, and acute to short acuminate in S. macrophylla). Swietenia Mahagoni has the smallest capsules of the three species (4-6 cm long; vs. 8-16 cm in S. humilis and 12-15 cm in S. macrophylla) and also the smallest seeds. These are brown in contrast to the pale orange-brown seeds of S. humilis and the usually dark brown ones of S. macrophylla. The calyx and corolla are reported to be ciliate in S. Mahagoni and entire in S. humilis and S. macrophylla (Styles in Pennington, 1981, family references), but this differ- ence was not apparent in numerous specimens of the three species that I studied. However, the flowers of S. humilis are slightly larger than those of the other two species (observations based on specimens in the combined herbaria of the Arnold Arboretum and Gray Herbarium). Although the diagnostic character- istics are largely quantitative and gradational, intermediates appear to be un- common (except for the hybrids discussed earlier). Leaflet size and shape are variable in all three species, but especially so in S. Mahagoni, in which they can vary from two to four times longer than broad. Elliptic to ovate leaflets appear to be the commonest expression, although the shape is difficult to describe precisely because the laminae, and particularly the leaf bases, are asymmetrical, sometimes greatly so. Swietenia and most other members of subfam. Swietenioideae have woody capsules containing a conspicuous columella and winged seeds. Species of Swietenia have ovoid woody capsules that open from the base to the apex (described by Johnson as starting in the middle of the capsule in S. macrophylla) and contain seeds with a large terminal wing, which is attached near the distal end of the columella. In Khaya A. Juss., the source of African mahogany and a genus thought by some (Lamb, 1966) to be close to Swietenia, the seeds have a complete, narrow marginal wing, and the capsules open from the apex to the base. Most other members of the tribe, except the poorly known Schmardaea Karsten, are restricted to the Old World. Chromosome numbers vary within species (Swietenia Mahagoni, 2n = 12, 18, 24, 36, 42, 46-48, 54, 60, 108) and between species (S$. macrophylla, 2n = 54 and S. humilis, 2n = 50, 52, 56; Datta & Samanta, Khosla & Styles, 480 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71] Styles & Vosa). The euploid series in S. Mahagoni was documented in plan- tation trees (Fiji Islands), and an additional number (” = 28) has been reported for a tree grown in India (Sareen & Kumar). These numbers indicate the existence of considerable karyological polymorphism, at least in plantation stock. Khosla & Styles counted 2” = 48 (S. macrophylla), and 2n = 56 (S. humilis) in plants obtained from within the native ranges of the species. The chromosome cytology of the genus is otherwise poorly known, particularly within naturally established populations. Few detailed studies of the morphology of species of Swietenia have been undertaken. Lee (1967) was the first to show that the flowers of S. Mahagoni and §. macrophylia were either staminate (styles long, ovaries slender) or car- pellate (styles short, ovaries globose). The trees are monoecious. The number of staminate and carpellate flowers per inflorescence is reported to vary, with staminate flowers being more abundant. Observations of the same tree over two years showed a large difference in the number of carpellate flowers formed on an annual basis. Pollen is released in sticky masses suitable for transport by insects, which presumably are attracted to the nectariferous disc. Flowers of S. Mahagoni and S. macrophylla are reported to be fragrant (Small; Lamb, 1966). Only the undersurface of the stigma is receptive to pollen (Tomlinson). Dichogamy may be the rule in members of subfam. Swietenioideae, including Swietenia (Styles, family references), since the sparse field observations indicate that the central carpellate flower of a three-flowered dichasium opens first, followed by anthesis of the staminate flowers, which occur in a pair below the carpellate flower. This asynchrony may promote outcrossing, but the hybrid- izations performed by Lee (1968) indicate that in S. Mahagoni, at least, the trees are self-compatible. Usually only one carpellate flower per inflorescence develops into a mature fruit. Growth of Swietenia Mahagoni follows a pattern similar to one typical of temperate trees, namely, articulate monopodial branching, scaly terminal bud formation, and suppression of secondary branches on primary shoots during the first year and release of the branches during the second year (Tomlinson). The xylem anatomy of the West Indian and Honduran mahoganies has been thoroughly characterized (Rock; under family references see Kribs; Panshin; Record, 1919, 1941, 1944). Annual growth increments are demarcated by bands of parenchyma. In comparison with the wood of Spanish cedar (Cedrela spp.), another economically important timber tree in the Meliaceae, intervascular pits in the mahoganies are smaller in diameter (2-3 um). In wood of the West Indian mahogany the perforation plates are simple, intervascular pitting of the vessel elements is alternate, pits of the fibers are simple and slitlike or vestigially as are uniseriate and multiseriate heterocellular rays (Rock). Minute foliar nectaries occur on the petiole, rachis, petiolules, and both leaflet surfaces in all three species of Swietenia (Lersten & Rugenstein). Their positions in live plants are marked by minute glistening balls of nectar. Extrafloral nectaries such as these perhaps function to attract ants, which may help rid the plants of insect pests, a relationship demonstrated in other flowering plants. However, evidence for it in Swietenia is lacking. 1990] MILLER, MELIACEAE 481 The seed biology of Swietenia is poorly documented. Seeds of S. macrophylla and probably those of the two other species have a short (several month) period of viability after ripening, unless the seeds are dried to five percent moisture content (Lamb, 1966). Viability is prolonged with refrigeration. The seeds have a prominent wing, and this structure presumably facilitates dispersal by wind, even though seedlings may be abundant under or near large trees. Lamb (1966) reported that seeds from a large tree of S. macrophylla can be scattered on the leeward side over an area of about 10 acres, but Johnson notes that seeds of this species are rarely found more than 100 yards from the parent tree. Capsules are borne erect in S. Mahagoni (and the other species). The seeds are weakly attached to the columella (central axis) and remain hanging loosely in place after the capsule valves have fallen away, which happens in the winter in Florida (C. E. Wood, Jr., obs.). Seeds are presumably dislodged by gusts of wind. Abundant aerenchymatous tissue occurs at the embryo end and in the wing along the raphe in seeds of S. Aumilis. It may increase the buoyancy of the seed in the air or possibly help to keep the seed floating in water. Aqueous leachates of leaves of S. Mahagoni inhibit the germination of its seeds in the laboratory (Andorfer & Teas), suggesting a possible allelopathic effect in nature. Seedlings of both S. Mahagoni and S. macrophylla are cryptocotylar. In the former the eophylls are alternate, whereas in S. macrophylla they are opposite (Duke, 1965). Swietenia Mahagoni was an important member of the highly diverse tropical hardwood forests that once were common in subtropical Florida (Craighead). Trees of the West Indian mahogany reaached a large size in hammocks (tree islands) that developed at places where the mineral soil or bedrock was slightly elevated (1 m or less) above the surrounding pinelands or glades (treeless wetlands). In such places fresh water remains year round in solution cavities and peat accumulates, both helping to maintain high humidity. West Indian mahogany was also common in a second kind of hammock that developed on ridges of marine marl deposited on the land by hurricane tides. Peat accu- mulation raised these low ridges further and isolated them from salt water. While charcoal preserved in the soil shows that fire swept through hammocks of both types, trees of S. Mahagoni can persist under such circumstances by forming root suckers. Phoradendron rubrum (L.) Grisebach grows on Swietenia in some existing hammocks in Florida, and this parasite has been implicated in the death of larger trees on Key Largo and perhaps on Rhodes and Sand Keys (Cooley). West Indian mahogany (Swietenia Mahagoni) and Honduran mahogany (S. macrophylla) are the sources of wood universally prized for cabinetry and fine woodworking for over 200 years. Honduran mahogany has been described as the “world’s premier cabinet wood” and ‘perhaps the most valuable timber tree in the whole of tropical Latin America” (Styles in Pennington, 1981, family references; p. 400). Harvestable trees are still available in Mexico, Guatemala, and South America. West Indian mahogany is rare on the commercial market, and writing in the mid-1960’s, Lamb commented that it had almost disappeared from commerce because the supply was exhausted. Its wood is thought to be superior to that of Honduran mahogany, which has a coarser grain and a less 482 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 rich color when finished. Both mahoganies are presently being grown in forest plantations. Under natural conditions S. macrophylla is reported to reach 60 m tall; trees of S. Mahagoni are generally shorter (to 20 m). The wood of Swietenia came to the attention of Europeans early in the 1500's during the Spanish domination of the West Indies. Over the next century and a half mahogany was used in the West Indies for shipbuilding and repair and in the construction of buildings because of its great resistance to dry rot and termites and to warping when in contact with water. Spanish shipbuilders, first at Santiago, Cuba, and later at Havana Arsenal, Cuba, constructed many ships for the Spanish Navy using logs cut in Cuba and Mexico (Lamb, 1966). The use of mahogany for furniture in England began about 1715. It rapidly sup- planted walnut and oak as the raw material for tables, chairs, and cabinets. Lamb named the period 1725-1825 the “golden age of mahogany,” in reference to the development of furniture styles during that 100 years by English master craftsmen, including Thomas Chippendale, George Hepplewhite, and Thomas Sheraton, who based their innovative designs on the strength, durability, out- standing working qualities, and excellent finished color and luster that are characteristic of mahogany. American cabinetmakers were also using mahog- any in Boston, New York, Philadelphia, and elsewhere during this period. Especially prized were boards with highly patterned grain. These came from the stumps, root crowns, and larger roots, as well as from logs that included the junction of the bole and a large branch. Such logs were sawed so that the boards had a Y-shaped grain pattern. Overharvesting has eliminated large trees of West Indian mahogany through- out its range, and what little of its wood comes on the market is from plantation- grown trees. Trees of Honduran mahogany are still available in parts of Central and South America. However, many of the largest trees have been cut from accessible locations. Concern has been expressed about the severe depletion of potentially useful genetic stock for breeding purposes (Styles in Pennington, 1981, family references), Mahogany shoot borers, Hypsipyla sp. (Lepidoptera), are serious insect pests in many areas and have limited the development of mahogany plantations and the use of Swietenia species in reforestation projects (Styles & Khosla). A decoction of the bark of Swietenia Mahagoni or of S. macrophylla has been used locally in the West Indies or Central America as a tonic and to treat inflamed mucous membranes and chest and other illnesses (Morton). REFERENCES: Under family references see ABDULLA; BAILEY ef ee BAILLON; BENTHAM & HOOKER; A. DE CANDOLLE; C. DE CANDOLLE; CARREIRA & SECCO; CORNER; DATTA & SAMANTA; Duke, 1965, 1969; ERDTMAN; GRUPMA & STYLES; ne 1896, 1940; Howarp; JussIEU; Narr, 1962; NARAYANA, 1958b; PANDEY; PANSHIN; PENNINGTON, 1981; PENNINGTON & SARUKHAN; PENNINGTON & STYLES; RECORD, 1919, 1941, 1944; RrpLeEY; SCHOLZ; SMITH, 1965: STYLES; — KHOSLA; STYLES & VOSA; TAYLOR, 1981, 1983, 1984; UMADEvVI et al.; and WILS 1990] MILLER, MELIACEAE 483 ALVARENGA, S., & E. M. Fores. Morfologia y germinacidn de la semilla de caoba, Swietenia macrophylla 1G ing (Meliaceae). (English abstr.) Revista Bio. Trop. 36 261-267. a [Cotyledons develop petiole-like bases during germination. ] ALVAREZ GarciA, L. A. A mahogany seedling blight in Puerto Rico. Carib. Forester 1(1): 23, 24, 1939. [Phyllosticta Sw ietenia sp. nov., causes leaf necrosis.] Amoros-Marin, L., W. I. Torres, & C. F. ASENJO. Isolation of cycloeucalenol from Wes endian mahogany wood. Jour. Organic Chem. 24: 411-413. 1959. [Sterol.] AnporFer, H., & H. Teas. Seed germination inhibition by leaf extracts of Swietenia Mahagoni Jacq. (Meliaceae). (Abstr.) ASB Bull. 18: 25. 1971. [Aqueous extract inhibits germination of seeds of S. Mahagoni.] AVILA HERNANDEZ, M. Ecological conditions of the mahogany tree (Swiefenia macro- phylla) in the jungles of the Yucatan Peninsula. (In Spanish.) Méx. Agr. 9(97): 29- 32.1962." Basak, S. P., & D. P. CHAKRABORTY. Scopoletin from the leaves of Swietenia Mahagoni Jacq. Jour. Indian Chem. Soc. a A 1970. [A coumarin.] BascopE VARGAS, F., A. L. BERNARDI, & H. LAMPRECHT. Descripciones de Arboles forestales. No. |. Swietenia ieee Bol. Inf. Divulg. Inst. Forest. Mérida. 18 pp. 1957.* BLAKE, S. F. Revision of the true mahoganies (Swietenia). Jour. Wash. Acad. Sci. 10: 286- 297. 1920. [Five spp.] Boone, R. S., & M. CHUDNOFF. Variations in wood density of the mahoganies of Mexico and Central America. Turrialba 20: 369-371. 1970. [S. humilis, S. macrophylla.] Briscoe, C. B., & F. B. Lams. Leaf size in Swietenia. (Spanish summary.) Carib. Forester 23: 112-115. 1962. [S. macrophylla and S. Mahagoni distinguished on leaflet size; leaflets intermediate in size in putative hybrids. Broscuat, T. K., & H. M. DonseLman. Effect of photoperiod on growth of West Indian mahogany. HortScience 18: 206, 207. 1983 Buscu, P. Die Mahagonisorten des Handels, geordnet nach den einzelnen Produktions- gebieten und ihrer botanischen Abstammung. Tropenpflanzer 15: 479-493. 1911. [Characteristics of “Cuban,” “Honduran,” and “Mexican” mahoganies, 482-487.] CatessBy, M. The natural history of Carolina, Florida and the Bahama Islands. 2 vols. London. 1731-1743. [“‘The mahogany tree” (= Swietenia Mahagoni), Vol. 2: 81, pl. 81. 1743; first description and illustration.] CHADHA, S. Y. R., Chief Editor. Swietenia. The wealth of India. Raw Materials 10: 81- 87. 1976. [S. Mahagoni, first introduced into India in 1795 as seedlings from Ja- maica; S. macrophylla, widely planted; useful summary of the economic botany of the two species in India. CHAKRABARTY, M. M., & D. K. CHowpuHuri. The fatty acid composition of the seed fat from Swietenia macrophylla. Jour. Am. Oil Chem. Soc. 34: 489, 490. 1958. [Six fatty acids identified.] CHALONER & FLEMING. The mahogany tree. 117 pp. + 7 pls. + J foldout table + 1 foldout map. Liverpool & London. 1851. [Commercial tract, but incl. interesting notes on distribution (Jamaica, Cuba, Hispaniola, Puerto Rico, Mexico, Central been harvesting methods in the West Indies and Belize (British Honduras), ee A, & T. CHAKRAVARTY. Swietenolide, the bitter principle of Swietenia macrophylla. Indian Sci. Cong. Assoc. Proc. 42(3, Abstr.): 135. 1955.* CHuDNOFF, M., & T. F. GEARY. On the heritability of wood density in ee macro- phylla. Turrialba 23: 359-362. 1973. [Little variation recorded over a wide range of growth conditions; no evidence that density is a heritable trait. CONNOLLY, J. D., . Lappé. Tetranortriterpenoids and related compounds. Part 24. The inienielation of swietenine and swi ietenolide, major tetranortriterpenoids from the seeds of Swietenia our. Chem. Soc. Perkin Trans. I. 1980: 529, 530. 1981. 484 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 Coo ey, G. Be Phoradendron rubrum in Florida. Rhodora 65: 190, 191. 1963. [On S. ponte F. C., Sr. The trees of south Florida. ee 1. xvi + 212 pp. Coral Gables, Florida. 1971. [Natural history notes on Swietenia Mahagoni, including develop- ment of sucker shoots, mistletoe (Phoradendron en en and occurrence in different vegetation types and at various locatio EcuLer, F. E. Mahogany: a potential resource of South Florida. Jour. Forestry 39: 725, 726. 1941. [Wide ecological amplitude of S. Mahagont; ecological notes GEARY, T. F., H. BARRES, & R. YBARRA-CORONADO. Seed source variation in Puerto Rico and Virgin Islands grown mahoganies. (Spanish summary.) U. S. Dep. Agr. Forest Serv. Res. Pap. ITF-17. 24 pp. 1973. [Based on seeds of S. humilis and S. macrophylla from Mexico and Central America and seeds of S. Mahagoni from St. roix, Virgin Islands. ] . W. Nostes, & C. B. Briscoe. Hybrid mahogany recommended for planting in the Virgin Islands. U.S. Forest Serv. Res. Pap. ITF-15. 4 pp. 1972. [S. macrophylla GLEASOoN, H. A., & A. J. PANSHIN. Swietenia Krukovii: a new species of mahogany from Brazil. Am. Jour. Bot. 23: 21-26. 1936. [= S. macrophylla; includes photographs and a ne of wood.] ahogany Os Mahagoni L.). Bull. Misc. Inf. Roy. Bot. Gard. Trinidad 2: 185-187. HEMSLEY, W. B. horas ee Jacq., var. praecociflora, Hemsl. Hooker’s Ic. PI. 28: pl. 2786 [+ 2 pp.]. 1905. [Flowers developed in plants at seedling stage in Trinidad. Hooker, W. J. Swietenia Mahagoni. Bot. Misc. 1: 21-32. pls. 16 & 17. 1829. [Descrip- tion; historical notes about wood and about methods used to harvest ‘““Honduras mahogany” aaa i macrophylla.)} IRMAY, H. DE. caoba, Swietenia macrophylla, en Bolivia. (English transl.; French Fee Sel Torestts 10: 43-57. 1949. [Distribution, characteristics. ] JoHNson, A. Studies on the fruit of Swietenia mac rophylla King. Malayan Forester 32: 180-186. 1969. [Dehiscence of fruit results as water is lost from pericarp, but “‘other factors” are involved also. KinG, G. Swietenia macrophylla. Hooker’s Ic. Pl. 16: pl. 1550 [+ 2 pp.]. 1886. [Sp. ov.; based on a plant growing in the Calcutta Botanic Garden; source: British KOEHLER, A. The identification of true mahogany, certain so-called Beers: and some common substitutes. U. S. Dep. Agr. Bull. 1050. 18 pp. figs. J-13. 1922. [Key to mahogany and mahogany-like er based on hand-lens characters; pce and illustrations of wood of Swietenia spp.] KUKACHKA, B. F. Mahogany (Swietenia macrophylla King). U.S. Dep. Agr. Forest Serv. Forest Prod. Lab. Rep. 2167. 11 pp. 1959. [Wood properties. ] Lamp, F. B. A selected, annotated bibliography on mahogany. Carib. Forester 20: 17- 37. 1959. [Headings include systematic botany, ecology, aera: etc. ] . An approach to mahogany tree improvement. /bid. 21: 12-20. 1960. [Reports on occurrence of intermediate progeny in areas where S. ei a and S. Ma- hagoni are being grown together, viz. Puerto Rico, St. Croix, Martinique.] n further defining mahogany. Econ. Bot. 17: 217-232. 1963. [Contends that the me mahogany 1s a corruption of the Yoruba word oganwo for African ma- hogany (Khaya); used for trees of S. Pee in Jamaica by slaves of West African (Nigerian) origin; see also LAMB (196 . Mahogany a America: ‘ ecaleny and management. x + 220 pp. Ann Arbor, ian —. Maho . Econ. Bot. 22: 84-86. 1968. [Answer to Malone’s criticism (Econ. Bot. 19: 286- 292, 1965) of Lamb (1963).] 1990] MILLER, MELIACEAE 485 , H. Y. A comparative study on the morphology of Swietenia Mahagoni and S. “rnarophsta, (In Chinese.) Quart. Jour. Forestry Taiwan 2(3): 76-80. 1966.* [Sto- t Mahagoni smaller and narrower than those of S. macrophylla. Stu dies i in Swietenia (Meliaceae): observations on the sexuality of the flowers. Jour, eee Arb. 48: 101-104. 1967. [S. Mahagoni, S. macrophylla.] minary report on the dase characters and heterosis of the hybrids be- tween met! Mahagoni x S. macrophylla. Taiwania 14: 43-52. 1968. [S. ma- ese (pollen parent) x S. Mahagoni successful; these hybrids grew taller than S. Mahagoni x S. Mahagoni progeny and had intermediate stomatal length to width ratios; sample size small, S. macrophylla parent perhaps itself a hybrid.] . Morphological variation of seedlings in Swietenia raised from open-pollinated seeds. Ibid. 15: 245-251. 1970. [In several characteristics seedlings in plantation of S. macrophylla and S. Mahagoni intermediate between those of the parental species. ] LersTEN, N. R., & S. R. RUGENSTEIN. Foliar nectaries in mahogany (Swietenia Jacq.). Ann. Bot. II. 49: 397-401. 1982. [S. Mahagoni, S. macrophylla, S. humilis.] LitTLE, E. L., ae Atlas of United States trees. Vol. 1. Conifers and important hardwoods. p. Agr. Misc. Publ. 1146. v + 9 pp. + base maps I & 2, maps 1-200, 9 overlay en 1971. [S. Mahagoni, map 192-E.] LosaTo, R. C. Novas analises biometricas e observacoes sobre a germinacao do Mo (Swietenia macrophylla King—Meliaceae) na Amaz6nia equatorial. Ciéncia Saree 20: 505, ae 1968 a M. Storage and germination of large-leaf mahogany seeds. Philip. Jour. Forestry 1: 397-409, 5 pls. 1938. [S. macrophylla; seed viability lengthened (to 138 days) by storage in sales charcoal and darkness. } MARIE, E. Note Swieteniq hyliak ing (English Ry Spanish transl.) Carib, Forest 10: 205-222. 1949. [Reforestation in Martinique.] Marauettl, J. R., M. A. GAtnzA, J. L. LEon Acosta, & R. MONTEAGO. Som e aspects of the genetics of Swietenia. ° Spanish.) Baracoa 5: 3-16. 1975.* [Abstr. in PI. Breed. Abstr. 47(1): 753. 1977.] Neusauer, H. F. Uber das Blatt von Swietenia macrophylla King. Ber. Deutsch. Bot. Ges. 73: 277-288. 1960. [Development of mature and seedling leaves; seedlings illustrated. ] Pittier, H. The Venezuelan mahogany, a hitherto undescribed species of the genus Swietenia. Jour. Wash. Acad. Sci. 10: 32-34. 1920. [S. Candollei (= 8. macrophylla).] PRASAD, S. S., & R. A. B. VERMA. Leafspot diseases of Swietenia macrophylla King and Swietenia Mahagoni (L.) Jacq. Sci. Cult. 32: 558, 559. 1966. Rice, C. H. The northern outpost of mahogany. Am. Forests 42: 266, 267, 293. 1936. [Includes notes on distribution in southern Florida, eer Madeira Hammock near Cape Rock, B. N. The woods and flora of the Florida Keys: “Pinnatae.”’ Smithson. Contr. Bot. 5: 1-35. 1972. [Swietenia Mahagoni, detailed description of wood anatomy, 11, figs. 4, 22.] Rotre, R. A. The true mahoganies. Kew Bull. Misc. Inf. 1919: 201-207. 1919. [S. eee is humilis, S. macrophylla; extensive and useful taxonomic and biblio- graphic n SAREEN, T. S., a Kumar. In A. LOvE , IOPB chromosome number reports XLII. Taxon 22: 651-652. 1973. [S. vue n= 28, SARGENT, C. 8. Swietenia. Silva N. Am. 1: 99-102. pls. 43, 44. Fat oe SMALL, J. K. Swietenia Mahagoni. Addisonia 15: 27, 28. pl. 494. STEHLE, H. Les mahoganys des Antilles francaises et le Swietenia ie Stehlé et Cusin, nov. spec. Mém. Soc. Bot. France 1956/57: 41-51. 1958 [1957]. [A putative hybrid of S. Mahagoni and S. macrophylla.) STYLes, B. T. Swietenia age (L.) N. J. Jacquin. The correct name and authority 486 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 for the West Indian or small- 3 mahogany. An example of procedure in botanical nomenclature. Jour. Oxf. Univ. Forestry Soc. VI. 7- ee 1968. TayLor, A. R. H., & D. A. H. TAYL or. Limonoid extractives from Swietenia macro- phylla. Phytochemistry 22: 2870, 2871. 1983. (Swietenine, swietenolide, and other limonoids from seeds.] TOMLINSON, P. B. The biology of trees native to tropical Florida. 1x + 480 pp. Publ. by the author. Allston, Massachusetts. 1980. [Meliaceae (Swietenia Mahagoni), 6 ae an map), 241-245 (description of plant and its growth characteristics, illustratio VERMA, R. A. B. aaa spot diseases of mahogany. Indian Phytopath. 25: 33-35. 1972. [See PRASAD & VERMA.] WapswortH, F. H. The development of Swietenia Mahagoni Jacq. on St. Croix. Carib. Forester 8: 161, 162. 1947. [Introduced in mid-1800’s.] Watts, D. Biogeographical variation in the mahogany (Swietenia Mahagoni (L.) Jacq.) woodlots of Barbados, West Indies. Jour. Biogeogr. 5: 347-363. 1978. [Forests of the introduced S. Mahagoni established in late 18th and early 19th centuries harbor a rich subcanopy flora of native forest species otherwise sparsely represented on the island. WHITMoRE, J. L., & G. Hinojosa. Mahogany ae dl hybrids. (Spanish summary.) U.S. Dep. Agr. Forest Serv. Res. Pap. ITF-23. 8 pp. 1977. [S. Aubrevilleana is the hybrid of S. macrophylla x Mahagoni, S. hun nilis x macrophylla and S. humilis x Mahagoni; reports discovery of natural S. humilis x macrophylla hybrids in northwestern Costa Rica. 1990] HU, TRADITIONAL CHINESE MEDICINE 487 HISTORY OF THE INTRODUCTION OF EXOTIC ELEMENTS INTO TRADITIONAL CHINESE MEDICINE SHIU YING Hu! An investigation into the history of using products derived from introduced species of plants and animals in traditional Chinese medicine (TCM) clearly demonstrates East-West cultural interaction. The research was performed in a framework arbitrarily set with the time limited to before 1911 and the material included in three important references: Zhong Yao Zhi(A New Chinese Materia Medica, Anonymous, 1959-1961), Zhong Yao Da Ci Dian (An En- cyclopedia of Chinese Medicines, Anonymous, 1977-1979), and Zhong Cao ao Xue (Chinese Pharmacology, Anonymous, 1976-1980). The work contains a list of 94 commercial products arranged alphabetically by scientific name, together with the equivalent Chinese and English names; a chronological ac- count of the first record of the species and imported products; an analysis of the reasons for using extra-Chinese elements in TCM; and an account of the principles and criteria for identifying exotic species. When I began to study the medicinal plants of China in the early 1930's, the knowledge of Chinese materia medica was scattered in various ancient herbals (ben-cao). I prepared a list of Chinese plant esculents used for the conservation of health (bupin), and bought the products from groceries and Chinese drug stores in Guangzhou (Canton). I went to Luo-fou Shan in Guang- dong repeatedly to learn about the sources of my collection of bupin from the monks and nuns residing in various Taoist temples and to collect voucher specimens with the help of the local herb collectors. During the entire period of the Sino-Japanese War (1937-1945), I taught at West China Union University in Chengdu. Together with several senior stu- dents, I conducted a survey of the medicinal plants in the herb shops of the area (Hu, 1945). During summer vacations I went with colleagues and students to live in the villages of the Giarong tribe; we explored the vegetation and studied the lives of migrant collectors of crude drugs in the alpine meadows of Min Shan and Qiong-lai Shan in western Sichuan. Between September 1946 and March 1968 I was working on the Chinese collection in the herbarium of the Arnold Arboretum, and my interest in medicinal plants was overshadowed by taxonomic research and the duties of a professional botanist. In connection with field work for the preparation of a Flora of Hong Kong, | taught in the Biology Department of Chung Chi College, Chinese University of Hong Kong, ‘Arnold Arboretum, Harvard University, 22 Divinity Avenue, Cambridge, Massachusetts 02138, U.S.A. © President and Fellows of Harvard College, 1990. Journal of the Arnold Arboretum 71: 487-526. October, 1990. 488 JOURNAL OF THE ARNOLD ARBORETUM [vov. 71 between March 1968 and October 1975. At first only a few students in that university were interested in the study of local medicinal plants. We collected specimens from herbalists in the area, identified the species, and ascertained their uses. Later, my colleagues Y. C. Kong (Biochemistry Department), H. M. Chang, and J. Ma (both of the Chemistry Department) showed a growing interest in research on the chemical composition of various Chinese crude drugs. To check the efficacy of the medicines, they tested them on animals. They asked me to identify the specimens and to contribute some botanical descriptions. Gradually, under the leadership of Chang, the Chinese Medicinal Material Research Centre was organized in the Chinese University of Hong Kong. This institution grew fast. With a strong staff, advanced instrumentation, and the support of the University, local institutions, and individuals, it soon attracted the attention of scientists and scientific institutions from around the world. Its computerized information on traditional Chinese medicine (TCM) is linked with medical centers in London, Tokyo, and Washington. It is a worldwide bridge for disseminating knowledge of Chinese materia medica both past and present. In connection with this research center, my interest in Chinese medicinal plants was revitalized. This report is one of the results of subsequent studies. The aim of this report is to provide a window through which historians of science can see the facts and the nature of material exchange and cultural diffusion between China and people in lands far and near. In addressing the questions, I began with the words what, when, where, why, and how about exotics used in TCM. A workable framework was arbitrarily formed by setting a time limit of 1911, the year when China was transformed from an empire to a republic, and by selecting three modern references: Zhong Yao Zhi(A New Chinese Materia Medica; Anonymous, 1959-1961), Zhong Yao Da Ci Dian (An Encyclopedia of Chinese Medicines, Anonymous, 1977-1979), and Zhong Cao Yao Xue (Chinese Pharmacology, Anonymous, 1976-1980). The com- prehensive treatise of An Encyclopedia of Chinese Medicines contains 5767 entries of crude drugs from 4161 species of plants, 442 species of animals, and 58 kinds of minerals. For each plant entry, the accounts include the recognized Chinese name, its synonyms, the source material with scientific identification, a morphological description illustrated by a line drawing, notes on distribution and ecology, and information on cultivation (if the species is not wild), har- vesting time, and procedure, identifying characteristics of the market product, chemical composition, pharmacology, special treatment needed before dis- ensing, properties, target organs and illnesses, selected recipes, and ancient references. The Supplement to this work (Anonymous, 1979) gives indexes to Chinese and scientific names and synonyms, Chinese and English equivalents of the chemical components, chemical formulas and properties, against phar- macological references, a classified list of diseases with cross references to remedies, a detailed bibliography for chemical composition, experimental and clinical reports for each entry, and three checklists for modern and ancient systems of measurements. Much of the information for this article is based on these references. 1990] HU, TRADITIONAL CHINESE MEDICINE 489 KINDS OF EXOTICS IN TCM The exotics in a flora of a country are either cultivated plants or adventives fully established in the area. Volumes 1-4 of Zhong Yao Zhi (Anonymous, 1959-1961) and volumes 2 and 3 of Zhong Cao Yao Xue (A 76- 1980) were carefully checked for extra-Chinese elements. Ninety-four com- mercial products prepared from 98 exotic species of plants and eight species of animals were selected. Several products are prepared from similar substances of different species. An alphabetic list of the botanical and zoological identi- fications of these crude drugs is given for quick reference, with the animal entries following those for plants (see TABLE |). When a crude drug is prepared from two or more related species, the epithet of the primary source 1s listed first. Included in the enumeration are the Chinese and English names, the place of origin, and the date the item was first recorded in a Chinese herbal or pharmacopoeia. In the list only five items (less than five percent) are of animal origin. These are ox bezoar, elephant hide and tusks, rhinoceros horns, and the ambergris of the sperm whale. Plant materials include 26 kinds of seeds, 18 leafy shoots, leaves, or whole herbaceous plants, 16 fruits, nine resins, six roots, six flowers, four fragrant woods, two extracts, one bark, and an exudate of a desert plant (alhagi). Botanically, these products are from plants belonging to 55 families of dicots and five of monocots. The largest contributing family is the Legu- minosae, with 15 species in 11 genera. The Cruciferae and the Cucurbitaceae each contributes four species, while the Euphorbiaceae, Malvaceae, Umbellifer- ae, Styracaceae, Labiatae, Solanaceae and Palmae are each represented by three species. The Piperaceae, Burseraceae, Guttiferae, Flacourtiaceae, Combreta- ceae, Apocynaceae, Rubiaceae, and Compositae are each represented by two species, and the remaining 33 families have only one species each. CHRONOLOGICAL ACCOUNT It is an undeniable fact that man occupied the habitable continents and large islands long before his existence was recorded by ancient and/or modern ex- plorers. Intentionally, man moved with his favorite animals and plants; un- intentionally, weed seeds and parasitic animals traveled with him. Dogs and the paper mulberry (Broussonetia papyrifera (L.) L’Hér.) in Polynesia are good examples of the former; many garden weeds, of the latter. Several of the extra- Chinese elements used in TCM arrived in China in a similar fashion, some during prehistoric time. We have material evidence that hemp arrived in China during the Neolithic period. Broomcorn (Panicum miliaceum L.) and hemp (Cannabis sativa L.) were reported by Q. R. Wang and Guo (1980) from an archaeological site of a Neolithic community, carbon 14 dated at 4850 B.p. + 180 years, on terraced land by Da-xia-he in Gansu. Archaeological evidence such as this is rare. For our purposes we have to use historical records to show the chronological sequence of the employment of exotic elements in TCM. Such records, however, should not be taken as the time of introduction, the exact date of which is unrecorded. Moreover, many ancient herbals and phar- TABLE |. Exotics in Chinese Materia Medica. RE- PORTION PLACE OF CORDED SCIENTIFIC NAME CHINESE NAME! ENGLISH NAME USED ORIGIN .D.) Abutilon theophrasti Medicus Qing-ma (1) Abutilon Seeds SW As 659 Acacia catechu (L illd Er-cha (2) Catechu Extract Tr As 1596 {/hagi pseudalhagi Desv Ci-mi (3) Alhagi Exudate Mid E 739 Aloé barbadensis Miller Lu-hui (4) loe Extract Afr 973 Amomum krervanh Pierre & Gagne Bai-dou-kou Siam Cardamon Fruits SE As 973 Andrographis paniculata com f.) Nees Chuan-xin-lian (6) Kiryat, creat Whole plant Tr As ? Aquilaria ses ha Chen-xiang Aloeswood ood SE As 540 Areca catec Bing-long (8 Betel nut Fruits Tr As 1300 Benincasa een (Thunb.) Cogn. ng-gua-zi (9) Winter gourd Seeds Tr As 100 Bombax ceiba Mu-mian (10) Tree cotton Roots SE As 1596 — carteri Birdw. Ru-xiang (11) Frankincense Resin Mid E 540 Brassica is rta Moench Bai-jie-zi (12) White mustard Seeds Eur $40 ee Yun-tai-zi (1 Field mustard Seeds Eur 659 Br anne pinnatum (L. f.) Oken Lu-di-shen-gen (14) ir plant Whole plant Afr 1848 Caesalpinia sappan Su-mu (15) Sappanwood Heartwood SE As 659 Canavalia — (Jaca. ) DC. &/or C. en- Dao-dou (16) word bean Seeds, roots Tr As 1596 siformis i Jack bean eeds WI 1596 Cannabis s Hu-ma-ren (17) Hemp s Fruits SW As 100 Capsicum | Pies L. La-jiao (18) Red pepper Fruits Tr Am 1621 Carica papaya an-mu-gua (19) apaya Fruits Tr Am 1848 Carthamus tinctori Hong-hua (20) Safflower orolla Med R 973 Cassia angustifolia Vail &/or C. acutifolia Fan-xie-ye (21) nn Leafy shoots Tr As 1894 Del. Cassia occidentalis L. Wang-jiang-nan (22) Coffee s Seeds Tr As 1407 Catharanthus roseus (L.) G. Don Chang-chun-han (23) Madagascar periwinkle Leafy shoots Mad 1848 Cinchona ledgeriana Moens. &/or C. succi- Jin-ji-na-pi (24) Qui Bark S Am 1765 rubra Pav X1-gua-cul (25) Watermelon Young fruits Afr 1596 Citrullus lanatus (Thunb.) Matsum. & Na- kal O6P WOLAYOUAV ATONAV AHL AO TYNANOFL IL “TOA] TABLE 1. Continued. PORTION PLACE OF CORDED SCIENTIFIC NAME CHINESE NAME! ENGLISH NAME USED ORIGIN? (A.D.) Cleome gynandra Bai-hua-cai (26) Spider-wisp Leafy shoots Tr Am 1596 Commiphora mir ei Mo-yao (27) Myrrh esi Mid E 973 Coriandrum sativ Yuan-sui-zi (28) Coriander Seeds SW As 1061 "rOCUS ) Fang-hong-hua (29) Saffron Styles Med R 1596 Cucumis melo Tian-gua-zi (30) Sweet melon Seeds C As 1061 Daemonorops dra aco Bl. Xue-jle Dragon’s blood Resin Indon 659 Da etc sissoo Roxb. &/or D. parviflora Jiang-chen-xiang (32) Sissoo, fragrant rosewood Heartwood Tr As 756 Rox Datura n metel L. Yang -Jin- hua (33) Datura Flowers Tr As 1596 cay aes aromatica Gaertner Bing- -pian (Long-nao- Borneo camphor Resin Indon 659 xiang) ( Euphorbia iol Le Huo-yang-le (35) Fleshy spurge Stems India 1848 ee lat Qian-jin-zi (36) Moleweed Seeds Mid E 973 erula ae Lz A-wel (37 Asafetida esin C As 659 a US CAFIC Wu-hua- 8 Fi Fruits Med R 1407 Foeniculum vulgare Miller X1ao0-hui-xiang (39) Fennel Fruits Med R 659 yarcinia hanburyi Hooker f. &/or G. mo- Teng-huang (40) Gamboge Resi SE As 756 r Gomphrena glo bosa L. Qian-ri-hong (41) Globe amaranth Flowers Tr Am 1688 Gossypium ees L. &/or G. hirsutum Mian-hua-gen (42) Upland cotton Roots Sw As 1596 Hordeum vulgare L. Mai-ya (43) Barley Fruits Sw As 540 Hydnocarpus eurzil Warb. &/or H. anthel- Da-feng-zi (44) Chaulmoogra Seeds SE As 1536 minthicus Gagnep Impatiens balsamina L. Ji-xing-zi (45) Garden balsam Seeds India 1407 Ipomoea nil (L.) Roth Qian-niu-zi (46) Morning glory Seeds Tr As 540 Tsatis tinctoria Da-qing-ye (47) W Leaves Eur 1061 Kochia scoparia (L.) Schrader Di-fu-zi ( Summer cypress Seeds Med R 100 Lablab purpureus (L.) Sweet Bai-bian-dou (49) Hyacinth bean Seeds Med R 540 ANIDIGAW ASANIHO TVNOILIGVUL “NH [0661 16P TABLE |. Continued. YEAR RE- PORTION PLACE OF CORDED SCIENTIFIC NAME CHINESE NAME! ENGLISH NAME USED ORIGIN? (A.D.) Lantana camara oe Ma-ying-dan (50) Lantana Flowers, roots Tr Am 1848 Lawsonia inerm Zhi-jia-hua (51) Henna Leaves ed R 300 Liquidambar dese Miller Su-he-xiang (52) Storax Resin S 547 Linum usitatissimum Hu-ma-zi (53) Flax Euras 100 Luffa aegyptica Miller Si-gua-luo (54) Sponge gourd Fiber of fruit Tr As 1596 Vimosa oe an-xiu-cao (55) Sensitive plant Whole plant Tr Am 1848 Mirabilis jala Zi-mo-li (56 Four-o’clock oots Tr Am 1708 Myristica pragrans Houtt. Rou-dou-kou (57) eds Indon 973 Nerium oleande Jia-zhu-tao (58) Oleander Leafy shoots ed 1688 Nicotiana ee L. Yan-cao (59) Tabacco ves Tr Am 1624 Ocimum basilicum L Luo-le (60) Sweet basil Leafy shoots Tr As 1061 Opuntia dillenii Haw Xian-ren-zhang (61) Prickly pear eshy stem Tr Am 1848 Panax quinquefolius L X1-yang-shen (62) American ginseng oots EN Am 1757 Papaver somniferum Ying-su-ke (63) Opium po Fruits AsM 973 Phoenix dactylifera L. Hai-zao (64) ate ruits Med R 300 Piper longum L Bi-ba (65) Long pepper Young fruits Tr As 973 Piper nigrum L Hu-jiao (66) Black pepper its Tr As 659 Pistacia ve Wu-ming-zi (67) Pistachio Seeds, bark SW As 739 Pogostemon cablin (Blanco) Bentham Guang-huo-xiang (68) Patchouli Leafy shoots Philip 1061 Populus diversifolia Schrenk &/or P. eu- Hu-tong-lei (69) Amber in Mid 89 Dhratica Oliver Portulaca oleracea L Ma-chi-xian (70) Purslane Whole plant Tr Am 9343 Prunus dulcis (Miller) D. A. Webb Bian-he-tao (71) Almond Seeds SW As 739 Psidium guajava L. Fan-shi-liu (72) Guava Leafy shoots C Am 1848 Psoralea corvlifolia L. Bu-gu-zhi (73) Bauchee seed Seeds India 973 nica granatum Shi-liu (74) Pomegranate Fruits As M 540 Quisqualis indica L Shi-jun-zi (75) Rangoon creeper SE As 973 aphanus sativus L Lai-fu (76) Radish Seeds, roots Med R 659 Ricinus communis Bi-ma-zi (77) Castor bean Seeds Afr 659 C6P WOLAYOUUV ATONYV AHL AO TVWNUNOFL IL ‘TOa] TABLE 1. Continued. YEAR RE- PORTION PLACE OF CORDED SCIENTIFIC NAME CHINESE NAME! ENGLISH NAME USED ORIGIN? (A.D.) Rosmarinus officinalis L. Mi-die-xiang (78) Rosemary Leafy shoots ed R 739 Ruta grave Yun-xiang (79) Rue Leafy shoots Med R 273 Santalum album L. Tan-xiang (80) White sandalwood Wood Indon 540 Saussurea lappa (Dene.) C. B. Clarke Guang-mu-xiang (81) Puchok, costusroot Roots Kashmir 100 Scaphium scaphigerum (G. Don) Guib. & Pang-da-hai (83) Pungtalai Seeds S 739 Planchon Sesamum orientale L. ae zhi-ma (82) Sesame Seeds Afr 1 Strychnos nux-vomica L. Ma-quian- Zi (84) C eeds Tr As 1596 Styrax benzoin Dryand &/or S. benzoides An-xi-xiang (85) Sumatran benzoin; Siam Resin SE As 659 Craib. or S. tonkinensis Craib. enzol Syzygium pals (L.) Merr. & Perry Ding-xiang (86) Cloves Flower buds Indon 973 Tamarindus indicus uan-jiao (87) Tamari Fruits fr 1370 Terminalia chebula R e-zi (88) Myrobalan Seeds India 659 Trigonella | aad ee L. Hu-lu-ba (89) Fenugreek eeds ed R 1061 Bos taurus Gm Niu-huang (90) x bez Gallstones India 1848 Elephas maximus cn Xiang-pi (91); Xiang- Elephant Hide, tusks As 1596 Physeter catodon L. Long-yang Sperm whale be Indon 1228 Rhinoceros unicornis L. Xi-jiao (94) Rhinoceros orns s l R. sondaicus Desmarest &/or R. suma- Rhinoceros Horns Indon 100 trensis Cuvier R. bicornus L. &/or R. simus Cottoni Rhinoceros Horns Afr 1840 [0661 ANIDIGAW ASANIHO TWNOLLIGVYUL ‘NH ' The numbers in parentheses in this column correspond to the numerical sequence of Chinese idiograms provided in APPENDIX I. > The abbreviations used are: Afr = Africa, As M = Asia Minor, C As = Central Asia, EN Am = Eastern North America, Eur = Europe, Euras = Eurasia, Indon = Madagascar, Med R = Meditterranean Region, Mid E = Middle a a = Philippines, S Am = South America, SE As = Southeast Asia, W As = Southwest Asia, Tr Am = Tropical America, Tr As = Tropical Asia, WI = t Indie 3 A native of the New World, this species was evidently dispersed to the Old by aa means, most likely through the dispersal of its dustl transported them to the upper atmosphere. This would explain its presence in TCM before Columbus had discovered the New World ik ds in storms that €6P 494 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 macopoeias are not properly dated, with some of them indicating just the reign of a certain emperor. The information traced from such records can only be broken down into broad categories of particular Chinese dynasties: the Han Dynasty and before, the Three Kingdoms Epoch, the North-South Division Epoch, the Tang Dynasty, the Song Dynasty, the Yuan Dynasty, the Ming Dynasty, and the Qing Dynasty. The designation of dynasty and epoch follows the chart given in the revised American edition of Mathews’ Chinese-English Dictionary (see Mathews, 1972). In the following explanations of the introduction of extra-Chinese elements, the romanization of the authors and the references follows the New China Dictionary (Xin-hua Zi-dan), hyphenated for the convenience of readers who do not know Chinese. The phrase ben-cao appears in many of the herbals and pharmacopoeias. This term has been translated as “materia medica.” From the Tang Dynasty onward, many ben-cao were prepared under imperial orders. Publications sponsored and supported by the central government, and those prepared by individuals approved by the emperor, are here translated as “phar- macopoeias.” Ancient Chinese herbals were hand-copied. Many of them were lost, others were changed by later workers with the original authorship retained to lend authority, and still others were incorporated into later works, giving credit to the original writers. Li Shi Zhen (1597) was the finest author of the third type. In his Ben-cao gan-mu, he first incorporated all known ancient records for each item entry, giving the name of the authors, and then adding his own observations and discoveries. The following account on the dating of the earliest records of the extra-Chinese elements in TCM represents a digest from Li (1597), Zhong Yao Zi (Anonymous, 1959-1961), and Zhong Yao Da Ci Dian (Anonymous, 1977-1979). These herbals and their respective authors have been included in the literature cited for this paper. For each entry an English translation of the title is followed by a pin-yin romanization, and as much pertinent publication information has been added as possible. These historical documents are available in research libraries in North America such as the collections of the Library of Congress and the Yenching Library at Harvard University. To assist scholars familiar with Chinese, a numerical list of Chinese idiograms for these titles is presented in APPENDIX IJ. The number of each reference in APPENDIX II follows the reference in the LITERATURE CITED section of this paper. HAN DyNaAstTy AND BEFORE (B.C. 206—A.D. 200) The first record of Chinese medicinal material is the Herbal Classics of the Divine Plowman (Shen-nong ben-cao jing). This work represents a crystalli- zation of the empirical knowledge of the prehistoric and early historical people of China about the healing efficacy of natural objects (plants, animals, water, earth, and minerals). The authorship is not known, but it is attributed to Shen- nong, a mythological figure who is said to have tried hundreds of kinds of plants, to have used himself as an experimental subject, to have been poisoned three times daily, to have been saved, and to have learned how to help the 1990] HU, TRADITIONAL CHINESE MEDICINE 495 people in their healthcare through his experiences. In this account, the exotic elements recorded included Benincasa hispida, Cannabis sativa, Kochia sco- paria, Linum usitatissimum, Saussurea lappa, Sesamum orientale, and Rhina- ceros unicornis. THREE KINGDOMS AND WESTERN JIN Epocus (A.D. 200-300) During this period there were two persons who wrote about their experiences in South China, where they had encountered Arabian traders with unusual objects. In Collections of a Man of Peace (Zi-an ji) Sima Sui (A.D. 231-273) mentioned that Ruta graveolens was used for keeping silverfish from books. Ji Han (A.D. 263-306) in A Description of Plants of the South (Nan-fang cao-mu Zhuang) recorded Phoenix dactylifera and Lawsonia inermis, introduced by Arabian merchants, and Areca catechu, from tropical Asia. NorTH SOUTH DIVISION EPOCH (A.D. 425-590) The first increase in recorded knowledge of Chinese materia medica occurred during this period. The outstanding recorder was Tao Hong-jing, a native of southern Jiangsu Province. He was a well-educated person who refused to serve the government, preferring to retreat into the mountains to learn from the people of their experiences in preserving life. He considered that by leading this type of life, he could achieve the manifestation of the heart of the sages in the service of the people. He prepared two treatises on materia medica, Unrecorded Materials of Eminent Physicians (Ming-yi bie-lu) and A Commen- tary on Shen-nong’s Herbal Classics (Shen-nong ben-cao jing ji-zhu). In the second work, he incorporated his findings of Unrecorded Materials of Eminent Physicians into a copy of Shen-nong’s Herbal, using red ink for the information taken from Shen-nong’s Herbal and black ink for his additions. Tao’s work was quoted repeatedly by Li Shi-zhen. Aqguilaria agallocha, Boswellia carteri, Brassica hirta, Lablab purpureus, Punica granatum, and Santalum album were recorded by Tao. TANG Dynasty (A.D. 618-905) Emperor Gao-zong ascended the throne in A.p. 650. He was concerned with having a complete pharmacopoeia and appointed the Minister of Works, Li Chi, to revise Tao’s Commentary on Shen-nong’s Herbal Classics. The com- plete work consisted of seven volumes and is known as Duke Ying’s Tang Pharmacopoeia (Ying-gong Tang ben-cao). A decade later, Su Gong pointed out the need for a revision of this pharmacopoeia. The Emperor appointed a team of 22 persons to cooperate with Su Jing. The complete work, the New Tang Pharmacopoeia (Tang ben-cao), consisted of 53 volumes with illustra- tions. The extra-Chinese elements in this work were Abutilon theophrasti, Brassica rapa, Caesalpinia sappan, Daemonorops draco, Dryobalanops aro- matica, Ferula assafoetida, Foeniculum vulgare, Piper nigrum, Raphanus sa- tivus, Ricinus communis, Styrax benzoin, and Terminalia chebula. In the Tang Dynasty, Japan and Korea had diplomatic relations with China. Soon after 496 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 71 this work was completed, it was copied and carried by returning students and/ or diplomatic or Buddhist missions to those countries. Several individuals interested in the materia medica of China did indepen- dent research on the subject. Chen Cong-qi completed the Unrecorded Materia Medica (Ben-cao shi-yi) in A.D. 739. He was a critical worker and added ap- proximately 400 items not mentioned in the New Tang Pharmacopoeia. These additions were quoted by Li Shi-zhen. The extra-Chinese material included in Chen’s work were Alhagi pseudalhagi, Pistacia vera, Rosmarinus officinalis, and Scaphium scaphigerum. By the middle of the eighth century, the first materia medica of subtropical and tropical China and adjacent areas of the Indo-Malayan Peninsula appeared. This was Li Xun’s 4 Register of the Medicine Imported via the Seas (Hai-yao ben-cao), completed in A.D. 756. Li’s grandfather was an Arab with a successful spice and medicine trade; he resided partially in Guangzhou, then known as Nan-hai. His riches brought him such fame and honor that he was elevated to share the surname of the imperial family, Li, during the Tang Dynasty. With this background, and knowing the family business, Li Xun was qualified to write on the imported medicines that came from the South Seas. In this work, a product prepared from Garcinia hanburyi or G. morella, Teng-huang, was first mentioned. By the close of the Tang Dynasty, an officer of the imperial government wrote about the strange things he saw beyond the Nanling Range. Liu Xun, in his Records of Strange Articles (Ling-biao lu yi, A.D. 889, 890), recorded a resin called tears of the Arabian tong (Hu-tong lei) and an almond that he described as a peach with a compressed kernel, both of which were seen in the homes of the Arabian traders. He gave a vivid description of the almond, saying in Chinese that this kind of peach was cultivated not for the flesh, but for the compressed kernels with thin shell that tasted like the seed of the northern peach; one end of the compressed kernel is round, and the other 1s pointed. These articles are products of Populus euphratica and Prunus dulcis, respec- tively. SONG Dynasty (A.D. 960-1200) In the history of Chinese materia medica, the Song Dynasty could be con- sidered the Golden Era for the publication of pharmacopoeias. The first Em- peror, Tai-zu, appointed the Imperial Pharmacist, Liu Han, and a Taoist monk, Ma Zhi, to lead a team of nine in revising the Tang pharmacopoeia. Using the works written during the Tang Dynasty as references, they organized the ma- terial, placed many former names into synonymy, and added 133 new items. It took them seven years to complete the revision, which filled 21 volumes, with the accounts from Shen-nong ben-cao written in white ink and all other references in black ink. This work is known as the Pharmacopoeia of the Kai- bao Reign (Kai-bao ben-cao, A.D. 973). Exotic elements recorded were Aloé barbadensis, Amomum krervanh, Carthamus tinctorius, Commiphora myrrha, Euphorbia lathyris, Myristica fragrans, Papaver somniferum, Piper longum, Psoralea corylifolia, Quisqualis indica, and Syzygium aromaticum. 1990] HU, TRADITIONAL CHINESE MEDICINE 497 In A.p. 1058 the fourth emperor, Ren-zong, had the imperial medical team, headed by Zhang Yu-x1 and Lin Y1, revise the 100-year-old Pharmacopoeia of the Kai-bao Reign. The working team added 82 new items and revised 17 entries. Known as the Revised Pharmacopoeia of Gia-you Reign (Gia-you bu- zhu shen-nong ben-cao, A.D. 1061) the work consists of 20 volumes. The new exotics were Coriandrum sativum, Cucumis melo, Isatis tinctoria, Ocimum basilicum, Pogostemon cablin, and Trigonella foenum-graecum. Emperor Ren- zong also issued a decree ordering officers of all the states and counties to prepare illustrations of their native natural objects used for healing purposes. On gathering the drawings, the Emperor appointed Su Song to prepare de- scriptions and to be the editor of the new work, which consists of 21 volumes and is known as An J/lustrated Pharmacopoeia (Tu jin ben-cao, A.D. 1062). This work has been quoted by all subsequent authors dealing with Chinese materia medica. In A.D. 1108, the second year of the Da-guan Reign, Emperor Hui Zong received a manuscript on materia medica prepared by Tang Shen-wel, a prac- titioner from Sichuan Province. In this manuscript, Tang modified the text of the Revised Pharmacopoeia of Giao-you Reign and incorporated the illustra- tions from An Illustrated Pharmacopoeia. To this combined work he added prescriptions and notes for dietary applications to each entry. He titled the manuscript Verified Identification of Materia Medica (Zheng-lei ben-cao). The mperor accepted the manuscript, which then became known as the Phar- macopoeia of the Da-guan Reign (Da-guan ben-cao). In this work the fragrant woods of Dalbergia sissoo and D. parviflora were first mentioned. Sinologists have characterized the Tang (A.D. 618-905) and the Song dy- nasties as the grand period of growth of Chinese culture. Contacts with nations far and near were frequent. Members of religious expeditions and diplomatic missions to India and southeastern Asia learned to use things unfamiliar to them, and they frequently brought samples and/or propagules back to China. Soldiers in war and traders in peace brought exotic things from nations near. With regard to the extra-Chinese elements used in TCM, the nature of the items recorded during the Tang Dynasty is very different from those of the Song Dynasty. In the Tang period medicines, food, spices, and resins and dyes, each with 25 percent of the total, comprised the material record. In the Song Dynasty the number of fragrant woods increased very prominently, consisting of one-half of the newly recorded exotics. During the Song Dynasty the central government supported the preparation of pharmacopoeias and evidently also encouraged the importation of fragrant woods and spices; furthermore, it es- tablished a Bureau of Spices and Medicines to oversee their importation. The reason for such governmental efforts will be explained below. YUAN Dynasty (A.D. 1206-1368) Historians often characterize this period as China under the Mongols. Al- though there was much communication and exchange of materials and ideas between the West and the East, very few new exotics were added to the Chinese materia medica. Practitioners of the Yuan Dynasty simply followed tradition, 498 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 used the phamacopoeias of the Sung Dynasty, prepared their prescriptions, and carried on with their business. Near the end of the Yuan Dynasty, Zhu Zhen-heng prepared 4 Supplement to Herbal Commentaries (Ben-cao yan-y1 bu-yi, A.D. 1358). In this work chaulmoogra (Da-feng-zi— Hydnocarpus kurzii and/or H. anthelminthicus) was first recorded. MinG Dynasty (A.D. 1368-1644) Countries such as Japan and Korea shared the heritage and information on Chinese medicine from the herbals and pharmacopoeias prepared in the Han, Tang, and Song dynasties. The Western World learned about Chinese medicine through the materia medica completed in the Ming Dynasty, particularly from Li Shi-zhen’s Materia Medica with Commentaries (Ben-cao gang-mu, 1596). Unlike the works completed during the Tang and Song dynasties, those done during the Ming Dynasty were realized through individual interests and efforts. Five treatises dealing with exotic elements used in TCM are discussed here. The Materia Medica of Southern Yunnan (Dian-nan ben-cao, A.D. 1370), by Lan Mao, contained 7amarindus indicus. Famine Herbal (Jiu-huang ben-cao, A.D. 1407), by Zhu Xiao, included Cassia occidentalis, Ficus carica, and Im- patiens balsamina. Ni Zhu-mo, in Organized Collections of Materia Medica (Ben-cao hui yan, A.D. 1624), first recorded tobacco (Nicotiana tabacum), in- troduced primarily for smoking, but in recent years rural people in Fujian have used it in the preparation of a poultice, which is mixed with rice gin to treat breast cancer. Li Shi-zhen, after failing the qualifying examinations for governmental of- ficers, entered the second-best profession of the time, becoming a practitioner in medicine. With his knowledge of the Chinese classics and his practical experience in diagnosing illnesses and prescribing remedies, he prepared a manual for practitioners of traditional Chinese medicine, Ben-cao gang-mu When Li used gang and mu, he implied arrangement and classification. Brom the approximately 800 references, he selected 1518 items; he also provided 374 new entries. For each of the 1892 objects, he gave a recognized name, which he called gang, and cited all the synonyms with his commentaries, which he called mu. To such a framework, he added information on properites, tastes, and efficacy against illness, as well as recipes. In his treatment the new extra- Chinese elements were Acacia catechu, Bombax ceiba, Canavalia ensiformis or C. gladiata, Citrullus lanatus, Cleome gynandra, Crocus sativus, Datura metel, Gossypium hirsutum, Luffa aegyptica, Strychnos nux-vomica, and the hide and tusks of Elephas maximus. During the Ming Dynasty, a famous treatise on plants and agriculture, Mono- graph of Flowers (Qun fang pu, A.D. 1621) was published by Wang Xiang-jin. In this work Capsicum frutescens was first recorded. QinG Dynasty (A.D. 1644-1910) Information on the exotic elements used in TCM before the Qing Dynasty was recorded largely in ancient herbals or pharmacopeias. During the Qin Dynasty both the elements employed and the source references changed. The 1990] HU, TRADITIONAL CHINESE MEDICINE 499 major portion of the information was recorded in treatises on horticulture and agriculture. The first such work was the famous Flower Mirror (Mi-chuan hua jing, A.D. 1688), by Chen Hua-zi. Gomphrena globosa and Nerium oleander were recorded in this work. Mirabilis jalapa first appeared in Enlarged Mono- graph on Flowers (Guang qun fang fu, A.D. 1708), by Liu Hao. Wu Yi-lo, in his Newer Version of Materia Medica (Ben-cao cong xin, A.D. 1757), first men- tioned Panax quinquefolius. Zhao Xue-min, in his Supplement of Ben-cao Gang Mu (Ben-cao gang mu shi-yi, A.D. 1765) included quinine (Cinchona succirubra or C. ledgeriana), and ambergris from Physeter catodon. The most outstanding work containing comprehensive information on the botany of China, An Illustrated Treatise on Plants (Zhi-wu ming shi tu kao, A.D. 1848), was prepared by Wu Qi-zhung. The extra-Chinese elements in- cluded in this treatise were Bryophyllum pinnatum, Carica papaya, Cathar- anthus roseus, Euphorbia antiquorum, Lantana camara, Mimosa pudica, Opuntia dillenii, Psidium guajava, and the gallstones of Bos taurus. BELATED RECORDS Two extra-Chinese elements commonly used in TCM before they were re- corded in herbals or botanical treatises are creat (Andrographis paniculata), and senna (Cassia angustifolia or C. acutifolia). Evidence of their introduction into China during the Qing Dynasty is convincing for creat and tangible for senna. Andrographis paniculata, a common weed in hedgerows throughout the plains of India, Bangladesh, and Burma, has been used locally in those countries in ethnomedicine for hundreds (thousands) of years. Some authors recorded it as an annual because it flowers during the first years of its life cycle. In the tropics new shoots emerge from the old plant after the dry season, and eventually the plant appears shrubby. I have seen the species growing in similar habitats in Hong Kong, flowering and fruiting the first year as an herb and growing as a perennial shrub in undisturbed sites. Andrographis paniculata has been reported from northern India southward to Sri Lanka and eastward to Java. George Watt (1889) reported over two dozen vernacular names for the species, including kiryat (Hindi), kalmegh (Gengali), kiriyatta (Malay), kirata (Sanskrit), and qasabhuva (Arabic), from that vast area and cited its ethnomedical properties as a febrifuge, stomachic, tonic, alterative, and anthelmintic. As a bitter tonic, stomachic, and febrifuge, the herb has been shared and spread by ethnic people throughout tropical Asia but has been unrecorded until recently. Its widespread use and cultivation in northern China 1s modern, but its actual introduction must have been prior to 1932, when it was mentioned in Records of Gathering Medical Herbs in South China (Ling-nan cai-yue lu, Xiao, 1932). From both the ethnobotanical practices and the vernacular names, it seems that Andrographis paniculata was introduced by the ethnic people living in the mountains of northern India, Balgadesh, Burma, and southern China. In the Bengal region the expressed juice of the leafy shoots is mixed with cardamon, cloves, and cinnamon, dried in the sun into small pills called alui, and used 500 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 to treat children for colic, loss of appetite, and irregular stools, and adults after a fever and for advanced stages of dysentery and general debility. Watt gave additional information that creat leaves mixed with the pulp of ripe tamarind were carried by wandering doctors as an antidote to the venom of poisonous snakes. The tribes in the mountains of Guangxi and Guangdong also used A. paniculata for fever, dysentery, and snake bite. Senna is a popular purgative in southern Asia. Its laxative principles are two glycosides, sennoside A and sennoside B. The commercial product consists of the dried leafy shoots of Cassia acutifolia, indigenous to northern Africa, and C. angustifolia, native to Yemen and southern Arabia. Dealers called the prod- uct from C. acutifolia Alexandrian senna, and that from C. angustifolia Arabian senna. Both species are cultivated in India, the source area of the senna imported into China, and both are shrubs | m tall with lanceolate leaves. Historically, senna was introduced by the Arabs to India, where the species has been extensively cultivated. We have no record of its introduction by the Arabs into China. This could have been because Chinese herbal medicine had a better and more effective purgative, rhubarb (da-huang, Rheum officinale Baillon and R. palmatum L.), which was widely accepted by the people before the arrival of the Arabian traders. Consequently, the Arabs were unable to create a market for senna in China. This situation continued until the indus- trialized European countries forced China to open her ports for international business. According to records in the list of Chinese medicines prepared under the instruction of the British Inspector General of Customs, Robert Hart, for the China Imperial Maritime Customs Department (see Anonymous, 1889), senna entered China by sea via India and Japan and overland through Yunnan- Sichuan. EFFECT OF TIME ON THE NATURE OF EXTRA-CHINESE ELEMENTS The chronological account of the recorded exotics used in TCM reveals several unique istics. First, they were obviously introduced into China a very long time ago. The record made about the beginning of the Christian era (Herbal Classics of the Divine Plowman) shows that seven exotics were already in use in Chinese medicine. Two of these items, puchok (root of Saus- surea lappa) and the rhinoceros horn, were imported overland from India. The remaining five substances were obtained from plants that were spread in as- sociation with adventurous prehistoric and/or ancient people who traveled extensively by the means of their time. Hemp arrived in China almost 5000 years ago in association with Neolithic man. The second outstanding characteristic is the continuous growth in the number of exotics, with gradual changes in what was imported. When the data are plotted in a simple graph (see FiGurE 1), the growth in number of exotics used in TCM becomes very obvious. Each point in the graph represents the accu- mulated number of exotics used in TCM at that particular time. The model is meaningful only when an item is used continuously from the time of its intro- duction. This condition is true with all the recorded exotics listed here because they are still in common use. The change in the nature of the exotics becomes obvious after an analysis 1990] HU, TRADITIONAL CHINESE MEDICINE 501 [om we ve | [emelesem | m@ [ [ RUT xz] aw oe } INTRODUCTION OF EXOTIC SUBSTANCES USED IN TRADITIONAL CHINESE MEDICINE 754 NUMBER OF ITEMS BCJAD 250 500 750 1000 1250 1500 1750 2000 TIME OF RECORDING Ficure |. Increase in number of exotics used in TCM. Broken lines indicate time before and after period investigated in this study. of their general function. Five of the earlier-mentioned seven items recorded from the Han Dynasty were from plants introduced primarily for daily ne- cessities—that is, for food and fiber. There was only one substance among them, mu-xing (dried root of Saussurea lappa), that was imported as an aro- matic medicine. At that early date, aromatic material comprised about 14 percent of the total imports. The increase in the proportion of aromatic products is very obvious. In the North-South Division Epoch, spices and aromatic materials were 37 percent of the total recorded items, in the Tang Dynasty 41 percent, and in the Song Dynasty 54 to 61 percent. The growth in exotic fragrant-wood species and aromatic substances reached a peak in the Song Dynasty due to the encouragement of the central government. During that period the State Department established a Bureau of Spices and Aromatic Medicine to deal with gifts and tributes of various diplomatic missions; the officers made requests and suggestions of articles to be brought into China in exchange for precious Chinese materials. In the Ming and Qing dynasties there was an obvious change of direction in importation. Substances for comfort and enjoyment of life other than spices and aromatic materials were brought in. Tobacco was recorded in the middle Ming, and cotton in the late Ming. Both these plants became extensively cultivated, one as a fumitory and the other for fiber. Their uses in medicine were secondary. During the Ming and Qing dy- nasties, introduction of plants for ornamental purposes increased very fast, from 27 percent of Ming importations to 50 percent of Qing records. New plants introduced for food comprised 18 percent of the species recorded in both the Ming and Qing dynasties. The third prominent feature of the exotic elements used in TCM is the effort made toward self-sufficiency in the supply of these materials. Of all the plant products, 70 percent of the species have been cultivated as important field crops in China, two thirds of these for food, fiber, and ornamentals and one third for medicine. Six of the exotic elements are adventive species that have 502 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 TABLE 2. Source areas and number of items. NUMBER OF SPECIES PLACE OF ORIGIN Plants Animals Total Tropical Asia 14 + % pes + * Indonesia India Europe Central Asia Asia Minor Philippines Madagascar Eastern North America — KM HOO BWW BUND ~) NN OO + % SSE EN BPWIDADAATNNNOW HL + ¥ COC COO COCO; Reem Om O Total (oe) \O wn + No * 94 + 2* * Indicates similar products (e.g., rhinoceros horn) from two different species of these areas are both present in Chinese markets. traveled with man unnoticed and have become weeds in China. Due to the climatic and ecological requirements of species from the humid tropics in southeastern Asia and the hot, dry arid lands of southwestern Asia, approxi- mately 30 percent of the exotic plant elements and all of the animal products from those areas used in TCM are still imported. GEOGRAPHICAL ANALYSES The geographic accounts provided here deal with the origins of the extra- Chinese elements used in TCM and the possible routes by which they were introduced. All the species are covered, whether they are cultivated, weedy, or foreign commercial products sold in Chinese markets. SOURCE AREAS AND SPECIES The names used for the source areas are loosely geographic, without any political significance. A summary of these areas and the number of species they have provided is given in TABLE 2. The areas contributing most of the exotics used in TCM are tropical Asia, the Mediterranean region, and tropical America. Next in numbers of source species are southeastern Asia, the Middle East, southwestern Asia, Central Asia, Indonesia, and India. Europe has three species, Asia Minor has two, and the remaining areas have one each. 1990] HU, TRADITIONAL CHINESE MEDICINE 503 atanen, CHINES gz} EMP] (SERICA) 2% : eo Ny oy AYO oe Map 1. Asiaca. A.D. 100 showing routes of introduction of exotic elements used in TCM. Numerals denote number of species in each area. a, indicates royal residence of Emperor Yu. b-h, collecting centers of 9 tributary principalities of Emperor Yu: b, Yu- zhou; c, Xu-zhou; d, Qing-zhou; e, Yang-zhou; f, Jing-zhou; g, Yung-zhou; h, Liang- zhou. E represents three exchange centers in Chang-an (Xi-an), Guang-zhou (Canton), and Luo-yang. ni ROUTES OF INTRODUCTION The possible routes by which exotic species were introduced for cultivation or naturalization and over which exotic market products were transmitted are shown on a map of the major national and international paths of communi- cation before and after the discovery of America at the end of the fifteenth century. In the foregoing discussion of the time of introduction of exotic ele- ments into China, the first recorded time was in the Han Dynasty. Map | shows the routes of communication within China and between nations of the East and the West. It represents a synthesis of information from Chinese and western sources. The explanation of possible routes for the introduction of exotic elements used in TCM follows several principles. The first is that plants follow man, and ancient man traveled extensively over the earth. There are many records concerning the travels of ancient peoples. In China, during the reign of Emperor 504 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Yu (2205-2198 s.c.), the supplies of the royal family were transported from the collecting centers of the nine tributary principalities (see Map |, a—h). These supply routes had become so effective that by the Tang Dynasty (A.D. 618- 905) the most perishable tropical fruit, the litchi, was transported fresh from Guangzhou to Xian (Changan). Ancient people did travel! When Confucius (B.c. 511-479) traveled with his disciples among the rulers of the feudal states, attempting to convince political leaders to accept his Doctrine of the Mean and to try his method of achieving national prosperity by starting with the well-being of the Chinese people, his contemporary, Herodotus, traveled in western Asia in Persia and many coun- tries in the Mediterranean region, Europe, and northern Africa. Historical records from this time give no indication of direct communication between East and West. Nevertheless, the intermediaries had been successful in spread- ing African species such as Sesamum orientale and Central Asian taxa like Cannabis sativa to China. The first recorded direct East-West contact that exerted obvious effects on the introduction of Chinese exotics was the famous Zhang Qian Mission to the West. However, before this Mission was possible, Emperor Wu had to make preparations by waging war to conquer rebelling chieftains and by sending gifts or negotiating marriages to appease less-resistant leaders of the ethnic peoples of the Central Asian steppes. In some of the military or diplomatic missions, thousands or even tens of thousands of people were involved. Even in Zhang’s Diplomatic Mission to the Middle East, sixty-thousand people went with gifts, tributes, and weapons, as well as provisions and food for his men and animals. Plants followed man in war and peace in both directions. In Chinese history, after Zhang’s mission, one of the captains responsible for safeguarding the communication route escaped with his men and settled in Armenia. He became an adviser to the Armenian king, helped him to achieve prosperity, and received high honors there. Consequently, together with the Chinese people and their lifestyle, plants were introduced to Armenia, the apricot (Prunus armeniaca L.) being the best-known example. By this time in Chinese history, a great agricultural civilization had devel- oped. Although kingdoms emerged and disappeared, palaces were built and destroyed, and wars occurred within and without, Chinese farmers continued working the land, discovering the secrets of nature, and trying to make the best use of all aspects of the environment to enable them to survive and to live better. One of the most outstanding features of Chinese agriculture was seri- culture — domesticating a worm that produced silk, which could provide clothing and comfort at home. This aspect of Chinese agriculture gradually became very successful, providing splendor in courts, a medium for exchange, and a commodity for international trade. Silk has played an important role in China’s foreign relations. It has been used as a medium of exchange, as gifts for all the foreign ambassadors that came to pay homage to the emperors, and as a commodity of trade with the West. Actually, China was first known to the West as Serica, the land of silk. The routes for silk export were many, and they radiated in all directions; over them, in turn, exotics were introduced into China. 1990] HU, TRADITIONAL CHINESE MEDICINE 505 Once a route of communication was opened, the traffic continued unless stopped by some natural or manmade calamity. From the earliest record of exotics, we know that in or before the Han Dynasty at least four routes of communication were opened: land routes to the Mediterranean region and Africa, to Central Asia, and to Eurasia, Kashmir, and tropical Asia; and a sea route by which rue, dates, and henna were imported to South China by Arabs. Three centuries later, in addition to the products introduced over the ancient Central Asian trade routes, aromatic products from southeastern Asia entered China. During the Tang and Song dynasties, the volume of imports and the number of items continued to increase by both the land and sea routes. It is worthy of note that in the early days, few items were introduced from India. It was only in the Tang Dynasty that myrobalan was imported, and in the Song Dynasty that bauchee seed was recorded for the first time. During the Ming Dynasty tamarind, a species widely cultivated in India, was introduced into unnan. Exotics continued to be imported via many of the ancient trade routes after the discovery of the New World and the emergence of ports for inter- national trade in the maritime provinces. The fifteenth century was an era of maritime exploration and colonization by European countries. Soon after Christopher Columbus discovered the New World and Spain colonized tropical America, Vasco da Gama of Portugal sailed from Lisbon for Asia and returned with pepper (Piper longum and P. nigrum) from Malabar. Twelve years later Portugal captured Malacca, and three years after this the Portuguese reached China. In all the main ports of southeastern Asia the Portuguese found that important businesses were managed by Chinese residents. For over one hundred years before the arrival of Europeans, the Chinese Empire of the Ming Dynasty had a vigorous maritime enterprise in the western Pacific and Indian oceans. In 1405 the King of the Straits of Malacca submitted to China and became a protectorate of the Ming Empire. Five years later a Chinese naval force conquered King Wijayababu VI of Sri Lanka and brought him as a captive to China. When Ferdinand Magellan arrived in the Philippines in 1519, he found Chinese vessels in all the important trading ports. From the middle of the sixteenth century onward, the Philippine Islands were Spanish, annexed to Mexico. The overseas Chinese are well known for their loyalty to China; in their return trips they brought back to Fujian and Guang- dong seeds and other propagules of plants they had learned to like. The red pepper and guava are good examples. At present, red pepper is cultivated in gardens throughout China, and the hot dishes of Sichuan and Hunan have won praises in metropolises worldwide. Guava is naturalized in Guangdong and Hong Kong, where people assume it is a native species. After 1842, China was forced to open her ports for international trade. British citizens were hired by the Qing emperors and the Empress Dowager to manage the Imperial Maritime Customs. British residents in Chinese ports lived in beautiful homes decorated with exotic ornamental plants. Some of these species were first brought from the Americas into hothouses in London and were later distributed to British citizens throughout the world. Some of them—for ex- ample, Lantana, Mimosa, Mirabilis, and Cleome—arrived in China. After they were grown by the Chinese people or escaped from cultivation, they were used 506 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 for medicinal purposes in China. Meanwhile, British businessmen brought medicinal products such as senna from India and ivory from Africa into Chinese markets for profit. The United States shared the economic interests of the western powers in China. American clippers sailed between Boston and Guang- zhou carrying American ginseng and other commodities— including opium — for China, and tea, cotton, silk, and porcelain were acquired in return. PRACTICAL ASPECTS In a span of four millennia and from places extending over all habitable continents of the earth, plant and animal products have been introduced into China, where they have been used for medicinal purposes. Why are these substances used in healthcare? Is it merely psychological and cultural, or are there certain physical and chemical bases for such practices? This is a com- plicated subject to be discussed in a limited space. However, a few examples can be given to illustrate the nature of the issue. PLANTS INTRODUCED FOR DAILY NECESSITIES An analysis of the nature of the extra-Chinese plants and their role in Chinese culture shows that a majority of the species were introduced by people as daily necessities (for example, food and fibers) or for pleasure (e.g., fumitories, mas- ticatories, or ornamentals). These species are still cultivated primarily for the same purpose for which they were introduced, with medicinal uses secondary. The best examples are the radish (Raphanus sativus), the hyacinth bean (Lablab purpureus), and sesame (Sesamum orientale). RaApIsH. The radish probably came with Neolithic people from the Mediter- ranean region as a fast-growing vegetable (the small fleshy root or just the greens). Through selection and cultivation by ingenious Chinese farmers, hundreds of varieties were selected to fit the climatic and ecological conditions of a country extending from the seashore in the east to the arid plateau in the west, so that people might have fresh or pickled radish throughout the country and at all times of the year. Radish seeds are used in Chinese medicine as a tonic for internal organs and for improving strength. In villages, after the seeds are harvested, the old dried roots are saved for making a tea to improve digestion and to cure gas pains. Phytochemists have isolated many organic acids from the radish, including coumaric, caffeic, ferulic, and gentisic acids. Radishes are very rich in calcium. HYACINTH BEANS. The hyacinth bean is a common vegetable in the frozen- food section of groceries in America, where it is sold as “Italian beans.” In China it is cultivated for the young fruit, which is used as a vegetable. However, the mature seed is prescribed for diabetes mellitus. It is rich in protein, calcium, phosphatide, pantothenic acid, phytin, and tyrosinase. SESAME. Sesame is cultivated for its seeds, which are an important source of edible oil. It is prescribed for the treatment of both external and internal ailments. Sesame is rich in protein and calcium and contains arachic, linoleic, and palimitic acids, sterols, sesamin, sesamolin, sesamol, and vitamin E. 1990] HU, TRADITIONAL CHINESE MEDICINE 507 Exotics INTRODUCED AS MEDICINE Approximately one-third of the exotics were imported into China for use as medicine. A few examples of the imported extracts, resins, aromatics, and animal products (aloé from Africa via Arabia, asafetida from Central Asia, gambier from India, gamboge from Thailand, and rhinoceros horn from south- eastern Asia) are selected to show the reasons for their importation. ALoé. Aloé was first introduced into South China in the form of a reddish brown extract by Arabian traders. Most of its Chinese names were direct translations from the Arabic. In the Pharmacopoeia of Kai-bao Reign (A.D. 973), it was recorded as a resinlike powder from Persia. When ancient Chinese consumers asked the trader the nature of the material, they were told “elephant bile,” due to its bitter taste. For this reason it was also recorded as xiong dan in Chinese literature. The material is an extract obtained from the leaves of Aloé barbadensis, a spiny succulent perennial attaining a height of 1 m (see FiGurRE 2). “Chinese aloé” entered China from the south and has become naturalized in many parts of Fujian. Although there is no record of the intro- duction of the plant to South China, we know that it has been grown, segregated and developed into special forms there. In 1817 a living aloé plant was taken from South China to London. Adrian Haworth, a specialist in succulent plants, recognized a distinctive leaf feature (spots on both surfaces) and provided the name A. barbadensis var. chinensis (Haworth, 1819). Individuals were culti- vated in British gardens under this name until J. G. Baker, on the basis of a flowering plant, renamed the taxon A. chinensis and determined it to be closely related to 4. abyssinica (Baker, 1877). In Chinese medical practice aloé has been used in combination with Pinellia tubers (Araceae), Atractylodes rootstock (Compositae), and licorice for the treatment of convulsions and epilepsy in children and adults. It contains al- oenin, aloin, alomicin, aloesin, barbaloin, p-coumaric acid, protein, and cal- cium-oxalate crystals. Recently an Aloé Association was organized in South China to promote the cultivation of aloé and the production of medicines and cosmetics from them. Among my houseplants there is an old aloé (see FIGURE 2a) with unspotted, grayish, lanceolate leaves, spiny along the margin. Two years ago I put it out in the yard. By the end of the summer, many suckers had arisen from beneath the surface of the soil. I transplanted the suckers and repotted the mother plant. The immature plants appeared very different from the old one in that their leaves were erect, terete, and spotted throughout. Two years have passed; these young plants all have spotted leaves, the character used by Haworth in the description of A/oé barbadensis var. chinensis. China really has no native aloé, and the feature used in identifying the Chinese aloé is unstable, changing with the age of the plants. ASAFETIDA. With respect to animal and plant resources for medicinal use, Central Asia is a land of poverty. In spite of geographic proximity, the only plant product imported from there into China for medicinal purposes 1s asa- fetida (Ferula assafoetida), known in TCM as a-wei. The crude drug is an irregular tearlike oleoresin with a powerful, foul odor and an acrid taste. The source species is a perennial herb that requires five to 18 years of vegetative 508 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 URE 2. Aloé chee, source riciae ey of commercial aloé: a, habit; b, flowering oar showing branched scape; c, portion of flowering branch, showing position of fully open owers, d, section of leat showing mucilaginous central portion; e, flower, showing outer } for basal half; f, flower, perianth split open and 4 stamens removed, showing ovary and free margin of inner segments; g, portions of 2 stamens, showing abaxial and adaxial views of anthers; h, vertical section of ovary; i, apical portion of trigonous style and stigma. growth before it sends up an erect flowering stem. The people of Central Asia use the fresh leaves in food as Europeans use parsley. The straight, stout, canelike flowering stem (see FiGure 3) can attain a height of two meters. The compound cauline leaves gradually become smaller upward, with the middle and upper ones subtending the flowering branches. The lower umbels bear staminate flowers only, while the upper ones include some perfect flowers as well. Each plant may produce hundreds of small, flat fruits; these split vertically 1990] HU, TRADITIONAL CHINESE MEDICINE a0y GURE 3. Ferula assafoetida: a, single erect stem bearing bipinnately compound umbel; g, fully cael fruit; h, mature fruit after separation of mericarps, and attached at the top to the carpophore; i, a transverse section of a fruit showing the broad wings and the longitudinal canals (vittae). into two winged mericarps, each containing one tiny seed. The mother plant dies after the seeds mature. Asafetida is collected by removing the soil around the crown before the plant flowers, then cutting off a portion of the crown and protecting the plant from the sun. A thick, milky juice exudes from the cortex of the fleshy rootstock, collects on the surface, and hardens. After ten days the hardened mass is gathered, and another portion of the stem is cut off to let fresh juice ooze out. This process is repeated every ten days until the plant is exhausted, which usually takes about three months. The rootstock eventually dies. Asafetida was introduced into China by Arabian traders over the sea route and first appeared in the pharmacopoeia of the Tang Dynasty (A.D. 659). A-wei was a Cantonese translation of the Arabian trade name. It is used for indigestion, lack of appetite, and gas pain. It contains volatile oils (10-71 percent), resin (40-64 percent), gum (25 percent), and ash (1.5-10 percent). Approximately 510 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 45 percent of the volatile oils is sec-butyl propenyl disulfide, which imparts the characterstic foul odor to the product. Ferulic acid and farnesiferol have also been identified. A-wei pills are available for pains of the chest and sides and lack of appetite. These are made from 7.5 g asafetida softened in vinegar into a paste, 15 g costusroot, 15 g betel nut, and 7.5 g black pepper. The powered costusroot, betel nut, and black pepper are mixed with the softened asafetida and cemented with cooked millet to make pea-sized pills. Taken with ginger tea, these are said to open up the internal physiological channels. The recipe is adapted from the well-known collection Life Saving Recipes. Recently, two species growing in Xin-jiang, Feru/a caspica and F. conocaula, have been identified as alternative source species for a-wel. GAMBIER. Gambier is prepared by decoction and evaporation from the heart- wood of a tropical tree, Acacia catechu (see FiGuRE 4). The plant, which may attain a height of 6-12 m, has bipinnately compound leaves, very tiny leaflets ciliate along the margin, small, yellowish flowers crowded in an elongate spike, and flat pods. In tropical Asia gambier is chewed with betel nut. It is exported for tanning and dyeing. In China it is combined with an equal amount of the dried tuber of Bletilla striata (Thunb.) Reichb. f. (bai-ji), pulverized, and applied to sores in the mouth and boils. Gambier is an astringent containing catechutannic acid, catechin, epicatechin, phlobatannin, fisetin, quercetin, quercetagetin, pro- tocatechuic tannins, and pyrogallic tannins. The tree has been introduced into Yunnan Province for commercial production. GAMBOGE. Gamboge is a resin prepared from Garcinia morella, a tropical tree 18 m tall with opposite leaves, small, yellow, unisexual flowers, and juicy, round fruits (see FiGURE 5). The material is prepared by making incisions in the bark. A viscid yellow juice oozes out and dries on exposure to air. The resin is collected in hollow bamboo sections, where it hardens into cylinders. It is soluble in water, alcohol, and oil and is much used by artists. It is fatally poisonous if taken internally. At present, it 1s only produced on a few small islands in southeastern Thailand and is taken to Singapore for dis- tribution. In China it is an important external medication for cutaneous dis- eases, especially carbuncles, abscesses, ulcers, scabby head, and cancerous tu- mors. From the material, phytochemists have isolated a-guttiferin, b-guttiferin, morellic acid, isomorellic acid, morellin, isomorellin, dihydroisomorellin, ethoxydihydroisomorellin, morelloflavone, and neomorellin. RHINOCEROS HORN. Whole horns have been brought into China from Thailand, Java, Sumatra, India, and adjacent tropical areas for medical use since pre- historic times. More recently, horns have also been imported from Africa for the same purpose. Rhinoceros horn was recorded in the first Chinese herbal, Herbal Classics of the Divine Plowman, and it is still in common usage. In TCM rhinoceros horn is used in the form of slices or powder and serves as a cooling and detoxifying agent, or as an ingredient in the preparation of emer- gency remedies for quick relief of heatstroke, high fever with delirium, or convulsions in children; there are many ancient formulas for it. After World War II medical scientists began to investigate its chemical composition and to study its pharmacological effects with animal assays. It has been found that 511 os 5 Be Wey os 8 \¢ yr. & WA Vie , ZZ, ore, 2 Ce”, Y Y \ a ge - ein fege ling ce ves a" (< & oe Poems A a ios Meg ti CRCERNT Figure4. Acacia catechu: a, a fruiting branch showing 3 fruits and | even-bipinnately compound leaf with gland on adaxial side at apex of petiole; b, portion of growing young shoot, showing unfolding leaf with petiolar gland and 2 axillary spikes, | with flowers and 1 with buds; c, a sessile flower, showing small nan and corolla and numerous stamens; d, flower bud subtended by linear bract, e, vertical section of flower, showing with sigmoid funiculus attached to adaxial suture of pericarp; j, pair of leaflets, showing ciliate margin. eee horn is very rich in protein (keratin), particularly in cystine, histi- e, lysine, and arginine. It also contains peptides, guanidine derivatives, Becacel and asparagic acid. Water extracts of rhinoceros horn excite the muscles of isolated toad hearts and of in situ rabbit hearts. It strengthens 512 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 FiGurE 5. Garcinia morella: a, flowering branch, showing opposite leaves and sta- minate flowers: b, a fruiting branch, showing a fruit with persistent calyx; c, mature € short filament and orbicular apex of connective; h, vertical section of stamen, showing thecae and the thin position for dehiscing; i, lateral view of dried immature seed with elevated raphe and germination heartbeats and increases blood flow. Intravenous injections reduce experimen- tally induced fever. In the hind legs of toads it causes temporary short con- tractions followed by relaxation; it also relaxes the smooth muscles isolated from the intestine and uterus of rabbits. EMERGENCY REMEDIES Hospitals and ambulance services were unknown in China before 1910 and are rarely available to the general public today. In emergencies, such as heat- rm (wan), pode (san), and red-coated elixir (dan)—that are available without a prescription. A random selection of ten of these emergency remedies from A Handbook of Chinese Pharmaceuticals in Traditional Medicine (Shong Yao Zhi Ji Shou Ce; Anonymous, 1974), reveals that each contains some imported material in amounts varying between 0.5 and 78 percent by weight. The number 1990] HU, TRADITIONAL CHINESE MEDICINE 513 of ingredients in these rescue remedies varies from six in a cardiovascular pill (guan-xin su-he wan) to 50 in a blood-flow activator (shen-xue huo-luo wan). The amount of herbal material taken is very small, with 90 percent of the dosages being less than | g. Translations of two samples are presented below to show the nature of the remedies and the importance of exotic elements in their preparation. In the recipes, only the ingredients now imported from abroad are marked with asterisks. Products from the cultivated extra-Chinese elements, such as fennel seed from Foeniculum vulgare and menthol from Mentha hap- localyx, are not marked. PEOPLE’S ELIXIR (ren-dan): A/pinia katsumadai Hayata,* Katsumada’s galanga, 31 g; Amomum villosum Lour., grain-of-paradise, 31 g; Areca catechu L.,* betel nut, 31 g; Canarium album (Lour.) Raeusch., Chinese olive, 31 g; Carthamus tinctorius L., saflower, 15.5 g; Cinnamomum loureiroi Nees, Saigon cinnamon, 31 g; Citrus reticulata Blanco, Mandarin orange peel, 31 g; Dryobalanops aro- matica Gaertner,* Borneo camphor 7 g; Foeniculum vulgare Miller, fennel seed, 31 g; Glycyrrhiza uralensis Fischer, licorice, 248 g; Mentha haplocalyx Briq. (or M. arvensis L.), menthol, 28 g; Poria cocos (Schw.) Wolf, Chinaroot, 31 g; Saussurea lappa (Dcne.) C. B. Clarke,* costusroot, 46.5 g; Syzygium aroma- ticum (L.) Merr. & Perry,* cloves, 15.5 g; Moschus moschiferus L., musk, 3.1 g. The finished product is packed in tiny glass bottles as small red pills coated with cinnabar, each weighing 0.03 g. The dosage varies from between three and 20 pills, depending on the situation. In this formula, 20.5 percent of the ingredients by weight are imported. ROYAL STORAX PILL (su-he-xiang): Aquilaria agallocha Roxb.,* aloeswood, 31 g; Boswellia carteri Birdw.,* frankincense, 31 g; Cyperus rotundus L., nut grass, 31 g; Dryobalanops aromatica Gaertner,* Borneo camphor, 15 g; Liguidambar orientalis Miller,* storax, 15 g; Piper longum L.,* long pepper, 31 g; Santalum album L.,* sandalwood, 31 g; Saussurea lappa (Dene.) C. B. Clarke,* costusroot, 31 g; Styrax benzoin Dryand. (or S. tonkinensis (Pierre) Craib),* benzoin, 31 g; Syzygium aromaticum (L.) Merr. & Perry,* cloves, 31 g; Terminalia chebula Retz,* myrobalan, 31 g; Rhinoceros bicornis L. (or R. sumatrensis Cuvier),* rhinoceros horn, 31 g; Moschus moschiferus L., musk, 23 g; and cinnabar, 31 g. This formula was first designed in the royal palace during the Song Dynasty and is still used. It calls for 14 ingredients, of which 11 (78 percent by weight) are exotics. The finished product consists of pills weighing ca. 3 g including the cementing material. Each dose contains ca. 0.16 g of herbal material. CRITERIA FOR DETERMINING EXOTICS In the foregoing list of exotics in Chinese materia medica (TABLE 1), items (e.g., Abutilon theophrasti and Cannabis sativa) are included that were formerly 514 JOURNAL OF THE ARNOLD ARBORETUM [voL. 7] recorded, on the basis of assumption, as native to China (Spencer, 1984; Wang & Guo, 1980). Various other native Chinese plants (e.g., Juglans regia) that used to be regarded as introduced on the basis of hearsay are absent. After extensive field observations of the ecological conditions in China, investigations of ethnobotanical practices within the country, and intensive phytogeograph- ical, systematic, and cytotaxonomic studies made in the herbarium and the library of the Harvard University Herbaria and the Chinese Library of the Harvard Yenching Institute, I used five criteria for determining which elements used in TCM are exotics. 1) Ecologically, the species is not adapted to grow in the natural flora or fauna of China and 1s thus dependent on man for survival. 2) Morphologically, it is a distinct taxon without closely related species in China. 3) Biogeographically, the center of species concentration 1s in a region far from China. 4) Cytologically, its chromosome complement is of high ploidy. The diploid progenitors are in distant phytogeographical areas. 5) Historically, plants and animals have moved with people and are changed by them; the routes of such movement are often traceable. Application of these criteria proves positively that Abutilon theophrasti and Cannabis sativa are exotic elements in the flora of China and establishes that teal ae sambac is not Arabic and Jug/ans regia is not Persian. The latter are indigenous to China, selected and developed by the ancient ethnic groups dwelling in areas where the wild species grew. ABUTILON In northern China 4buti/on constitutes a minor fiber crop or—particularly in neglected disturbed areas of large cities such as Beijing, Nanjing, and Wuhan— is a weed. It was repeatedly introduced from China into the United States as a potential source of fiber for cordage. It has been called Indian mallow, stamp- weed, buttonweed, and Chinese velvetleafand has become a troublesome weed in fields of corn, cotton, and soybeans. Spencer (1984) has estimated that the annual economic loss due to velvetleaf is approximately $343 million. In response to the request of American weed scientists, in 1984 I searched for the natural enemies of Abuti/on in China, hoping to find some special animal or plant pathogens that I could bring back for the biological control of velvetleaf in the United States. From Hong Kong, I entered China and traveled all the way to Guangzhou, Lanzhou, Qinghai, Inner Mongolia, Datong in Shanxi, Beijing, Wuhan, Nanjing, and many areas in Jiangsu (including where Abuti/on was formerly cultivated on my grandmother’s farm) but failed to find any specific natural enemy. I studied the ecological conditions of the plants, took photographs, and collected voucher specimens. Subsequently, I concluded that Abutilon is not native to China. Occuring only in association with man, it is an isolated tetraploid with the related diploid progenitors growing in south- western Asia. Apparently, what is known as the Chinese velvetleaf (4. theo- phrasti) was unintentionally introduced into China, having attached itself to man or his animals; it later became naturalized and was gradually domesticated and improved by Chinese farmers to furnish the bast fibers used in cordage 1990] HU, TRADITIONAL CHINESE MEDICINE 15 and for making sandals. The medicinal use of the seed was first recorded in a seventh-century herbal. It is worthy of note that in the vicinity of Hong Kong and Guangzhou, the tetraploid perennial 4. indica (L.) Sw. grows as a weed. Apparently the species of Abuti/on in China are all adventives. CANNABIS OR HEMP The worldwide distribution of Cannabis sativa is mostly man related. Hu- mans have captured and capitalized on the genetic variability and ecological plasticity of the species and have used it to meet their special needs. In India and now in Latin America, it has become a cash crop. Young fruits and leaves are harvested and used as a narcotic drug. Large quantities of the drug are sold in the United States as marijuana, the use of which has caused serious social problems. In China it is extensively cultivated as a minor crop for the useful bast fiber, which is stronger but less plentifully yielded than that of Abutilon. The medicinal use of its by-products (seeds, roots, bark, leaves, and young fruits) has not been important to Chinese farmers, at least not before the 1930's when I lived among them during the harvest season. The origin and the systematics of the species of Cannabis have puzzled botanists throughout the history of botanical science. In China hemp appears in the earliest plant records as ma, which has become a generic term for different types of bast fibers and the names of their source species—for example, huang- ma for jute (Corchorus capsularis L.), qging-ma for velvetleaf (Abutilon theo- phrasti), and zhou-ma for ramie or China grass (Boehmeria nivea (L.) Gaud.). Archaeological evidence shows that Cannabis was associated with Neolithic dwellers in Gansu, the cradle of Chinese civilization (Wang & Guo, 1980). In this case, antiquity does not prove that Cannabis 1s indigenous to China, for the seeds were kept in an earthenware jar. Cannabis is an exotic element in the flora of China because its existence in the country is completely dependent on man. It is noteworthy that the young fruit, the most potent organ of the plant used in marijuana, was recorded in ancient Chinese classics, Er-ya, as ma-fen (ma grave) or ma-lan (blue ma or ma cradle). As a medicine, 0.6 g of ma-fen is given to patients with the warning that the material is poisonous, and when taken in excess it a a person, making pun violent and nay, The active principle isa I 1(THCQ), which gradually disappears as the fruit matures. The Chinese people use ripe seeds as feed for pigeons fattened to be served in special restaurants. JASMINE The origin of the source species of mo-li-hua, used in the Chinese tea industry to produce the delicate flavor of jasmine tea, has been both ¢ ly recorded in botanical literature and misinterpreted in Chinese references. In classic bo- tanical works, Linnaeus (1753) and Aiton (1789) cited two records based on plants introduced from Arabia and cultivated in European gardens under the common name “sambac.” In 1753 Linnaeus proposed Nyctanthes sambac L., 516 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 which Aiton transferred to Jasminum L. in 1789 to validate the currently used scientific name, J. sambac (L.) Aiton. Aiton also added the English name Arabian jasmine. In ancient Chinese writings, mo-/i-hua first appeared in Nan- fang cao mu zhong (A Description of Plants from the South, A.p. 306), by Ji Han. Ji was from a famous educated family of the time. Through family connections he was appointed Awestricken General and Prefect of Xiang-yang. Xiang-yang (long. 122°8’ E, lat. 32° N) was an important communications post where diplomatic missions and merchants from southern China stopped to change means of transportation from boats to horses to continue along their journeys to the capital. At that time valuable tropical natural products such as ivory, rhinoceros horn, spices, and perfumes were sent as tributes to the emperor, kings, dukes, and local officers, or as material for exchange gifts. In Ji Han’s office, a record of these products was made under the title Nan-fang cao mu zhong. When Li was assassinated in 306, the list contained about 80 items, each with a name and with the information supplied by the bearer. The second item on the list covers two species of Jasminum. The information given provided several facts. At the beginning of the fourth century, two kinds of white-flowered jasmine were transplanted from the tropics to the warm-tem- perate city of Xiang-yang in northern Hubei Province. Both types produced very fragrant flowers, for which the species had been domesticated in southern China (where Guangzhou was the major city). The species were native to the land west of Guangzhou—an area occupied by the Hu people, who had not adopted the Chinese culture. One species was called mo-li-hua; the other, ye- XI-ming, apparently a translation of the Persian-Arabic word “‘yasmin” (‘‘jas- mine” in English). The latter species has been identified as J. officinale L.., which has compound leaves and is rarely cultivated in China. Based on the above Latin and English names and the ancient Chinese records, early Chinese botanists began to draw the conclusion that mo-/i-hua was native to the area between India and Arabia. Chen (1937, p. 1031) further explained, “It seems to have been introduced from southern Asia to China, first planted in South China, and now cultivated throughout the country for ornamental purposes. In Fuzhou, it is planted extensively, for there it is used to scent tea. The annual production reaches 2,000,000 catties.”” Actually, the areas in the vicinities of Guangzhou and Kunming both have large acreages planted with the crop. Like other botanists of my generation, I used to believe that the mo-li-hua in Chinese gardens was introduced from Arabia. Then I saw the wild type of Jasminum sambac in Yunnan and studied the species in the natural flora of China. On 30 October, 1980, I was stranded in Si-mao after a lecture tour to the Tropical Botanical Garden, Academia Sinica, in Xi-shuong-ban-na, situ- ated in the area of the Chinese Tai minority on the Yunnan-Thailand border. Since the small plane could not take off, I spent the extra time studying the plants in the gardens and in a nursery. Among the numerous pots of mo-li-hua, I saw some plants sending out one or two elongate vinelike shoots, which the nurseryman had trimmed off to produce a shrublike appearance. Suddenly I realized that I was looking at the prototype of the mo-li-hua that I had seen 1990] HU, TRADITIONAL CHINESE MEDICINE 517 throughout China as shrubs in pots and in gardens. Actually, I was standing on the land where early cultivators began to domesticate J. sambac, trans- forming it from its natural clambering habit to a shrubby one. To check the validity of this insight, I examined herbarium specimens of species of the genus Jasminum on a worldwide basis and studied the morpho- logical characters and distributional patterns of the Chinese species in partic- ular. Species of Jasminum have a pantropic distribution. The Chinese species can be divided into two major groups: the temperate deciduous ones, such as yin-chun (J. nudiflorum Lindley) and tan-chun (J. floridum Bunge), with shrub-like habits and yellow flowers that open early in the spring before the leaves, and the tropical evergreen ones, such as su-xin (J. sinense Hemsley), gia-su-xin (J. amplexicaule Buch.-Ham.), and su-xin-hua (J. officinale L. var. grandiflorum (L.) Kobuski), with vinelike habits and white fragrant flowers. Some of the species have simple leaves, while others have compound ones. Mo-li-hua 1s an evergreen plant with simple leaves and fragrant white flowers that often turn purplish pink with age. On plotting the distributions of the wild simple-leaved Chinese species of Jasminum, I was indeed surprised to find that six of them all closely related to /. sambac grow in an area within a 100 km radius of Si-mao (101°02' E, 22°24’ N). These little-known species (/. anastomasans Wall., J. coarctulum Roxb., J. coffeinum Hand.-Mazz., J. dunic- olum W. W. Sm., J. nintooides Rehder, and J. sequinii Léveillé) are not yet known in cultivation. This area where the morphological diversity and species concentration of Jasminum occurs, called Hu in the time of Ji Han, has been termed a “favored area for evolution” by one modern biogeographer and Cathaysia by Takhtajan (Hu, 1971, p. 222). It is the homeland of mo-li-hua. Nine species of Jasminum, including J. sambac, are used in traditional Chinese medicine in the form of roots, stems, and leafy shoots. Apparently J. sambac was domesticated in Cathaysia and gradually spread by man to Arabia, from where it was introduced to Europe. All former records that ascribed its origin to Arabia were made on assumptions, and they are botanically unrealistic: J. sambac and closely related species grow in the humid tropics; the arid Arabian climate does not support such plants except in gardens. WALNUT The origin of Juglans regia L. has puzzled early researchers interested in Chinese economic botany and the cultural exchange between East and West. Unfortunately, at the time when Laufer (1919) conducted his research into the ancient literature dealing with walnuts, the biology of Juglans and the ethno- botany of the ancient people who first cultivated, selected, and improved the wild progenitor into a form like J. regia (which has thin-shelled nuts with edible kernels) were yet unknown. Consequently, his conclusion that the English wal- nut has a Persian origin contradicts modern scientific findings. Moreover, under the influence of ancient Chinese and more recent foreign scholars, early bot- anists in China also considered J. regia to have been introduced. For example, a widely circulated reference (Chen, 1937, p. 136) stated, “According to tra- nee: JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 dition, the species is a native of Arabia, was introduced into China in the Han Dynasty, and is extensively cultivated in central and northern China today.” The concept that the edible walnut is not Chinese should be rectified. Current botanical data indicate that the walnut family (Juglandaceae) is old, having originated in the Northern Hemisphere. The present-day center of di- versity is eastern Asia (Manchester, 1988), particularly China, where living species of Platycarya Sieb. & Zucc., Engelhardtia Leschen. ex BI., Plerocarya Kunth, Cyclocarya Iljinskaja, Carva Nutt., and Juglans L. coexist. In Europe there are no native species of these genera, and in America only species of Carya and Juglans are indigenous. (Actually, these two genera have wider geographic distribution and more ecological amplitude in America than in China.) The natural species of Jug/ans are all monoecious, with small, green, unisexual flowers; the staminate ones are in pendulous catkins on second year growth, the carpellate ones in upright spikes terminating current year’s growth. Additionally, all have lamellate pith in the young branchlets; alternate, pin- nately compound, estipulate leaves; drupelike fruits, each developed from an inferior ovary, and nuts with thick shells that are very hard to break. The species with large, thin-shelled, relatively easy-to-crack nuts and tasty kernels, like J. regia, are of anthropogenic origin, having been selected through culti- vation and/or hybridization. Botanical literature of the last two centuries has recorded many such species and horticultural forms (see Rehder, 1940, 1949). Juglans x intermedia Carr. was developed through hybridization between an American species (J. nigra L.) and an Asian one (J. regia); the thin-shelled horticultural forms Juglans ‘Paradox’ and Jug/ans ‘Royal’ were selected in the _S. A. from hybrids of Juglans hindsii (Jepson) ee aa J. regia, and Juglans ‘James River Hybrid’ from Juglans nigra x J. reg. Ancient Chinese literature and recent ethnobotanical findines together con- firm the fact that Juglans regia was cultivated from native wild species in China at the beginning of the Christian era. The earliest Chinese records concern the cultivation of walnuts in the capital (now called Nanjing) of the Kingdom of Wu (A.D. 222-277) and the quality of the kernel in the Eastern Jin Dynasty (ca. A.D. 326). In the royal garden called Hua-lin-yuan 84 walnut trees were in cultivation. Eastern Jin (A.D. 317-419) conquered Western Jin and used Nan- jing as the capital. Fifty years later, Su Jun tried to overthrow the Jin govern- ment. The women of some rich families escaped the rebellion by going from Nanjing to Lin An Shan for safety. It was recorded that a messenger was sent to bring supplies from the capital to Lady Liu witha letter explaining the special fruits included. Regarding the walnut, the letter stated that it was originally grown in the farther land of Xi-qiang, that the outer shell was hard but the inner kernel sweet, and that it was sent as a special tribute because it was easy to carry (Laufer, 1919). The ethnobotanical significance of the above references has never been grasped by subsequent scholars (Chinese or foreign) who wrote about the edible walnut in China because they were uninformed concerning the land where the cultivated and wild species of walnut grow, or about the life, language, and culture of the people there, until China became involved in World War II. 1990] HU, TRADITIONAL CHINESE MEDICINE sake, The Sino-Japanese War began in 1937 and ended in 1945. During this period I taught in the Department of Biology, West China Union University, Chengdu. During summer vacations I hired some native hunters or medicine gatherers to help me with field work and with the study of vegetation and ethnobotany. We collected specimens of different species of Juglans, which they called he- tao (the cultivated trees in the villages) or shan-he-tao (the wild trees on the hillsides). In the summer of 1941, C. C. Liu (China’s leading herpetologist), his associates, my assistant, and I followed a tributary of the Min River from Li-fan up the Mong-dong Valley. Opposite this valley in the mountains south of Li-fan lived the Qiang ethnic group. These people cultivated barley and corn, depended on wild onions and fern fiddleheads for vegetables, and kept dried, wild persimmons and walnuts in their houses. I was given some walnuts and shown how to crack them by placing two nuts in one hand and giving a quick squeeze. I was very surprised to know that walnuts could have such thin shells. Those that I had seen elsewhere in China had hard, thick shells and were rather rare; they were broken by a stone or a hammer. The ethnic people living in the mountain valleys of the Qiong-lai Range (Chung-lai on Map in Hu, 1956) had early selected and developed walnuts, and they continued to improve the quality to the state I had seen. I ascended the Qiong-lai Range from different directions and reached several peaks covered by snow in August. At the level where broad-leaved deciduous forests fluorished, there were wild Chinese walnuts (Juglans cathayensis Sarg.) growing on the hillsides and cultivated trees of J. regia in the villages. In the Mong-dong Valley, the people were very proud of their special thin-shelled nuts. The Qiong-lai Range and the adjacent land mass to the north and west were occupied by the ancestors of the Xi-qiang and by the Rong or Gia-rong. In the history of China, the name of this area has been changed many times. During the Three Kingdoms Epoch and the Eastern Jin Dynasty (A.b. 222-419) when Juglans regia was first mentioned in Chinese literature, the region was called “The Further-Land of the Western Qiang” by scholars. When the famous European explorers A. David and E. H. Wilson collected animal and botanical specimens there, they used “Eastern Tibet” on their labels and in their pub- lications. When I collected in the area between 1938 and 1942, the western portion was administered by the Sikang provincial government, and the eastern portion was Sichuan (Szechuan), as printed on my field labels. On today’s map of China, the entire region is in Sichuan. Early historians and sinologists mis- interpreted the ancient records about Xi-qiang because they used a map of the twentieth century with the western Chinese boundary on the Pamir Mountains and took the phrase “farther land of Xi-qiang” to mean Persia. They thus shifted the origin of the walnut from western China to Persia. A botanist from Iraq, I. A. Al-Shehbaz, told me that all the walnuts in Iraq and Iran are cultivated. There is no wild species of Jug/ans in the Middle East. The dispersal of Juglans regia was also through man’s activities. Before 1950 the agriculture of the Qiang and Yong ethnic minorities in western Sichuan was for self-sustenance, as it was for 95 percent of all Chinese farmers. One or 220 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 a few walnut trees may have been planted near the family house to provide nuts for family use, not for sale. They were for festivities, or to be used as a rich source of energy, especially for the men on their hunting trips or during religious pilgrimages. On such trips away from home, the rations were zhan- ba (roasted barley meal) and walnuts. In a wayside rest station the traveler just mixed the zhan-ha with hot tea and butter provided by the host, supplemented by a few nuts to make a meal. Often the walnut was given in exchange for favor or for gifts. In ancient times the Qiang people were politically and militarily involved with the people using the Chinese language and writings to their east, and the Rong minorities were religiously associated with the people on the Tibetan plateau to the west. It was evidently through such activities that walnuts were carried eastward to the people in the plains of the Yellow and Yangtze rivers, and westward into Tibet, thence to northwestern India and Persia. The common Sanskrit terms for walnut, akhota, aksota, and aksosa, seem to confirm the movement of the Chinese he-fao to India via Tibet. Generally, when a cultural trait passes between two peoples living in adjacent areas, it goes unrecorded, and the people receiving it take it as a matter of fact. This was exactly the case with walnuts in the lower Yangtze provinces. Ap- parently, ancient scholars like Ji Han of the fourth century were familiar with walnuts. In his A Description of Plants from the South, when he described nuts brought to Xiang-yang by special envoy, he compared them with the taste of walnuts. For example, in recording coconut from southern China, he stated that the white meat tastes like walnut, but it has a richer flavor. CONCLUSIONS AND DISCUSSION Having worked with plants for over sixty years, | am habitually inclined to treat a historical problem as I do a living organism: I investigate its develop- ment, structure, interaction within a community, and reaction to time, space, and human activities. I believe that scientific methods are applicable to the study of history as well as to work with biological objects. Guided by this, I started researching the introduction of exotic elements used in TCM. Some extra-Chinese elements were introduced into China a very long time ago. Hemp, for example, entered China with Neolithic man, according to recent archaeological evidence. There has been a continuous growth in the number of plant and animal products introduced. Meanwhile, there has been a gradual change in the nature of the importations: daily necessities such as food and fiber gave way to aromatics and spices used in the preparation of emergency remedies in the Song Dynasty, and then to substances for comfort and enjoy- ment of life in the Ming and Qing dynasties. The importation of 99 percent of the items can be traced in ancient records. However, when a species is used only in folk medicine, shared by people living in adjacent areas, and spread through intermediaries, the time of its introduction into China may never be known. The creat of India (4ndrographis paniculata) in South China is a good example. 1990] HU, TRADITIONAL CHINESE MEDICINE 521 The geographic analysis shows that tropical Asia, the Mediterranean region, tropical America, and southeastern Asia have been the primary contributors of the exotic elements used in TCM. Over 56 percent of the 94 items listed came from these four areas. Most of the species were introduced intentionally, although a few plants—including the velvetleaf (Abutilon theophrasti)—arrived in China as adventives following the movement of man. Since China is rich in natural resources, why are animal and plant products imported into the country for healthcare purposes? First, it should be under- stood that adventure is in the nature of mankind, and plants and animals move with man. A majority of the extra-Chinese plant species was first introduced into China as daily necessities such as food and fiber for the ancient travelers. Today these species are still cultivated for the same purposes. Ingenious and inventive people who tried to use the agricultural by-products effectively dis- covered their medicinal uses long after their introduction into China. Radish, hyacinth bean, black sesame, and cannabis seed are good examples. Second, the geographic position and the general climatic conditions of China limit the production of certain source species that require humid tropical or warm xerophytic conditions. Certain products from the tropics are needed for soothing the unbearable pain of abscesses, ulcers, and carbuncles, and for curing tenacious cutaneous diseases such as scabby head. Gambier from India and gamboge from Thailand are imported via Singapore for these purposes. Other products are needed in the preparation of rescue remedies used in such emer- gencies as heatstroke, epilepsy, convulsions in children, loss of consciousness, and delirium. Frankincense, borneol, cloves, and rhinoceros horns are importe primarily for such preparations. In the popular and nationally available people’s elixir, 20.5 percent of the weight of the 15 ingredients is of extra-Chinese origin. In the more expensive royal storax pill, 78 percent of the total weight of the 14 ingredients consists of imported material. Recently, Chinese pharmaceutical manufacturers have tried to use synthetic materials and/or local products to substitute for imported elements obtained from endangered species—for ex- ample, using horns of the buffalo to replace those of the rhinoceros in pre- scriptions. However, these efforts have not stopped the flow of exotic natural products into the country for use in TCM. In researching the origin of extra-Chinese elements used in TCM, I have encountered numerous incorrect and untrue statements both in the botanical literature and in ancient Chinese documents. Reports of Abutilon and Cannabis as indigenous to China, of Jasminum sambac as a native to Arabia, and of Juglans regia having a Persian origin are incorrect. After months of research, I have been able to rectify these erroneous concepts; for example, Abutilon and Cannabis have been added to the list of introduced species, and jasmine and walnut have been removed from my original list of exotics used in TCM. ACKNOWLEDGMENTS My special thanks are due to Thomas C. Hu and Luke L. Wang for their unfailing assistance in the preparation of the manuscript, and to Beverly Huck- a2 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 ins and David E. Boufford for reading it and offering many helpful suggestions. I am deeply indebted to the staff of the Harvard University Herbaria and the Chinese Library of the Harvard Yenching Institute, Harvard University. With- out the unique herbarium collections and library facilities, it would have been impossible for me to carry out this research. My interest in Chinese medicinal plants was rekindled by the enthusiasm of my colleagues on the Faculty of Science, the Chinese University of Hong Kong, and particularly by the en- couragement and support of H. M. Chang, Y. C. Kong, and P. P. H. But LITERATURE CITED Arron, W. 1789. Hortus Kewensis. Vol. |. London ANONYMOUS. 1889. List of Chinese ne enue Imperial Maritime Customs, Miscellaneous Series No. 17. Shanghai. 1959-1961. Zhong yao zhi (A new Chinese materia medica). (In Chinese.) 4 ‘iol. Health Press, Beijing. [1. 1974. Zhong yao zhi ji shou ce (A handbook of Chinese pharmaceuticals in traditional medicine). (In Chinese.) People’s Health Press, Beijing. 1976-1980. Zhong cao yao xue (Chinese pharmacology). (In Chinese. ) Vols. 2, 3. Jiangsu People’s Press, Beijing. [3.] . 1977-1979. Zhong yao da ci dian (An encyclopedia of Chinese medicines). (In Chinese.) 2 vols. Shanghai People’s Press. Shanghai. Supplement. Shanghai Science Technology Press, Shanghai. [4.] 1980. Xinhua zidian (New China dictionary). revised ed. Commercial Press, Beijing. (In aes with pin-yin.) [5.] Baker, J. G. 1877. Aloé chinensis. Bot. Mag. 103: ¢. 6301. CHEN, C. Q. a6 penal materia medica (Ben-cao shi-yi). (In Chinese, incorpo- rated in S. Z. L1.) [6.] Cuen, H. Z. 1688. Flower mirror (Mi-chuan hua jing). 6 Fascicles. Honzhou. (In Cuen, Y. 1937. Illustrated manual of Chinese trees and shrubs. (In Chinese.) Nena HawortnH, A. H. 1819. Supplementum plantarum succulentarum. xi + 158 pp. I. J. Harding, London Hu, S. Y. 1945. Medicinal plants of Chengtu herb shops. W. China Border Res. Soc. ISB: 95-178. 1956. Climbing the trails of the giant panda. Appalachia. 1956: 164-172. 1971. The Orchidaceae of China (II). The composition and distribution of orchids in China. Quart. Taiwan Mus. 24: 181-223. Ji, H. A.p. 306. Nan-fang cai mu zhong (A description of plants from the south). (In Chinese), [7.] Lan, M. 1370. Materia medica of southern Yunnan (Dian-nan ben-cao). 3 Fascicles. (In Chinese. Reprint, 1914, Kunming; Revision, 1975-1978.) [8.] LAuFER, B. 1919. Sino-Iranica, Chinese contribution to the history of civilization in ancient Iran with special reference to the history of cultivated plants and products. Field Mus. Nat. Hist., Anthropol. Ser. 15: i-iv, 185-630. Li, C. 650? Duke Ying’s Tang pharmacopoeia. (Ying-gong Tang ben-cao). 7 Fascicles. (In om incorporated in J. Su’s work, the original lost.) [9.] Li, CHANG- 1966. Zhong-guo zhi min shi (History of Chinese emigration). 352 pp. Cane Press, Taiwan. [10.] Li, S. Z. 1596. Materia medica with commentaries (Ben-cao gang- mu). 52 Fascicles. (In Chinese. 1930 reprint, oie Commercial Press; 1967 reprint in 2 vols., Hong Kong Commercial Press.) [1 1.] 1990] HU, TRADITIONAL CHINESE MEDICINE a2) Li, X. 756. A register of the medicine imported via the sea (Hai-yoa ben-cao). 6 Fascicles. (In Chinese. Incorporated in S. Z. Li, a lost.) [12.] LinnaeEus, C. 1753. Species Plantarum. 2 vols. Holm Lr, H. 1708. Enlarged monograph on flowers (Guane aaa fang pu). 32 Fascicles. Beying. (In Chinese.) [13.] Liu, H., Z. MA, et al. 973. Pha armacopoeia of the Kai-bao Reign (Kai-bao ben-cao). 2 Fascicles. (In Chinese. Incorporated in S. Z. Li’s work, the original lost; ee in many forms.) [14.] Liu, X. 889-890. Records of strange articles (Ling biao lu yi). (In Chinese.) [15.] MANCHESTER, S. R. 1988. The fossil history of the Juglandaceae. Monogr. Syst. Bot. Missouri Bot. Gard. 21: 1-137 MarTuHews, R. H. 1972. Mathews’ Chinese-English eee Revised American edi- tion. Harvard University Press, Cambridge, Massachuset Ni, Z.M. 1624. Organized collection of materia medica ae -cao hui yan). 20 Fascicles. (In Chinese.) [16.] Renper, A. 1940. Manual of cultivated trees and shrubs hardy in North America. Macmillan, New York. . 1949. Bibliography of cultivated trees and shrubs hardy in the cooler temperate of the Northern Hemisphere. Arnold Arboretum, Jamaica Plain, Massachu- re S 231-273. Collections of a man of peace (Zi-an ji). (In Chinese.) [17.] Spencer, N. R. 1984. Velvetleaf, Abutilon theophrasti (Malvaceae), history and eco- nomic impact in the United States. Econ. Bot. 38: 407-416. Su, J. 659. New Tang pharmacopoeia (Tang xin ben-cao). 53 Fascicles. (In Chinese. Original lost in China, 10 fascicles ah and taken to Japan by visiting Japanese scholars. Now available in Japan.) [18.] Su, S. 1062. An illustrated pharmacopoeia (Tu jin ben-cao). 21 Fascicles. (Original lost, text incorporated in S. Z. Li). [19. Tana, S. W. 1108. Verified or aoa of materia medica (Zheng-lei ben-cao). 31 Fascicles. 30 Fascicles. (In Chinese. A commonly used name for Jing-shi zheng-lei bi-ji ben-cao; also known as Pharmacopoeia of the Da-guan Reign, Da-guan ben- cao; several later editions with changed names for the reigns of different emperors, the 1116 edition called Zheng-he Jing-shi zheng-lei bei-yong ben-cao. Reprinted in Nanjing in 1579.) [20 Tao, H. J. 500? Unrecorded materials of eminent physicians (Ming yi bie-lu). (In Chinese. Hand-copied manuscript lost, text incorporated in next entry.) [21.] 500? A commentary on Shen-nong’s herbal classics (Shen-nong ben-cao jing ji- sh. 7 Fascicles. (In Chinese. Incorporated in S. W. Tang, also in S. Z. Li. Wana, Q. R., & D. Y. Guo. 1980. Studies on the hemp seeds and millet unearthed at Linjia femaitis in Dong-xiang county, Gansu province. (In Chinese.) Gansu Teach- ers University, Lanzhou WANG, X. G 21. Monograph of flowers (Chun fang pu). 30 Fascicles. (In Chinese.) [ an 2.] Watt, G. 1889-1893. A dictionary of the ic products of India. 6 vols. Calcutta. oo 1: 239-241. 1889.) Wu, C 1848. An illustrated treatise on plants (Zhi-wu ming shi tu kao). 38 Fascicles. ae aan Shanxi. (In Chinese. Every species illustrated. See also, Zhi-wu ming shi tu kao chang bian. 22 Fascicles. In Chinese.) [23 Wu, P. 220-260. Herbal classics of the divine plowman (Shen-nong ben-cao jing). 3 Fascicles. (In Chinese. 799 edition by Sun et Sun.) [24. Newer version of materia medica (Ben-cao cong xin). 6 Fascicles. ‘Beijing. (In Chinese) ) [25.] x1a0, B. D. 1932. Records of gathering medicinal herbs in South China. 2 Vols. (In Chinese. Revised ed., 1936, incorporated in ZHUANG, 1954-1963). [26.] 524 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Xu, G. J., et al. 1960. Yao Cai Xue. (Chinese Pharmacognosy). 1416 pp. People’s Health Press, Beying. [27.] Yan, Y.H. 1253. Life saving recipes (Ji sheng fang). i Fascicles. ae Chinese.) [28.] ZHANG, C. S. 1985. Han Dai Shi Chou Zhi Lu Kai Tuo Yu Fa Zhan. (The opening and development of the silk route in Han ae Shi- hua eas 15(1-2): 1l- l Fike, Y.X. 1061. Revised pharmacopoeia of Gia-you Reign (Gia-you bu-zhu shen- nong ben-cao). 20 Fascicles. (In Chinese. Original lost, incorporated in S. Z. Li.) [30. ZHAO, X. M. 1765. Supplement to Ben-cao gang mu (Ben-cao gang mu shi yi). 10 Fascicles. (In Chinese.) [31.] ZHu, X. 1407. Famine herbal (Jiu-huang ben-cao). 4 Fascicles. (In Chinese. Every species illustrated.) [32.] ZHu, Z. H. 1358. A supplement to herbal commentaries (Ben-cao yan-yi bu yi). (In Chinese. Incorporated in S. Z. Li, 1597.) [33.] ZHUANG, X. B. 1954-1963. Revision of Xiao’s records of gathering medicinal herbs in South China. 2 Vols. 1963. [34.] Lm 1 AH 19 GRA 20 SRF 37 Fy 4B BFF 09 ee 0 ie 73 FRA 22 RE o1 RRR APPENDIX I. Chinese ideograms of names of exotics used in TCM.* SOB 6 Pik 15 RE io TRH isd Roe We 2 $e uF 20 1 2é, 29 & Sade 38 BAGS A a 56 RSH sip 746 BB 83 EK 2 RT 3H) 2a TF 21 HB of 30 SDF 39 BE 48 HEARF 57 ABA 65 ig A 15 ad 54 6 $4 93 76, SES ‘RE 13 REF 22 Wu 31 ato oB 0 eS 49 fy 58 RHE, 67 BAF "Big keg 1B 23 RAL 2 AF “14a 50% RA so # “6 EE 17 Hes oT & *The numbers correspond to those given in column two in TaBLe |. 24 BOAR 25 BB 33 Ka 42 4B FOR 51 4B PHL, 60 F By 36 ACR oy 52 HAS 61 WATE 69 By aR 70558 18 ERS 37 BE ) 19 38 ty F «A a LT ABER 26 BH 35 XARA ss AMF 53 FAR £2 BHR 71 4B RAL 50 1. $ 8958 Be, ANIOIGAW ASANIHO TVNOLLIGVUL ‘NH [0661 SCS APPENDIX II. Chinese ideograms of names of ancient herbals and pharma- copoeias.* 1 PRB 2 Hy ael -a8 3 pRee ee Lee > EAH 6 RARE 1 PABREAR 6 BY BRAF 9 At RARE 10 FEE PR GRE 1 Fash koa a 12 $34 SREB 13 Ql eh RES 14 $B # PEAS 15 Bi {4 HERE 16 Kee ARES 17 AER TE 18 GH AS (BPE) 19 Rig ASL AY RRC sehr *The numbers correspond to those given in LirERATURE CiTep. The names of authors are underlined. 1 Fj DH LEASE 22 Re ELS > aga ce 24 : (5k) Rites 4 , (RK) ay 25 BRERA rs ALAMOS 27 BAS RAGE 29 «Ok SEN H ZBEWA Hine 28 A 4e HEF 30 GS Fab ae TARA § BE 31 GR ARR At 2 RRM ERS 33 REAM Re AWE 34 BUS thy HR 5% 9C¢ WOLAYOUUV GCIONUV AHL JO TVNANOL [L “1OA] 1990] VARADARAJAN, GEOGRAPHIC PATTERNS IN PUYA 527 PATTERNS OF GEOGRAPHIC DISTRIBUTION AND THEIR IMPLICATIONS ON THE PHYLOGENY OF PUYA (BROMELIACEAE) G. S. VARADARAJAN! Puya (+ 185 species) is widespread at altitudes from sea level to 5000 m and has an extensive range of distribution from Costa Rica to Chile. Previous studies of this genus have been almost entirely taxonomic in nature due to the paucity of field collections and field data. Using extensive field and her- barium data, the present work investigates Puya phylogeny based on geographic distributions and cladistic relationships within the genus. Ten geographical ndemic, or relatively narrow distributions in mountain ranges or Norse tm and long-distance dispersal. Several lineages appear to have radiated both vertically (especially in the central Andes) and horizontally (in the northern Andes) during the Pliocene and Pleistocene when climatic and vegetational that while arid cycles generally favored speciation, some episodes in conjunc- tion with uplift of tectonic units caused extinction of lineages. Puya Molina (+ 185 species) is the second largest genus of subfamily Pit- cairnioideae of the Bromeliaceae (Varadarajan, 1988). It is widely distributed from Costa Rica to Chile, especially in the Andean Cordilleras. A very small number of species extend eastward into the Guayana Highlands (Guayana) and Amazonian Brazil. Only three other bromeliad genera— Aechmea Ruiz & Pa- von, Pitcairnia L’Hérit., and Tillandsia L.—are more widely distributed than Puya (L. B. Smith & Downs, 1974, 1977, 1979). Growth forms of Puya range from small tuberous herbs to massive palmlike plants (FIGURE 1). Habitats of different species are diverse, occurring from sea level to nearly 5000 m (Vara- darajan, 1986, 1988, 1989a; Varadarajan & Gilmartin, 1987). It has been shown that analysis of geographic distributions often provides interesting insights that aid in discerning the phylogeny of a group (e.g., Hum- phries, 1981; Platnick, 1981). Despite the extensive distribution of Puya, re- search has not progressed much beyond the level of basic taxonomy (L. B Smith & Downs, 1974). This situation is due to the paucity of field collections rvard University Herbaria, 22 Divinity Avenue, Cambridge, Massachusetts 02138, U. S. A. Current Address: School of Agriculture, Tuskegee University, Tuskegee, Alabama 36088, U. S. A. © President and Fellows of Harvard College, 1990. Journal of the Arnold Arboretum 71: 527-552. October, 1990. 528 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 ) and Ecuador. c, P. chilensis, a species with sympatric populations widespread in several arid localities of Chile. d, P. brittoniana, a species attaining 50-60 cm in height; previously 1990] VARADARAJAN, GEOGRAPHIC PATTERNS IN PUYA 52) and field data. The complexity within Puya is evident from a few recent studies (e.g., Varadarajan, 1987a, 1987b, 1988, 1989a; Varadarajan & Brown, 1988; Varadarajan & Gilmartin, 1988a, 1988b). These studies also emphasize the importance of data from biogeography, pollination studies, ecology, and eco- physiology for elucidating evolutionary phenomena. The principal objectives of this work are to investigate the geographic distributions of various species and examine their inferences on the phylogeny of Puya. METHODS AND OBSERVATIONS FIELD AND HERBARIUM RESEARCH Field work for this study has been conducted in Argentina, Bolivia, Chile, Colombia, Ecuador, and Venezuela. The data collected (Varadarajan, 1987a, 1987b, 1988, 1989a) include three major components: descriptive details of the 57 species of Puya collected (e.g., growth habit, caudex and foliage type, indumentum morphology, inflorescence morphology, petal color, mode of cap- sule dehiscence, and estimates of seed number; local distributions of species within their habitats and geographic range (Varadarajan, unpubl. MS); and descriptions of the new species encountered (Varadarajan, 1989b; Varadarajan & Flores, 1990). Field collections consist of plant parts preserved in liquid fixatives and voucher specimens deposited in a number of herbaria (TABLE 1). Herbarium collections of 140 species of Puya from 25 major herbaria in South America and the United States have also been studied. In many instances my field collections have supplemented the original author’s observations and subsequent herbarium data (Varadarajan, 1988, 1989a). Herbarium and field studies were useful for determining precise habitat types, clarifying the tax- onomy of some problematical species, estimating overall geographic and/or altitudinal range of species, and understanding the interrelationships between habitats and phenology and their phylogenetic implications. Habitats Most species of Puya are specific to particular habitat types that range from semixeric to extremely xeric. At high elevations dry mountain slopes, rock outcrops, open grassy and boggy meadows, cloud forests, and paramos include semixeric to moderately xeric habitats. At low elevations savannas and thorn woodlands contain some semixeric habitats. Habitats associated with monte, sierras, punas, and coastal desert vegetation are extremely xeric. A compre- hensive account of the habitats of Puya is in preparation (Varadarajan, MS). Phenology Phenology was studied in selected species of Puya and included recognition of distinct stages of morphological change in plants from seed germination known only from the type collection, this taxon is sparsely distributed in the punas of Bolivia. A total of only three populations of this species has been encountered by the author during his explorations. (All photographs by the author.) TABLE |. Field collections of Puya made by Varadarajan and associates. ' ISTRI- BUTIONAL COLLECTION COUNTRY OF COLLECTION UB- TAXA NUMBERS (KNOWN DIVERSITY CENTER CLASS P. aequatorialis André var. aequatorialis André 1423, 1426, 1430 Ecuador (IV, V) lb P. aequatorialis André var. albiflora André 1418 Ecuador (IV, V) la P. aristeguietae L. B. Smith 1188 Venezuela (II) Ic P. nat ns L. B. Smith? 1257 Argentina (X) la P. atra L. B. Smith? 1446, 1453 Bolivia (IX) la P. berteroniana Mez 1480, 1482, 1486, Chile (XI) 2c P. brittoniana Baker? 1466 Bolivia (VIII) la P. cardenasii L. B. Smith? 1442, 1463 Bolivia (IX) 2a P. castellanosii L. B. Smith? 1476 Argentina (X) la P. chilensis Molina 1484, 1488 ile (XI) 2c P. clava-herculis Mez . - ae 1435, 1436 Ecuador (IV, V) 2b P. coerulea Lindley va rulea 149] Chile (XI) 2b P. coerulea Lindley var. ey (Smith & Looser) 1493 Chile (XI) 2b S Looser EB. eS : - 1438 Ecuador (V la Pic ) mith? 1469 Bolivi la P. yee eae ae var. dyckioides 1256, 1475 Bolivia, Argentina (IX, X) lb . dyckioides (Baker) Mez var. novare Varadarajan* Argentina la P. ferruginea (Ruiz & Pavon) L. B. Smith 1272, 1278, 1295 Bolivia (VI-LX) ld 311, 1448, 1467 P. floccosa (Linden) E. Morren ex Mez 1170, 1172, 1174 Venezuela (I-III) Id , 1185 P. fosteriana L. B. Sm Bolivia (VIII) la P. gilmartinii Varadarajan & Flores’ 481 Chile (XI 2a P. glabrescens L. B. Sm oe 1300, 1303, Bolivia (IX) Ic De 1307, 1309, 449 OCs WOLAYOUUV ATONUV FHL AO TWNUANOF IL “TOA] TABLE 1. Continued ISTRI- BUTIONAL COLLECTION COUNTRY OF COLLECTION UB- TAXA UMBERS (KNOWN DIVERSITY CENTER) CLASS P. oe Na e Sodiro? 1417 Ecuador (IV, V) Ic 1422 Ecuador (IV-VI) 2b P. lee (Castellano Castellanos? 1245 entina (X) lb herzogli Wit 1301, 1302, 1304 Bolivia (1X) 2a P. humilis Mez 1298, 1444, 1450, Bolivia (1X) 2a 1451 P. lanata (HBK.) oe 1424 Ecuador (V) lb P. leptostachya L. B. Smit 5 Bolivia (VIII, [X) lb P. lilloi Castellanos 1229, 1235, 1264, Argentina (X) lb 1236, 1474 P. deity L. B. Smith? 1434 Ecuador (V) la P. meziana Wittmack? Bolivia (VIII) 1b P. bs (Mez) L. B. Smith 1223, 1226, 1230, Argentina (IX, X) ld 1233, 1237, 1263 P. mollis Baker ex Mez? 1470 Sees (VI-VII) ld P. nana Wittmack? 1461 = pes la P. nutans L Smith 1429, or (V 2a P. pearcei (Baker) Mez? 1286, 1291, 1447, aera (VIII, IX) lb 472 P. pygmaea L. B. Smith? on 1437 Ecaudor (V 2a P. raimondii Harms 1465 Bolivia (VIJ-IX) lb P. retrorsa Gilmartin? 1419, 1420, 1440, Ecuador (between IV & V) lb 1441 P. riparia L. B. Smith 1274, 1275, 1276 Bolivia (VIII) la P. sanctae-crucis (Baker) L. B. Smith var. sanctae-crucis 1454, 1457, 1458 Bolivia (VIII, Tx) lb VANd NI SNYALLVd OIHdVADOAD ‘NV[VUVAGVUVA [0661 Tes TABLE |. Continued DISTRI BUTION COLLECTION COUNTRY OF COLLECTION SUB TAXA NUMBERS (KNOWN DIVERSITY CENTER) CLASS P. sanctae-crucis (Baker) L. B. Smith var. verdensis Varadarajan? 1459 Bolivia (IX) la P. smithti Castellanos 1479 Argentina (X) Ib P. sodiroana Mez? 1427, 1431 Ecuador (V) 2a P. solomonii Varadarajan° 1471 Bolivia (VIII) Ib P. spathacea recat s — 1240, 1243, 1268 Argentina oo of X, XI) lb P. stenothyrsa (Baker) M 1310, 1312, 1468 Bolivia (VIII) lb P. trianae Baker? 1189, Venezuela, Ecuador (II, III, V) ld P. tristis L. B. Smith? 1306, 1308 Bolivia (IX) 2a P. tuberosa Mez? 1455 Bolivia (IX) la P. tunarensis Mez 1296, 1452 Bolivia (IX) 2b P. ultima L. B. 1277 Bolivia (VIII) la P. ushae Varadarajan 1460 Bolivia (IX) la P. venusta Philip 1483, 1487 Chile (XI) 2b P. yakespala Castellanos 1478 Argentina (X) la P. zakiana Varadarajan? 1425 Ecuador (V) la oucher sana housed in one or more of the following herbaria: GH, LAP, LPB, MO, regs aoe et > Tax eX usly sie ‘con five specimens or fewer. 3 ee taxa ponents NY, PORT, SALT, SEL, US, VEN, ws; for details of diversity centers and ces WOLAYOUUV GIONYUV AHL AO TVNANOfL IL “TOA] 1990] VARADARAJAN, GEOGRAPHIC PATTERNS IN PUYA of) TABLE 2. Most commonly observed patterns of phenology in polycarpic species of Puya.! PATTERNS OF SEASONAL PHENOLOGICAL CHANGES PHENOLOGICAL CTERISTICS N RECORDED LOWER LATITUDES? SOUTHERN LATITUDES? Seed germination and Completed in 1-5 years Completed in 1-5 years Initiation of scape February September Differentiation of inflorescence Initiation of flowers Differentiation of flowers nthesis Pollination September February Fruit set October March Seed dev elopment Capsule maturation and dehisc Seed dispersal January August ' Nonseasonal phenological cosa are irregular, not synchronized with seasons; some years are characterized by much longer periods of vegetative growth and omer by greatly exlended peneds of flowering and fruiting (examples: se pearcel, P. stenothyrsa) changes occur in sympatric groups where the changes are not synchronized with ne seasons eee Puya ca herculis, P. nutans, and P. pygmaea). Ex e: Puya 3 aaa Puya ae through seed dispersal (TABLE 2). There appears to be an interesting correlation ween phenology and geographic distribution, especially in the following contexts: (a) sympatric species groups; (b) overlapping species distributions within the same geographical Be and (c) occasional sympatric distribution of predominantly isolated species. Species of Puya are elt and with a few exceptions all are polycarpic. Phenology of monocarpic species is well illustrated by P. raimondii (FIGURE 1). This gigantic species continues to grow vegetatively for seven to ten years. The transition from a vegetative to the reproductive phase appears to be some- what abrupt, yet the flowering period lasts for up to three years. The nearly twelve-foot tall, indeterminate inflorescence tends to grow rather rapidly after it attains three to four feet in height. The final phase of inflorescence growth, however, appears to be gradual. Although seeds are abundantly produced (about 400 seeds estimated per capsule), the germination rate is extremely low (< 1%). The seed remains dormant for nearly two years Polycarpic species exhibit at least three distinct phenological patterns (TABLE 2). In some species (e.g., Puya spathacea and P. mirabilis) phenological cycles synchronize with yearly seasonal changes. While the phenological pattern is nonseasonal in some widespread species (e.g., P. ferruginea), it is complex in others. The following group of co-occurring species in Ecuador illustrates one complex pattern. Within the assemblage of P. clava-herculis, P. nutans, and P. fa’) 534 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 7] pygmaea, P. pygmaea begins its early reproductive phase (inflorescence dif- ferentiation) when P. clava-herculis enters the final reproductive phase (fruit and seed setting), but P. nutans remains in a vegetative phase. ANALYTICAL RESEARCH Results of the analysis of herbarium and field data are presented in the following section, and some are summarized on the Map and in Ficures 2-4. Phylogenetic interpretations of Puya are founded on the individual species distributions, species subsets, their putative cladistic relationships, and similar data known for organisms occurring in the same region. Geographic Distributions The principal objective in analyzing the distributions of Puya was to rec- regions where multiple species are concentrated. This was done by mapping the sites of occurrence of every species. Regions that contain five to forty species are circumscribed here as centers of diversity (see Map). The area of each diversity center ranges approximately from 12,000- 25,000 km?. Cladistic Analysis A series of procedural steps were involved in the cladistic analysis of various species of Puya. The first step was to delimit tentative monophyletic subsets or groups of species and analyze these subsets together in order to reconstruct the phylogeny of the genus. In addition to species subsets, several isolated taxa that did not associate with other individuals or species groups became evident. These isolated taxa displayed only autapomorphies. Monophyletic species sub- sets recognized here by synapomorphies are of two categories: (1) allied pairs of species that occupy putative terminal branches within a larger (somewhat unresolved) cladogram, and (2) alliances of three to several species that con- stitute a principal lineage. Three principal sources aided in my initial choice of species subsets for the analysis. First, the diagnostic key to species of Smith and Downs (1974) sug- gested a number of pairs or groups of taxa that share multiple characters. Several of these polythetic groups were tested (Sanders, 1981; Wiley, 1981) and were found to be monophyletic (e.g., Puya subgenus Puya). Second, my continued observations of various developmental stages of species in the field, supple- mented by herbarium studies, provided a better definition of some morpho- logical traits (e.g., inflorescence, indumentum). This approach allowed me to recognize additional monophyletic species subsets (e.g., the Puya tuberosa com- plex). Third, a re-examination of several previously neglected, unusual char- acters (e.g., lustrous bracts, caudex) suggested a few more monophyletic species assemblages Data bases with usually less than ten characters were prepared for the subsets of Puya. The ancestral and derived states of these characters were determined 1990] VARADARAJAN, GEOGRAPHIC PATTERNS IN PUYA 535 Equator Brazil 400 800 1 mi Ct i fo km 600 1200 Map. Geographic locations of centers of diversity of Puya. The approximate lati- tudinal and longitudinal limits of each center are given in parentheses. For other details see text. I, Sierra Nevada de Santa Marta (10°5S’-11°2' N, 73°2’-74°2' W); II, eastern cordillera of Colombia and Mérida Andes (5°8’—8°5' N, 71°6'-73°3’ W); III, Cundina- marca (4°%-5°5’ N, 73°3'-75° W); IV, southwestern Colombia and northern Ecuador (0° 36’ N, 76°-78°3' W); V, south-central Ecuador (2°-4° S, 78°5'-79°5’ W); VI, northern Peru (6°-8° S, 77°-79°2’ W); VII, central Peru (9°2’-12° S, 75°-77°2' W); VIII, Titicaca (13°6’-17° S, 67°5'-72° W); IX, southern Bolivia (16°5’-21°5’ S, 63°3'-66°7' W); X, northwestern Argentina (22°3’-28° S, 64°-67°8’' W); XI, Chile (26°-36°5'’ S, 69° 72°8' W). by the outgroup method (Stevens, 1980; Watrous & Wheeler, 1981). Clado- grams were manually constructed in accordance with the procedures discussed by Sanders (1981). Provisional outgroup for the species of Puya subgenus Puya could be any species of the putative sister taxon, Puya subgenus Puyopsis. Morphological and/or distributional criteria provided the basis for the choice of outgroups for the other species subsets recognized. 536 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 RESULTS Results of the distributional data and of the cladistic analysis address three important questions. Which geographic regions are characterized by relatively high concentration of species? What patterns do these distributions underlie? What is the correlation between phic di d suggested cladistic affinities of species? Each acon is discussed below. CENTERS OF DIVERSITY All the geographic regions that constitute the following eleven centers of diversity of Puya are in South America (see MAp). Center I (Sierra Nevada de Santa Marta) is the most northern and isolated unit of the Andes in Magdalena region in northern Colombia. Center II (eastern cordillera of Colombia and Mérida Andes) covers the northeastern section of Colombia between Norte de Santander and Boyaca, and neighboring Venezuela including Tachira, Mérida, and Trujillo regions. Center III (Cundinamarca) consists of the central part of the eastern cordillera of Colombia. Center IV (southwestern Colombia and Northern Ecuador) includes the territory approximately between the area where the three Andean Cordilleras are divided (Valle del Cauca, Narino in Colombia) and the volcanic sites of Imbabura in northern Ecuador. Center V (south- central Ecuador) extends between the Canar and Zamora regions. Center VI (northern Peru) encompasses part of the lowlands of the Amazonas and San Martin regions and the Cordillera Oriental of the Peruvian Andes in the Ca- jamarca and La Libertad regions. Center VII (central Peru) lies between the Ancash and Ayacucho regions, mainly along the Cordillera Oriental. Center VIII (Titicaca) includes the Cordillera Oriental and Altiplano. It stretches across Lake Titicaca between the Apurimac-Arequipa regions in Peru and the La Paz region in Bolivia. Center IX (southern Bolivia) consists of the eastern part of the Altiplano and the adjoining Cordillera Oriental covering the Cochabamba, Chuquisaca, Potosi, and Tarija regions. Center X (northwestern Argentina) covers the Salta and Jujuy regions. Center XI (Chile) extends between the Antofagasta in the north to Bio Bio in the sout Center I contains the lowest number of species (five), while centers VII, VIII, and IX harbor the highest species number (nearly forty). All but center XI encompass extensive altitudes (800-5000 m), while center XI includes a much smaller range (sea level to approximately 2000 m). Each center is separated from the closest adjacent one by a distance of at least 200 km. Nearly 75% of the species of Puva occur within one or more of the diversity centers circum- scribed here, and some species occupy two to three adjacent centers (e.g., P. clava-herculis, P. dyckioides, and P. hamata) as well as occurring in the inter- vening areas (e.g., P. ferruginea.) Rarely, a few species occur only in the in- terdiversity regions (e.g., P. retrorsa and P. spathacea). DISTRIBUTION PATTERNS Geographical distributions of species of Puya are allopatric and sympatric (TABLE 3) TABLE 3. Key characteristics of the distributional classes of Puya. 0.0 INDIVIDUALS DISTRIBUTION CLASS PER STAND OCCURRENCE WITHIN AND SUBCLAS STANDS (AVERAGE) DIVERSITY CENTER(S) EXAMPLES |. Allopatric a. Locally endemic Usually single <15 Confined to 1 center P. castellanosii b. Locally widesprea ultiple 20-30 1 or more centers . dilloi c. Predominantly allopatric, widespread, oe eee and oO. a Cc E > 2 6 ® @ wo 5 2 a a « a a o oO o a oe — Figure 2. Cladogram of the Puya tuberosa complex. The outgroup (P. mitis Mez) and the ingroup share an apomorphy (dwarf growth habit) against P. wurdackii (plants 1 m tall or taller). Roman numerals indicate diversity centers, arabic numerals signify apomorphies of characters; single asterisks indicate homoplasies and double asterisks, outgroups. Data matrix and characters for taxa included are presented in TABLE 4. berteroniana, P. boliviensis Baker, P. gilmartinii, and P. chilensis (FIGURE 1) occur 1n various combinations and are encountered in several localities (see below). Puya berteroniana and P. chilensis are the most frequent species in any sympatric assemblage. Similar replicate sympatric groups appear to be rare in Puya. CLADISTIC ANALYSIS AND DISTRIBUTIONS The results of cladistic analysis may be examined from a distributional perspective of the various members of a monophyletic group. Conversely, the distributional data may also be described from a cladistic perspective of sym- patric groups. The Puya tuberosa complex (FIGURE 2; TABLE 4) illustrates a monophyletic group in which the individual sister taxa are allopatric and geographically separated by great distances. This subset of Puya is found in several diversity centers, although the individual species are locally endemic to particular centers (subclass la). Sister species of this subset as well as other groups are geograph- ically disjunct by distances of a few to several hundred kilometers (TABLE 5). The monophyletic group Puya subgenus Puya (FIGURE 3; TABLE 6) reveals some interesting distribution pattern types. A part of this group in center XI consists of replicate sympatric species (subclass 2c; e.g., P. alpestris and P. 1990] VARADARAJAN, GEOGRAPHIC PATTERNS IN PUYA 541 TABLE 4. Data matrix of Puya tuberosa complex and its outgroup. CHARACTERS! SPECIES l 2 3 4 5 6 7 P. mitis A A A A A A A P. eryngioides A A B B A B A P. depauperata B A B A B B A P. paupera B A B A A B A B B B A B B A . reducta B A B A B B B P. tuberosa A A B A B B A ' List of characters: |, underground rhizome/bulb (A) eeu (B) absent; 2, scape (A) lax and conspicuous, (B) condensed and short; 3, inflorescence (A) simple, (B) compound; 4, flowers (A) pedicellate, (B) subsessile; 5, floral bracts (A) toothed, (B) entire; 6, flowers (A) nodding, (B) erect; and 7, scape bracts (A) membranous to coriaceous, (B) chartaceous. chilensis in the Bio Bio region; P. berteroniana and P. chilensis in the Coquimbo and Valparaiso regions; P. berteroniana, P. boliviensis, P. chilensis, and P. gilmartinii in the Atacama-Antofagasta regions). Puya berteroniana may also be allopatric along the dry slopes of the Andes (ca. 2000 m). Puya castellanosii and P. weddeliana are allopatric, local endemics (subclass 1a) while the allo- patric P. raimondii consists of disjunct multiple stands. Locally endemic sympatric groups and non-replicate assemblages do not include cladistically closely related taxa. Non-replicate sympatric groups differ from site to site. Cladistic analysis of the sympatric group including Puya bicolor, P. goudotiana, and P. lineata (subclass 2b) indicates affinities for each taxon with allopatric species characterized by narrow distributions and wide geographical disjunctions (FIGURE 4). TasBLe 5. Examples of geographically allied disjunct species pairs in Puya. SPECIES DIVERSITY CENTER(S) Species disjunct in contiguous centers aristeguietae and P. goudotiana Il & Ill A killipii and P. nitida P. micrantha and P. pearcei IX & X P. fosteriana and P. weberiana P. castellanosii and - chilensis xX & XI P. coerulea and P. smithit “oe with major disjunctions P. floccosa and -IIl P. spathacea outside of X & XI P. nivalis and I P. yakespala x P. nutans and P. venezuelana II 542 JOURNAL OF THE ARNOLD ARBORETUM [vov. 71 P. coerulea (X\)** P. weddeliana (IX) P. alpestris (Xl) P. berteroniana (Xl) raimondii (VIII, IX) P. chilensis (Xl) P. gilmartinii (Xl) P. castellanosii (Xl) P. boliviensis (Xl) P. oe Ficure 3. Cladogram of Puya subgenus Puya. The outgroup (P. coerulea) and the ingroup share an apomorphy (spirally twisted postanthesis petals) against eeu paniculata (Ruiz & Pavon) Ruiz & Pavon (straight postanthesis petals). Roman numerals indicate diversity centers, arabic numbers signify apomorphies of characters; ae as- terisks indicate homoplasies and double asterisks, outgroup. Data matrix and characters for taxa included are presented in TABLE 6 DISCUSSION Results of the analyses of geographic distributions and cladistic relationships among species of Puya provide insights into the phylogenetic history of di- vergence lineages and some likely causes and consequences. In addition, these results suggest that a number of species were allopatrically derived. Restricted occurrence and habitat specificity of many species indicate that the isolating mechanisms are geographical. The paraphyletic or polyphyletic nature of most sympatric groups reflects local radiations of several independent phyletic lin- eages within a geographic center. There is no direct evidence for hybridization of species within the sympatric complexes. Mechanisms that isolate these sym- patric species are likely to be genetic. PHYLOGENETIC HISTORY OF DIVERGENCE OF SPECIES OF PUYA The literature on the neotropical climatic and vegetational history in con- of various species lineages. This phylogenetic model emphasizes some likely 1990] VARADARAJAN, GEOGRAPHIC PATTERNS IN PUYA 543 TABLE 6. Data matrix of Puya subgenus Puya and its outgroup. CHARACTERS! SPECIES | 2 3 4 5 6 7 8 9 P. coerulea A A A A A A A A A P. raimondii B B B A B B A B B P. weddelian A A A A A A A A B P. castellanosii A A A B ? A B A B P. chilen A A A B B B A B B P. alpestris A A A A A A A A B P. boliviensis A A A B B A A A B P. gilmartinit A A A B B B A B B P. berteroniana A A A A A B A A B ' List of characters: 1, caudex (A) noncolumnar, often prostrate, (B) stout and columnar; 2, plants (A) polycarpic, (B) monocarpic; 3, leaf rosette (A) at ground level, (B) above ground level: 4, leaf blades (A) with strongly contrasting surfaces (at least in part), (B) concolorous; 5, petals (A) blue/ blue-green, (B) ey ca green/greenish white; 6, inflorescence branches (A) numbering 30 or less, (B) adie: 80 or more; 7, inflorescence indument (A) nonferruginous, (B) ferruginous; 8, floral bracts (A) tly eded by sepals, (B) equaling or exceeding sepals; and 9, sterile apex in jnflorescence (A) absent, ‘(B) pr present. routes of divergence from a previously proposed geographical region of primary differentiation. Initial Divergence Diversity centers of Puya from I to X are confined to the Andes, and center XI is in Chile (MAp). Characteristic species habitats within these centers indicate that divergence of many lineages occurred principally in paramos (northern Andes), punas (central Andes), and coastal deserts (Chile). We know from recent studies (Varadarajan, 1986; Varadarajan & Gilmartin, 1988a) that the genera of the Pitcairnioideae evolved in the Guayana, and some of them (e.g., Pit- cairnia and Puya) later expanded in the Andes. It is intriguing, however, that species of Puya are abundant only in the moderately to extremely xeric habitats of the Andes and not in the semixeric habitats of the Guayana. At least two explanations warrant consideration in this connection. The first explanation contends that only a few species evolved in the Guayana prior to a major proliferation in the Andes. This idea is concurrent with the preponderance of more diverse xeric habitats in the Andes than in the Guayana (Sarmiento, 1975). According to the second explanation, the ancestral lineages of Puya in at high elevations. This hypothesis also rests on the assumption that several semixeric or moderately xeric lineages became extinct in the Guayana as a result of a series of catastrophic environmental changes (van der Hammen, 1982), some of which were reflected in the frequent contraction and expansion of savannas (Huber, 1982). During comparable time periods, however, the Andean regions experienced more or less equitable environments that were on E P. glaucovirens ** (VI) santosii** (III) P. medica** (VII) nD ne P. lehmanniana (IV) P. angulonis (VII) P. goudotiana (Il & Ill) P. bicolor (Ill) P. lineata (I & Ill) P. aristeguietae (Il) prs WOLAYOUUV GIONUYV AHL AO TWNANOL IL “I0A] 1990] VARADARAJAN, GEOGRAPHIC PATTERNS IN PUYA 545 congenial to the survival and proliferation of the descendant lineages (see van der Hammen, 1982, for related discussions). Colonization and Diversification in the Northern Andes Two migratory routes may have been important for the radiation of Puya from Guayana into the Andes. The first was via the eastern highlands of Colombia into the northern Andes of Colombia and Ecuador (centers II to IV); the second was via the Peruvian Amazon region into the central Andes (centers VI to VIII). Guayana, an expansive floristic province of South America (Steyermark, 1982), is characterized by its underlying sedimentary mantle (Roraima) that constitutes most of the highlands. The mantle stretches out into eastern Co- lombia in a series of isolated “‘mesitas,” adjoining the Cordillera Macarena and the flanks of the Andes (Maguire, 1970). The physical contiguity between Guayana and the Andes may have facilitated a direct migration of some Guay- anan species of Puya into the Colombian Andes (eastern Cordillera, centers II and III). The distribution of a sizable number of species of Puya in the paramos and the adjoining subparamos and montane habitats (elevations 2000 m and above) would be important in the context of divergence. A plausible explanation may be that at lower elevations species initially invaded the montane forest habitats and subparamos, from which they migrated to the paramos at high elevations. These migrations were probably best suited to glacial periods. During glacial times paramo areas were geographically much expanded, and the intervening distances between them were much reduced. This physio-geographic situation, combined with a massive production of wind-dispersed seeds, facilitated ex- tensive colonization by several species. During interglacial periods, on the Ficure 4. Cladograms of the members of a sympatric species group (Puya bicolor, P. goudotiana, and P. lineata). a, Alliance of Puya bicolor. The outgroup (P. glaucovirens) and the pair of sister species share an apomorphy (stout pedicels) against P. adscendens L. B. Smith (slender pedicels). Polarity is established in the following characters: 1, dry infloresence axes: angled to cylindrical; 2, racemes: pols mchously flowered to densely flowered towa rd apex; 3, 4, flowers: erect to downwardly cund; 5, primary br: acts: shorter ee or equalling sterile bases of racemes to longer ae ere bie of racemes. b, Alliance of P. goudotiana. The outg sg (P. medica) and the pair of sister species share an ae cule 4 cm long or longer) against P. phelpsiae L. B. Smith (petals 2-5 cm long or smaller). Polarity is established in the following characters: 1, scape bracts and primary bracts: persistent to disintegrating to simple; 2, floral bracts: entire to pectinate-serrate; 3, floral bracts: not lustrous to numbers signify apomorphies of Characters: single asterisks indicate homoplasies and double ane outgroups. 546 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 contrary, paramos were fragmented into discrete “islands” in which isolation was followed by speciation (for related discussions see Cleef, 1979; Cuatrecasas, 1957, 1979; van der Hammen, 1974, 1979, 1982) The Sierra Nevada de Santa Marta deserves mention in this discussion. While it constitutes the northernmost and geographically the most ancient structural unit of the Andes (Simpson, 1975), the Sierra Nevada de Santa Marta may not necessarily have been a region of initial invasion by species of Puya. This suggestion is based on the contemporary idea of Guayanan differentiation of the genus and on the lack of direct geological and ecological continuity between Guayana and the Sierra Nevada de Santa Marta. These floristic provinces are separated by a region of predominantly mesic communities—the “Ilanos”’ and the adjoining mosaics of low vegetation —that historically separated xeric species of some plant families between the Andes and Guayana (see Steyermark, 1982). The latter situation is true for Puya as well. This suggests that the Sierra Nevada de Santa Marta may be a refuge area for species of Puya where taxa were dispersed from centers II and/or III. Subsequently, they became isolated as a result of unusually long dry periods and unique geological events. Preliminary assessments of phyletic relationships suggest relictual status for the species of Puya native to the Sierra Nevada de Santa Marta. Expansion in the Central Andes The Amazonian lowlands, including pockets of xeric habitats, have perhaps provided geographical continuity between Guayana and the eastern slopes of the Peruvian and Bolivian Andes. The lineages that initially colonized these low Andean habitats further extended into high elevation punas and adjoining arid regions. A substantial number of lineages seem to have proliferated in the alpine regions of Peru and Bolivia (centers VII and VIII). Present evidence suggests that at least some of these high elevation lineages have descended directly from the lower elevation species (TABLE 7). The process of radiation and differentiation in the central Andes probably occurred at different times during glacial cycles than did similar events in the northern Andes. Specifically, glacial periods in the central Andes saw the development of significant water barriers across the Altiplano as a result of the presence of a massive system of lakes and rivers; these water barriers fragmented habitats and led to population isolation and speciation (Simpson, 1975; Varadarajan, 1986; Varadarajan & Gilmartin, 1988a). In the northern Andes, by contrast, speciation events oc- curred during interglacial times. Radiation into Chile There are at least three central Andean ancestral sources for the species lineages now native to Chile. The present distributional and cladistic analyses of Puya subgenus Puya (FiGuRE 3) suggest that these ancestral taxa (similar to P. raimondii, P. weddeliana, and P. castellanosii) may have been from centers VIII, IX, or X (Map). A vicariance model (see below) 1s conceivable for the differentiation of Puya subgenus Puya in light of some available evidence (TABLE 4). TABLE 7. Examples of high altitude species of Puya and their lower altitude allies. HIGH ALTITUDE TAXA ALLIED LOWER ALTITUDE TAXA Altitudinal Altitudinal Vegetational Diversity Species (in m) (in m) type center P. harmsii (Castellanos) 3000-3600 800-2000 Transitional forests Xx Castellanos P. dyckioides (Baker) Mez 1300-4000 1800-2000 Low scrub forests Xx P. weberbaueri Mez 2800-4000 500-2300 Moist forests VIII VANd NI SNYALLVd OIHdVUDOAD ‘NVIVUVAVUVA [0661 LPS 548 JOURNAL OF THE ARNOLD ARBORETUM [VoOL. 71 CONVERGENT EVOLUTION IN PUYA The distribution of habitats (punas, coastal deserts, etc.) within diversity centers suggests evolution of most species of Puya under the regime of arid environments. This notion is well attested to by a number of structural (Rob- inson, 1969; Tomlinson, 1969; Varadarajan, 1986; Varadarajan & Gilmartin, 1987) and physiological (Medina, 1974; Medina et a/., 1977; Griffiths, 1984; Grifhths & J. A. C. Smith, 1983; J. A. C. Smith et a/., 1986) features. It is important to note that these features are by no means confined to a single phyletic lineage. Climatic history of the Andes (e.g., Simpson, 1975; van der Hammen, 1974) indicates the occurrence of several independent episodes of arid cycles during Pliocene and Pleistocene times when various regions expe- rienced frequent climatic and vegetational pulsations and the uplift of tectonic plates. For example, in the intermountain valleys of the northern Andes xeric environments developed as a result of mountain barriers, which produced rain shadow effects by intercepting the movement of wet air masses (Gilmartin, 1973; Sarmiento, 1975). In coastal areas, however, aridity resulted from per- sistent strong winds associated with increased ocean currents and coastal up- welling (Simpson, 1975; van der Hammen, 1974, 1982). Thus, it is possible to conclude that different species lineages of Puya evolved in response to various, often similar environmental pressures. Similar xeric adaptations dis- played by these lineages probably illustrate convergent evolution. RADIATION OF SPECIES IN RELATION TO ALTITUDE Expansion of species lineages of Puya in the Andes was probably vertical as well as horizontal. The idea of vertical evolution has been postulated for several paramo taxa (Ericaceae, Melastomataceae, Rubiaceae; Chardon, 1938). Mod- ern geographic distributions and phyletic alliances of various lineages of Puya suggest that only species confined to the punas (in Peru, Bolivia, and Argentina) could have evolved vertically (TABLE 7). Paramo species of Puya generally lack ae at lower elevations. These species radiated by a horizontal mechanism milar to some other paramo-dwelling groups (e.g., Espeletia, Jamesonia, Dip- eee Cuatrecasas, 1957, 1979; A. C. Smith & Koch, 1935; Tryon, 1962). DISJUNCTIONS Vicariance and dispersal models explain the origin of disjunctions (Nelson & Platnick, 1981; Platnick & Nelson, 1978; Roe, 1967, 1972; Rosen, 1978; Whalen, 1983). There are apparently various degrees of disjunctions among the species of Puyva (TABLE 5). Vicariance models postulate the appearance of barriers fragmenting the range of ancestral species. The congruence between the hypothesized cladistic relationships and the area of occurrence of species provides a useful indication of vicariance (see also Croizat et al., 1974; Wiley, 1980, 1981). Such a congruence usually translates into parallel phylogenetic histories involving two or more lineages. Vicariance is a plausible mechanism especially for some high altitude lineages and putatively allied species pairs of Puya (TABLE 7). Replicate patterns of occurrences of sympatric groups and allied species pairs also suggest vicariance (TABLE 5). 1990] VARADARAJAN, GEOGRAPHIC PATTERNS IN PUYA 549 Dispersal models, on the other hand, portray disjunction in light of dispersal across barriers, which is entirely dependent on unique dispersal abilities within the individual lineages rather than parallel patterns. For instance, the theory of long distance dispersal is compatible with the distribution of the Puya tu- berosa complex across several intermountain valleys (FIGURE 2). Present evi- dence indicates that the disjunction of various species in this assemblage is probably an isolated event, not matched by any other subgroups in terms of parallel geographic patterns or cladistic relationships. Major disjunctions also occur as unique patterns in a few independent lineages (TABLE 5). These are undoubtedly an outcome of long distance dispersal. PHENOLOGY AND GEOGRAPHIC DISTRIBUTION Distinct phenological trends are correlated with some of the subclasses of geographic distributions of Puya. For example, phenological cycles synchro- nizing with climatic seasons are noted especially in the allopatric taxa (sub- classes la and |b). This trend may suggest a relatively long-term local stability of the species. When the range of a species expands over time, it may include environments differing from the parent one(s). This type of distributional change might induce specific (local) modifications in phenological traits. Thus, in species distributed widely across various macroclimates (e.g., species of subclass 14d), the relationship between phenology and the annual seasons is locally altered, revealing no apparent correlations. Phenology in sympatric taxa appears to be complex and remains as yet an intriguing, unexplored subject. To understand this complexity, we need data concerning socio-ecologic relationships, and the evolutionary histories of individual taxa as well as those of the habitats of sympatric species. SUMMARY AND CONCLUSIONS A majority of the species of Puya are allopatric and narrow endemics. They are most likely allopatrically derived and are isolated by geographic barriers. Speciation in many lineages primarily in the northern and central Andes and Chile resulted from episodes of aridity. Species appear to have colonized high altitudes directly from the differentiation of low elevation lineages as well as by a horizontal mechanism. Allied species became disjunct due to vicariance and dispersal events. Species extinctions probably occurred in a few lineages as a result of the obliteration of habitats brought about by localized arid periods and the uplift of tectonic plates. From Guayana, Puya immigrated into the Andes and attained its present range. ACKNOWLEDGMENTS This paper is dedicated to the memory of my late Ph.D. advisor, Amy Jean Gilmartin. Her encouragement, support, discussion, and advice helped me in various aspects of bromeliad research. I also wish to acknowledge the support of the individuals who facilitated my field explorations in South America, particularly M. K. T. Arroyo, D. Diaz, T. Di Fulvio, A. Hunziker, R. Neumann, 550 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71] L. Novara, F. Oliva, F. Ortega, J. Solomon, B. Stergios, and J. Steyermark. N. Platnick, P. F. Stevens, and J. Utley offered constructive comments on the earlier drafts of the manuscript. M. K. T. Arroyo, P. S. Ashton, and L. B. Smith are thanked for their useful discussions, and E. B. Schmidt and S. A. Spongberg gave helpful editorial advice. Also, I wish to extend special thanks to my wife, Usha, for her assistance in the field and herbarium and for typing the manu- script. The curators of GH, MO, NY, SEL, US, and ws are also acknowledged for permission to examine material in their herbaria and for sending specimens on loan. My field research was supported by a National Science Foundation doctoral dissertation improvement grant (BSR-8306999) and a National Geo- graphic Society Research Grant (#3463-86). LITERATURE CITED CHARDON, C. E. 1938. Apuntaciones sobre el origen de la vida en los Andes. Bol. Soc. Venez. Ci. Nat. 5: 1-47. CieeF, A. 1979. The phytogeographic position of the Neotropical vascular paramo flora with special reference to the Colombian Cordillera Oriental. Pp. 175-184 in K. Larsen & L. B. HoLM-NIELSEN, eds. Tropical botany. Academic Press, London. Croizat, L., G. J. NELSON, & D. E. Rosen. 1974. Centers of origin and related concepts. Syst. Zool. 23: 265-287. Cuatrecasas, J. 1957. Asketch of the vegetation of the North-Andean province. Proc. VIII Pacific Sci. Congr. 9: 167-173. . Growth forms of the Espeletiinae and their correlation to vegetation types of the high tropical Andes. Pp. 397-410 in K. LARSEN & L. B. HOLM-NIELSEN, eds. Tropical botany. Academic Press, London. Gitmartin, A. J. 1973. Transandean distributions of Bromeliaceae in Ecuador. Ecology 54: 1389-1393. GriFFitHs, H. 1984. Ecological distribution of bromeliads in Trinidad and their O 3C values: implications for the use of O values to indicate carboxylation pathways in plants. Pp. 145-174 in E. Mepina, ed. Physiological ecology of CAM plants. In- ternational Center for Tropical Ecology (UNESCO- tae Caracas, Venezuela. & J. A. C. SmitH. 1983. Photosynthetic pathways in the Bromeliaceae of Trinidad: relations ge life-forms, habitat serene and the occurrence of CAM. Occologia 60: -184. Huser, O. 1982. ia ae of savanna vegetation in the Amazon territory of Ven- ezuela. Pp. 221-244 in G. T. Prance, ed. Biological diversification in the tropics. Columbia University Press, New York. Humpuries, C. J. 1981. Biogeographic methods and the southern beeches (Fagaceae: Nothofagus). Pp. 177-208 in V. A. Funk & D. A. Brooks, eds. Advances in cla- distics. New York Botanical pine Bronx, New Maauire, B. 1970. On the flora of the Guayana Highland. Biotropica 2: 85-100. Mepina, E. 1974. Dark CO, fixation, habitat preference and evolution within the Bromeliaceae. Evolution 28: 677-686. ,M. DeLGapo, & J. H. TRoUGHTON. 1977. Physiological ecology of CO, fixation in ‘Bromeliaceae. Flora 166: 137-152. Neson, G. J., & N. I. PLatnick. 1981. Systematics and biogeography: cladistics and vicariance. Columbia University Press, New York. Pratnick, N. [. 1981. Widespread taxa and fice se eee congruence. Pp. 223-236 in V. A. Funk & D. A. Brooks, eds. Advances in cladistics. New York Botanical Garden, Bronx, New York. 1990] VARADARAJAN, GEOGRAPHIC PATTERNS IN PUYA a5 — &G.]J. Netson. 1978. A method of analysis for historical biogeography. Syst. Zool. 27: 1-1 Rosinson, H. 1969. A monograph on foliar anatomy of the genera Connellia, Cotten- dorfia, and Navia (Bromeliaceae). Smithsonian Contr. Bot. 2: 1-41. Rog, K. E. 1967. A revision of Solanum section Brevantherum (Solanaceae) ; in North and Central America. Brittonia 19: 353-37 1972. A revision of Solanum section Brevantherum (Solanaceae). [bid. 24: 239- 278 Rosen, D. E. 1978. Vicariant patterns and historical explanation in biogeography. Syst. Zool. 27: 159-188. SANDERS, R. 1981. Cladistic analysis of Agastache (Lamiaceae). Pp. 95-114 in V. A. Funk & D. R. Brooks, eds. Advances in cladistics. New York Botanical Garden, onx, New Yor a G. 1975. The dry plant formations of South America and their floristic connections. J. Biogeography 2: 233-251. Smmpson, B. 1975. Pleistocene changes in the flora of the high tropical Andes. Palaeo- biology 1: 273-294. Smit, A. C., & M. F. Kocu. 1935. The genus Espeletia: a study in phylogenetic taxonomy. Brittonia 1: 479-54 SmitH, J. A. C., H. Grirrirus, & U. Luttce. 1986. Comparative ecophysiology of CAM and C3 bromeliads I. The ecology of the Bromeliaceae in Trinidad. Plant Cell and Env. 9: 359-376 SMITH, L. B., & R. J. Downs. 1974. eae aeers ere Flora Neotropica Monograph 14(1): 1-658. Hafner Press, N & . 1977. Tillandsioideae ae Flora Neotropica Monograph 14(2): 663-1492. Hafner Press, New Yor . Bromelioideae (Bromeliaceae). Flora Neotropica Monograph 14(3): 1493- 2142. New York Botanical Garden, Bronx. Stevens, P. F. 1980. Evolutionary polarity of character states. Ann. Rev. Ecol. Syst. 11: 333-358. STEYERMARK, J. 1982. Relationships of some Venezuelan forest refuges with lowland tropical forests. Pp. 189-219 in G. T. PRANCcE, ed. Biological diversification in the tropics. Columbia University Press, New a P. B. 1969. Commelinales—Zingiberales iii. Pp. 192-294 in C. R. MET- ALFE, ed. Anatomy of Monocotyledons. Oxford University Press, London. eae A. F. 1962. A monograph of the fern genus /Jamesonia. Contr. Gray Herb. 191: 109-203. VAN DER HAMMEN, T. 1974. The Pleistocene changes of vegetation and climate in tropical South America. J. Biogeography 3: 3-26. 1979. History of flora, vegetation, and climate in the Colombian Cordillera Oriental during the last five million years. Pp. 24-32 in K. LARSEN & B. Hotm-NIELSEN, eds. Tropical botany. Academic Press, on 982. Paleoecology of tropical South America. Pp. 60-66 in G. T. PRANCE, ed. Biological diversification in the tropics. Columbia University Press, New York. VARADARAJAN, G.S. 1986. Taxonomy and evolution of the subfamily Pitcairnioideae (Bromeliaceae). Ph. D. dissertation. Microfilms International, Ann Arbor, Michigan. 87a. Explorations for Pitcairnioideae in South America. Part 1. J. Bromeliad Soc. 37: 16-25. 1987b. Explorations for Pitcairnioideae in South America. Part 2. [bid.: 62- . 1988. Genus Puya Molina: relocation of several rare species and some prelim- inary remarks on geographic distributions and species divergence. Part 1. Ibid. 38: 243-247, 254-256 1989a. Genus Puya Molina: relocation of several rare species and some pre- 552 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 liminary remarks on geographic distributions and species divergence. Part 2. Ibid. 39: 3-7. —. 1989b. Novelties of Puya Molina I: a new species from Bolivia. /bid.: 121- 123 ROWN. 1988. Morphological variation of some floral features of the sibfamily Pitcaimioideae (Bromeliaceae) and their role in pollination biology. Bot. oar (Crawfordsville) 149: 82-91. LORES. 1990. Novelties of Puya Molina II: a new species from Chile. J. sie a Soc. 40: (in press). . GILMARTIN. 1987. Foliar scales of the subfamily Pitcairnioideae (Bro- meliaceae). Syst. Bot. 12: 562-571. & 1988a. Phylogenetic relationships of groups of genera within the subfamily Pitcairnioideae (Bromeliaceae). Syst. Bot. 13: 283-293. l Seed morphology of the subfamily Pitcairnioideae (Brome- msaia and its systematic implications. Am. J. Bot. 75: 808-818. Watrous, L., & Q. WHEELER. 1981. The outgroup method of phylogeny reconstruc- tion. ae Zool. 30: 1-16. WHALEN, M. D. 1983. Centers cael sympatry, and historical biogeography in the tropical plant genus So/anum. The Biol. 65: 78— Witey, E.O. | _ eee se and vicariance biogeography. Syst. Bot. 5: 19 lw hs ——. 1981. Phylogenetics: the theory and practice of phylogenetic systematics. John Wiley & Sons, New York 1990] CHANG, ACER PALMATUM COMPLEX 553 A RECONSIDERATION OF THE ACER PALMATUM COMPLEX IN CHINA, TAIWAN, AND KOREA CHIN-SUNG CHANG! Several varieties of Acer palmatum in China, Taiwan, and Korea are reevalu- ated based on morphology and flavonoid chemistry and a new taxonomic treatment is presented. The new name Acer duplicatoserratum var. chinense and new combinations for Taiwanese and Korean taxa are proposed. Distri- bution maps of A. duplicatoserratum and A. duplicatoserratum var. chinense are also presented. During a comprehensive systematic study of Acer L., section Palmata Pax, series Palmata, nomenclatural problems were encountered in the Acer pal- matum complex in China, Taiwan, and Korea. Data from morphology, ge- ography, and flavonoid chemistry indicate that a new name is needed for the Chinese. individuals of 4A. palmatum Thunberg. New combinations for the Taiwanese and Korean Acer pa/matum are also required to clear up the taxo- nomic and nomenclatural confusion in this complex. Thirteen species of series Palmata (sensu de Jong, 1976) in central and eastern China, Taiwan, Korea, Japan, and western North America are currently recognized. Taxonomic treat- ment of the Acer palmatum complex in eastern Asia is further complicated by the existence of over 250 cultivars of A. palmatum sensu lato, developed and selected over more than 300 years (Vertrees, 1978). In Japan, Ogata (1965) recognized two taxa in this complex, Acer palmatum and A. amoenum Carriére, the latter with three varieties: 4. amoenum var. amoenum, A. amoenum var. nambuanum (Koidz.) Ogata, and 4. amoenum var. matsumurae (Koidz.) Oga- ta. Recent Japanese floras (Kitamura & Murata, 1984; Ohwi, 1984) simply treat A. amoenum as a variety or a subspecies of A. pa/matum, although one of the former varieties of 4. amoenum, subsp. matsumurae (Koidz.) Koidz., is still recognized as a distinct taxon. However, the name A. palmatum var. matsumurae (Koidz.) Makino is illegitimate since it was published as a nomen nudum (Ogata, 1965). The Acer palmatum complex is widely distributed from Japan to Korea and to the coastal regions of the Yellow Sea and the East China Sea in continental China and in Taiwan (Ogata, 1965). The complex consists of the following taxa: 4. palmatum var. palmatum, A. palmatum var. amoenum (Carriére) Ohwi, and A. palmatum subsp. matsumurae from Japan, the problematical A. palmatum var. palmatum from China, A. duplicatoserratum Hayata from Tai- wan, and A. palmatum var. pilosum Nakai from Korea. The purpose of this study was to reevaluate the taxonomic status of the taxa ‘Botany Department, University of Georgia, Athens, Georgia 30602, U.S. A. © President and Fellows of Harvard College, 1990. ee of the Arnold Arboretum 71: 553-565. October, 1990. 554 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 in this complex using morphological, flavonoid, and field data. In particular, flavonoid data had already been applied with considerable success to systematic problems in Acer by Delendick (1981, 1984, 1985). MATERIALS AND METHODS More than 50 individuals of the Acer palmatum complex and A. pseudosie- boldianum (Pax) Komarov were collected from natural populations in Korea, apan, and Taiwan (only two individuals) mainly during the summers of 1986, 1987 and 1988 (see Appendix). Taxa of more limited distribution in mainland China were examined, with permission, using herbarium specimens from 4, GH, KYO, PE, NF, TAI, TI, and TUS. The extraction of foliar flavonoids followed the methods of Mabry and colleagues (1970) and Giannasi (1975). The flavonoid survey employed two- dimensional paper chromatography using Whatmann 3 MM paper. The sol- vents used were TBA (tertiary butanol: acetic acid: water, 3:1:1, v/v) as the first solvent and 15% HOAc (acetic acid: water, 15:85, v/v) as the second solvent. The flavonoid profiles were viewed under UV light and colors were recorded before and after fuming with ammonia vapor. Comparisons of the spectral and Rf data were made with published data (Delendick, 1981; Mabry et al., 1970). Structural identification of the flavonoids was accomplished using UV spectroscopy and enzymatic and acid hydrolysis (trifluoroacetic acid) for sugar analysis (Wilkins & Bohm, 1976). Recovered sugars were cochromato- graphed with standard sugars in one-dimension using ethyl acetate : pyridine: water (5:1.6:1, v/v) as the solvent. The sugars were visualized by spraying with p-anisidine hydrochloric acid (Pridham, 1956). RESULTS AND DISCUSSION TAIWANESE ACER PALMATUM The Taiwanese member of the Acer palmatum complex was named A. du- plicatoserratum by Hayata (1911) and was thought to be endemic to Taiwan (Formosa). Hayata used a Konishi specimen (collected in June, 1902) as the type specimen of 4. duplicatoserratum. Koidzumi (1911) treated this taxon as a variety of 4. palmatum, while Nemoto (1936) regarded it as a variety of A. ornatum Carriére. Sasaki (1928) described the taxon as A. matsumurae var. Jformosum Sasaki but did so without a description or reference, thereby creating an illegitimate nomen nudum (Li, 1952). Since the name A. ornatum had already been applied to one of the cultivated forms of A. palmatum var. amoenum (Hara, 1954; Ogata, 1965) by Carriére (1867), Nemoto’s A. ornatum var. mat- sumurae B. spontaneum subvar. formosanum (Koidz.) Nemoto was illegitimate. Li (1952) renamed it A. palmatum var. pubescens and characterized it as having five-lobed leaves and pubescent young branches, inflorescences, mature leaves, and petioles. He cited Suzuki 17859 (PE) as the holotype. Fang (1981) agreed with Hayata’s concept concerning the status of this taxon and thus treated it 1990] CHANG, ACER PALMATUM COMPLEX moe) as a distinct species, 4. duplicatoserratum. Despite these discernible differences, most Taiwanese botanists (Liu, 1962; Lu & Liao, 1988) follow Li’s treatment and use the name A. palmatum var. pubescens for A. duplicatoserratum. Indeed, pubescence on the ovary, mature leaves, and petioles of the Tai- wanese taxon distinguishes it from its putatively related Japanese and Korean varieties of Acer palmatum. Pubescence is present on another distantly related Japanese endemic, A. sieboldianum Miguel, but the 9- to 11-lobed leaves of A. sieboldianum easily distinguish it from the Taiwanese taxon, which has 5- to 7-lobed leaves (TABLE |). The flavonoid complement of the Taiwanese taxon (Acer duplicatoserratum in TABLE 2) is quite distinct in comparison to Japanese and Korean A. palmatum var. palmatum and A. palmatum var. amoenum due to the absence of flavones and glycoflavones (see TABLE 2, compounds | 1-16) characteristically found in the Japanese taxa. Delendick’s (1981) extensive flavonoid survey of Acer doc- umented two chemical types within section Pal/mata. The “palmatum” type dominated by the O- and C-glycosylflavones, and the “japonicum” type dom- inated by flavonol-O-glycosides (TABLE 2, compounds 1-10), but with C-gly- cosylflavones absent. As shown in TABLE 2, the Taiwanese A. duplicatoserratum clearly belongs to the “japonicum” type, and not the “pa/matum” type as suggested by previous morphological studies alone. ince the Taiwanese taxon is distinct morphologically, chemically and geo- graphically, it is treated as a distinct species, Acer duplicatoserratum, rather than as a variety of 4. palmatum. Acer dang aces Hayata, J. Coll. Sci. Imp. Univ. Tokyo 30: 65. 1911. n. Taitung Hsien, Mataianchi, N. Konishi s.n., June 1902. ee S holotype, T1!). Acer palmatum Thunb. subsp. matsumurae Koidz. var. se aoa Koidz. subvar. formosanum Koidz., J. Coll. Sci. Imp. Univ. Tokyo 32: 50. 1911. Type: Konishi g.n., June, 1902. (11, as photograph’). Acer matsumurae Koidz. var. formosanum Sasaki, List of Plants of Formosa p. 324. 1928, nomen nudum. Acer ornatum Carr. var. ae Koidz. 8. spontaneum (Koidz.) Nemoto subvar. ae mosanum (Koidz.) N , Flora Ce a Supplement p. 454. 1936. TyPeE: ishi s.n., June, 1902. eae '). Acer palmatwm Thunb. var. pubescens Li, "Pacific Sci. 6: 293. 1952. Type: Sasaki 17859 Ae PE!). er pce subsp. pubescens (Li) E. Murray, Kalmia 8(2-3): 20. 1978. Type: “Sasaki 17859 (A!, PE!). SPECIMENS EXAMINED: Taiwan. HUALIEN Hsien: Huang 166 (Tat, Tus), Liu & Liu 135 (tat), Sasaki 9478 (TAI), Tang 084 (TAI). ILAN Hsien: Sasaki s.n. (A; Herb. no. 071527, TAI), Hibino et al. s.n. (Herb. no. 082788, TAI). NANTOW Hsien: Chang 1374, Chang 1376 (GA), Kuo 14642 (GA). TAICHUNG HSIEN: Tang 437 (TAI), Ohwi s.n. (KYO). TAIPEI Hsien: Suzuki 17859 (A, PE), Suzuki 16490 (PE), Ohwi s.n. (KYO). DISTRIBUTION: endemic to north and central Taiwan (Map 1); distributed very locally in deciduous forests at altitudes of 1000-2000 m. TABLE 1. Morphological comparison of the Acer palmatum complex and related taxa.* VAR PETIOL NUMBER OF LEAF WIDTH NUTLET LENGTH WING LENGTH PUBES- PUBES- TAXA LEAF LOBES (cm) (mm) (cm) CE CENCE A. duplicatoserratum var. duplicatoserratum 7 4.9-7.1 (6.0) 2.5-5.0 (3.4) 1.3-2.0 (1.7) + + A. duplicatoserratum var. chinense 5-9 (6.6) 4.8-8.1 (6.2) 4.0-5.0 (4.5) 1.7-2.3 (2.0) + 7 A. ceriferum 5-7 (6) 4.4-5.0 (4.7) 4.0 2.0—2.1 (2.05) ? + A, sieboldianum 9 5.8-9.1 (6.5) 2.5-5.0 (3.95) 1.4-2.4 (1.9) + + A. pseudosieboldianum 9-11 (9.4) 7.5-10.3 (8.8) 3.0-5.5 (4.1) 1.4-2.4 (1.9) + + A, pubipalmatum =) 5.3-6.7 (5.7) 3.0 1.7-1.8 (1.75) ~ - A. palmatum var. palmatum 5-7 (6.6) 4.0-7.8 (6.1) 3.0 1.3-1.6 (1.4) — = A. palmatum var. amoenum 5-9 (7.0) 6.0-10.0 (8.5) 3.0-5.5 (4.3) 1.6-2.3 (2.0) — a * Range of variation with mean values in parentheses. 9¢¢ WOLAYOUAV AIONAYV AHL AO TVNUNOFL IL “10A] 1990] CHANG, ACER PALMATUM COMPLEX 557 CHINESE ACER PALMATUM Fang (1932a, 1932b, 1939, 1981) recognized two varieties of Acer palmatum in China, viz. var. thunbergii Pax and var. palmatum. The type specimen of A. palmatum consists of two different collections mounted on the same sheet. They are currently recognized as two different taxa, A. palmatum var. amoenum and var. palmatum. However, Pax (1886) applied a single name, A. palmatum var. thunbergii, to Thunberg’s type specimen. Since A. palmatum var. thun- bergii was applied to two different taxa, it was later rejected as an illegitimate name by Hara (1954) and Ogata (1965). Fang (1932a), apparently unaware of his error, nonetheless applied the illegitimate name var. thunbergii to specimens with deeply lobed leaves. He included not only the native Chinese plants under var. palmatum and var. thunbergii but also included many cultivated plants in China (probably introduced from Japan), which clearly belong to the two currently recognized Japanese taxa, var. pa/matum and var. amoenum. Fang’s (1932a) description of Chinese var. pa/matum is further contradictory to con- cepts of the Japanese and Korean var. pa/matum (Koidzumi, 1925; Nakai, 1932; Hara, 1954; Lee, 1979). For example, Fang described Chinese var. pal- matum as having globose nutlets 7 mm in diameter, wings 2—2.5 cm long and 1 cm broad (real measurements indicate that nutlets are 4.5 mm in diameter, wings 1.7—2.3 cm; see TABLE 1). In fact, Japanese and Korean var. pal/matum have slender nutlets 3 mm in diameter, with wings 1.4 cm long (see TABLE | and Ficure). In addition, the ovaries of the Japanese and Korean individuals of var. pa/matum are distinctly glabrous, unlike Fang’s Chinese individuals of var. palmatum, which have densely villous ovaries (see FiGuRE). The fruit shape and size of Chinese var. pa/matum are more like those of var. amoenum. However, the leaves of var. amoenum (6-10 cm in diameter; based on mea- surements of over 50 specimens) are larger than those of Chinese var. pa/matum (4.8-8.1 cm, TABLE | as A. duplicatoserratum var. chinense; based on mea- surements of 14 specimens). This Chinese taxon like the Taiwanese Acer duplicatoserratum belongs to the typical “japonicum” flavonoid type, i.e., no O- and C-glycosylflavone (see TABLE 2). However, the Chinese 4. pa/matum var. palmatum sensu Fang (now recognized as A. duplicatoserratum var. chinense in TABLE 2) is distinguished chemically from Taiwanese 4. duplicatoserratum var. duplicatoserratum by the resence of two kaempferol 3-O-glycosides and myricetin 3-O-rhamnoside (compound 10). These compounds are absent in typical Taiwanese A. dupli- catoserratum var. duplicatoserratum (TABLE 2). It is obvious that while distinct, these two taxa are closely related. The main difference between them 1s that A. duplicatoserratum has very pubescent leaves and petioles, while the Chinese taxon lacks pubescence on mature leaves. This is now reflected in being ac- corded only varietal status. Based on these chemical and morphological data, however, it appears that the Chinese taxon is more appropriately treated as a variety of A. duplicatoserratum rather than as a variety of A. palmatum. Acer duplicatoserratum Hayata var. chinense C. S. Chang, var. nov. Figure. Acer palmatum var. palmatum auct. non Thunb.; Fang, Contr. Biol. Lab. Chin. Assoc. Advacem. Sci., Sect. Bot. 8: 163, 164. 1932. TABLE 2. Flavonoid chemistry of the Acer palmatum complex and related taxa. FLAVONOL FLAVONES K Q M_ L-O A-C L-C UNKNOWN Ie 862 3 4 5 6 a 8 9 10 11 12 13 14 15 16 17 18 19 Japonicum type A. palmatum var. pilosum x x x x x x xX xX xX XxX 4. pseudosieboldianum x xX 4 Xx x x x xX xX X 4. duplicatoserratur var. duplicatoserratum XxX (X)> X (X) X xX XxX A. gnc il var. chinense x (X) x xX x XK xX xX XxX A, pubipalmatum x xX x »4 x x xX xX A. ceriferum x xX Xx x x x x 4 xX x Palmatum type A. palmatum ar. palmatum (X) (X) (X) (X) (Xx) (X) xX xX x %X (X) xX xX xX xX A. palmatum var. amoenum (X) (X) Xx x xX xX Xx x xX xX xX = kaempferol: a he = quercetin; M = myricetin; L-O = Ie 3- a rhamnoside; 2, luteolin O-glycoside; A-C = apigenin C-glycoside; L-C = luteolin C-glycoside K 3-O-xyloside; 3, K 3-O-arabinoside; 4, K 3-O-xylose-rhamnose; 5, Q 3-O-glucoside; 6, Q 3-O-xyloside; 7, Q 3-O- chasnaetie: 8, 0 3- Uungas, O- glucose-xylose; 9, Q 3-O-glucose-rhamnose; 10, M 3-O-rhamnoside; 11, L 7-O-glucoside; 12, A 6-C-glycoside; 13, A 8-C-glycoside; 14, A 6,8-di-C-glycoside; 15, L 6-C-glycoside; 16, L 8-C-glycoside; 17, unknown; 18, unknown; 19, e »(X) compound detected in trace quantities only and of sporadic occurrence. ellagic acid. BSS WOLAYOUAUV AIONAV AHL AO TWNANOfL IL “Toa] 1990] CHANG, ACER PALMATUM COMPLEX eb) ~24 120 Map 1. Distribution of Acer duplicatoserratum Hayata in Taiwan. Ramuli glabri; folia ambitu orbicularia 6.5-7.5 cm diametro, palmatim 7- 9-loba, lobis lanceolatis, acuminatis, argute et dupliciter serratis, 4.5-5 cm longa, petiolis 4-5 cm longis; flores 3-4 mm diametro purpurei 2 mm long); ovarium dense et longe lanatum; samarae loculo globoso 5 mm diametro glabro, incluso 2.2 cm longae, 8 mm latae, alis angulo obtuso divergent. Type: China. ANHUI PROVINCE: Mt. Chang-gon-shan, Wu-yuan, alt. ca. 2500 ft (ca. 820 m), a big shrub (30 ft, ca. 10 m), DBH 10” (25 cm), common in wood along the stream, R. C. Ching 3243, August 17, 1925 (holotype, A). ADDITIONAL SPECIMENS EXAMINED: China. ANHUI PRovINCcE: Mt. Chu-hwa-shan, Ching 2828 (A), Mt. Hwang-shan, Liou & Tsoong 293] (PE). HUNAN PROVINCE: Mt. Mo-fou- shan (Mu-fu-shan), Hsiung 5824 (GH). JIANGxI Province: Mt. Lu-shan, Hu 24/4 (a, nas), Ip 1784 (aA), Kuling, Wilson 1504 (a, us), Wilson 1505 (A). ZHENANG PROVINCE: Tien-mu-shan Fang 2126] (pe), Tien-tai-shan, Ching 1515 (A), Ching 1417 (aA, GH), Chiao 14317 (A, PE), Chiao 14379 (aA, us), Song 674 (NAS), Chang-huwa, Hu 22768 (NAS). DISTRIBUTION: endemic to eastern China (Map 2); distributed in deciduous forests at altitudes of 800-1500 m. 560 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 hy 1cm GuRE. Acer duplicatoserratum var. chinense: a, habit of branchlet with fruit (based on Ching 3243). b, fruit (based on Ching 3243), c, fruit of A. palmatum for comparison (based on Chang 1018), d, flower of A. duplicatoserratum var. chinense (based on Ching 1515). Acer pubipalmatum Fang, which is also known primarily from Zhejiang Province, is closely related to A. duplicatoserratum var. chinense. Their close relationship is reflected by their similar flavonoid complements (TABLE 2) as well as by their similar morphology. However, 4. pubipalmatum has smaller fruits, very deeply five-lobed leaves, villous hairs on the leaf veins and petioles, and densely pubescent ovaries and sepals. Morphologically, the Chinese Acer ceriferum Rehder (Rehder, 1911), which has thus far been found only in western Hubei Province (Wilson 1934), appears to be closely related to the Taiwanese 4. duplicatoserratum var. duplicatoser- ratum. The morphological differences between these two taxa are mainly in the degree of pubescence on leaves and twigs (TABLE |). However, the flavonoid chemistry indicates that A. ceriferum is most closely related to A. duplicato- serratum var. chinense rather than var. duplicatoserratum because of the pres- ence of myricetin 3-O-rhamnoside. These relationships are certainly logical from a phytogeographical standpoint since both A. duplicatoserratum var. chi- nense and A. ceriferum are mainland Chinese taxa (see below). A difference in flower color is a significant distinguishing character that separates A. pseudo- 1990] CHANG, ACER PALMATUM COMPLEX 561 32 Le -30 A “~L. : é @ HUNAN } os JIANGXI w: -26 % 114 17 120 Map 2. Distribution of Acer duplicatoserratum var. chinense C. S. Chang in eastern China. sieboldianum and A. sieboldianum (see Key). Because flower color of A. ceri- jferum and members of the 4. duplicatoserratum complex is unknown, no- menclatural changes involving A. ceriferum await further study. KOREAN ACER PALMATUM Several varieties of Acer palmatum have been described from Korea, al- though some have not been accepted. For example, the name 4. palmatum var. coreanum Nakai (Nakai, 1914) was rejected by Ogata (1965) based on Article 71 of the Jnternational Code of Botanical Nomenclature (Lanjouw, 1961) since this name was applied to an abnormal form that lacked petals in female flowers. This Article, however, was deleted from the Code at the Len- ingrad Congress in 1985 (Greuter, 1988). Nakai (1919) characterized A. pal- matum var. pilosum Nakai as having leaves with 7 to 9 lobes, doubly serrate margins, and pedicels and peduncles densely covered with white hairs. This variety, along with 4. pa/matum var. palmatum, is still listed in the Korean flora (Lee, 1979). The known localities for A. palmatum var. pilosum include Un-san City in Pyong-an-buk-do Province (Wilson 8649, A, TI), Kyong-sung- gun, Nam-ha-dan in Ham-kyong-buk-do Province (Nakai 7235, T1) and Mt. Chil-bo-san in Ham-kyong-buk-do Province in North Korea (Nakai 7236, T1). 562 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Flavonoid data (TABLE 2), however, clearly indicate that this taxon is chemically identical to A. pseudosieboldianum rather than A. palmatum var. palmatum (sensu lato), which also occurs in southeastern Korea. It appears that this taxon is merely an extremely small-leaved and small-fruited form of A. pseudosie- boldianum erroneously identified as a variety of A. palmatum. It is therefore proposed that 4. palmatum var. pilosum be treated as conspecific with 4. pseudosieboldianum, which is common in the Korean peninsula and north- eastern China. Synonymy for A. pseudosieboldianum is presented below. Acer ace a (Pax) Komarov, Trudy Glavn. Bot. Sada 22: 725. ee Type: China. Northeast (Manchuria) Port Bruce, Maximowicz , 1860 ae NY). Acer circumlobatum var. eal auncni ke Pax, Bot. Jahrb. Syst. 7: 199. 1886. _ sieboldianum var. mandshuricum Maxim., Mélanges Biol. Bull. Phys.-Math. Acad. mp. Sci. Saint- oS 12: 433. 188 36. "Tyre: not seen ie palmatum var. pilosum Nakai, Bot. Mag. Tokyo 33: 59. 1919. Type: Kore Pyong-an-buk-do Province, Wilson 8649 (lectotype, here chosen, A; opened TI). DISTRIBUTION AND KEY TO THE ACER PALMATUM COMPLEX AND RELATED TAXA It appears that Acer palmatum var. palmatum (sensu stricto) 1s restricted to southwestern Korea (Cheol-la-nam-do, Cheol-la-buk-do, and Je-ju-do Prov- inces) and Japan on Honshu (Fukushima to Yamaguchi Prefectures), Shikoku, and Kyushu. The related 4. palmatum var. amoenum, however, is common in Japan from Hokkaido to Kyushu (Ogata, 1965). The other taxa of the complex—A. duplicatoserratum var. duplicatoserratum, A. duplicatoserratum var. chinenseand A. pseudosieboldianum —that are often linked to A. palmatum appear, in fact, to be distinct and/or with close ties to other sympatric taxa in China, Korea or Taiwan. The flavonoid data have been especially helpful in delimiting inter- and intraspecific relationships. The following key is provided to aid in the identification of individuals in the Acer palmatum complex and related taxa in eastern Asia. Most of the taxa can be separated based on vegetative characters. However, without reproduc- tive characters, it is extremely difficult to distinguish between A. sieboldianum and 4. pseudosieboldianum and between 4. palmatum var. amoenum and A. duplicatoserratum var. chinense (TaBLe 1). Nonetheless, knowledge of geo- graphical distribution, in addition to floral and other morphological characters, can be effectively employed for the identification of the taxa KEY TO THE ACER PALMATUM COMPLEX AND RELATED TAXA 1. Ovary glabrous; petioles and peduncles glabrous or nearly so. 2. Nutlet of fruit ca. 2.5 mm wide and ca. 3 mm long; wings of fruits ca. 6 mm wide and ca. 14 mm long; leaf blades 4-7 cm wide; plants of southwestern Korea and JADA faces sacgae nea oes ae oka ete 4. palmatum var. palmatum. 1990] CHANG, ACER PALMATUM COMPLEX 563 2. Nutlet of fruits ca. 4 mm wide and ca. 4.5 mm long; wings of fruits ca. 8 mm wide and ca. 20 mm long; leaf blades 6-10 cm wide; plants from Japan. ...... A, palmatum var. amoenum. 1. Ovary villous; petioles and peduncles pubescent at least when youn 3. Leaf blades 9-11-lobed; margins deeply serrated with teeth ca. 3 mm long. 4. Flowers yellow; leaf blade pubescent; petioles hairy; plants of dese eens A. sieboldianum. 4. Flowers purple; leaf blade pubescent only on veins, eer a hairy; plants of Korea, northeastern China, and southern Siberia. ..................... Suh ease eo encanta atone ears ee es eesalee Gee a ee Se ene cag ee 3. Leaf blades 5—7-lobed (rarely 9-lobed); margins with teeth less than . Leaf blades ene, deeply 5-lobed; nutlet of fruit small = 3 nr aba of Caster @Mina.. x5 (ec se Bee es Bene et 2 ved A. pubipalmatum. a blades eee 7-lobed (rarely 9-lobed); nutlet of fruit large (ca. 3.44.5 1a) wn m). 6. Petioles of mature leaves persistently pubescent; plants of Taiwan. ..... A. duplic atoserratum var. duplicatoserratum. 6. nia of hee leaves pubescent, ee ng glabrous with age; plants of astern China. .................... Pee ratum var. chinense. ACKNOWLEDGMENTS I would like to express my appreciation to David E. Giannasi for his support and guidance. I also gratefully acknowledge the assistance of the following people: Samuel B. Jones, David E. Boufford, Nancy Coile, Hiroyoshi Ohashi, Mike Moore, Derek Focho, Chen-Meng Kuo, Tzen-Yuh Chiang (for field work in Taiwan), Jin Murata (for photographs of type specimens), Rupert Barneby (for correcting the Latin description), and Tchang-Bok Lee (for helpful dis- cussions of Korean taxa). I would also like to thank T. Delendick and J. Gurevitch for their critical reviews of the manuscript. I am gratefully indebted to the directors and the curators of the following herbaria for the loan of specimens: A, GH, KYO, NF, PE, TAI, TI, and Tus. This study was supported b four Student Research Grants from the Botany Department, University of Georgia. LITERATURE CITED CARRIERE, E. A. 1867. Plantes nouvelles, rares ou peu connues. Rev. Hort. 3: 300. DE i P.C. 1976. Flowering and sex oo in Acer L.: a biosystematic study. M. Landbouwhogeschool 76(2): 1-20 DELENDICK, T. J. 1981. A systematic review of the Aceraceae. Unpubl. Ph.D. disser- n. 693 pp. City University of New York. 4. Reconsideration of two ee taxa of the fullmoon maple, Acer japonicum — Brittonia 36: 49- 1985. “Acer buergerianum” ‘Jakokae’, member of the Verbenaceae. Taxon 34: - 101. Fane, W. P. 1932a. Preliminary notes on Chinese Aceraceae. Contr. Biol. Lab. Chin. Assoc. Advancem. Sci., Sect. Bot. 7: 143-188. 1932b. Further notes on Chinese Aceraceae. /bid. 8: 162-182. —. 1939. A monograph of Chinese Aceraceae. Ibid. 11: 1-346. 564 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 1981. Aceraceae. Pp. 66-273 in Flora Republicae Popularis Sinicae. Vol. 46. (In Chinese.) Science Press, Being. GiannasI, D. E. 1975. The flavonoid oo of the genus Dahlia (Compositae). Mem. New York Bot. Gard. 26(2): | GreuTer, W., ed. 1988. International ae oF Seesanan nomenclature. Regnum Ve- getabile. Vol. 118. pot Scientific Books, oo nigste Hara, H. 1954. E re aponicarum. Vol. 3. Iwanami Shoten Co., Tokyo. HAYATA, 3 1911. Materials for a flora of Formosa. J. Coll. Sci. Imp. Univ. Tokyo 30: 1-47 Kran S., & G. Murata. 1984. Color illustrations of woody plants of Japan. Vol. 1. (In Japanese.) Horikusha Publ. Co., Osaka. KOIDZUMI, a 1911. Revisio Aceracearum Japonicarum. J. Coll. Sci. Imp. Univ. Tokyo 32: |l- : vn Contributiones ad Cognitionem Florae Asiae Orientali. Bot. Mag. Tokyo 39: 299-318. Lanjouw, J., ed. 1961. International code of botanical nomenclature. Regnum Vege- tabile. Vol. 23. International Bureau for Plant Taxonomy and Nomenclature of the International Association for Plant Taxonomy, Utrecht. Lee, T. B. 1979. Illustrated flora of Korea. (In Korean.) Hyang-mun Publ. Co., Seoul. Li, H. L. 1952. The eae Acer (maples) in Formosa and the Liukiu [Ryukyu] Islands. Pacific Sci. 6: 288-2 Liu, C.C. 1962. Illustrations of native and introduced ligneous plants of Taiwan. Vol. II. (In Chinese.) National Taiwan University Press, Taipei. Lu, T. S., & S. C. Liao. 1988. Dendrology. Vol. 3. (In Chinese.) Taiwan Commercial Masry, T. J.. K. R. MARKHAM, & M. B. THomaAs. 1970. The systematic identification of flavonoids. Springer-Verlag, New York. Nakal, T. 1914. Plantae novae Japonicae et Koreanae II. Bot. Mag. Tokyo 28: 301- 305 —. 1919. Notulae ad plantas Japoniae et Coreae XX. Jbid. 33: 39-61. 1932. Notulae ad plantas Japoniae et Koreae XLII. /bid. 46: 608-613. Nemoto, K. 1936. Flora of Japan, Suppl. (In Japanese.) Shunyoso Shoten Co., Tokyo. OcaTA, K. 1965. A dendrological study on the Japanese Aceraceae, with special ref- erence to the geographical distribution. Bull. Tokyo Univ. Forest 60: 1-99. Ouwi, J. 1984. Flora of Japan. Smithsonian Institution, Washington, D.C. Pax, F. 1886. Monographie der Gattung Acer. Bot. Jahrb. Syst. 7: 77-2 63. PripHaM, J.B. 1956. Determination oe on paper chromatograph with p-anisidine hydrochloride. Anal. Chem. 28: 1967- Renper, A. 1911. Aceraceae. Pp. 83-98 in re 'S. SARGENT, ed. Plantae Wilsonianae. ol. |. Harvard University Press, Cambridge SASAKI, S. 1928. List of plants of Formosa. Natural History Society of Formosa, Tai- h iwan, VERTREES, J. D. 1978. Japanese maples. Timber Press, Forest Grove, Oregon WILKINS, C. K., & B. A. Boum. 1976. Chemotaxonomic studies in the Saxifragaceae s. |. 4. The flavonoids of Heuchera micrantha var. diversifolia. Canad. J. Bot. 54: 2133-2140. 1990] CHANG, ACER PALMATUM COMPLEX 565 APPENDIX. Material used for chemical survey. Acer duplicatoserratum Hayata: Taiwan. NANTow Hsien: Chang 1376, Chang 1374, Kuo 14642; Tarpe! Hsien: Ohwi s.n. (KYO); HUALIEN Hsien: Huang 166 (Tus), Liu & Liu 135 (Tal), Tang 84 (TAI); TAICHUNG HSIEN: Ofhwi s.n. (KYO), Tang 437 (TAI); ILAN HSIEN: Hibino et al. s.n. (Herb. no. 082788, TAI). Acer duplicatoserratum Hayata var. chinense C. S. Chang: China. ZHEJIANG PROVINCE: Ching 1417 (A), Chiao 14379 (us); ANHUI Province: Ching 3243 (a), Liou & Tsuong 2931 (PE); JIANGXI Province: Hu 24/4 (NAS). Acer pseudosieboldianum (Pax) Komarov (as 4. palmatum var. pilosum Nakai): Korea. NG-BUK-DO PRovINcE: Nakai 7235 (11); PYONG-AN-NAM-DO PRovINCE: Wilson 8649 (TI). Acer pubipalmatum Fang: China. ZHESANG PROVINCE: Chen 493 (PE). PENG & CHEN, BEGONIA S07 BEGONIA AUSTROTAIWANENSIS (BEGONIACEAE), A NEW SPECIES FROM SOUTHERN TAIWAN CHING-I PENG! AND YUNG-KUAN CHEN! A new species of Begonia, B. austrotaiwanensis Chen & Peng from southern Taiwan, is described and illustrated. This species is usually found on moist, steep, rocky slopes at 200-800 m alt. Diagnostic features include an acaules- cent, deciduous habit; moniliform rhizomes consisting of several ellipsoid to subglobose units with persistent stipules; inflorescences arising directly from the rhizome; flowers of both sexes scented, the staminate ones with four tepals, the pistillate ones with three tepals and two styles; two-locular ovaries; axile placentation; unequally three-winged capsules, with the abaxial wing narrowly triangular, erect, and much protruded; and a chromosome number of n = 18. The systematic position of B. austrotaiwanensis is in the Asiatic section Pla- tycentrum. During the past decade three new species of Begonia have been added to the flora of Taiwan (T. S. Liu & Lai, 1977; Y.-C. Liu & Ou, 1982; Peng et al., 1988). Most recently, Ying (1988) proposed Begonia fenchihuensis as new; however, an examination of his original description and type specimens reveals species of Begonia, we discovered a distinct new species that brings the total to ten; it 1s described here. Begonia austrotaiwanensis Y.-K. Chen & C.-I Peng, sp. nov. FIGURE |. Herba acaulis rhizomatibus reptantibus praedita, rhizomate aspectu grosse moniliforme, stipulis persistentibus triangularibus, 7-22 mm longis, 4-9 mm latis, induto. Folia oblique ovata, ad 38 cm longa, 32 cm lata. Inflorescentiae d 55 cm longae, axillares ex rhizomate statim orientes. Flores omnes rosei, fragrantes. Flos 6: tepala 4, decussata, duobus exterioribus rotundis ad orbi- culares, 13-35 mm longis, 9-32 mm latis. Flos ?: tepala 3, 2-seriatim disposita, nt exterioribus oppositis, rotundis ad oblata, 16-32 mm longis, 7-23 mm latis, uno interiore oblanceolato ad obovato, 12-23 mm longo, 7-23 mm lato; styli 2. Fructus capsulares, inaequaliter alato, corpore ellipsoideo, ca. 15 mm longo, 7-8 mm in diametro, ala abaxiali recta, conspicue producta, anguste triangulari, 27-46 mm longa, 9-15 mm lata, alis lateralibus multo angustiori- bus, 9-14 mm longis, 2.3-3.5 mm latis. Numerus chromosomatum, 7 = ‘Institute of Botany, Academia Sinica, Nankang, Taipei, Taiwan 11529, Republic of China. © President and Fellows of Harvard College, 1990. Journal of the Arnold Arboretum 71: 567-574. October, 1990. 568 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 Figure |. Begonia austrotaiwanensis: A, habit, x 0.3; B, bract, x 1.2; C, stipule, x 0.6; D, stamen, x 6; E, pistillate flower, x 0.6; F, staminate flower, x 0.6; G, G’, style, x 3: H, fruit, x 0.6; I, fruit, cross section, showing placentation, x 0.9; J, seed, x 60. (A-I from Chen 100, J from Chen 522.) 1990] PENG & CHEN, BEGONIA 569 Perennial, deciduous, acaulescent herb with creeping rhizome; rhizome mo- niliform, consisting of several ellipsoid to subglobose units each 9-50 mm long, 7-25 mm in diameter, densely covered by persistent stipules, these triangular, 7-22 mm long, 4-9 mm wide, acuminate at apex. Leaves with petiole to 40 cm long, 9 mm in diameter, greenish or sometimes reddish; blade obliquely to broadly ovate, 13-38 by 10-32 cm, acuminate to cuspidate at apex, unequally cordate at base (often lobed in adult leaves), irregularly dentate-serrate, green, sometimes with white foveolate spots on upper surface and/or broad brown belts along adaxial veins, chartaceous to slightly succulent, venation palmate, the veins 6 to 10. Bracts caducous, in pairs, narrowly ovate to elliptic, 8-12 by 3.5-6.5 mm, acute at apex, serrate, pale green, papery to nearly succulent, glabrous. Inflorescences in androgynous dichasial cymes to 55 cm long arising directly from rhizome; peduncle erect, to 42 cm long, 8 mm in diameter. Flowers of both sexes pinkish, scented. Staminate flowers with tepals 4, de- cussate, spreading, margin sometimes reflexed, the outer 2 rotund to orbicular, 13-35 by 9-32 mm, the inner 2 oblanceolate to narrowly obovate, 10-32 by 6-18 mm; stamens 51 to 66, golf club shaped, yellow, the filaments free, 0.9- 2.7 mm long, the anthers 1.8—2.3 mm long, 0.8—1.2 mm in diameter. Pistillate flowers with tepals 3 (very rarely 2 or 4, the fourth much reduced when present), the outer 2 opposite, rotund to oblate, 16-32 by 14-37 mm, the inner | ob- lanceolate to obovate, 12-23 by 7-23 mm; styles 2, Y-shaped, 4.2-5.3 mm long, yellow, the base separate to short-connate, 2.1—2.6 mm long, each arm covered with a spiral, papillose stigmatic band; ovary 2-locular, unequally 3-winged, obovoid, white to pale green, glabrous; placentae axile, bilamellate, the lamella forked. Infructescences to 72 cm long, the fruit-bearing pedicels 28-47 mm long; mature fruit a dry, trigonous, very unequally 3-winged capsule, the body ellipsoid, 9-15 mm long, 7-8 mm in diameter, the abaxial wing erect, much protruded, narrowly triangular, 27-46 by 9-15 mm, the lateral wings much narrower, 2.3-3.5 by 9-14 mm. Seeds numerous, narrowly ovoid, 0.39-0.43 mm long, 0.19-0.25 mm in diameter, rounded at apex, constricted at micro- pylar end, yellowish brown. Chromosome number, 7 = 18. DISTRIBUTION. Presently known only from Kaohsiung Hsien, Taiwan (see MAP 1), 200-1000 m alt. Type: Taiwan, Kaohsiung Hsien, Mawlin District, Shanping, 22°58'N, 120°41’E, elev. ca. 700 m, on moist, steep, rocky slope, Y.-K. Chen 100 (living collection made on 27 July, 1987; type specimens pressed from cultivated plants in October 1988) (holotype, HAST; isotypes, GH, K, MO, NY, TAI, TAIF, US). ADDITIONAL SPECIMENS EXAMINED. Taiwan. KAOHSIUNG Hsien. Maolin Distr.: Shanping, 22°58'N, 120°41'E, elev. ca. 700 m, Chen 522 (Hast), Peng 10091 (HAsT); Shanping logging trail, 22°58’N, 120°40’E, elev. ca. 600 m, Chen 354 (Hast); Tona Hot Spring, 22°55'N, 120°42’E, elev. ca. 350 m, Chen 638 (Hast); Tona logging trail, 22°54'N, 120°43’E, elev. ca. 750 m, Chen & Peng 375 (HAST), 22°53'N, 120°44’E, elev. ca. 800 m, Chen & Peng 381] (HAST); Tachin Waterfall, 22°52'N, 120°39’E, elev. ca. 300 m, Chen 619 (HAST). Liukuei Distr.: Wangshanchiao, 22°55'N, 120°38’E, elev. ca. 200 m, Chen 344 (HAST). 570 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 | | 120° 121 ~25° 25— ~24° 24~ a Pg Pre { g i | og / o 25 0° / ) ° S / PY 7 9 -23° ‘ 23 N 0 30 60km | co al -22° 22° 120 Map |. Distribution of Begonia austrotaiwanensis (stars) and B. laciniata (dots) in Taiwan. HABITAT AND ASSOCIATED PLANTS. Begonia austrotaiwanensis often grows in moist, shallow soil on steep, rocky slopes and in somewhat shady habitats. The following species are commonly seen associated with it: A/ocasia macrorrhiza (L.) Schott, Arachnioides rhomboidea (Wallich ex Mett.) Ching, Asparagus cochinchinensis (Lour.) Merr., Aspidistra daibuensis Hayata, Aster taiwanensis 1990] PENG & CHEN, BEGONIA 571 Kitam., Begonia laciniata Roxb., Carex baccans Nees, Colocasia formosana Hayata, Cyanotis kawakamii Hayata, Elatostema sessile Forster var. cuspi- datum Wedd., Oplismenus undulatifolius (Ard.) Roemer & Schultes, Peperomia japonica Makino, Pilea kankaoensis Hayata, Pilea trinervea Wight, Pollia se- cundiflora (Bl.) Bakh., Selaginella delicatula (Desv.) Alston, Setaria palmifolia (Koenig) Stapf, and Tricyrtis formosana Baker. KEY TO THE SPECIES OF BEGONIA IN TAIWAN The following ae 1S s provided to aid in the identification of the ten species of Begonia indigenous to Taiw . Plant acaulescent; inflorescences arising directly from rhizome. 2. Tepals of pistillate flowers 3 (very rarely 2 or 4); rhizome moniliform; stipules persistent, plant deciduous. ............ 0.0.00... 000s . austrotaiwanensis. 2, en an pistillate flowers 5; rhizome not moniliform; seule deciduous; plant B. fenici — EVETOTCON., au Hs Hob aS eos Hoo pa pe ee Se en . a. ee inflorescences arising from axils of cauline leaves 3. pals 2 in both staminate and Berm flowers; tubers and stolons produced a the erect stem base; plant deciduous. ....................... . ravenil. 3. Tepals 4 in staminate ae 5 to 6 (to 8) in pistillate flowers; tubers and stolons lacking; plant ev 4. Plant with distinct ecoies rhizome. 5. Plant densely pubescent See Deke ie ae eye tae ae B. laciniata. 5. Plant glabrous to subglabrou 6. Leaves irregularly finely cane abaxial wing on capsules rotund to — Orbicular, 19-26 11M 1ONBY wae hee ceca ede sees us a ae 6. aa laciniate; abaxial wing on capsules nearly uae 7-16 m NGS ce eee aes ee ane eae See ae B. formosana. 4. Plant aahout creeping rhizome. 7. Plant densely pubescent. ..........0...00 0000.00 aee B. buimontana. 7. Plant glabrous to subglabrous. 8: Capsules wingless. 2:6: 2.04<$ist eb ivied eieide bed edad B. aptera. 8. Capsules unequally 3-win 9. Mature leaves narrowly lanceolate to lanceolate, 7-14 by 2-4 cm, width/length quotient ca. %—’; petiole 2-4 cm long. ............ b pitecacaha ecg cy ate pie ee oo eee ne B. taiwaniana. 9. Mature leaves broadly lanceolate to narrowly ovate, 12-33 by 6-14 cm, width/length quotient ca. %—-%; petiole 6-26 cm long. ....... B. lukuana. CYTOLOGY Cytological studies of Begonia austrotaiwanensis (see FIGURE 2) reveal n = 18 or 19 in pollen mother cells and 2n = 36-38 in root-tip cells. Some inter- bivalent connections are observed in diakinesis, but the segregation of chro- mosomes at anaphase appears normal. Three small bivalents and either 15 or 16 larger ones are observed at diakinesis. This is corroborated by the somatic complement that consistently comprises six small chromosomes and 30 to 32 larger ones. It is possible therefore that the somatic chromosome number of B. austrotaiwanensis 1s 2n = 36, with the additional one or two large chro- mosomes observed in some cells being B-chromosomes. The B-chromosomes eye JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 FIGURE 2. at rai of Begonia austrotaiwanensis (from Chen 375): A, diaki- nesis, H = 18; B, diakinesis, nm = 19, possibly 18 + 1B. Scale bar = 10 um; sb = small bivalent; ic = vnterb valent connection. appear to be euchromatic and are not readily distinguishable from somatic chromosomes. DISCUSSION Begonia austrotaiwanensis is unique in having the combination of an acau- lescent, deciduous habit; a moniliform rhizome consisting of several ellipsoid to subglobose units with persistent stipules; inflorescences arising directly from the rhizome; flowers of both sexes scented, the staminate flowers with four tepals, the pistillate ones with three tepals and two styles; two-locular ovaries; axile placentation; and very unequally three-winged capsules, with the abaxial wing narrowly triangular, erect, and much protruded. Plants of B. austrotai- wanensis usually grow on steep, bare, rocky slopes or rock faces with shallow soils, often near seasonal waterfalls or by streams. They are found at elevations between 200 and 800 m along the forest margin in southern Taiwan, an area with a monsoon climate. It is a very distinct species that is often sympatric with B. laciniata Roxb., which has a much wider distribution in Taiwan (see Map 1). Begonia laciniata is usually found in the forest or at the forest margin at elevations between 600 and 2100 m. When the two species occur together, B. austrotaiwanensis tends to grow higher up on the rocky slopes, whereas B. laciniata is usually found at the base of the slope and 1s frequently intermixed with the abundant herbaceous vegetation. An examination of Begonia specimens deposited at all herbaria in Taiwan reveals that B. austrotaiwanensis was not collected until 1986. It was overlooked in the past presumably due to its restricted distribution and because it often occurs together with the widespread B. /aciniata, from which it is not readily distinguishable at a distance. Since B. austrotaiwanensis tends to grow higher up on steep rocky slopes that are relatively inaccessible, plant collectors are likely to ignore it and collect instead the abundant B. /aciniata at their feet. The fact that B. austrotaiwanensis sheds its leaves in the winter also makes it less visible than the evergreen B. /aciniata. 1990] PENG & CHEN, BEGONIA 573 Traditionally, the number of tepals in staminate and pistillate flowers, the number of styles and ovarian locules, and the placentation type are important characters in delimiting the sections of Begonia (Baranov, 1981; Baranov & Barkley, 1974; Barkley, 1972a; De Candolle, 1864; Warburg, 1894). Begonia austrotaiwanensis appears to fit in section Weilbachia A. DC., which otherwise includes ten species from Mexico and two from Central America (one from El Salvador, one from Guatemala and El Salvador) (Baranov, 1981; Barkley, 1972b, 1972c; Barkley & Baranov, 1972; Barkley & Golding, 1974; Smith et al., 1986; Warburg, 1894). Dieter C. Wasshausen (pers. comm.) suggested placing this new species in the Asiatic section Platycentrum, which character- istically has 4 to 6 tepals in the pistillate flowers (B. austrotaiwanensis usually has three, very rarely two or four, the fourth much reduced when present). Other than this character, which is evolutionarily labile (L. Brouillet, pers. mm.), B. austrotaiwanensis appears to be appropriately placed in sect. Pla- tycentrum. Studies of leaf microcharacters of Begonia conducted in Brouillet’s laboratory showed that sects. Platycentrum and Weilbachia differ in several traits related to indumentum and anatomy (Brouillet, pers. comm.). Section Platycentrum characteristically has leaf emergences with processes and pneu- mathodes (stomata in lines) on the petiole. By contrast, sect. Weilbachia lacks the above characters but has T-shaped glandular hairs and oxalacetic crystals in the leaves. In our study we found that B. austrotaiwanensis agrees perfectly in leaf microcharacters with sect. Platycentrum and may be appropriately placed in this Asiatic section. ACKNOWLEDGMENTS This work was supported in part by a research grant from the Academia Sinica, Taipei, Taiwan, to Ching-I Peng. We thank Shu-Chuan Shaw, Biology Department, National Chengkung University, who first collected a living spec- imen of Begonia austrotaiwanensis from Shanping, Taiwan, in July, 1986, and presented it to the senior author; Nai-Hang Chang, Taiwan Forestry Research Institute, for field assistance at the type locality; and Hsin-Fu Yen, Horticultural Department, National Taiwan University, who identified some of the plants associated with B. austrotaiwanensis. We are indebted to Hongya Gu (Center of Life Sciences, Peking University) for providing useful information concern- ing the mainland Chinese species of Begonia. We are most grateful to John D. Dwyer (Missouri Botanical Garden) for assistance with botanical Latin. Luc Brouillet (Institut Botanique de l'Université de Montréal), Dieter C. Wasshau- sen (Smithsonian Institution), Bernice G. Schubert, and David E. Boufford (both at the Arnold Arboretum, Harvard University) provided useful com- ments in their reviews of the manuscript. LITERATURE CITED BaRANOV, A. I. 1981. Studies in the Begoniaceae. Phytologia Mem. 4: 80-83. — & F. A. BARKLEY. 1974. The sections of the genus Begonia. 28 pp. Northeastern University, Boston 574 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 BARKLEY, F. A. 1972a. Key to the sections of the Begoniaceae. Buxtonian 1(Suppl. 3): 1-7. . 1972b. oe The genera, sections, and known species of each. Ibid. 1(Suppl. 4): 1-20 —. 1972c. The species of the Begoniaceae. /bid. 1(Suppl. 5): 1- —- & A. BARANOv. 1972. The sections of the Begoniaceae. /bid. ee 1): 1-8. J. Gotpinc. 1974. The species of the Begoniaceae. ed. 2. Northeastern University, Boston. CANDOLLE, A. DE. 1864. Begoniaceae. Prodr. 15(1): 266-408. Liu, T.-S., & M.-J. Lar. 1977. Begoniaceae. Pp. 791-798 in H.-L. Li, T.-S. Liu, T.-C. HuAana, T. KoyaAMa, & C. E. DeVor, eds. Flora of Taiwan. Vol. 3. Epoch Publishing Co., Taipei Liu, Y.-C., & C.-H. Ou. 1982. Contributions to the rane ae plants of Taiwan Bs Bull. Exp. Forest Natl. Chung Hsing Univ. 4: 1-16. G, C.-I, Y.-K. CHEN, & H.-F. YEN. 1988. Baie. ravenii (Begoniaceae), a new ele from Taiwan. Bot. oo Acad. Sin. 29: 217- on SmiTH, L. B., D. C. WASSHAUSEN, J. GOLDING, & C. E REGEANNES. 1986. Begoni- aceae. Part 1: Ulustrated i. Part 2: Annotated ris list. Smithson. Contr. Bot 60: 1-584. WARBURG, O. 1894. Begoniaceae. Jn: A. ENGLER & K. PRANTL, Nat. Pflanzenfam. III. -150. Osivor Y. 1933. Observationes ad floram Formosanam VIII. J. Soc. Trop. Agr. 346-354 YinG, S.-S. 1988. Coloured illustrated flora of Taiwan, with the introduced plants. Published by the author, Taipei 1990] BOUFFORD & YING, MALLOTUS 575 A NEW SUBSPECIES OF MALLOTUS OREOPHILUS (EUPHORBIACEAE) FROM SOUTH-CENTRAL CHINA DAvipD E. BOUFFORD! AND TSUN-SHEN YING? Mallotus oreophilus subsp. /atifolius (Euphorbiaceae), which differs from subspecies oreophilus in having leaves consistently wider than long, is described as new from Sichuan Province, China Mallotus oreophilus Muell.-Arg. subsp. latifolius Boufford & T. S. Ying, subsp. nov. FIGURE Mallotus oreophilus similis sed in foliis latioribus subreniformis vel trans- verse oblongis differt. Shrubs or small trees 1.5—13 m tall. Leaves subreniform to oblong, the blades 6-14 cm long, 6.5—20 cm wide, base truncate to very broadly cuneate, apex abruptly short acuminate, upper surface glabrous or with sparse stellate hairs along the main veins, lower surface densely stellate tomentose. Petioles, rachis of inflorescence, flower buds and fruiting capsules densely stellate tomentose. Branchlets glabrescent with stellate hairs. Type: China. Sichuan Province: Dujiangyan Municipality (formerly Guan Xian); between Guanmengzi and Guihuashu, upstream from Longxi along the Longx1 River, elev. 1120 m, tree ca. 20 m tall, on moist slope, 3 Sep 1988, D. E. Boufford & B. Bartholomew 24496 (holotype, PE; isotypes, A, CAS, HAST, IBSC, KYO, L, TI). ADDITIONAL SPECIMENS EXAMINED. China. SICHUAN: Vena Xian, Chuanjingyi sald O56] (pe); Emei Xian, C. Y. Chiao & C. S. Fan 184 (aA), C. L. Chow 6148 (a), W. P. Fang 7842 (pe), 7843 (PE), 7914 (A, PE), 7915 (PE), 12690 (A, PE), 16497 (A), 16640 (PB), 16897 (A), 18863 (A), 18928 (PE), K. Kuan 710 (re), 1521 (PE), Sino-U. SS. R. Exp Team 1845 (pe), 2472 (pE), T. H. Tu 189 (PE), F. T. Wang 23148 (A, Pe), K. H. Yan ng 54677 (PE), 54897 (PE), 55595 (PE), ee (PE); Guan Xian, S. Chang 00525 (PE), Y. S. Liu 1954 (a), S. S. Chien 5692 (a), W. P. Fang 2103 (a), F. T. Wang 20909 (a); Hongya Xian, S. Chang 1723 (PE), T. W. Yao 4060 (Pe); Mabian Xian, F. 7. Wang 23096 (a); Pingshan Xian, Chuanjingyi Team 0607 (pe), 0952 (PE); Tianquan Xian, S. Chang 01762 (PE), Chuanjingyi Team 00411 (pe); Ya’an Xian, C. Y. Chiao 2049 (a), T. P. Wang 8494 (PE); Yibin Xian, Chuanjingyi Team 1247 (PE); Yingjing Xian, 7. W. Yao 3762 (PE). Mallotus oreophilus subsp. /atifolius differs from the typical subspecies in the shape of the leaf. Subspecies oreophilus has the blades ovate to widely ovate, acute to acuminate, and clearly longer than wide; in subspecies /atifolius ‘Harvard University Herbaria, 22 Divinity Avenue, Cambridge, Massachusetts 02138, U.S. A Institute of Botany, Chinese Academy of Sciences, Beijing 100044, People’s Republic of China. © President and Fellows of Harvard College, 1990. Journal of the Arnold Arboretum 71: 575-578. oe. 1990. JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 576 ce ch ee iy 2G i“ Mallotus oreophilus subsp. latifolius: a, habit; b, lower surface of leaf showing FIGURE. stellate hairs; c, mature capsule; d, seed. Based on Boufford & Bartholomew 24496 (PE). 1990] BOUFFORD & YING, MALLOTUS af Map of China showing distribution of Mallotus oreophilus subsp. latifolius. the leaves are subreniform to oblong and consistently wider than long. The relatively short, acuminate leaf tip in subspecies /atifolius arises abruptly from the otherwise rounded to subtruncate apical edge of the blade making it quite pronounced. Subspecies /atifolius occurs in the mountains on the southern and western edge of the Sichuan Basin (see Map). Mallotus oreophilus subsp. oreo- Dhilus ranges from south-central China through the Himalayan region to the eastern Himalaya (Shaw, 1968). We have seen a single specimen from Yunnan Province, Mengse [Mengzi] Xian (A. Henry 10925 (A)), well south of the range of subspecies /atifolius, that appears intermediate between typical subspecies oreophilus and subspecies /a- ifolius. ACKNOWLEDGMENTS We wish to thank Bruce Bartholomew, Li Gang, Zhu Guanghua, who par- ticipated in the fieldwork, Chen Minghong for making the arrangements in western Sichuan in 1988, C. Z. Ji for his beautiful illustration, and the National Geographic Society for its generous support of fieldwork in China. 578 JOURNAL OF THE ARNOLD ARBORETUM [voL. 7] LITERATURE CITED SHaw, H. K. Airy. 1968. Notes on Malesian and other Asiatic Euphorbiaceae XC: new or noteworthy species, transferences, etc., in Ma/lotus Lour. Kew Bull. 21: 379- 400. 1990] BOOK REVIEW Sie! BOOK REVIEWS Tropical Woody Rubiaceae, by E. Robbrecht. Opera Botanica Belgica Volume 1.272 pp. Meise: Nationale Plantentuin van Belgié. 1988. ISBN 90-72619- 02-1. Softcover. No price given. The premier issue of Opera Botanica Belgica is the first family-wide summary of the systematics of the Rubiaceae to appear in almost twenty-five years, bringing us up to date in virtually all areas of research bearing on the taxonomy of the fourth largest family of angiosperms (637 genera and roughly 10,700 species). Most of this well organized study describes and evaluates “‘charac- teristic features” (especially morphological and anatomical features) of Rubi- aceae that have been, or might be, employed by taxonomists. Discussion ex- tends to idioblasts, exudates, crystals, arilloids, ant and mite associations, bacterial and fungal interactions, pollination biology, dispersal biology, bio- eography at the generic and tribal levels, and much more. The new classifi- cation follows from a re-evaluation of the characters as well as the taxonomic schemes proposed by Schumann (1891), Verdcourt (1958), and Bremekamp (1966). Dr. Robbrecht convincingly argues that division of the family into four subfamilies and more than forty tribes yields an improved alignment of mor- phological and anatomical data that is also supported by chemical and cyto- logical information, as scanty as that often is. Woody Rubiaceae are emphasized, but herbaceous (and mostly temperate) taxa are included in the proposed classification, which extends to the entire family. Herbaceous species are present in less than 20 percent of rubiaceous genera, and their repeated derivation from woody ancestors is evident. Another admitted bias is toward Old World (especially African) genera, and the subfam- ily Ixoroideae, which is where the author has much research experience, and where our knowledge of the family is thus relatively far advanced. Robbrecht’s reliance on floristic as well as monographic research is evident throughout, and one senses that years of “routine” herbarium identifications have contributed substantially to his comprehensive view of character distribution in the family. The discussion of “‘characteristic features” is presently the most reliable and accessible exposition of morphological and anatomical variation in the Ru- biaceae. Terminology pertaining to trichomes, the wall structure of exotesta cells, colleters, and inflorescences 1s expanded and clarified. For example, the anomalous “hairs” on Hillia seeds are shown to be highly divided wings. The discussion is thorough and unambiguous, an exception perhaps being the dis- tinction between ‘“‘fang”’ and “climbing” hooks in some lianas (if indeed they are different structures), and the characterization of Airosperma as having co- rolla lobes contorted to the right, an error apparent in Schumann’s original illustration where both left- and right-contorted corollas are figured. The author often suggests the adaptive significance of characters, although © President and Fellows of Harvard College, 1990. Journal of the Arnold Arboretum 71: 579-582. October, 1990. 580 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 where data are slim the conclusions are necessarily tentative. Robbrecht does not over-speculate, and for a few systematically significant features—such as ovule number or exotesta cell anatomy—he advances no adaptive explanation. Is there a correlation between seed coat anatomy and fruit type, perhaps? The taxonomic value or “reliability” of characters in distinguishing genera, tribes, and subfamilies is illustrated with reference to their use in prior system- atic studies. The many characters deemed useful only at lower taxonomic ranks are, of course, those that exhibit strong homoplasy among the tribes. Other potentially useful characters, such as stipule ontogeny and vascularization, pollen, and fruit-production strategies, are still insufficiently known to be of much systematic value at the tribal level. Homoplasious evolution is also strongly suggested for many characters when they are mapped against the author’s proposed classification in a series of Venn diagrams. This mapping method, where tribes are represented by circular shapes, allows the “placing” of taxa in the system without implying a Probable instances of convergence include reduction in the number of ovules per ovary locule and presence of raphid crystals—characters that served as the basis for subfamilial circumscription by Schumann (in the case of ovule num- ber) and by Verdcourt and Bremekamp (for raphides). Other instances of con- vergent or parallel evolution include tetramerous flowers, a tendency for in- crease in flower parts (pleiomery), corolla lobe aestivation, dioecism and heterodistyly, and development ofa stylar pollen-presentation mechanism. The parallel evolution of small fruits in both Old and New World Gardeniineae is an example of how data from floristic studies contribute to evolutionary hy- potheses. Robbrecht’s classification purports to delimit “natural”? taxa, and is “evo- lutionary” in the sense that paraphylesis is accepted while polyphylesis is not; he considers phylogenetic classification of Rubiaceae to be an eventual goal. (Garcia Kirkbride’s (1982) phylogeny of Rubiaceae subfamilies based on two characters, and Bremer’s 287) any of Arposiemmialea’ and Hamelieae are among the very few explicitly cladistic d out for Rubiaceae. ) Robbrecht emphasizes taxa that are “intermediate, ” linking higher groups, when discussing relationships. He accepts the supposition (F. Hallé, 1967) that Ixoroideae are the most primitive subfamily, but further phylogenetic specu- lation is limited (indeed, ““~phylogeny” does not appear in the index), although evolutionary trends are hypothesized for nodal anatomy, stipules (four per node is primitive and rare), and habit (herbs are advanced). Placental evolution is outlined for Hedyotideae (from phyllospory to stachyospory) and for Ru- bioideae (from pluriovulate to uniovulate placentas). For broader phylogenetic hypotheses, the available data base seems yet inadequate. Looking for overall correlations among characters, Robbrecht abandons the single-character criteria previously employed to align genera and tribes. The result is a scheme that resembles Schumann’s classification in some important respects, but cannot be considered a return to the ‘‘classical’’ systems of the nineteenth century. Bremekamp’s small, segregate subfamilies are rejected, as is Verdcourt’s apparently unwarranted reliance on raphid crystals as the pri- 1990] BOOK REVIEW 581 mary defining feature of Rubioideae. Two examples give an idea of the extent of Robbrecht’s rearrangements, as well as the rationale informing them. Subfamily Rubioideae contains, ‘“‘without doubt, the most advanced mem- bers of the family,” i.e., the largely herbaceous tribes. The subfamily includes all tribes assigned there by Bremekamp, except Knoxieae and Craterispermeae (moved to Antirheoideae), as well as Ophiorrhizeae (placed by Bremekamp in Urophylloideae), and Pomazoteae (placed in its own subfamily by Breme- kamp). While possession of raphids was essential for membership in Rubioi- deae in the systems of Verdcourt and Bremekamp, Robbrecht points out that it cannot stand alone as a defining feature, since species possessing raphides are manifestly related to others without aa Robbrecht identifies multiovu- late and uniovulate tribes as g two major groups within the subfamily, the uniovulate group being the more ae (monophyletic?), but some Morindeae— usually with one ovule per cell—may sometimes have two. The Hedyotideae “‘indicates that a link between the two [tribal groupings] seems plausible. This same tribe seems to link the subfamily Rubioideae with the Cinchonoideae, viz. the Isertieae.” The tribes Anthospermeae, Paederieae, The- ligoneae, Spermacoceae, and Rubieae show a trend toward dry fruits as well as herbaceous habit Subfamily Antirheoideae is the most heavily emended of Bremekamp’s and Verdcourt’s subfamilies—by them restricted to tribe Guettardeae— but here greatly expanded to include Alberteae, Knoxieae, Vanguerieae, and Chio- cocceae, these last tribes distributed to other subfamilies by Verdcourt and Bremekamp on account of presence of raphides or appreciable endosperm. Robbrecht’s Antirheoideae is essentially a reconstruction of Schumann’s [su- pertribe] ““Guettardinae,’’ members of which “show so striking similarities in their flowers and fruits, that Bremekamp’s and Verdcourt’s dispersal of the ‘“Guettardinae”’ over different subfamilies resulted, no doubt, in a more arti- ficial classification.” To this re-expanded subfamily Robbrecht also assigns the Retiniphylleae, Cephalantheae, and Craterispermeae. The gaps in our rubiaceous knowledge are many and wide, and Robbrecht calls for survey-type studies to establish the limits of tribes Cinchoneae, Con- damineae, Rondeletieae, and Isertieae. In the Antirheoideae, the tribal position of many genera 1s uncertain, and this is also true for Psychotrieae, where delimitation of the enormous Psychotria is a problem. Modern revisions are lacking for many genera, and our knowledge is uneven geographically. The least well known Rubiaceae are in Malesia and Madagascar, where groups often have an important bearing on an understanding of African and Asiatic taxa. Concerning Madagascar, the study of Rubiaceae is “‘practically dormant” while relentless deforestation results in the continuing loss of species. The study of Australian and New Caledonian Rubiaceae is said to be in a similarly quiescent state. In the Flora Neotropica series, a single monograph (on the three genera of Henriquezieae) has appeared—more quiescence. Perhaps only a fragment of living tropical woody Rubiaceae will be saved for future investigation, but an enormous and hardly exploited resource of evolutionary data 1s already present in systematic collections. Robbrecht shows 582 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 us how those data can be extracted and what we should be looking for. As a minimal data set that should be assembled for the whole family (presumably for all genera) he suggests the following: presence or absence of raphides; anat- omy of external hairs; corolla aestivation; data on floral biology; anatomy of seed exotesta; and pollen type (position, number, and type of apertures and features of the exine). This is, of course, far from what might be achieved with concerted and coordinated effort, namely an exhaustive, accessible, perfectly comparable, and thoroughly vouchered data base for Rubiaceae worldwide. The present study makes it plain that a lot of information can be gathered using the simplest techniques. As a reference work, Dr. Robbrecht’s monograph 1s excellent. In the expo- sition of his proposed classification, he gives the principal features of each suprageneric taxon, references to previous studies bearing on tribal classifica- tion, and a summary of geographical distribution. All included genera are listed for each tribe and subtribe. Among the appendices is a list of all accepted generic names and synonyms with tribal assignment—the first such published index in nearly a century. The bibliography of about 300 references includes the pages where each is mentioned in the text (a very useful feature), and in the subject index the main place of discussion is set in bold type. The taxonomic index also includes page references to Schumann, Verdcourt, and Bremekamp. Typographical errors are few, and I found only two cited references omitted from the bibliography. Commenting on the task of assembling and arranging so much material, the author admits that the Rubiaceae are “so vast a family that a single researcher cannot hope to gather all essential taxonomic data for an in-depth analysis.” But some researchers are evidently better equipped to do so than others. Addressing Volken’s question of how Schumann’s great productivity might be explained, Stafleu and Cowan (1985) suggest that “‘a really convincing and comprehensive answer is not possible.” Schumann’s industry, and that of his successors in Rubiaceae systematics— Verdcourt, Bre- mekamp, and Robbrecht—is an inspiration.—S. P. DARwin, Department of Biology, Tulane University, New Orleans, Louisiana 7011 LITERATURE CITED BREMEKAMpP, C. E. B. 1966. Remarks on the position, the delimitation and the sub- division of the Rubiaceae. Acta Bot. Neerl. 15: 1-33. Bremer, B. 1987. The sister group of the paleotropical tribe Argostemmateae: a re- defined neotropical tribe Hamelieae (Rubiaceae, Rubioideae). Cladistics 3: 35-51. GarciA Kirksripe, M. C. 1982. A preliminary phylogeny for the neotropical Rubi- aceae. Pl. System. Evol. 141: 115-121. HALLe, F. 1967. Etude biologique et Se de la tribu des Gardeniées (Ru- 8. T. O. M. 22: 1-146. SCHUMANN, K. 1891. Rubiaceae. /7: A. oe K. PRANTL, Nat. Pflanzenfam. IV. 4: 1-156. STAFLEU, F. A., & R.S. Cowan. 1985. Taxonomic Literature, vol. 5. Regnum Vegetable vol. | ol. , VeRDcouRT, B. 1958. Remarks on the classification of the Rubiaceae. Bull. Jard. Bot. Belg. 28: 209-281. 1990] BOOK REVIEW 583 Wayside Trees of Malaya, by E. J. H. Corner. ed. 3. Kuala Lumpur: The Malayan Nature Society, 1988. 2 volumes, 861 pp., 236 plates, & 260 figures. ISBN 967-99906-0-5. Hardcover. No price given. Out of print for decades, that triumph of natural history, the Wayside Trees of Malaya, has now appeared in a new edition. The Wayside Trees describes the commonly cultivated species and the fre- quently encountered indigenous trees of one equatorial country, but its value is global. Professor Corner pioneered the objective description of tree shape, bark morphology, and mode of growth. His observations on phenology and reproductive biology are still unexcelled; his narrative remains provocative and totally absorbing The form and content of the original 1940 edition remain essentially un- changed, though it is a little sad that the quality of printing has declined, particularly with respect to the illustrations. Nomenclature has been meticu- lously checked. The accounts of several families are now illuminated by the author’s ‘““Durian Theory,” to which his researches on Malaysian trees gave birth. Other new information includes Hallé and Oldeman’s on tree branching patterns, Koriba’s on modes of growth, and Medway’s on phenology. The essay on the natural history of figs is embellished with fascinating additions. The section on trees of local interest has been brought up to date and expanded with the author’s astute eye and acerbic wit. By and large the family and species descriptions stand firm also. This is because they remain based on opinion derived from Corner’s own field ob- servations within one country. Generally, his family and generic concepts are broader, and his species narrower than are currently common in the Far East, concepts that are under the spell of “Flora Malesiana.”’ Nevertheless, the Bom- bacaceae are now separated from the Malvaceae, and the Moraceae and UI- maceae from the Urticaceae (where Streblus becomes four genera); the Ani- sophylleaceae are retained; but xonanthes and Erythroxylum are kept together on the basis of their seed structure; /rvingia remains in the Simaroubaceae; the Chrysobalanaceae in the Rosaceae; and Duabanga and Sonneratia in Lythra- ceae. Various satellites of the Euphorbiaceae are retained in that family. The original generic definitions and names in such families as the Rosaceae (Chry- sobalanaceae), Apocynaceae, Rubiaceae, and Leguminosae, where they have been under recent fissiparous scrutiny, have been maintained. Langsats and dukuns are placed in one species, as are the forms of A//ophyllus, but not those of Pometia. This is the work of someone who knows the plants and only makes changes when personally convinced. Everywhere, Corner’s continuing love of Malaya, its people, and its forests shines through the text. Considering the more than forty years of change of old forests to plantations, lanes to superhighways, and villages to cities that have elapsed, it is not surprising that many of the trees originally illustrated have gone. Amazingly, some also remain—it is always fascinating to learn how they © President and Fellows of Harvard College, 1990. Journal of the Arnold Arboretum 71: 583, 584. October, 1990. 584 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 have survived. This should challenge the Malayan Nature Society to search for others that are still alive. As testimony of the continuing value and majesty of this masterpiece to the public, the embarking graduate student, and the experienced scientist alike, the following is taken from a new passage in the introduction to the Dipterocar- paceae: ““When we look upon the lowland forest or gaze up into its vaults, we see the canopy of dipterocarps whose sombre crowns compose very largely the ocean of trees that once covered the Malay Peninsula. .. . This glorious specta- cle has been whittled away in the course of civilization, decimated this century by commercial logging, and now, with urban demand for agriculture, it is in danger of disappearing. Vast trunks thunder along highways to saw-mills, ap- parently from nowhere, and revenue accrues, but where can the citizen, the biologist, or the visitor see these fabulous giants—the most majestic trees that any land produces? One would have thought that such a national heritage should have been guarded zealously. There remain, fortunately, some tracts of this lowland forest preserved in catchment areas, national parks, forest reserves and game reserves though, as the shortage of timber increases, they may be deprived.”’—P. S. Asuton, Harvard University Herbaria, 22 Divinity Avenue, Cambridge, Massachusetts 02138. 1990] INDEX 585 INDEX Abutilon, 514, 515, 521 Acrolophus maculosus, 442 — indica, 515 Acroptilon, — theophrasti, 490, 495, 513-515, 521 — repens, 443 Acacia eae 490, 498, 510, 511 Acrotome, 359 — farnes — angustifolia, 337 Acrymia, Acanthaceae i 360-363 Acantholepis, 44 Acer Senate: Complex in China, Tai- wan, and Korea, A Reconsideraien of — — var. nam — ceriferum, 556, 558, 560, 5 — circumlobatum var. aoe. lanum, 56 — duplicatoserratum, 553-555, 557, 559, 561, — — var. chinense, 553, 556-563, 565 — — var. duplicatoserratum, 556-558, 560, 562, 563 — matsumurae var. formosanum, 554, 555 — ornatum, — — var. matsumurae # spontaneum sub- 55 ro) tome wn 5 Cc — ~ ao Wn — — — var. spontaneum subvar. formo- sanum, 555 — — subsp. pubescens, 555 — — var. amoenum, 2 e 558, 562, 563 — — var. coreanum, — — var. matsumura ee — — var. palmatum, 353, 555-558, 561, 562 — — var. pilosum, 553, 558, 561, 562, 565 — — var. earns 554, 555 — — var. thunbergii, 557 ~ pecudosebolclena 554, 556, 558, 560-563, 565 — pubipalmatum, 556, 558, 560, 563, 565 — rubru - eeiccaae 555, 556, 561-563 — — var. pap eaaaaa 562 Aceracea Ach a en 359 Acosta, 441, 442 Actinostrobus, 278, 280, 281, 321 Adadirachta excelsa, 471 Additions to the Flora of China, 119-127 Adesmia atacamensis, 13 Aechmea, 527 Aetheopappus, 440 Aglaia, 457 Ajuga, 343, 359, 364 AL-SHEHBAZ, IHSAN A. A Note on the Chil- ean Endemic Draba thlaspiformis (Bras- sicaceae), 385-387 L-SHEHBAZ, IHSAN A. A Revision of We- berbauera (Brassicaceae), 221-250 AL-SHEHBAZ, IHSAN A. Generic Limits and Taxonomy of Brayopsis and Eudema (Brassicaceae), 93-109 AL-SHEHBAZ, IHSAN A., and CLODOMIRO MARTICORENA. Menonvillea rollinsii (Brassicaceae), a New Shrubby Species from Chile, 135-138 Alnus, res ee 38, 39, 41, 43, 44, 55, 6 ae oe 7, 19, 21, 22, 25, 26, 38,4 — ne an 6, 21, 25, 26 — — sect. Alnus, 22 — — sect. Phyllothyrsus, 21, 24, 26 9, 21, 23-25 4 — subg. Gymnothyrsus, 22 — sect. Clethra, 22 — acuminata, 22, 27 — cordata, 27, 28 — subsp. tenuifolia, 22 — japonica _ jorullensis, 22, 27 586 JOURNAL OF THE ARNOLD ARBORETUM — — subsp. lutea, 24 — maritima, 23, 25, 2 — oblongifolia, 22 — rubra, 22, 27, 28 — rugosa, 23 See. 20, 22-24, 26, 27 , 24, 26 — — var. ah neasis. 507 Alpinia katsumadai, 513 Altingia, 114, 116 Altingiaceae, 113, 116 428 — argotaenia, 72, 74, 75 — formosana, 72 Amethystea, 323, 364 Ammophila, 156 Amomum krervanh, 490, 496 3 e, 8 Andrographis paniculata, 490, 499, 500, 520 Anisomeles, We 359 Anisopogon, 149 Apocynacea ae Aquilaria ‘agaloca, 490, 495, 513 Arabis —_ Se ania 241 — spathulata, 227 Arachnioides rhomboidea, 570 Araucariaceae, 74, 280, 281 s, 308 Arctium, 392, 393, 396, 397, 405, 431- 436 — sect. Arctium, 433 — sect. ecu 433 — sect. Glandulosa, 433 — lappa, 432-435 — minus, 432-434 [voL. 71 — x mixtum, 433 — nemorosum, 432, 434 4 Argythamnia candicans Aristida, 146, 147,149, 150, 155, 170-174 — sect. Chaetaria, 172 — sect. Pseudarthratherum, 172 — sect. Streptachne — adscensionis, 172 | io” Dp fae = =| No = tuberculosa 172 — wrightil Peter ae a (Gramineae) 1n the South- eastern United States, The Genera of, 145-177 Arundo, 146, 147, 149, 150, 155, 156 arenaria, | — bambos, | calamagrostis, 156 — donax, 149, 156, 157 — ee 15 — formosa 56 = phragmites. 156, 157 — plini fe eee ren Seeder 105 Asparagus cochinchinens Aspidistra daibuensis, 370. Aster taiwanensis, 570 Asteraceae; Compositae: The Genera of Cardueae in - Southeastern United States, 391-4 Athrotaxis, ae Athrotaxus, 75 Atractylis, 396 Atractylodes, 507 ATRAN, ScorTT, et al. Histoire du Concept d’Espéce dans les Sciences del la Vie [re- view], 139-142 1990] Austrocedrus, 276, 281 e, 70 a, Azadirachta, 459, 471 — indica, 458, 460, 471, 473 Balansia hypoxylon, 150 Bald cypress, 293 Bambusa, 15 Begonia austrotaiwanensis (Begoniaceae), a New Species from Southern Taiwan, 567-574 Begonia parviflora, Notes on a Novel Ab- axial Leaf Epidermis in Ecuadorian, 259- 264 ae 259-262, 567, 571, 573 a pear 567, 573 ie: Weilbachi 3 — aptera, 571 — austrotaivanensis, 567-574 — parviflora, 259-264 1 Begoniaceae: Begonia taiwanensis, a New Species from Southern Taiwan, 567-574 Behen, 441 Benincasa hispida, 490, 495 Berriochloa, 15 Betula, 3-8, 10, 55, 18, 21, 24-26, 32-51, — a Betula, 41 , 36 — sect. Humiles, 7, 33, 38, 41 — subsect. Nanae, 38 — alba, 33, 39, 4 — alleghaniensis, 35, 41-44 — caerulea, 37, 38 INDEX 587 — caerulea-grandis, 37, 38 — cordifolia — excelsa, 35 — fontinalis, 37, 42 — glandulifera, 38, 41 — glandulosa, 37, 38, 41 — japonica, 42 — lenta, 36. 41, 42, 44 a, 38 — nigra, 34, 35, 42, 45 — occidentalis, — papyrifera, ve Lg 41-45 — — var. cordifolia, 36 — pendula, 39, 40, 42-45 — populifolia, 36-38, 42, 43, 45 — — var. glandulifera, 38 — resinifera, 37 — rupestris, 38 — saxophila, 37 — uber, 36 — verrucosa, 39, 40 Betulaceae in the Southeastern United States, The Genera of, 1-6 Betulaceae, 1-67 — subfam. Betuloideae, 4, 5, 7, 8, 18, 55, 6 0 — — tribe Betuleae, 4, 6, 18 eee Coryloideae, 4, 5, 7, 8, 18, 51, 4,65 —— tribe Carpineae, 4, 6, 51 ribe Coryleae, 4, 6, 56, 61 Sea ster, 33 Bignoniaceae, 360-362 Biota orientalis, 304 3 Boc_-e, A. LINN. Multilacunar Nodal Anat- omy in Mytilaria (Hamamelidaceae), 111-118 Bombax ceiba, 490, 498 52, 253 Boswellia eae, 490, 495, 513 588 ease 339, 359 ORD, Davip E., ZHAN-HUO TsI, and eeu WANG. Additions to the Flora of China, 119-127 BOUFFORD, caer E., and TsuN-SHEN Yinc. A New Subspecies of Mallotus oreophilus Euphosiacsac) from South- Central China, 575-578 Brassica hirta, 490, 495 — rapa, 4 Pius icdosde: A Note on the Chilean En- demic Draba thlaspiformis, 385-387 Brassicaceae: A Revision of Weberbauera, 221-250 Brassicaceae: Generic Limits and Taxon- omy of Brayopsis and Eudema, 93-109 Brassicaceae: Menonvillea rollinsii, a New Shrubby Species from Chile, 135-138 eerie 93, 223, — tribe Sisymbrieae subtribe Pachycladi- nae, 221 Braya densiflora, 227 — imbricatifolia, 247 — pusilla, 241 Brayopsis and Eudema (Brassicaceae), Ge- neric Limits and Taxonomy of, 93-109 Brayopsis, 93-96, 99, 100, 103 — — subs — — subsp. ecuadoriana, 93, 97-99 — diapensioides, 93-95, 99, 108 — grandiflora — skottsbergeri, 106 — trichocarpa, 240 Brazoria, 323, 344, 345, 360, 363 Bromeliaceae: Patterns of Geographic Dis- tribution and Their Implications on the Phylogeny of Puya, 537-552 Bromeliaceae subfam. Pitcairnioideae, 527 Bromuniola, 148 Broussonetia papyrifera, 489 Bryophyllum pinnatum, 490, 499 Bucklandia, 113, 114, 116 Bumelia obovata, 131 Burdock, 4 Burretiodendron Sensu Lato (Tiliaceae), On the Genus, 371-380 Burretiodendron, 371-380 — brilletu, 371, 373, 376-378 — combretoides, 372 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 — esquirolu, 371-376, 378-380 — hsienmu, 372-378 — kydiifolium, 371- 376,378 — longistipitatum, 379, 380 — obconicum, 373, 376, 377 — siamense, 371, He 376, 378, 379 Pierccae. 489 Cabralea, 455 Caesalpinia eta 490, 495 Calamagros 6 Calcitrapa, a 443 — stellata, 443 Calderonella, 148 Calliandra portoricensis, 131 Callicarpa, 344 Callitris, 277, 278, 280, 281, 292, 320-322 — columellaris, 321 — — var. campestris, 320, 321 — — var. columellaris, 320, 321 — — var. intratropica, 321 — glauca, 320 — Seis. 320 — hugeli, 320 — intratropica, 320 a, 321 Galocednux "396, 278, 281 — decurrens, 313 Calycobolus, 251, 252 Cambodia and Vietnam, a New Spec m, Desmodium schubertiae (Leu. 4 m, 513 Canavalia ensiformis, 490, 498 — gladiata, 4 98 Cannabis, 515, 521 — sativa, 489, 490, 495, 504, 513-515 CANTINO, Puiuip D. The Phylogenetic Sig- nificance of Stomata and Trichomes in the Labiatae and Verbenaceac, 323-370 Caoba, Capsicum frutescens, 490, 498 Capuronianthus, 456 Carapa, 455, 459 — guianensis, 458 Carbenia, 450 1990] INDEX 589 Cardamine, 223 Cedrela, 455, 457, 459, 460, 480 — colchaguensis, 241, 243 Cedrelaceae, 454, 455 Cardiochlamys, 251, 252 Cedrelopsis, 459 Cardueae (Compositae; Asteraceae) in the sane 392-398, 405, 427, 431, 436- Southeastern United States, The Genera 51 1 — ae Acrolophus sect. Maculosae, 443 Carduncellus, 397 — subg. Amberboa, 44 Carduus, 392, 393, 396, 397, 405-407,422— — subg. Calcitrapa, 439, 440, 443, 450 , 450 — sect. Eucalcit — subg. Afrocarduus, 422, 423 — subg. Centaurea, 442 — — sect. Acaulon, 423 — subg. Centaurium, 440 — — sect Afrocarduus 423 — subg. Chartolepis, 439 —s innatisquama, 423 — subg. Cheirolophus, 440 — subg. Alfredia, 422, 423, 427 — — sect. Eucheirolophus, 440 — — sect. Alfredia, 423 — subg. Crocodilium, 44 — — sect. Apteron, 423 — subg. Crocodylium, 440 — — sect. Pterocaulon, 423 — subg. Crupina, 440 — subg. Carduus, 422, — subg. Cyanus, 439, 440 — — sect. Carduus, 423 — — sect. Eucyanus, 440 — — sect. Leptocephali, 423 — — sect. Pannophyllum, 440 — acanthoides, 423, 424 — subg. Hymenocentron, 439 — crispus, 423 — ns Jacea, 439, 440, 442 — getulus, 424 — — sect. Eujacea, 440, 442 — heterophyllus, 406 — — sect. Nigrescentes, 442 — lanceolatus, 414 — subg. Lopholoma, 439 — marianus, 426 — subg. Microlophus, 440 — nutans, 422-424 — subg. Odontolophus, 440, 442 — x orthocephalus, 424 — subg. Phrygia, 440 — pycnocephalus, 423, 424 — subg. Plectocephalus, 440 — tenuiflorus, 423 — subg. Seridea, 440 Carex baccans, 571 — subg. Solstitiaria, 443 Carica papaya, 490, 499 — subg. Stoebe, 440 Carlina, 396 — subg. Verutum, 440 Carpinaceae, 3, 4 — sect. Acrolophus, 3 Carpinum, 54 — sect. Calcitrapa, 439, 440, 443, 450 Carpinus, 4-8, 10, 18, 51-60, 64, 65 — sect. Centaurea, 439, 440 — sect. Distegocarpus, 54, 55 — sect. Crocodylium, 439, 440 — betulus, 53, 54, 56 — sect. Crupina, 440 — caroliniana, 53-56, 59 — sect. Cyanus, 439, 440, 442 — — subsp. caroliniana, 54 — sect. Jacea, 442, 443 — subsp. virginiana, 54 — sect. Lepteranthus, 442 — ostrya, 5 — sect. Mesocentron, 443 — tropicalis, 54 — sect. Plectocephalus, 442 — virginiana, 58 — sect. Rhaponticum, 439 Carthamus, 396, 397, 427 — sect. Seridia, 439, 440 — tinctorius, 490, 496, 513 — americana, 437, 438, 442, 444 Carya, 518 — calcitrapa, 437, 438, 442-444 Caryopteris, 340 — centaurium, 437 — sect. Pseudocaryopteris, 364, 365 — cyanus, 438, 442, 444 Casselia, 358 — dubia, 437, 438, 442, 443 Cassia acutifolia, 490, 499, 500 _ uate 437, 442-444 — angustifolia, 490, 499, 500 a x Centaurea ee 438 _ eae 490, 498 — mace stale. Catadysia, 93 — maculosa, 438, “ Ca a aihis roseus, 490, 499 — melitensis, 438 590 — nigra, 437, 442-444 — nigrescens, 438, 443 Cephalotaxaceae, 70, oF 73-75 Cephalotaxus, 70-73, 75, 83 ati ies 276, a 281; 282, 292, 299-303 — formosensis, 299 | = 35 - ees 299, — pisifera, 280, oon 301 — thyoides, 299-302 — var. henryae, 301 Cainaesycs articulata, 131 CHANG, CHIN-SunG. A Reconsideration of the Acer palmatum Complex in China, Taiwan, and Korea, 553-565 Chasmanthium, 147-149, 155, 176 — gracile, 176 eee ee Pa 177 = 76, —-—*x Ce ornithorhynchum, 177 — nitidum, 177 — omnithorhynehum, 177 — sessiliflor CHEN, tae nae and CHING-I PENG. Begonia austrotaiwanensis (Begoni- ew Species from Southern 74 Chevalierella, 148 Chile, a New Shrubby Species from, Men- onvillea rollinsii (Brassicaceae), 135- 138 ophilus (Euphorbiaceae) from South- Central, 575-578 China, Additions to the Flora of, 119-127 China, Notes on the Species of Selaginella from Guizhou, 265-270 China, Taiwan, and Korea, A Reconsid- JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 eration of the Acer palmatum Complex n, 553-565 Chinaberry, 470 Chinese Medicine, History of the Intro- duction of Exotic Elements into Tradi- tional, 487-526 Chionochloa, 149, 150 Chisocheton, 456, 458 Chunia, 111, 114, 116, 117 Cinchona ledgeriana, 490, 499 — succirubra, 490, 4 San ene loureiroi, 513 Cipadessa, 47 Cirsium, 392, a 396, 397, 405-422, 427, 431, 442, 450 — subg. Cirsium, 409 . Erythrolaena, 409 — — sect. Onotrophe, 409 — sect. Breea, 409 _ ae Chamaeleon, 409 — sect. Corynotrichum, 407 — sect. Dubia, 407 — sect. Epitrachys, 407, 409 — sect. Lamyra, 40 , 410 — — subsect. Acanthophylla, 410, 411 — — — ser. Altissima, 4 — — — ser. ae 4i2 — — — ser. Virginiana, 412 — — subsect. eon 410, 411 — — subsect. Odorata, — sect. Orthocentron, 407 — sect. Ptilostemon, 409 1990] — sect. Solitaria, 407 — acarna, 4 — altissimum, 407, 411, 412 — arvense, 407, 409, 412-414 — brevistylum, 41 — carolinianum, 411, 412 — discolor, 407, 411, 412 — edule, 414 — engelmannii, 411, 412 — nelenioides 406, 409 — heterophyllum, 406, 409 — horridulum, 407, 409, 410 — lanceolatum, 414 — lecontei, 409, 410 — muticum, 407, 411, 412 Cirsium discolor, 41 1 — nuttallii, 410, 411 | = oO an 2 Qa fo) <= 3 a oa — palustre, 412 — smallii, — terrae-nigrae, 411, 412 texanum, 414 undulatum, 414 — virginianum, 407, 411, 412 — vittatum 0 — vulgare, 405, 409, 413, 414 Citharexylum, 344, 356, 359 Citrullus lanatus, 490, 498 Citrus reticulata, 513 Cladostigma, 255 CLaytTon, W. D., and Genera Graminum [review], 271-273 Cleome, — gynandra, 491, 498 Cleonia Clerodendrum, 34 9 — subg. anc: 339, 364 Clusia rosea, 131 Cneoraceae, 459 Cnicus, 392, 393, 397, 405-407, 427, 431, 442, 449-45] — benedictus, 450 Coccoloba microstachya, 131 Colebrookea Colocasia formoeand 571 3 Colymbada, 441 Comanthosphace, 324, 357, 359 Commiphora myrrha, 491, 496 Common reed, 158 Compositae, 393, 489, 507 S. A. RENVOIZE. INDEX 591 — subfam. Asteroideae, 393, 395 — subfam. Cichorioideae, 394, 395 — subfam. Cynaroideae, 393 — subfam. Lactucoideae, 395 — tribe Anthemideae, 394 — tribe Arctoteae, 304, 395, 439 — tribe Carlineae, 392, 393, 438, 442 — tribe Centaureeae, 392, 393, 438, ass, 441 — tribe Cynareae, 392, 393 — tribe Echinopseae, 392, 393, 395, 442 — tribe Echinopsideae, 39 — tribe Eremothamneae, 395 — tribe Eupatorieae, 394, 395 — tribe Lactuceae, 394, 395 — tribe Liabeae, 395 _— tribe Mutisieae, 394, 396, 431, 442 — subtribe Cardopateae, 393 — subtribe Carduinae, 302- 397, 406, 407, 2 — subtribe Carlininae, 397 — subtribe Centaureinae, 392, 494-497, 436, 441, 442, 444, 450 — subtribe Echinopsinae, 395, 397 — subtribe Gochnatinae, 394 enera of Cardueae in the Southeastern United States, 391-451 Convolvulaceae: Preliminary Taxonomic Consideration of the Poraneae, 251-258 Convolvulaceae, 251, 254, 255 — tribe Cresseae, 2 1, 2 2 — tribe Hildebrandtia, 251, 255 tribe Poraneae, 251-258 Gariomes: 372 — crenata, 374 Corchorus capsularis, 515 Cordisepalum, 251, 252 Coriandrum sativum, 491, 497 Wayside Trees of Malaya 4 — selloana, 147 ou2 Corylaceae, 3, 4, 54 Corylus, 3. 10, 18, 44, 54-56, 60-67 — sect. Acanthochlamnys, 62, 65 — sect. Avellana, 62 — sect. Tubo-Avellana, 62 ubulosa, 62 Cotton hale 430 Cousinia, 430 Craigia, 37 72 — yunnanensis, 372, 374 Craniotome, 344, 359 Cremolobus, 138 Cristaria, 138 rocus sativus, 491, 498 Crossopetalum rhacoma, 131 Cruciferae, 93, Cryptantha ig ae 138 — parviflor 8 Cryptomeria, at 281, 295 0 CULLEN, S. P., P. F. Stevens. Linnae- us, the Cortex- "Medulla Theory, and the Key to His Understanding of Plant Form and Natural Relationships, 179-220 Cunninghamia, 8 Cupressaceae (Including Taxodiaceae) in the Southeastern United States, The Genera of, 275-322 Cupressaceae, 73, 74, 275-322 bfam. Callitroideae, 278 — — tribe Cupresseae, 278 — — tribe Junipereae, 278 tribe Thujopsideae, 278 Cupressinae, 275 x Cupressocyparis leylandii, 301 Cupressus, 276-278, 281, 299, 301, 312 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 71 Cupuliferae, 4 Cyanostegia, 363 Cyanotis kawakamiu, 571 Cynara, 396, 397, 442 _ es 397, 398 — scolymus, 398 Cyperus aca 513 Cypress Cypress pine, 320 Dactylocardamum, 94 Dactylostigma, 251, 252, 254, 255 — linearifolia, 254, 255 Daemonorops draco, 491, 495 Dalbergia erat 491, 497 — sissoo, 491, Danthonia, ee 149, 150, 155, 164-170 | Q $8 oO nn se) S .o) wn g. an’ — californica, 166 — — var. pinetor — tenuior, | — unispicata, | Datura metel, 491, 498 Deinostema, 126 = gaenoeaulon. 119, 125, 126 — violacea, 126 Desmodium schubertiae (Leguminosae), a New Species from Cambodia and Viet- nam, 381-384 Desmodium, 381 — subg. Sagotia sect. Nicolsonia, 383, 384 — rubrum 381, 383, 384 a es 381-384 1990] — strigillosum, 384 — toppinii, 384 Draba thlaspiformis (Brassicaceae), A Note on the Chilean Endemic, 385-387 — philippii, 385 = thlaspiformis, 385-387 Dregeochloa, | Dryobalanops aromatica, 491, 495, 513 Dryopteridaceae, Dryopteris kweichowicola, 119-121 — wallichiana, 119, 12 Duranta, 341, 344, 345, 361, 363 — fraseranum, 459 Echenais, 427 Echinops, 392, 396, 398, 405, 442 — bannaticus, 398 — exaltatus, 398 — ritro, 398 So a abe 398 Ecuadorian Begonia parviflora, Notes ona Novel Leaf Epidermis in, 259-264 INDEX 593 Elatostema sessile var. cuspidatum, 571 Eleocharis congesta, 126 Elytrophorus, 149 ris, 93 Entandrophragma, 459, 460 Ephelis borealis, 150 Equisetum, 73 Ericaceae: A New Variety of Lyonia from Puerto Rico, 129-133 Ericaceae, 548 — subfam. Vaccinioideae tribe Andro- medeae, 129 Eriocaulon sieboldianum, 126 Eriolaena esquirolii, 379, 380 simum pusillum, 241, 243 Eudema (Brassicaceae), Generic Limits and Taxonomy of Brayopsis and, 93-109 Eudema, 93-96, Le 101, 103, 108, 385 — microphylla, 108 — monantha, 108 — nubigena, 95, 100, 101, 103-106 — — subsp. nubigena, 96, 104-106 — — subsp. remyana, 93, 104-106 — pycnophylloides, 108 — remyana, 99, 105 — rupestris, 95, 96, 100-103 — thlaspiforme, 108, 385 — trichocarpa — trichocarp m, 240 — werdermannu, 100, 101, 106-108 Eugenia sessiliflora, 131 — woodburyana, 131 Euphorbia eae 491, 499 — lathyris, 4 96 Euphorbiaceae: ‘A New S of Mal- lotus oreophilus from South- Central China, 575-578 Euphorbiaceae, 489, 575 594 a 324, 343 Eustigma, | sae bonplandii 104 — humboldti 3 Bonontiee aa — populnea, |1 Exeentrodendron, 371, 372, 375, 376 6 113-117 — chombifolinin: - 377 — tonkinense, 371 Fagaceae, 3, 6, 8-10 Fagus grandifolia, 260 False cypress, 299 Faradaya, 345 Ferula assafoetida, 491, 495, 507, 509 IC 10 Fitzroya cupressoides, 280 Flacourtiaceae, 427, 489 Flindersia, 459 Flora of China, Additions to the, 119-127 Foeniculum vulgare, 491, 495, 513 aii hodginsii, 279 Fomes, 44 — ae 27, 44 had 113,114 Frankia Prins ornus, 470 Frenela, — hugelii, 320 J. The Genera of Betula- ceae in the Southeastern United States, Galactites, 397, 427 Galeobdolon, 359 Garcinia hanburyi, 491, 496 — morella, 491, 496, 510, 512 Gastrochilus nanus, 119, 121-123 Genera Graminum [review], 271-273 Genera of Arundinoideae (Gramineae) in the Southeastern United States, The, -177 Genera of Betulaceae in the Southeastern United States, The, 1-67 Genera of Cardueae (Compositae; Aster- aceae) in the Southeastern United States, The, 391-451 Genera of Cupressaceae (Including Taxo- JOURNAL OF THE ARNOLD ARBORETUM [VOL. 71 diaceae) in the Southeastern United States, The, 275-322 Genera of Meliaceae in the Southeastern United States, The, 453-486 Genera of Taxaceae in the Southeastern United States, The, 69-91 ete ae and Taxonomy of Bray- opsis (Brassicaceae), 93-109 Genus es Sensu Lato (Tili- aceae), On the, 371-380 Geographic Distribution, Patterns of, and Their Implications on the Phylogeny of Puya (Bromeliaceae), 527-55 Gesneriaceae, 360-362 Giant reed, 156 Ginkgo, 72, 74 Glycyrrhiza uralensis, 513 Glyptostrobus, 279, 294, 295 Gmelina, — moluccana, 343 Goprrey, Ropert K. Trees, Shrubs, and Woody Vines of Northern Florida and Adjacent Georgia and Alabama [re- view], 389, 390 Gomphostemma, 357, 359 Gomphrena aia 491, 499 Goniocaulon Gossypium nen 491 — hirsutum, 491, 498 Gouldochloa, 147, 148, Gramineae: The Genera ef aneinedee in the Southeastern United States, 145- 177 Gramineae subfam. Arundinoideae, 145, 145-177 — — tribe Aristideae, 146, 147, 170, 171 — — tribe Arundineae, 146, 156, 157 — — tribe Centotheceae, 146, 147, 171, 176 7 — — tribe Micraireae, 14 — — tribe Stipeae, 146, a. 171, 174, 175 — — tribe Thysanolaeneae, 147 — subfam. Bambusoideae, 148 — subfam. Centothecoideae, 146, 148 — subfam. Chloridoideae, 147, 176 48 Grass Genera of the World [review], 271- Gratiela, 126 Guarea, 455, 457-460 — glabra, 457 —_ eurdonia: 458 1990] — rhopalocarpa, 457, 458 Guetterda scabra, 131 Guizhou, China, Notes on the Species of Selaginella from, 265-270 Guttiferae, 489 ymnosporangium, 313 Gynerium, — sagittatum, 149 Hakonechloa, 149 Hamamelidaceae: Multilacunar Nodal Anatomy in Mytilaria, 111-118 Hamamelidaceae, 10, 64, 111, 113, 114 116 — subfam. Chunioideae, | 17 — subfam. Disanthoideae, 116, 117 — pte Exbucklandioideae, 111, 113, 116, _ a Hamamelidoideae, 117 — subfam. (iguidambaroidese, 116, 117 — subfam. Mytilarioideae, 11 — subfam. Rhodoleioideae, 117 i 16 Hamamelis, 113, 116 — japonica, 116 — virginiana, 116 Harlanlewisia, 324 HArT, JEFFREY A., and Rospert A. PRICE. The Genera of Cupressaceae (Including Taxodiaceae) in the Southeastern Unit- ed ea 275-322 Hazel, 62 Hemiandra, 345 Hesperis colchaguensis, 241 — imbricatifolia — nubigena 104 — orophila, 227 — planchoniana, 103 i l , 131 Hildebrandtia, 251, 385, 257 — africana, 257 — austinil, 255-257 — linearifolia, 255, 257 Histoire du Concept d’Espece dans les Sci- ences de la Vie [review], 139-142 History of the Introduction of Exotic Ele- INDEX 595 ments into Chinese Traditional Medi- cine, 487- Holmskioldia, 345, 362 — sanguinea, Hoover, W. Scott. Notes ona Novel Leaf Epidermis in Ecuadorian Begonia par- viflora, 259-264 Hop hornbeam, 58 Hordium vulgare, 491 Horminum, 358 Hornbeam Hu, Suu Yinc. History of the Introduc- tion of Exotic Elements into Traditional Chinese Medicine, 487-526 Hydnocarpus, 427 — anthelminthicus, 491, 498 — kurzi, 491, 498 Impatiens balsamina, 491, 498 Ipomoea nil, , 54, 58 Isatis nciorie. 491, 497 Jacea, 44 1, 442 — — var. grandiflorum, 517 — sambac, 516, 517, 521 sequinil, 517 sinense, 517 Jupp, WALTER S. A New Variety of Lyonia (Ericaceae) from Puerto Rico, 129-133 Juglandaceae, 8-10, 518 Juglans, 517-519 — cv. James River Hybrid, 518 — cv. Paradox, 518 — cv. Royal, 518 — cathayensis, 519 — hindsii x Juglans regia, 518 — x intermedia, — nigra, 518 — — x Juglans regia, 518 — — 514, 517-519, 521 Juni ee 276 596 se ea 275-278, 280, 281, 292, 307- 19 — me Caryocedrus, 308 _ salsa. 307, 308 — horizontalis, 309, 311-313 — mexicana, — nana, 308 _ ee. 276, 308, 313 — saltillensis, 312 — scopulorum, 309, 311, 312 8 Juriniea. 397, 407.420, 431 Khaya, 460, 479 Kinostemon, 343, 364 Knapweed, 437 Kochia scoparia, 491, 495 Korea, China, and Taiwan n, A Reconsid- eration of the Acer palmatum Complex Labiatae and Verbenaceae, The Phyloge- netic Significance of Stomata and Tn- chomes in the, 323-370 Labiatae, 323-370, 489, 507 — subfam. Lamioideae, 324, 325, 340, 343, 344, 357, 359 — tribe Ajugeae, 323, 324, 339, 340, 344, 345, 356, 362, 365 — — tribe Lamieae, 324, 339, 344, 359, 364 — — — subtribe Melittidinae, 323, 359 — — tribe Pogostemoneae, 324, 339, 344, 364 — — tribe Prostanthereae, 323, 324, 339, 340, 343, 345, 356, 362-365 — — tribe Scutellarieae, 324, 339, 356, 364 — subfam. Nepetoideae, 324, 325, 339, 340, 343-345, 357-359, 361, 364 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 — — tribe Mentheae, 343, 344, 358, 362 — — tribe Nepeteae, — — tribe Ocimeae, 343, 344, 358 — — — subtribe Hyptidinae, 358 — — tribe Salvieae, 358 Lablab purpureus, 491, 495, 506 148 pine ines cs 149 341 Lawsonia inermis, 492, 495 Leaf Epidermis in Ecuadorian Begonia parviflora, Notes on a Novel, 259-264 Leguminosae: Desmodium schubertiae, a New Species from Cambodia and Viet- nam, 381-384 Leguminosae, 489 — subfam. Papilionoideae, 381 Lentibulariaceae, 360-362 Lepidothamnus, 281 Lesquerella, 385 — thlaspiformis, 385 Men 443 Leacosceptrum, 324, 357 Leuzia, 397 ere 276, 278 a eee vulgare, 260 s, the Cortex-Medulla Theory, and the Key to His Understanding of Plant Form and Natural Relationships, 179- 220 Linnaeus, 179-220 Linum usitatissimum, 492, 495 me mera var. helleri, 131 Lophatherum, 148 Lovoa, 459 Luehea, 372 — speciosa, 374 Luffa aegyptica, 492, 498 Lycopodium repandum, 268 Lyonia (Ericaceae) from Puerto Rico, A New Variety of, 129-13 3 — heptamera, 129, 13 1990] — lippoldii, 129 — microcarpa, 129, 131 — obtusa, 129 — — var. longipes, 129 — rubiginosa, 132 — — var. stahlii, 129, 132 — truncata, 129, 131, 132 — — var. montecastina. 129, 132 — — var. proctoril, 129-132 — — var. truncata, 129, 132 Macbridea, 323, 345, 360, 363 Machaonia portoricensis, 131 3 Mallotus ei (Euphorbiaceae) from South-Central China, A New Subspecies Mallotus oreophilus subsp. latifolius, 575— 578 sp. oreophilus, 575, 577 Malvaceae, 489 Mansonia gagei, 380 Mariana, 427 MARTICORENA avast and IHSAN A. AL-SHEHB Menonvillea rollinsii pene a New Shrubby Species from Chile, 135-138 Matudaea, 116 Medicine, Traditional Chinese, History of the Introduction of Exotic Elements into, 26 Megastachya, 148 e, 548 Melia, 454, 455, "457, 459, 469-476 — australasica, 471 — azedarach, 455, 458, 470-473 Umbraculifera, 471 Meliaceae in the Southeastern United States, The Genera of, 453-486 Meliaceae, 453-486 — subfam. Capuronianthoideae, 456 — subfam. Melioideae, 455-459, 469 — — tribe Melieae, 455, 459 — subfam. Quivisianthoideae, 456 INDEX 597 subfam. Swietenioideae, 455, 458, 459, 476 479, 480 — tribe Swietenieae, A Nigloeachis intortus, | Menonvillea vollinsi ees a New Shrubby Species from Chile, 135-138 Menonvillea, 135, 1 — linearis, 135 — nordenskjoeldii, 135 — pinnatifida, 135 — haplocalyx, 513 Meriandra, 358 Metaporana, 251-253 ss ee sepa ala, 252, 253 Metasequoia, 276, 277, 279, 281, 294, 295 Microcachrys, 277 ea or minima, 126 Milium Milk- ae 426 MiILLer, Norton G. The Genera of Me- liaceae in the Southeastern United States, Molinia, 149 Moluccella, 359 oraceae, Multilacunar Nodal Anatomy in Mytilaria (Hamamelidaceae), 111-118 , 10 Myristica fragrans, 492, 496 ie as (Hamamelidaceae), Multilacu- r Nodal An eee. in, 111-118 iitilaria 111-118 — laosensis, - 112, 114 Nassella, 146, 148, 149, 155, 174, 175 — leucotricha, 150, 174 — pungens, 174 Nectria galligena, 44 598 Needlegrass, 175 seers, 281, 321 ium oleander, 492, 499 Rcyeaee 323, 362, 363 — dupont, 343 New Subspecies of Mallotus oreophilus (Euphorbiaceae) from South-Central China, A, 575-578 New Variety of Se on from Puerto Rico, A, 129- Neyraudia, 146, 147, a 155, 164 — arundinacea, 164 — madagascariensis, 150, 164 — reynaudiana, Nicotiana tabacum, 492, 498 Nodal Anatomy in Mytilaria (Hamameli- daceae), Multilacunar, 111-118 Note on the Chilean Endemic Draba thlas- piformis (Brassicaceae), A, 385-387 Notes on a Novel Abaxial Leaf Epidermis in Ecuadorian Begonia parviflora, 259- 264 Notes on the 2 ee of ee from Guizhou, China, 265-270 ac, 51 Nyssa aquatica, 294 — sylvatica, 294 Oat grass, 165 Ocimum basilicum, 492, 497 Octoclinis, 320 O , Hrroyosuit. Desmodium § schu- New Species from Cambodia and Viewann., 381-384 Oleaceae, 8 Onopordum 392, 393, 396, 397, 405, 427, — be: er 430 — subg. Onopordum, 430 — — sect. Echinata, 430 — — sect. Erecta, 430 | | n Q QO = — leptolepis, 431 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 — sibthorpianum, 431 — tauricum, 4 Oplismenus undulatifolius, 571 Opuntia dillenii, 492, 499 — repens, 13 Orchidaceae, 119 — asperifolia, 148 — canadensis, 148 — hymenoides, 148 Ouratea littoralis, 131 Oxera, 340 Seal 56 Palmae, 489 Panax ‘quinquefotius, 492, 499 fonlanenas. res 377 Paro, Ae 16 — persica, | : ee 116 — jacquemontiana, 116 Pastinaca, 2 Patterns of Geographic Distribution and Their Implications on the Phylogeny of ya (Bromeliaceae), 527-552 Pedaliaceae, 360-362 YUNG- ee CHEN. Begonia austrotaiwanen s (Begoni- w Species oe Southern Pentameris, 149 1990] Peperomia japonica, 571 Peronema, 35 Perovskia, 357, 358, 363 Petitia, See ns 340, 364 Petrea, 359 Paco ek 440 Philippiamra Psieiaia: 138 — pachyphylla, 138 Phlomis, 343, 357, 358, 363 Phragmites, 146, 147, Pe ae 155-163 — australis, 149, 150, 157- 8 — s, Phyla, 323, 339-341, 357, 359, 361, 363 Phyllocladus, 71 Phyllostegia, 359 Phylogenetic Significance of Stomata and paca in the Labiatae and Verbena- eae, The, 323-370 oe 323, 360, 363 Picnomon, Pilea kankacensis, $7] — trinervia, 57 Pilgerodendron, 276 Pinaceae, 73, 74 Pinellia, 507 Pinus, 54 — clausa, 320 Piper longum, 492, 496, 505, 513 — nigrum, 492, 4 5 Piperaceae, aH Piptatherum, 148 Piptochaetium, 148, 155, 175, 176 5 — paniculata, 542 Pena 360, 362 Platycarya, 51 Platycladus, 281, 292, 304 — orientalis, 292, 304, 305 Plectocephalus, 442 Podocarpaceae, 71, 72, 74, 277, 280, 281 Pogostemon, 324, 343 — cablin, 492, 497 INDEX Pohlidium, 148 Pollia secundiflora, 571 Polygala penaea, 131 Populus diversifolia, 492 — euphratica, 492, 496 3 — densiflora, 253, 254 — obtusa, 252, 253 Poraneae (Convolvulaceae), Preliminary Taxonomic Consideration of the, 251- 258 Poria, 44 — cocos, 513 Portulaca oleracea, 492 Preliminary Taxonomic Consideration of the Poraneae (Convolvulaceae), 251-258 Premna, 339, 359, 364 — octonervia, 343 PRICE, RosertT A. The Genera of Taxaceae and JEFFREY A. Hart. The Genera of Cupressaceae (Including Taxodiaceae) in the Southeastern Unit- ed States, 275-322 iva, 358 Prostanthera, 340, 344, 356 Prunella, 359 Prunus, 262 — armeniaca, 504 — dulcis, 492, 496 Psephellus, 440 Pseudotaxus, 69-75, 88 — chienii, 74 Psidium guajava, 492, 499 Psora, 441 Psoralea corylifolia, 492, 496 Ptaeroxylaceae, 455, 459 Ptaeroxylon, 459 Pterocarya, 518 Ptilostemon, 397 Puccinia invenusta, 150 o, A New Variety of Lyonia (Er- een fou 129-133 Punica granatum, 492, 495 Puya (Bromeliaceae), Patterns of Geo- graphic Distribution and Their Impli- cations on the Phylogeny of, 527-552 Puya, 527-552 — subg. Puya, 534, 535, 540, 542, 546 — aequatorialis var. aequatorialis, 530 600 — — var. albiflora, 530 — alpestris, 539, 540, 542, 543 — angulonis, 544 = aristexiietad, 530, 539, 541, 544 — assurgens, 530, 547 — atra, 530 — berteroniana, 530, 537, 540-543 — bicolor, 537, 539, 541, 544, 545 boliviensis, 540-543 rdonae, 539 — castellanosii, 530, 537, 538, 541-543, 546 — chilensis, 528, 530, 537, 539-543 — clava-herculis, 530, 533, 534, 536, 539 — cleefii, 539 _ anes a, 541-543 co — dykioides, 536, 547 — — var. sme 530 — — var are, 530 — ce. 540, 541 — ferruginea, 530, 533, 536, 538, 539 — floccosa, 530, 537-539, 541 — - hattiata, 531. 536, 539 — mollis, 531 — nana, 531, 540, 541 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 — nitida, 539, 541 — nivalis, 541 — nutans, 531, 534, 537, 539, 541 — paupera, 54 0, — pearcei, 531, = 538, 541 — phelpsiae, 545 — pygmaea, 531, 533, 534, 537, 539 — raimondii, 528, 531,533, 538, 541-543, 546 — reducta, 540, 541 rectrorsa, ae 533, 536 — riparia, — sanctae-crucis var. sanctae-crucis, 531 — — var. verdens 32 — santanderensis, ai — santosii, 539, 544, 545 2,5 _ eadibann 532, 537, 539 — solomonil, 532 — spathacea, 532, 533, 536, 538, 541 — stenothyrsa, 532, 533 — trianae, 528, 532, 539 — tristis, 532 — tuberosa, 532, 534, 540, 541, 549 — tunarensis, 532, 538 — ultima, 532 — ushae, — venezuelana, 541 — venusta, 532, 537, 539 , 41 _ weddeliana, 541-543, 546 — wurdacku, 5 — yakespala, 532, 538, 541 — zakiana, 532 Pyrrhanthera, 149 Quercus coccinea, 260 — muehlenbergii, 260 — prinus, 260 — rubra, 26 — sessiliflora, 260 Quisqualis ingica, 492, 496 Quivisianthus, 456 Raphanus sativus, 492, 495, 506 Rapona, 251, 252 ee of the Acer palmatum n China, Taiwan, and Korea, 55 08 Revision of Weberbauera (Brassicaceae), , 221-250 Reynaudia, 164 1990] Rhaponticum, 397, 442 Rheedia mariquitensis, 131 Rheum officinale, 5 Rosmarinus omieinaligs 493, 496 Rostrinucula, 324, 357, 359 Rotala mexicana, 126 Ruagea, 455 Rubiaceae 489, 548 Rubiteucris, 364 Rubus fanjingshanensis, 119, 123, 124 — treutleri, 123 Ruta graveolens, 493, 495 Rutaceae, 459, 460 Rytidosperma, 149, 165 Sabia swinhoei var. parvifolia, 119, 126 Sabiaceae, 119 Sabina, 308 — salicicola, 309 Sagmen, 441 Salazaria, 324 Salvia, 3 Sandoricum koetjape, 460 Santalum album, 493, 495, 513 Sapindaceae tribe Aitonieae, 454 Sarcodraba, 237 — herzogii, 236 Sartidia, 147 ee 396, 397, — lappa, 493, 495, on 501, 513 Bete scaphigerum, 493, 496 6 Schoutenia, 371, 372, 375 — ovata 4 Sciadopityaceae, 73, 277, 278 Sciadopitys, 277-280 Scotch thistle, oe Scott, RANDA . The Genera of Car- dueae (ence. Asteraceae) in the Southeastern United States, 391-451 rophulariaceae, 119, 345, 360-362 Scutellaria, 324, 344 INDEX 601 Selaginella from Guizhou, China, Notes on the Species of, eas Selaginella, 265-27 _ albociliata, 267, ve , 268 — chaetoloma, 67: 269, 270 — compta, 270 — davidii, 269 — delicatula, 266, 268, 571 — doederleinii, 266 — drepanophylla, 267, 270 — effusa, — elephantopus, 268 — flagellifera, 265, 268 — gebaueriana, 267, 269 — helferi, 265 _ heterostachys, 266, 268 , 268 _ kouycheensis, 265, 267, 268 — labordei — TbDERS is. 267 _ ieee 265, 268 6 = omeiensis, 267, 269 a, 26 6 — prostrata, 267, 269, 270 | so) = ue =) © = m — repanda, 267-269 trachyphylla, 266 — uncinata, 267 — xipholepis, 267, 270 2 , 395 Sequoia, 276, 279-282, 295 — sempervirens, 276, 280 Sequoiadendron, 276, 279-281, 295 Serratula, 397 Sesamum orientale, 493, 495, 504, 506 Setaria palmifolia, 57 Sicrea, 371, 372, 375 Sida jamaicensis, 131 Sideritis, 343, 359 Sieglingia, 149, 165 Silk-reed, 164 Silybum, 392, 393, 396, 397, 405, 407, 426- 429, 442 — eburneum, 427 — marianum, 426-428 602 Simaroubaceae, 460 , 227 spathulaefolium, 225, 227 suffruticosum, 247 Se ctcbere aathan, 100 Skottsbergiella, 100 Solanaceae, 489 Solanum, 262 — sitiens, 138 Southeastern United States, The Genera of Arundinoideae (Gramineae) in the, 145- 177 Southeastern United States, The Genera of Betulaceae in the, 1-6 Southeastern United States: The Genera of Cardueae (Compositae; Asteraceae) in the, 391-451 Southeastern United States, The Genera of Cupressaceae (Including Taxodiaceae) in the, 275-322 Southeastern United States, The Genera of Meliaceae in the, 453-486 Southeastern United States, The Genera of Taxaceae in the, 69-91 Sphenodesme, 340 Stachytarpheta, 339, 345, 358, 359, 363 STAPLES, GEORGE W. Preliminary Taxo- nomic Consideration of the Poraneae Si aa 251-258 Star-thistle, 4 Stenodraba, o 223, 225, 243, 248, 249 — sect. Elatia, 22 — sect. Stenodraba, 223 — andina, 241-243 — — var. hirticaulis, 241, 243 — — var. patagonica, 241, 243 — var. stylosa, 241 — chillanensis, 225, 244 — f. laxa, 244 — colchaguensis, 241-243 — imbricatifolia, 247 — — var. glabrata, 247 — parvifolia, 248 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 — patagonica, 241, 243 - sane 241, 243 r. patagonica, 241 - tee — — var. leyboldii, 245 — suffruticosa, 246, 247 pee 372, 380 ns, P. F., and S. P. CULLEN. Linnae- us, the Cortex-Medulla Theory, and the Key to His Understanding of Plant Form and Natural Relationships, 179-220 Stomata and Trichomes in the Labiatae and eee The Phylogenetic Sig- nificance of, 3 70 Streptanthus, na Strychnos nux-vomica, 493, 498 Stylosanthes viscosa, 13 Swamp cypress, 293 Swietenia, 454-457, 460, 469, 476-486 — humilis, 477, 479-481 — macrophylla, 477, 479-482 — mahagoni, 458, 477-482 Symphorema, 340 Syzygium a a 493, 496, 513 Taiwan, Begonia austere: Neal goniaceae), a New So 567-574 Taiwan, Korea, and China, A Reconsid- eration of the Acer palmatum Complex in, 553-565 Taiwania, 279-281 Tamarindus indicus, 493, 498 27 Taxaceae in the Southeastern United States, The Genera of, 69-91 Taxaceae, 69-91 — tribe Taxeae, 71, 87, 88 — tribe Torreyaceae, 71, 83, 84 Taxineae, 70 Taxodiaceae, 73, 74, 276-278, 292 Taxodium, 276, 279, 292-299 1990] ascendens, 293, 294 distichum, 293, 296 — var. distichum, 294, 296 Brayopsis and Eudema (Brassicaceae), Generic Limits and, 93- 9 Taxus, 69-75, 82, 87-91 — baccata, 74, 87-89 — subsp. canadensis, 87 — — var. canadensis, 87 — — var. min — brevifolia, 87, 89 — x hunnewelliana, 89 — x media, 89 — subfam. Excentrodendroideae, 372 — tribe Enteleeae, 37 Tectona, 35 Terminalia chebula, 493, 495, 513 Tetraclea, 323, 340, 362, 364, 365 Tetramicra canaliculata, 131 aa 43 4 Thistle, Three-awn grass, 172 Thrinax morrisii, 131 Thuja, 276, 280, 281, 292, 303-307 — koraiensis, 304 — occidentalis. 304, 305 , 30 Thujopsis, 280, 281, 304, 305 Thuspeinanta, 3 Thysanolaena, 146 — maxima, 147 Tilia tuan, 374 Tiliaceae: On the Genus Burretiodendron 0 Tinnea, 323, 344, 364 — aethiopica, 363 — apiculata, 337 Toona, 456 INDEX 603 Bee 10-75, 82-88 = texifolia, 74, 83- 85, 88 Torreyaceae, Trees, Shrubs, and Woody Vines of North- ern Florida and Adjacent Georgia and eee [review], 389, 390 Triboliu Trichilia, 455, 459, 460 — havanensis, 457 — poeppigii, 457 aay carat crinitis, 116 — elliptica, 116 Trichocting caulescens, 138 Trichomes in the Labiatae and Verben ceae, The Phylogenetic Significance a Stomata and, 323-370 chiana, 89 Trichostema, 303, 340, 362, 364, 365, 340 — lanceolatum, 343 Tricyrtis formosana, 71 Trigonella foenum- -graecum , 493, 497 Tropical Woody Rubiaceae [r review], 579- 582 Ts1, ZHAN-HUO, PEISHAN WANG, and Da- vip E. Bourrorp. Additions to the Flora of China, 119-127 TUCKER, GORDON C. The Genera of Arun- dinoideae (Gramineae) in the South- eastern United States, 145-177 Tumion, 83 — taxifolium, 84 Turraea, 457, 459 Tyrimnus, 397, 427 Ulmaceae, 10 Umbelliferae, 489 Uniola, 1 — latifolia, 171 Urachne sect. Nassella, 174 Urochlaena, 149 Uromyces danthoniae, 150 Varadarajan, G. S. Patterns of Geographic Distribution and Their Implications on the Phylogeny of Puya (Bromeliaceae), 527-552 Verbena, 358 604 Verbenaceae, The Phylogenetic Signifi- ance of Stomata and Trichomes in the Labiatae and, 323-370 Verbenaceae, 323-370 — subfam. Avicennioideae, 324 — subfam. niente ees 341, 360, 363, — ee Chloanthoideae, 323, 324, 340, 341, 345, 356, 358, 360, 362, 363, 365 — subfam. sein ones 324 — subfam. Phrymoideae, 324 — subfam. Stilboideae, 304, 358, 365 — subfam. Symphorematoideae, 324, 340, 341, 343 — subfam. Verbenoideae, 323, 345, 356, 360, 363, 364 — — tribe Citharexyleae, 341, 343, 345 — — tribe Lantaneae 343 — — tribe Petreeae, 341 — — tribe Verbeneae, 341, 345 — subfam. Viticoideae, 343, 345, 360, 363, 64 3 — — tribe Callicarpeae, 34 — — tribe Clerodendreae, a 340, 365 — — tribe Viticeae, 341, 343 Vernonia proctoril, 131 Vietnam, a New Species from Cambodia and, Desmodium schubertiae (Legum1- nosae), 381-384 Volutaria, 442 Wagenitzia, 441 WANG, PEISHAN. Notes on the Species of Selaginella from Guizhou, China, 265-— 270 WANG, PEISHAN, DAvip E. BOUFFORD, and ZHAN-HuO Ts}. Additions to the Flora of China, 119-127 ATSON, L., and M. D. DALLwitz. Grass Genera of the World [review], 271-273 Wayside Trees of Malaya [review], 583, 584 Weberbauera (Brassicaceae), A Revision of, 221-250 Weberbauera, 221-250 — bracteata, 223, 224, 226, 237-239 — chillanensis, 221, 224, 227, 243-245 — colchaguensis, 221, 224, 226, 241-246 — cymosa, 221, 224, 226, 237-241 JOURNAL OF THE ARNOLD ARBORETUM [voL. 71 — densiflora, 225, 227 — densifolia, 221, 224, 226, 229-231 — herzogi, 221, 223, 224, 226, 236, 237 — imbricatifolia, 221, 224, 226, 246-248 — lagunae, 221, 224, 226, 246 — minutipila, 221, 222,224,227, 231-233 — parvifolia, 221, 224, 226, 248, 249 — pusilla, 241-243 — retropila, 221, 222, 224, 226, 234-236, 240 — smithii, 221, 224, 227, 233, 234 — spathulaefolia, 221-229, 231, 232, 236, 237, 239, 240, 242 — — var. integrifolia, 227 — stenophylla, 221, 224, 227, 243-245 — suffruticosa, 221, 224, 226, 246-248 — trichocarpa, 108, 222, 224, 226, 239- 24] Westringia, 365 White cedar, 304 Widdringtonia, 278, Wiedemannia, 359 280, 281, 321 Xanthopsis, 441 Xeranthemum, 394, 396 Xerodraba, 94, 100 _ colobanthoides, 100, 108 — glebaria, 108 — lycopodioides, 108 — monantha, 108 — lenge 108 — pectina — prcnophslloies 108 var. microphylla, 108 Xslocarpus chngr 460 — mekon Xylosma buxifolia, 131 ry Yew, 87 Yew Family, 69 YING, TsuN-SHEN, and Davip E. BouFForD. A New Subspecies of Mal- lotus oreophilus (Euphorbiaceae) from South-Central China, 575-578 Zeugites, 148 HUGE, REN. On the Genus Burretioden- dron Sensu Lato (Tiliaceae), 371-380 Zhumeria, 358 Zugilus, 59 JOURNAL OF THE ARNOLD ARBORETUM HARVARD UNIVERSITY VOLUME71 1990 Dates of Issue No. | (pp. 1-144) issued 25 January 1990. No. 2 (pp. 145-274) issued 5 April 1990. No. 3 (pp. 275-390) issued 9 July 1990. No. 4 (pp. 391-604) issued 11 October 1990, Contents of Volume 71 Important Notice WASSIOH: SIQtEIMONE o:o.443c ace ec cis esa ae Oe Ahad ho RE ea The Genera of Betulaceae in the Southeastern United States. JOHN J. FURLOW 1.0.00 ccc cent ene en tenns The Genera of Taxaceae in the Southeastern United States. RoBERT A. PRICE .... 0... eee ete ees Generic Limits and Taxonomy of Brayopsis and Eudema (Brassi- caceae). IHsAN A. AL-SHEHBAZ ............. 0.000 cece cece uceeuuueens Multilacunar Nodal Anatomy in Myti/aria (Hamamelidaceae). 5 MSTNIN- BOGIES secctence fan ananek e 4 istic he tag. on ie ge cece ane he er een a Additions to the Flora of China. Davip E. BOUFFORD, ZHAN-HUO TSI, AND PEISHAN WANG ..... A New Variety of Lyonia (Ericaceae) from Puerto Rico. WALTER ©; JUDD: 3434625350260 cst bod oud feanedpiedaee nes Menonvillea rollinsii (Brassicaceae), a New Shrubby Species from Chile. IHSAN A. AL-SHEHBAZ AND CLODOMIRO MARTICORENA POO RC WIOW ude oa uusey ocean beckon ewen dae eure eee canes The Genera of Arundinoideae (Gramineae) in the Southeastern United States. GORDON C. TUCKER... 00.1 Linnaeus, the Cortex-Medulla Theory, and the Key to His Under- standing of Plant Form and Natural Relationships. P. F. STEVENS AND S. P. CULLEN .......000 0000 cece A Revision of Weberbauera (Brassicaceae). ITHSAN A. AL-SHEHBAZ ... 0000000000 eee Preliminary Taxonomic Consideration of the Poraneae (Convol- vulaceae). GEORGE: We SIAPOES. pict doo ee denen eeu es peas Notes on a Novel Abaxial Leaf Epidermis in Ecuadorian Begonia parviflora. W. Scott HOOVER 135-138 139-142 145-177 179-220 221-250 251-258 259-264 Notes on the Species of Selaginella from Guizhou, China. PEIGHAN. WANG 2.056402 4th idd i be eHm ee oi RUS ky RG pO RY Book Review ............. 00. c cc eee eee teen e nee The Genera of Cupressaceae (Including Taxodiaceae) in the South- eastern United States. JEFFREY A. HART AND RoBERT A, PRICE ...............002055 The Phylogenetic Significance of Stomata and Trichomes in the Labiatae and Verbenaceae. Petite DACANTING: 26354062 d pease xed eet ed ee tees aa On the Genus Burretiodendron Sensu Lato (Tiliaceae). REN UGE ta daurdee te. nner bo ke haere wena wns Rome ee eyes Desmodium schubertiae (Leguminosae), A New Species from Cam- bodia an ietnam. HIRGYOSET OMAGH oa.4 cd adie ba eice eh KONE Re 8 So Bebe A Note on the Chilean Endemic Draba thlaspiformis (Brassicaceae). IHSAN A. AL-SHEHBAZ .... 000000000 cc een es Book Review ......... 0.0.0 ccc eben tent e nee e es The Genera of Cardueae (Compositae: Asteraceae) in the South- eastern United States. RANDALL Wo SCOTT 12495201400-ceie to dunew he twa ee eae eee ee The Genera of Meliaceae in the Southeastern United States. Norton G. MILLER ......... 000. e ees History of the Introduction of Exotic Elements into Traditional Chinese Medicine. HO VNC oe aeons ae eee ae eRe Ree eee ee Patterns of Geographic Distribution and Their Implications on the Phylogeny of Puya (Bromeliaceae). (i. 5) VARADARATAN 2c aeao obits eee cen seudebisineaendes A Reconsideration of the 4cer palmatum Complex in China, Tai- wan, and Korea CHIN-SUNG CHANG 2000. n teen eens Begonia austrotaiwanensis (Begonianeae), a New Species from Southern Taiwan. CHENG-I PENG AND YUNG-KUAN CHEN ...........00 0000000 265-270 271-273 275-322 323-370 371-380 381-384 385-387 289, 390 391-451 453-486 487-526 527-552 553-565 567-574 A New Subspecies of Mallotus oreophilus (Euphorbiaceae) from South-Central China. DAvip E. BOUFFORD AND TSUN-SHEN YING ..............-0.- 575-578 Book Reviews .......... 0.00 ccc cece eect eee n eee ens 579-584 TMGOX: ces cr se ne he ee pea Eales SD ee ee 585-604 Now Available and Complete FLORA OF THE LESSER ANTILLES by Richard A. Howard and Collaborators The Flora of the Lesser Antilles, a long-term project of Dr. Richard A. Howard, former director of the Arnold Arboretum, has been brought to a conclusion with the publication of volumes 5 and 6 during 1989. Other volumes in the series are still available, either individually or as part of a complete set: VOLUME |. Orchidaceae $20 VOLUME 2. Pteridophyta a5 VoLUME 3. Monocotyledoneae 35 VoLuME 4. Dicotyledoneae, Part | 75 VoLuME 5. Dicotyledoneae, Part 2 85 VOLUME 6. Dicotyledoneae, Part 3 85 Special price for the complete 6-volume set $260* Special price for volumes 4—6 S205" Also Available ARNOLD ARBORETUM PLANT INVENTORY A comprehensive plant inventory of the living collections of the Arnold Arboretum, together with a statement of the Arboretum’s accession policy and a fee schedule for obtaining propagating material from the collections. s2 > Orders with payment in U. S. funds, including a $2 ($4 foreign) shipping and handling charge per book, should be addressed to the attention of Frances Maguire, Arnold Arboretum, 125 The Arborway, J amaica Plain, Massachusetts 02130, U. S. A. Checks should be made payable to the Arnold Arboretum. *Prices marked with an asterisk include postage and handling. Journal of the Arnold Arboretum October, 1990 CONTENTS OF VOLUME 71, NUMBER 4 The Genera of Cardueae (Compositae; Asteraceae) in the South- eastern United States. PANGALE WOT 642442 tA 44a $tadaehie esd darksdenees 391-451 The Genera of Meliaceae in the Southeastern United States. Pigerom GG. MICRER | occ ks bie dace ena 60h b aw ve SCR RE ORR 453-486 History of the Introduction of Exotic Elements into Traditional Chinese Medicine. EE 2g eax oes ak eas Heke dis besa edad 487-526 Patterns of Geographic Distribution and Their Implications on the Phylogeny of Puya (Bromeliaceae). 5 VARADARAIAN ciichena pueveukededutenveondeeeieesaa be §27-552 A Reconsideration of the Acer palmatum Complex in China, Tai- wan, and Korea. CHIN-SUNG CHANG ........ 0000.0 cece ccc c eee nen nenenaanue 553-565 Begonia austrotaiwanensis (Begoniaceae), a New Species from Southern Taiwan. CHING-I PENG AND YUNG-KUAN CHEN ..............0 000000 567-574 A New Subspecies of Ma/lotus oreophilus (Euphorbiaceae) from South-Central China. Davip E. BOUFFORD AND TSUN-SHEN YING .................. 575-578 IS 8 [oy ail St catd (oe ee ar ee 579-584 PIV. ecieve: a hae ceni ae GR eee bea Seed Ge hed, as RACES 585-604 Volume 71, Number 3, including pages 275-390, was issued 9 July 1990.